Vijay K. Garg - St.Mary's

Loading...
WIRELESS COMMUNICATIONS AND NETWORKING

The Morgan Kaufmann Series in Networking Series Editor, David Clark, M.I.T. Wireless Communications and Networking Vijay K. Garg Ethernet Networking for the Small Office and Professional Home Office Jan L. Harrington Network Analysis, Architecture, and Design, 3e James D. McCabe IPv6 Advanced Protocols Implementation Qing Li, Tatuya Jinmei, and Keiichi Shima Computer Networks: A Systems Approach, 4e Larry L. Peterson and Bruce S. Davie Network Routing: Algorithms, Protocols, and Architectures Deepankar Medhi and Karthikeyan Ramaswami Deploying IP and MPLS QoS for Multiservice Networks: Theory and Practice John Evans and Clarence Filsfils Traffic Engineering and QoS Optimization of Integrated Voice & Data Networks Gerald R. Ash IPv6 Core Protocols Implementation Qing Li, Tatuya Jinmei, and Keiichi Shima Smart Phone and Next-Generation Mobile Computing Pei Zheng and Lionel Ni GMPLS: Architecture and Applications Adrian Farrel and Igor Bryskin Network Security: A Practical Approach Jan L. Harrington Content Networking: Architecture, Protocols, and Practice Markus Hofmann and Leland R. Beaumont Network Algorithmics: An Interdisciplinary Approach to Designing Fast Networked Devices George Varghese Network Recovery: Protection and Restoration of Optical, SONET-SDH, IP, and MPLS Jean Philippe Vasseur, Mario Pickavet, and Piet Demeester Routing, Flow, and Capacity Design in Communication and Computer Networks Michał Pióro and Deepankar Medhi Wireless Sensor Networks: An Information Processing Approach Feng Zhao and Leonidas Guibas Communication Networking: An Analytical Approach Anurag Kumar, D. Manjunath, and Joy Kuri The Internet and Its Protocols: A Comparative Approach Adrian Farrel Modern Cable Television Technology: Video, Voice, and Data Communications, 2e

Walter Ciciora, James Farmer, David Large, and Michael Adams Bluetooth Application Programming with the Java APIs C Bala Kumar, Paul J. Kline, and Timothy J. Thompson Policy-Based Network Management: Solutions for the Next Generation John Strassner MPLS Network Management: MIBs, Tools, and Techniques Thomas D. Nadeau Developing IP-Based Services: Solutions for Service Providers and Vendors Monique Morrow and Kateel Vijayananda Telecommunications Law in the Internet Age Sharon K. Black Optical Networks: A Practical Perspective, 2e Rajiv Ramaswami and Kumar N. Sivarajan Internet QoS: Architectures and Mechanisms Zheng Wang TCP/IP Sockets in Java: Practical Guide for Programmers Michael J. Donahoo and Kenneth L. Calvert TCP/IP Sockets in C: Practical Guide for Programmers Kenneth L. Calvert and Michael J. Donahoo Multicast Communication: Protocols, Programming, and Applications Ralph Wittmann and Martina Zitterbart MPLS: Technology and Applications Bruce Davie and Yakov Rekhter High-Performance Communication Networks, 2e Jean Walrand and Pravin Varaiya Internetworking Multimedia Jon Crowcroft, Mark Handley, and Ian Wakeman Understanding Networked Applications: A First Course David G. Messerschmitt Integrated Management of Networked Systems: Concepts, Architectures, and their Operational Application Heinz-Gerd Hegering, Sebastian Abeck, and Bernhard Neumair Virtual Private Networks: Making the Right Connection Dennis Fowler Networked Applications: A Guide to the New Computing Infrastructure David G. Messerschmitt Wide Area Network Design: Concepts and Tools for Optimization Robert S. Cahn For further information on these books and for a list of forthcoming titles, please visit our Web site at http://www.mkp.com.

WIRELESS COMMUNICATIONS AND NETWORKING Vijay K. Garg

Amsterdam • Boston • Heidelberg London • New York Oxford Paris • San Diego • San Francisco Singapore • Sydney • Tokyo Morgan Kaufmann is an imprint of Elsevier

Senior Acquisitions Editor Publishing Services Manager Senior Project Manager Associate Editor Editorial Assistant Cover Design Composition Illustration Copyeditor Proofreader Indexer Interior printer Cover printer

Rick Adams George Morrison Brandy Lilly Rachel Roumeliotis Brian Randall Alisa Andreola diacriTech diacriTech Janet Cocker Janet Cocker Distributech Scientific Indexing Sheridan Books Phoenix Color

Morgan Kaufmann Publishers is an imprint of Elsevier. 500 Sansome Street, Suite 400, San Francisco, CA 94111 This book is printed on acid-free paper. © 2007 by Elsevier Inc. All rights reserved. Designations used by companies to distinguish their products are often claimed as trademarks or registered trademarks. In all instances in which Morgan Kaufmann Publishers is aware of a claim, the product names appear in initial capital or all capital letters. Readers, however, should contact the appropriate companies for more complete information regarding trademarks and registration. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means—electronic, mechanical, photocopying, scanning, or otherwise—without prior written permission of the publisher. Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone: (⫹44) 1865 843830, fax: (⫹44) 1865 853333, E-mail: [email protected] You may also complete your request on-line via the Elsevier homepage (http://elsevier.com), by selecting “Support & Contact” then “Copyright and Permission” and then “Obtaining Permissions.” Library of Congress Cataloging-in-Publication Data Garg, Vijay Kumar, 1938Wireless communications and networking / Vijay K. Garg.–1st ed. p. cm. Includes bibliographical references and index. ISBN-13: 978-0-12-373580-5 (casebound : alk. paper) ISBN-10: 0-12-373580-7 (casebound : alk. paper) 1. Wireless communication systems. 2. Wireless LANs. I. Title. TK5103.2.G374 2007 621.382’1–dc22 2006100601 ISBN: 978-0-12-373580-5 For information on all Morgan Kaufmann publications, visit our Web site at www.mkp.com or www.books.elsevier.com Printed in the United States of America 07 08 09 10 11 5 4 3 2 1

Working together to grow libraries in developing countries www.elsevier.com | www.bookaid.org | www.sabre.org

The book is dedicated to my grandchildren — Adam, Devin, Dilan, Nevin, Monica, Renu, and Mollie.

This page intentionally left blank

Contents About the Author Preface 1 An Overview of Wireless Systems

xxiii xxv 1

1.1

Introduction

1

1.2

First- and Second-Generation Cellular Systems

2

1.3

Cellular Communications from 1G to 3G

5

1.4

Road Map for Higher Data Rate Capability in 3G

8

1.5

Wireless 4G Systems

14

1.6

Future Wireless Networks

15

1.7

Standardization Activities for Cellular Systems

17

1.8

Summary

19

Problems

20

References

20

2 Teletraffic Engineering

23

2.1

Introduction

23

2.2

Service Level

23

2.3

Traffic Usage

24

2.4

Traffic Measurement Units

25

2.5

Call Capacity

30

2.6

Definitions of Terms

32

2.7

Data Collection

36

2.8

Office Engineering Considerations

36

2.9

Traffic Types

38

2.10 Blocking Formulas

39

2.10.1 Erlang B Formula

40

2.10.2 Poisson’s Formula

41

2.10.3 Erlang C Formula

41

2.10.4 Comparison of Erlang B and Poisson’s Formulas

42

2.10.5 Binomial Formula

42

vii

viii

Contents

2.11 Summary

43

Problems

44

References

45

3 Radio Propagation and Propagation Path-Loss Models

47

3.1

Introduction

3.2

Free-Space Attenuation

48

3.3

Attenuation over Reflecting Surface

50

3.4

Effect of Earth’s Curvature

53

3.5

Radio Wave Propagation

54

Characteristics of Wireless Channel

58

3.6

3.6.1 3.7

Multipath Delay Spread, Coherence Bandwidth, and Coherence Time

47

60

Signal Fading Statistics

62

3.7.1

Rician Distribution

63

3.7.2

Rayleigh Distribution

64

3.7.3

Lognormal Distribution

64

3.8

Level Crossing Rate and Average Fade Duration

65

3.9

Propagation Path-Loss Models

66

3.9.1

Okumura/Hata Model

67

3.9.2

Cost 231 Model

68

3.9.3

IMT-2000 Models

72

3.10 Indoor Path-Loss Models

75

3.11 Fade Margin

76

3.12 Link Margin

79

3.13 Summary

81

Problems

82

References

83

4 An Overview of Digital Communication and Transmission

85

4.1

Introduction

85

4.2

Baseband Systems

87

4.3

Messages, Characters, and Symbols

87

4.4

Sampling Process

88

4.4.1

Aliasing

91

4.4.2

Quantization

93

Contents

ix

4.4.3

Sources of Error

94

4.4.4

Uniform Quantization

95

4.5

Voice Communication

97

4.6

Pulse Amplitude Modulation (PAM)

4.7

Pulse Code Modulation

100

4.8

Shannon Limit

102

4.9

Modulation

103

98

4.10 Performance Parameters of Coding and Modulation Scheme

105

4.11 Power Limited and Bandwidth-Limited Channel

108

4.12 Nyquist Bandwidth

109

4.13 OSI Model

112

4.13.1 OSI Upper Layers

112

4.14 Data Communication Services

113

4.15 Multiplexing

115

4.16 Transmission Media

116

4.17 Transmission Impairments

118

4.17.1 Attenuation Distortion

118

4.17.2 Phase Distortion

118

4.17.3 Level

118

4.17.4 Noise and SNR

119

4.18 Summary

120

Problems

121

References

121

5 Fundamentals of Cellular Communications

123

5.1

Introduction

123

5.2

Cellular Systems

123

5.3

Hexagonal Cell Geometry

125

5.4

Cochannel Interference Ratio

131

5.5

Cellular System Design in Worst-Case Scenario with an Omnidirectional Antenna

134

5.6

Cochannel Interference Reduction

136

5.7

Directional Antennas in Seven-Cell Reuse Pattern

137

5.7.1

Three-Sector Case

137

5.7.2

Six-Sector Case

138

5.8

Cell Splitting

141

x

Contents

5.9

Adjacent Channel Interference (ACI)

144

5.10 Segmentation

144

5.11 Summary

145

Problems

146

References

147

6 Multiple Access Techniques

149

6.1

Introduction

149

6.2

Narrowband Channelized Systems

150

6.2.1

6.3

Frequency Division Duplex (FDD) and Time Division Duplex (TDD) System

151

6.2.2

Frequency Division Multiple Access

152

6.2.3

Time Division Multiple Access

154

Spectral Efficiency

156

6.3.1

156

Spectral Efficiency of Modulation

6.3.2

Multiple Access Spectral Efficiency

159

6.3.3

Overall Spectral Efficiency of FDMA and TDMA Systems

160

6.4

Wideband Systems

163

6.5

Comparisons of FDMA, TDMA, and DS-CDMA (Figure 6.7)

166

6.6

Capacity of DS-CDMA System

168

6.7

Comparison of DS-CDMA vs. TDMA System Capacity

171

6.8

Frequency Hopping Spread Spectrum with M-ary Frequency Shift Keying

172

Orthogonal Frequency Division Multiplexing (OFDM)

173

6.9

6.10 Multicarrier DS-CDMA (MC-DS-CDMA)

175

6.11 Random Access Methods

176

6.11.1 Pure ALOHA

176

6.11.2 Slotted ALOHA

177

6.11.3 Carrier Sense Multiple Access (CSMA)

178

6.11.4 Carrier Sense Multiple Access with Collision Detection

180

6.11.5 Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA)

181

6.12 Idle Signal Casting Multiple Access

184

6.13 Packet Reservation Multiple Access

184

6.14 Error Control Schemes for Link Layer

185

6.15 Summary

188

Contents

xi

Problems

189

References

190

7 Architecture of a Wireless Wide-Area Network (WWAN)

193

7.1

Introduction

193

7.2

WWAN Subsystem Entities

194

7.2.1

User Equipment

194

7.2.2

Radio Station Subsystem

196

7.2.3

Network and Switching Subsystem

197

7.2.4

Operation and Maintenance Subsystem (OMSS)

198

7.2.5

Interworking and Interfaces

199

7.3

Logical Channels

199

7.4

Channel and Frame Structure

201

7.5

Basic Signal Characteristics

203

7.6

Speech Processing

203

7.7

Power Levels in Mobile Station

208

7.8

GSM Public Land Mobile Network Services

209

7.9

Summary

212

Problems

213

References

213

8 Speech Coding and Channel Coding 8.1 8.2

8.3

215

Introduction

215

Speech Coding

215

8.2.1

Speech Coding Methods

216

8.2.2

Speech Codec Attributes

217

8.2.3

Linear-Prediction-Based Analysis-by-Synthesis (LPAS)

218

8.2.4

Waveform Coding

219

8.2.5

Vocoders

220

8.2.6

Hybrid Coding

221

Speech Codecs in European Systems

222

8.3.1

GSM Enhanced Full-Rate (EFR)

222

8.3.2

Adaptive Multiple Rate Codec

224

8.4

CELP Speech Codec

227

8.5

Enhanced Variable Rate Codec

230

8.6

Channel Coding

233

xii

Contents

8.7

8.6.1

Reed-Solomon (RS) Codes

234

8.6.2

Convolutional Code

237

8.6.3

Turbo Coding

241

8.6.4

Soft and Hard Decision Decoding

244

8.6.5

Bit-Interleaving and De-Interleaving

245

Summary

246

Problems

247

References

247

9 Modulation Schemes

249

9.1

Introduction

249

9.2

Introduction to Modulation

249

9.3

Phase Shift Keying

257

9.3.1

Quadrature Phase Shift Keying (QPSK), Offset-Quadrature Phase Shift Keying (OQPSK) and M-PSK Modulation [5,7,11]

260

9.3.2

␲/4-DQPSK Modulation

264

9.3.3

MSK and GMSK Modulation

268

9.4

Quadrature Amplitude Modulation

272

9.5

M-ary Frequency Shift Keying

275

9.6

Modulation Selection

278

9.7

Synchronization

278

9.8

Equalization

282

9.9

Summary

284

Problems

284

References

285

10 Antennas, Diversity, and Link Analysis

287

10.1 Introduction

287

10.2 Antenna System

287

10.3 Antenna Gain

288

10.4 Performance Criteria of Antenna Systems

293

10.5 Relationship between Directivity, Gain, and Beam Width of an Antenna

295

10.5.1 The Relationship between Directivity and Gain

296

10.5.2 Relation between Gain and Beam Width

297

10.5.3 Helical Antennas

298

Contents

xiii

10.6 Diversity 10.6.1 Types of Diversity 10.7 Combining Methods

300 301 302

10.7.1 Selection Combiner

303

10.7.2 Switched Combiner

306

10.7.3 Maximal Ratio Combiner

306

10.7.4 Equal Gain Combiner

309

10.8 Rake Receiver

310

10.9 Link Budgets

312

10.10 Summary

314

Problems

315

References

315

11 Spread Spectrum (SS) and CDMA Systems

317

11.1 Introduction

317

11.2 Concept of Spread Spectrum

317

11.3 System Processing Gain

321

11.4 Requirements of Direct-Sequence Spread Spectrum

328

11.5 Frequency-Hopping Spread Spectrum Systems

329

11.6 Operational Advantages of SS Modulation

333

11.7 Coherent Binary Phase-Shift Keying DSSS

335

11.8 Quadrature Phase-Shift Keying DSSS

337

11.9 Bit Scrambling

339

11.10 Requirements of Spreading Codes

341

11.11 Multipath Path Signal Propagation and Rake Receiver

342

11.12 Critical Challenges of CDMA

347

11.13 TIA IS-95 CDMA System

347

11.13.1 Downlink (Forward) (BS to MS)

348

11.13.2 Uplink (Reverse) (MS to BS)

351

11.14 Power Control in CDMA 11.14.1 Open Loop Power Control

356 357

11.15 Softer and Soft Handoff

361

11.16 Summary

364

Problems

364

References

366

xiv

Contents

12 Mobility Management in Wireless Networks

369

12.1 Introduction

369

12.2 Mobility Management Functions

370

12.3 Mobile Location Management

371

12.3.1 Mobility Model 12.4 Mobile Registration

372 376

12.4.1 GSM Token-Based Registration

379

12.4.2 IMSI Attach and IMSI Detach (Registration and Deregistration) in GSM

381

12.4.3 Paging in GSM

381

12.5 Handoff

384

12.5.1 Handoff Techniques

386

12.5.2 Handoff Types

387

12.5.3 Handoff Process and Algorithms

387

12.5.4 Handoff Call Flows

389

12.6 Summary

393

Problems

394

References

394

13 Security in Wireless Systems

397

13.1 Introduction

397

13.2 Security and Privacy Needs of a Wireless System

399

13.2.1 Purpose of Security

399

13.2.2 Privacy Definitions

399

13.2.3 Privacy Requirements

400

13.2.4 Theft Resistance Requirements

402

13.2.5 Radio System Requirements

403

13.2.6 System Lifetime Requirements

404

13.2.7 Physical Requirements

404

13.2.8 Law Enforcement Requirements

405

13.3 Required Features for a Secured Wireless Communications System

407

13.4 Methods of Providing Privacy and Security in Wireless Systems

407

13.5 Wireless Security and Standards

409

13.6 IEEE 802.11 Security

409

13.7 Security in North American Cellular/PCS Systems

411

13.7.1 Shared Secret Data Update

412

Contents

xv

13.7.2 Global Challenge

412

13.7.3 Unique Challenge

414

13.8 Security in GSM, GPRS, and UMTS

415

13.8.1 Security in GSM

415

13.8.2 Security in GPRS

417

13.8.3 Security in UMTS

419

13.9 Data Security

420

13.9.1 Firewalls

420

13.9.2 Encryption

421

13.9.3 Secure Socket Layer

427

13.9.4 IP Security Protocol (IPSec)

427

13.9.5 Authentication Protocols

427

13.10 Air Interface Support for Authentication Methods

429

13.11 Summary of Security in Current Wireless Systems

430

13.11.1 Billing Accuracy

431

13.11.2 Privacy of Information

431

13.11.3 Theft Resistance of MS

431

13.11.4 Handset Design

431

13.11.5 Law Enforcement

431

13.12 Summary

432

Problems

432

References

433

14 Mobile Network and Transport Layer

435

14.1 Introduction

435

14.2 Concept of the Transmission Control Protocol/Internet Protocol Suite in Internet

436

14.3 Network Layer in the Internet

439

14.3.1 Internet Addresses

441

14.3.2 IP Adjunct Protocols

442

14.3.3 QoS Support in the Internet

443

14.4 TCP/IP Suite

446

14.5 Transmission Control Protocol

448

14.5.1 TCP Enhancements for Wireless Networks

452

14.5.2 Implementation of Wireless TCP

455

14.6 Mobile IP (MIP) and Session Initiation Protocol (SIP)

457

xvi

Contents

14.6.1 Mobile IP

458

14.6.2 Session Initiation Protocol (SIP)

464

14.7 Internet Reference Model

464

14.8 Summary

465

Problems

465

References

466

15 Wide-Area Wireless Networks (WANs) — GSM Evolution

469

15.1 Introduction

469

15.2 GSM Evolution for Data

470

15.2.1 High Speed Circuit Switched Data

472

15.2.2 General Packet Radio Service

473

15.2.3 Enhanced Data Rates for GSM Enhancement

483

15.3 Third-Generation (3G) Wireless Systems

489

15.4 UMTS Network Reference Architecture

495

15.5 Channel Structure in UMTS Terrestrial Radio Access Network

497

15.6 Spreading and Scrambling in UMTS

504

15.7 UMTS Terrestrial Radio Access Network Overview 15.7.1 UTRAN Logical Interfaces 15.7.2 Distribution of UTRAN Functions 15.8 UMTS Core Network Architecture

506 508 516 518

15.8.1 3G-MSC

520

15.8.2 3G-SGSN

520

15.8.3 3G-GGSN

521

15.8.4 SMS-GMSC/SMS-IWMSC

522

15.8.5 Firewall

522

15.8.6 DNS/DHCP

522

15.9 Adaptive Multi-Rate Codec for UMTS

523

15.10 UMTS Bearer Service

524

15.11 QoS Management

526

15.11.1 Functions for UMTS Bearer Service in the Control Plane

526

15.11.2 Functions for UMTS Bearer Service in the User Plane

527

15.12 Quality of Service in UMTS

528

15.12.1 QoS Classes

528

15.12.2 QoS Attributes

528

15.13 High-Speed Downlink Packet Access (HSDPA)

530

Contents

xvii

15.14 Freedom of Mobile multimedia Access (FOMA)

536

15.15 Summary

537

Problems

538

References

539

16 Wide-Area Wireless Networks — cdmaOne Evolution

541

16.1 Introduction

541

16.2 cdma2000 Layering Structure

544

16.2.1 Upper Layer

544

16.2.2 Lower Layers

545

16.3 Forward Link Physical Channels of cdma2000

550

16.4 Forward Link Features

553

16.4.1 Transmit Diversity

553

16.4.2 Orthogonal Modulation

555

16.4.3 Power Control

556

16.4.4 Walsh Code Administration

558

16.4.5 Modulation and Spreading

558

16.5 Reverse Link Physical Channels of cdma2000 16.5.1 Reverse Link Power Control 16.6 Evolution of cdmaOne (IS-95) to cdma2000

562 565 568

16.6.1 cdma2000 1X EV-DO

574

16.6.2 cdma2000 1X EV-DV

581

16.7 Technical Differences between cdma2000 and WCDMA

586

16.8 Summary

587

Problems

592

References

592

17 Planning and Design of Wide-Area Wireless Networks

595

17.1 Introduction

595

17.2 Planning and Design of a Wireless Network

596

17.3 Radio Design for a Cellular Network

600

17.3.1 Radio Link Design

600

17.3.2 Coverage Planning

601

17.4 Receiver Sensitivity and Link Budget

602

17.4.1 Link Budget for the GSM1800 System

602

17.4.2 Pole Capacity of a CDMA Cell

605

xviii

Contents

17.4.3 Uplink Radio Link Budget for a CDMA System

606

17.4.4 Downlink Radio Link Budget for a CDMA System

609

17.5 cdma2000 1X EV-DO

615

17.5.1 1X EV-DO Concept

615

17.5.2 Details of cdma2000 1X EV-DO

617

17.6 High-Speed Downlink Packet Access 17.6.1 HSDPA SINR Calculation

620 623

17.7 Iub Interface Dimensioning

624

17.8 Radio Network Controller Dimensioning

624

17.9 Summary

626

Problems

626

References

629

18 Wireless Application Protocol

631

18.1 Introduction

631

18.2 WAP and the World Wide Web (WWW)

631

18.3 Introduction to Wireless Application Protocol

632

18.4 The WAP Programming Model

633

18.4.1 The WWW Model

634

18.4.2 The WAP Model

634

18.5 WAP Architecture

636

18.5.1 Wireless Application Environment

637

18.5.2 Wireless Telephony Application

638

18.5.3 Wireless Session Protocol

639

18.5.4 Wireless Transaction Protocol

640

18.5.5 Wireless Transport Layer Security

641

18.5.6 Wireless Datagram Protocol

641

18.5.7 Optimal WAP Bearers

642

18.6 Traditional WAP Networking Environment

643

18.7 WAP Advantages and Disadvantages

645

18.8 Applications of WAP

646

18.9 imode

647

18.10 imode versus WAP

649

18.11 Summary

650

Problems

650

References

650

Contents

xix

19 Wireless Personal Area Network — Bluetooth

653

19.1 Introduction

653

19.2 The Wireless Personal Area Network

654

19.3 Bluetooth (IEEE 802.15.1)

656

19.4 Definitions of the Terms Used in Bluetooth

659

19.5 Bluetooth Protocol Stack

660

19.5.1 Transport Protocol Group

660

19.5.2 Middleware Protocol Group

661

19.5.3 Application Group

663

19.6 Bluetooth Link Types

663

19.7 Bluetooth Security

666

19.7.1 Security Levels

667

19.7.2 Limitations of Bluetooth Security

669

19.8 Network Connection Establishment in Bluetooth

669

19.9 Error Correction in Bluetooth

670

19.10 Network Topology in Bluetooth

671

19.11 Bluetooth Usage Models

671

19.12 Bluetooth Applications

672

19.13 WAP and Bluetooth

673

19.14 Summary

673

Problems

673

References

674

20 Wireless Personal Area Networks: Low Rate and High Rate

675

20.1 Introduction

675

20.2 Wireless Sensor Network

675

20.3 Usage of Wireless Sensor Networks

678

20.4 Wireless Sensor Network Model

678

20.5 Sensor Network Protocol Stack

683

20.5.1 Physical Layer

683

20.5.2 Data Link Layer

684

20.5.3 Network Layer

685

20.5.4 Transport Layer

687

20.5.5 Application Layer

687

20.5.6 Power, Mobility, and Task Management Planes

688

xx

Contents

20.6 ZigBee Technology 20.6.1 ZigBee Components and Network Topologies 20.7 IEEE 802.15.4 LR-WPAN Device Architecture

688 689 691

20.7.1 Physical Layer

692

20.7.2 Data Link Layer

694

20.7.3 The Network Layer

697

20.7.4 Applications

702

20.8 IEEE 802.15.3a — Ultra WideBand

703

20.9 Radio Frequency Identification

707

20.10 Summary

710

Problems

710

References

711

21 Wireless Local Area Networks

713

21.1 Introduction

713

21.2 WLAN Equipment

716

21.3 WLAN Topologies

717

21.4 WLAN Technologies

719

21.4.1 Infrared Technology

719

21.4.2 UHF Narrowband Technology

719

21.4.3 Spread Spectrum Technology

721

21.5 IEEE 802.11 WLAN

721

21.5.1 IEEE 802.11 Architecture

722

21.5.2 802.11 Physical Layer (PHY)

723

21.5.3 IEEE 802.11 Data Link Layer

735

21.5.4 IEEE 802.11 Medium Access Control

736

21.5.5 IEEE 802.11 MAC Sublayer

742

21.6 Joining an Existing Basic Service Set

744

21.7 Security of IEEE 802.11 Systems

747

21.8 Power Management

747

21.9 IEEE 802.11b — High Rate DSSS

748

21.10 IEEE 802.11n

749

21.11 Other WLAN Standards

752

21.11.1 HIPERLAN Family of Standards

752

21.11.2 Multimedia Access Communication — High Speed Wireless Access Network

758

Contents

xxi

21.12 Performance of a Bluetooth Piconet in the Presence of IEEE 802.11 WLANs

759

21.12.1 Packet Error Rate (PER) from N Neighboring Bluetooth Piconets

760

21.12.2 PER from M Neighboring IEEE 802.11 WLANs

761

21.12.3 Aggregated Throughput

762

21.13 Interference between Bluetooth and IEEE 802.11

763

21.14 IEEE 802.16

765

21.15 World Interoperability for MicroAccess, Inc. (WiMAX)

767

21.15.1 WiMAX Physical Layer (PHY)

770

21.15.2 WiMAX Media Access Control (MAC)

771

21.15.3 Spectrum Allocation for WiMAX

772

21.16 Summary

772

Problems

774

References

775

Appendix A

777

Acronyms

787

Index

806

The following Bonus Chapters can be found on the book’s website at http://books.elsevier.com/9780123735805: 22 Interworking between Wireless Local Area Networks and 3G Wireless Wide Area Networks

22-1

22.1 Introduction

22-1

22.2 Interworking Objectives and Requirements

22-2

22.3 Interworking Schemes to Connect WLANs and 3G Networks

22-3

22.4 De Facto WLAN System Architecture

22-5

22.5 Session Mobility

22-7

22.6 Interworking Architectures for WLAN and GPRS

22-8

22.7 System Description with Tight Coupling

22-9

22.7.1 Protocol Stack

22-12

22.7.2 WLAN Adaptation Function

22-13

22.7.3 GIF/RAI Discovery Procedure

22-15

22.8 System Description with Loose Coupling

22-17

xxii

Contents

22.8.1 Authentication

22-20

22.8.2 User Data Routing and Access to Services

22-23

22.8.3 3GPP-based Charging for WLAN

22-23

22.8.4 Session Mobility

22-26

22.9 Local Multipoint Distribution Service

22-26

22.10 Multichannel Multipoint Distribution System

22-29

22.11 Summary

22-31

Problems

22-32

References

22-32

23 Fourth Generation Systems and New Wireless Technologies

23-1

23.1 Introduction

23-1

23.2 4G Vision

23-2

23.3 4G Features and Challenges

23-3

23.4 Applications of 4G

23-7

23.5 4G Technologies

23-7

23.5.1 Multicarrier Modulation

23-7

23.5.2 Smart Antenna Techniques

23-10

23.5.3 OFDM-MIMO Systems

23-14

23.5.4 Adaptive Modulation and Coding with Time-Slot Scheduler

23-14

23.5.5 Bell Labs Layered Space Time (BLAST) System

23-15

23.5.6 Software-Defined Radio

23-18

23.5.7 Cognitive Radio

23-20

23.6 Summary

23-21

Problems

23-21

References

23-22

Appendix B

Path Loss over a Reflecting Surface

B-1

Appendix C

Error Functions

C-1

Appendix D

Spreading Codes Used in CDMA

D-1

Appendix E

Power Units

E-1

About the Author Vijay K. Garg has been a professor in the Electrical and Computer Engineering Department at the University of Illinois at Chicago since 1999, where he teaches graduate courses in Wireless Communications and Networking. Dr. Garg was a Distinguished Member of Technical Staff at the Lucent Technologies Bell Labs in Naperville, Illinois from 1985 to 2001. He received his Ph.D. degree from the Illinois Institute of Technologies, Chicago, IL in 1973 and his MS degree from the University of California at Berkeley, CA in 1966. Dr. Garg has co-authored several technical books including five in wireless communications. He is a Fellow of ASCE and ASME, and a Senior Member of IEEE. Dr. Garg is a registered Professional Engineer in the state of Maine and Illinois. He is an Academic Member of the Russian Academy of Transport. Dr. Garg was a Feature Editor of Wireless/ PCS Series in IEEE Communication Magazine from 1996–2001.

xxiii

This page intentionally left blank

Preface During the past three decades, the world has seen significant changes in the telecommunications industry. There has been rapid growth in wireless communications, as seen by large expansion in mobile systems. Wireless communications have moved from first-generation (1G) systems primarily focused on voice communications to third-generation (3G) systems dealing with Internet connectivity and multi-media applications. The fourth-generation (4G) systems will be designed to connect wireless personal area networks (WPANs), wireless local area networks (WLANs) and wireless wide-area networks (WWANs). With the Internet and corporate intranets becoming essential parts of daily business activities, it has become increasingly advantageous to have wireless offices that can connect mobile users to their enterprises. The potential for technologies that deliver news and other business-related information directly to mobile devices could also develop entirely new revenue streams for service providers. The 3G mobile systems are expected to provide worldwide access and global roaming for a wide range of services. The 3G WWANs are designed to support data rates up to 144 kbps with comprehensive coverage and up to 2 Mbps for selected local areas. Prior to the emergence of 3G services, mobile data networks such as general packet radio service (GPRS) over time division multiple-access (TDMA) systems and high-speed packet data over IS-95 code-division multiple access (CDMA) systems were already very popular. At the same time, after the introduction of Bluetooth and imode technology in 1998, local broadband and ad hoc wireless networks attracted a great deal of attention. This sector of the wireless networking industry includes the traditional WLANs and the emerging WPANs. Multi hop wireless ad hoc networks complement the existing WLAN standards like IEEE 802.11a/b/g/n and Bluetooth to allow secure, reliable wireless communications among all possible hand-held devices such as personal digital assistances (PDAs), cell-phones, laptops, or other portable devices that have a wireless communication interface. Ad hoc networks are not dependent on a single point of attachment. The routing protocols for ad hoc networks are designed to self-configure and self-organize the networks to seamlessly create an access point on the fly as a user or device moves. Provisioning data services over the wireless data networks including ad hoc networks requires smart data management protocols and new transaction models for data delivery and transaction processing, respectively. While personalization of data services is desired, over personalization will have ramifications on scalability of wireless networks? As such, mobile computing not only poses challenges but also opens up an interesting research area. It is redefining existing business models xxv

xxvi

Preface

and creating entirely new ones. Envisioning new business processes vis-à-vis the enabling technologies is also quite interesting. Over the past decade, wireless data networking has developed into its own discipline. There is no doubt that the evolution of wireless networks has had significant impact on our lifestyle. This book is designed to provide a unified foundation of principles for data-oriented wireless networking and mobile communications. This book is an extensive enhancement to the Wireless & Personal Communications book published by Prentice Hall in 1996, which primarily addressed 2G cellular networks. Since then, wireless technologies have undergone significant changes; new and innovative techniques have been introduced, the focus of wireless communications is increasingly changing from mobile voice applications to mobile data and multimedia applications. Wireless technology and computing have come closer and closer to generating a strong need to address this issue. In addition, wireless networks now include wide area cellular networks, wireless local area networks, wireless metropolitan area networks, and wireless personal area networks. This book addresses these networks in extensive detail. The book primarily discusses wireless technologies up to 3G but also provides some insight into 4G technologies. It is indeed a challenge to provide an over-arching synopsis for mobile data networking and mobile communications for diverse audiences including managers, practicing engineers, and students who need to understand this industry. My basic motivation in writing this book is to provide the details of mobile data networking and mobile communications under a single cover. In the last two decades, many books have been written on the subject of wireless communications and networking. However, mobile data networking and mobile communications were not fully addressed. This book is written to provide essentials of wireless communications and wireless networking including WPAN, WLAN, WMAN, and WWAN. The book is designed for practicing engineers, as well as senior/first-year graduate students in Electrical and Computer Engineering (ECE), and Computer Science (CS). The first thirteen chapters of the book focus on the fundamentals that are required to study mobile data networking and mobile communications. Numerous solved examples have been included to show applications of theoretical concepts. In addition, unsolved problems are given at the end of each chapter for practice. After introducing fundamental concepts, the book focuses on mobile networking aspects with several chapters devoted to the discussion of WPAN, WLAN, WWAN, and other aspects of mobile communications such as mobility management, security, and cellular network planning. Two additional “Bonus” chapters on inter-working between WLAN and WWAN and on 4G systems (along with several helpful appendices) are available free on the book’s website at http://books. elsevier.com/9780123735805.

Preface

xxvii

Most of the books in wireless communications and networking appear to ignore the standard activities in the field. I feel students in wireless networking must be exposed to various standard activities. I therefore address important standard activities including 3GPP, 3GPP2, IEEE 802.11, IEEE 802.15 and IEEE 802.16 in the book. This feature of the book is also very beneficial to the professionals who wish to know about a particular standard without going through the voluminous material on that standard. A unique feature of this book that is missing in most of the available books on wireless communications and networking is to offer a balance between theoretical and practical concepts. This book can be used to teach two semester courses in mobile data networking and mobile communications to ECE and CS students. Chapter 4 may be omitted for ECE students and Chapter 14 for CS students. The first course — Introduction to Wireless Communications and Networking can be offered to senior undergraduate and first year graduate students. This should include first fourteen chapters. Chapters 4 and 14 may be omitted depending on the students’ background. The second course — Wireless Data Networking should include Chapters 15 through 23. The first course should be a pre-requisite to the second course. The student should be given homework, two examinations, and a project to complete each course. In addition, this book can also be used to teach a comprehensive course in Wireless Data Networking to IT professionals by using Chapters 2, 3, 5, 6, 7, 11, 15, 16, and 18 to 22. During the preparation of this manuscript my family members were very supportive. I would like to thank my children, Nina, Meena, and Ravi. Also, I appreciate the support given by my wife, Pushpa. In addition, I appreciate the support of the reviewers, Elaine Cheong, Frank Farrante, and Pei Zhang in providing valuable comments on the manuscript. Finally, I am thankful of Rachel Roumeliotis for coordinating the reviews of the manuscript. Vijay K. Garg Willowbrook, IL

This page intentionally left blank

CHAPTER 1 An Overview of Wireless Systems 1.1

Introduction

The cellular system employs a different design approach than most commercial radio and television systems use [1,2]. Radio and television systems typically operate at maximum power and with the tallest antennas allowed by the regulatory agency of the country. In the cellular system, the service area is divided into cells. A transmitter is designed to serve an individual cell. The system seeks to make efficient use of available channels by using low-power transmitters to allow frequency reuse at much smaller distances. Maximizing the number of times each channel can be reused in a given geographic area is the key to an efficient cellular system design. During the past three decades, the world has seen significant changes in the telecommunications industry. There have been some remarkable aspects to the rapid growth in wireless communications, as seen by the large expansion in mobile systems. Wireless systems consist of wireless wide-area networks (WWAN) [i.e., cellular systems], wireless local area networks (WLAN) [4], and wireless personal area networks (WPAN) (see Figure 1.1) [17]. The handsets used in all of these systems possess complex functionality, yet they have become small, lowpower consuming devices that are mass produced at a low cost, which has in turn accelerated their widespread use. The recent advancements in Internet technology have increased network traffic considerably, resulting in a rapid growth of data rates. This phenomenon has also had an impact on mobile systems, resulting in the extraordinary growth of the mobile Internet. Wireless data offerings are now evolving to suit consumers due to the simple reason that the Internet has become an everyday tool and users demand data mobility. Currently, wireless data represents about 15 to 20% of all air time. While success has been concentrated in vertical markets such as public safety, health care, and transportation, the horizontal market (i.e., consumers) for wireless data is growing. In 2005, more than 20 million people were using wireless e-mail. The Internet has changed user expectations of what data access means. The ability to retrieve information via the Internet has been “an amplifier of demand” for wireless data applications. More than three-fourths of Internet users are also wireless users and a mobile subscriber is four times more likely to use the Internet than a nonsubscriber to 1

2

1

Short Range: Low Power, Wireless Personal Area Network (WPAN) Bluetooth (1 Mbps) Ultra Wideband (UWB) (100 Mbps) Sensor Networks IEEE 802.15.4, Zigbee

An Overview of Wireless Systems

Long Distance: High Power, Wireless Wide Area Networks (WWAN) 2G GSM (9.6 kbps) PDC GPRS (114 kbps) PHS (64 kbps, up to 128 kbps 3G (cdma2000, WCDMA) (384 kbps to 2 Mbps)

Middle Range: Medium Power, Wireless Local Area Network (WLAN) Home RF (10 Mbps) IEEE802.11a,b,g (108 Mbps) [802.11a based proprietary 2x mode]

PDC: Personal Digital Cellular (Japan) GPRS: General Packet Radio Service PHS: Personal Handy Phone System (Japan)

Figure 1.1

Wireless networks.

mobile services. Such keen interest in both industries is prompting user demand for converged services. With more than a billion Internet users expected by 2008, the potential market for Internet-related wireless data services is quite large. In this chapter, we discuss briefly 1G, 2G, 2.5G, and 3G cellular systems and outline the ongoing standard activities in Europe, North America, and Japan. We also introduce broadband (4G) systems (see Figure 1.2) aimed on integrating WWAN, WLAN, and WPAN. Details of WWAN, WLAN, and WPAN are given in Chapters 15 to 20.

1.2

First- and Second-Generation Cellular Systems

The first- and second-generation cellular systems are the WWAN. The first public cellular telephone system (first-generation, 1G), called Advanced Mobile Phone System (AMPS) [8,21], was introduced in 1979 in the United States. During the early 1980s, several incompatible cellular systems (TACS, NMT, C450, etc.) were introduced in Western Europe. The deployment of these incompatible systems resulted in mobile phones being designed for one system that could not be used with another system, and roaming between the many countries of Europe was not possible. The first-generation systems were designed for voice applications. Analog frequency modulation (FM) technology was used for radio transmission. In 1982, the main governing body of the European post telegraph and telephone (PTT), la Conférence européenne des Administrations des postes et des

1.2

First- and Second-Generation Cellular Systems

Spectral Efficiency

0.30 bps/Hz

0.15 bps/Hz Max. rate 64kbps

Max. rate 2 Mbps

TDMA & CDMA

TDMA, CDMA and WCDMA

FDMA 1G Analog

AMPS TACS NMT C-450

2G Digital modulation Convolution coding Power control

2.5G/3G Hierarchal cell structure Turbo-coding

PDC GSM HSCSD GPRS IS-54/IS-136 IS-95/IS-95A/IS-95B PHS

3

3 4 bps/Hz (targeted) Max. rate ~ 200 Mbps WCDMA 4G Smart antennas? MIMO? Adaptive system OFDM modulation

EDGE cdma2000 WCDMA/UMTS 3G 1X EV-DO 3G 1X EV-DV

PHS: Personal handy phone system (Japan) MIMO: Multi-input and multi-output OFDM: Orthogonal Frequency Division Multiple Access

Figure 1.2

Wireless network from 1G to 4G.

télécommunications (CEPT), set up a committee known as Groupe Special Mobile (GSM) [9], under the auspices of its Committee on Harmonization, to define a mobile system that could be introduced across western Europe in the 1990s. The CEPT allocated the necessary duplex radio frequency bands in the 900 MHz region. The GSM (renamed Global System for Mobile communications) initiative gave the European mobile communications industry a home market of about 300 million subscribers, but at the same time provided it with a significant technical challenge. The early years of the GSM were devoted mainly to the selection of radio technologies for the air interface. In 1986, field trials of different candidate systems proposed for the GSM air interface were conducted in Paris. A set of criteria ranked in the order of importance was established to assess these candidates. The interfaces, protocols, and protocol stacks in GSM are aligned with the Open System Interconnection (OSI) principles. The GSM architecture is an open architecture which provides maximum independence between network elements (see Chapter 7) such as the Base Station Controller (BSC), the Mobile Switching Center (MSC), the Home Location Register (HLR), etc. This approach simplifies the design, testing, and implementation of the system. It also favors an evolutionary growth path, since network element independence implies that modification to one network element can be made with minimum or no impact on the others. Also, a system operator has the choice of using network elements from different manufacturers.

4

1

An Overview of Wireless Systems

GSM 900 (i.e., GSM system at 900 MHz) was adopted in many countries, including the major parts of Europe, North Africa, the Middle East, many east Asian countries, and Australia. In most of these cases, roaming agreements exist to make it possible for subscribers to travel within different parts of the world and enjoy continuity of their telecommunications services with a single number and a single bill. The adaptation of GSM at 1800 MHz (GSM 1800) also spreads coverage to some additional east Asian countries and some South American countries. GSM at 1900 MHz (i.e., GSM 1900), a derivative of GSM for North America, covers a substantial area of the United States. All of these systems enjoy a form of roaming, referred to as Subscriber Identity Module (SIM) roaming, between them and with all other GSM-based systems. A subscriber from any of these systems could access telecommunication services by using the personal SIM card in a handset suitable to the network from which coverage is provided. If the subscriber has a multiband phone, then one phone could be used worldwide. This globalization has positioned GSM and its derivatives as one of the leading contenders for offering digital cellular and Personal Communications Services (PCS) worldwide. A PCS system offers multimedia services (i.e., voice, data, video, etc.) at any time and any where. With a three band handset (900, 1800, and 1900 MHz), true worldwide seamless roaming is possible. GSM 900, GSM 1800, and GSM 1900 are second-generation (2G) systems and belong to the GSM family. Cordless Telephony 2 (CT2) is also a 2G system used in Europe for low mobility. Two digital technologies, Time Division Multiple Access (TDMA) and Code Division Multiple Access (CDMA) (see Chapter 6 for details) [10] emerged as clear choices for the newer PCS systems. TDMA is a narrowband technology in which communication channels on a carrier frequency are apportioned by time slots. For TDMA technology, there are three prevalent 2G systems: North America TIA/ EIA/IS-136, Japanese Personal Digital Cellular (PDC), and European Telecommunications Standards Institute (ETSI) Digital Cellular System 1800 (GSM 1800), a derivative of GSM. Another 2G system based on CDMA (TIA/EIA/IS-95) is a direct sequence (DS) spread spectrum (SS) system in which the entire bandwidth of the carrier channel is made available to each user simultaneously (see Chapter 11 for details). The bandwidth is many times larger than the bandwidth required to transmit the basic information. CDMA systems are limited by interference produced by the signals of other users transmitting within the same bandwidth. The global mobile communications market has grown at a tremendous pace. There are nearly one billion users worldwide with two-thirds being GSM users. CDMA is the fastest growing digital wireless technology, increasing its worldwide subscriber base significantly. Today, there are already more than 200 million CDMA subscribers. The major markets for CDMA technology are North America, Latin America, and Asia, in particular Japan and Korea. In total, CDMA has been adopted by almost 50 countries around the world. The reasons behind the success of CDMA are obvious. CDMA is an advanced digital cellular technology, which can offer six to eight times the capacity of analog

1.3

Cellular Communications from 1G to 3G

5

technologies (AMP) and up to four times the capacity of digital technologies such as TDMA. The speech quality provided by CDMA systems is far superior to any other digital cellular system, particularly in difficult RF environments such as dense urban areas and mountainous regions. In both initial deployment and long-term operation, CDMA provides the most cost effective solution for cellular operators. CDMA technology is constantly evolving to offer customers new and advanced services. The mobile data rates offered through CDMA phones have increased and new voice codecs provide speech quality close to the fixed wireline. Internet access is now available through CDMA handsets. Most important, the CDMA network offers operators a smooth migration path to third-generation (3G) mobile systems, [3,5,7,11].

1.3

Cellular Communications from 1G to 3G

Mobile systems have seen a change of generation, from first to second to third, every ten years or so (see Figure 1.3). At the introduction of 1G services, the mobile device was large in size, and would only fit in the trunk of a car. All analog components such as the power amplifier, synthesizer, and shared antenna equipment were bulky. 1G systems were intended to provide voice service and low rate (about 9.6 kbps) circuit-switched data services. Miniaturization of mobile devices progressed before the introduction of 2G services (1990) to the point where the size of mobile phones fell below 200 cubic centimeters (cc). The first-generation handsets provided poor voice quality, low talk-time, and low

2G

3G

PDC

ARIB (WCDMA)

TDD

WCDMA

Direct Spreading

3G 2G

FDD

WCDMA/UMTS

GSM

GPRS EDGE 2.5G

1G

2G

AMPS

IS-54

UWC-136

IS-136 IS-136

cdma2000 1x

IS-95/95A IS-95B

Figure 1.3

3G

Cellular networks (WWAN) evolution from 1G to 3G.

Multiple Carriers CDMA

6

1

An Overview of Wireless Systems

standby time. The 1G systems used Frequency Division Multiple Access (FDMA) technology (see Chapter 6) and analog frequency modulation [8,20]. The 2G systems based on TDMA and CDMA technologies [6] were primarily designed to improve voice quality and provide a set of rich voice features. These systems supported low rate data services (16–32 kbps). For second-generation systems three major problems impacting system cost and quality of service remained unsolved. These include what method to use for band compression of voice, whether to use a linear or nonlinear modulation scheme, and how to deal with the issue of multipath delay spread caused by multipath propagation of radio waves in which there may not only be phase cancellation but also a significant time difference between the direct and reflected waves. The swift progress in Digital Signal Processors (DSPs) was probably fueled by the rapid development of voice codecs for mobile environments that dealt with errors. Large increases in the numbers of cellular subscribers and the worries of exhausting spectrum resources led to the choice of linear modulation systems. To deal with multipath delay spread, Europe, the United States, and Japan took very different approaches. Europe adopted a high transmission rate of 280 kbps per 200 kHz RF channel in GSM [13,14] using a multiplexed TDMA system with 8 to 16 voice channels, and a mandatory equalizer with a high number of taps to overcome inter-symbol interference (ISI) (see Chapter 3). The United States used the carrier transmission rate of 48 kbps in 30 kHz channel, and selected digital advanced mobile phone (DAMP) systems (IS-54/IS-136) to reduce the computational requirements for equalization, and the CDMA system (IS-95) to avoid the need for equalization. In Japan the rate of 42 kbps in 25 kHz channel was used, and equalizers were made optional. Taking into account the limitations imposed by the finite amount of radio spectrum available, the focus of the third-generation (3G) mobile systems has been on the economy of network and radio transmission design to provide seamless service from the customers’ perspective. The third-generation systems provide their users with seamless access to the fixed data network [18,19]. They are perceived as the wireless extension of future fixed networks, as well as an integrated part of the fixed network infrastructure. 3G systems are intended to provide multimedia services including voice, data, and video. One major distinction of 3G systems relative to 2G systems is the hierarchical cell structure designed to support a wide range of multimedia broadband services within the various cell types by using advanced transmission and protocol technologies. The 2G systems mainly use one-type cell and employ frequency reuse within adjacent cells in such a way that each single cell manages its own radio zone and radio circuit control within the mobile network, including traffic management and handoff procedures. The traffic supported in each cell is fixed because of frequency limitations and little flexibility of radio transmission which is mainly optimized for voice and low data rate transmissions. Increasing

1.3

Cellular Communications from 1G to 3G

7

traffic leads to costly cellular reconfiguration such as cell splitting and cell sectorization. The multilayer cell structure in 3G systems aims to overcome these problems by overlaying, discontinuously, pico- and microcells over the macrocell structure with wide area coverage. Global/satellite cells can be used in the same sense by providing area coverage where macrocell constellations are not economical to deploy and/or support long distance traffic. With low mobility and small delay spread profiles in picocells, high bit rates and high traffic densities can be supported with low complexity as opposed to low bit rates and low traffic load in macrocells that support high mobility. The user expectation will be for service selected in a uniform manner with consistent procedures, irrespective of whether the means of access to these services is fixed or mobile. Freedom of location and means of access will be facilitated by smart cards to allow customers to register on different terminals with varying capabilities (speech, multimedia, data, short messaging). The choice of a radio interface parameter set corresponding to a multiple access scheme is a critical issue in terms of spectral efficiency, taking into account the everincreasing market demand for mobile communications and the fact that radio spectrum is a very expensive and scarce resource. A comparative assessment of several different schemes was carried out in the framework of the Research in Advanced Communications Equipment (RACE) program. One possible solution is to use a hybrid CDMA/TDMA/FDMA technique by integrating advantages of each and meeting the varying requirements on channel capacity, traffic load, and transmission quality in different cellular/PCS layouts. Disadvantages of such hybrid access schemes are the high-complexity difficulties in achieving simplified low-power, low-cost transceiver design as well as efficient flexibility management in the several cell layers. CDMA is the selected approach for 3G systems by the ETSI, ARIB (Association of Radio Industries and Business — Japan) and Telecommunications Industry Association (TIA). In Europe and Japan, Wideband CDMA (WCDMA/UMTS [Universal Mobile Telecommunication Services]) was selected to avoid IS-95 intellectual property rights. In North America, cdma2000 uses a CDMA air-interface based on the existing IS-95 standard to provide wireline quality voice service and high speed data services at 144 kbps for mobile users, 384 kbps for pedestrians, and 2 Mbps for stationary users. The 64 kbps data capability of CDMA IS-95B provides high speed Internet access in a mobile environment, a capability that cannot be matched by other narrowband digital technologies. Mobile data rates up to 2 Mbps are possible using wide band CDMA technologies. These services are provided without degrading the systems’ voice transmission capabilities or requiring additional spectrum. This has tremendous implications for the majority of operators that are spectrum constrained. In the meantime, DSPs have improved in speed by an order of magnitude in each generation, from 4 MIPs (million instructions per second) through 40 MIPs to 400 MIPs.

8

1

An Overview of Wireless Systems

Since the introduction of 2G systems, the base station has seen the introduction of features such as dynamic channel assignment. In addition, most base stations began making shared use of power amplifiers and linear amplifiers whether or not modulation was linear. As such there has been an increasing demand for high-efficiency, large linear power amplifiers instead of nonlinear amplifiers. At the beginning of 2G, users were fortunate if they were able to obtain a mobile device below 150 cc. Today, about 10 years later, mobile phone size has reached as low as 70 cc. Furthermore, the enormous increase in very large system integration (VLSI) and improved CPU performance has led to increased functionality in the handset, setting the path toward becoming a small-scale computer.

1.4

Road Map for Higher Data Rate Capability in 3G

The first- and second-generation cellular systems were primarily designed for voice services and their data capabilities were limited. Wireless systems have since been evolving to provide broadband data rate capability as well. GSM is moving forward to develop cutting-edge, customer-focused solutions to meet the challenges of the 21st century and 3G mobile services. When GSM was first designed, no one could have predicted the dramatic growth of the Internet and the rising demand for multimedia services. These developments have brought about new challenges to the world of GSM. For GSM operators, the emphasis is now rapidly changing from that of instigating and driving the development of technology to fundamentally enable mobile data transmission to that of improving speed, quality, simplicity, coverage, and reliability in terms of tools and services that will boost mass market take-up. People are increasingly looking to gain access to information and services whenever they want from wherever they are. GSM will provide that connectivity. The combination of Internet access, web browsing, and the whole range of mobile multimedia capability is the major driver for development of higher data speed technologies. GSM operators have two nonexclusive options for evolving their networks to 3G wide band multimedia operation: (1) they can use General Packet Radio Service (GPRS) and Enhanced Data rates for GSM Evolution (EDGE) [also known as 2.5G] in the existing radio spectrum, and in small amounts of new spectrum; or (2) they can use WCDMA/UMTS in the new 2 GHz bands [12,15,16]. Both approaches offer a high degree of investment flexibility because roll-out can proceed in line with market demand and there is extensive reuse of existing network equipment and radio sites. The first step to introduce high-speed circuit-switched data service in GSM is by using High Speed Circuit Switched Data (HSCSD). HSCSD is a feature that enables the co-allocation of multiple full rate traffic channels (TCH/F) of GSM into an HSCSD configuration. The aim of HSCSD is to provide a mixture of

1.4

Road Map for Higher Data Rate Capability in 3G

9

services with different user data rates using a single physical layer structure. The available capacity of an HSCSD configuration is several times the capacity of a TCH/F, leading to a significant enhancement in data transfer capability. Ushering faster data rates into the mainstream is the new speed of 14.4 kbps per time slot and HSCSD protocols that approach wire-line access rates of up to 57.6 kbps by using multiple 14.4 kbps time slots. The increase from the current baseline 9.6 kbps to 14.4 kbps is due to a nominal reduction in the error-correction overhead of the GSM radio link protocol, allowing the use of a higher data rate. The next phase in the high speed road map is the evolution of current short message service (SMS), such as smart messaging and unstructured supplementary service data, toward the new GPRS, a packet data service using TCP/IP and X.25 to offer speeds up to 115.2 kbps. GPRS has been standardized to optimally support a wide range of applications ranging from very frequent transmissions of medium to large data volume. Services of GPRS have been developed to reduce connection set-up time and allow an optimum usage of radio resources. GPRS provides a packet data service for GSM where time slots on the air interface can be assigned to GPRS over which packet data from several mobile stations can be multiplexed. A similar evolution strategy, also adopting GPRS, has been developed for DAMPS (IS-136). For operators planning to offer wide band multimedia services, the move to GPRS packet-based data bearer service is significant; it is a relatively small step compared to building a totally new 3G network. Use of the GPRS network architecture for IS-136 packet data service enables data subscription roaming with GSM networks around the globe that support GPRS and its evolution. The IS-136 packet data service standard is known as GPRS-136. GPRS-136 provides the same capabilities as GSM GPRS. The user can access either X.25 or IP-based data networks. GPRS provides a core network platform for current GSM operators not only to expand the wireless data market in preparation for the introduction of 3G services, but also a platform on which to build UMTS frequencies should they acquire them. GPRS enhances GSM data services significantly by providing end-to-end packet switched data connections. This is particularly efficient in Internet/intranet traffic, where short bursts of intense data communications actively are interspersed with relatively long periods of inactivity. Since there is no real end-to-end connection to be established, setting up a GPRS call is almost instantaneous and users can be continuously on-line. Users have the additional benefits of paying for the actual data transmitted, rather than for connection time. Because GPRS does not require any dedicated end-to-end connection, it only uses network resources and bandwidth when data is actually being transmitted. This means that a given amount of radio bandwidth can be shared efficiently between many users simultaneously.

10

1

An Overview of Wireless Systems

The significance of EDGE (also referred to as 2.5G system) for today’s GSM operators is that it increases data rates up to 384 kbps and potentially even higher in good quality radio environments that are using current GSM spectrum and carrier structures more efficiently. EDGE will both complement and be an alternative to new WCDMA coverage. EDGE will also have the effect of unifying the GSM, DAMPS, and WCDMA services through the use of dual-mode terminals. GSM operators who win licenses in new 2 GHz bands will be able to introduce UMTS wideband coverage in areas where early demand is likely to be greatest. Dual-mode EDGE/ UMTS mobile terminals will allow full roaming and handoff from one system to the other, with mapping of services between the two systems. EDGE will contribute to the commercial success of the 3G system in the vital early phases by ensuring that UMTS subscribers will be able to enjoy roaming and interworking globally. While GPRS and EDGE require new functionality in the GSM network with new types of connections to external packet data networks, they are essentially extensions of GSM. Moving to a GSM/UMTS core network will likewise be a further extension of this network. EDGE provides GSM operators — whether or not they get a new 3G license — with a commercially attractive solution for developing the market for wide band multimedia services. Using familiar interfaces such as the Internet, volume-based charging and a progressive increase in available user data rates will remove some of the barriers to large-scale take-up of wireless data services. The move to 3G services will be a staged evolution from today’s GSM data services using GPRS and EDGE. Table 1.1 provides a comparison of GSM data services. Table 1.1 Comparison of GSM data services.

Service type

Data unit

Max. sustained user data rate

Technology

Resources used

Short Message Service (SMS)

Single 140 octet packet

9 bps

simplex circuit

SDCCH or SACCH

CircuitSwitched Data

30 octet frames

9.6 kbps

duplex circuits

TCH

HSCSD

192 octet frames

115 kbps

duplex circuits

1-8 TCH

GPRS

1600 octet frames

115 kbps

virtual circuit packet switching

PDCH (1-8 TCH)

EDGE (2.5G)

variable

384 kbps

virtual circuit/ packet switching

1-8 TCH

Note: SDCCH: Stand-alone Dedicated Control Channel; SACCH: Slow Associated Control Channel; TCH: Traffic Channel; PDCH: Packet Data Channel (all refer to GSM logical channels)

1.4

Road Map for Higher Data Rate Capability in 3G

11

The use of CDMA technology began in the United States with the development of the IS-95 standard in 1990. The IS-95 standard has evolved since to provide better voice services and applications to other frequency bands (IS-95A), and to provide higher data rates (up to 115.2 kbps) for data services (IS-95B). To further improve the voice service capability and provide even higher data rates for packet and circuit switched data services, the industry developed the cdma2000 standard in 2000. As the concept of wireless Internet gradually turns into reality, the need for an efficient high-speed data system arises. A CDMA high data rate (HDR) system was developed by Qualcomm. The CDMA-HDR (now called 3G 1X EV-DO, [3G 1X Enhanced Version Data Only]) system design improves the system throughput by using fast channel estimation feedback, dual receiver antenna diversity, and scheduling algorithms that take advantage of multi-user diversity. 3G 1X EV-DO has significant improvements in the downlink structure of cdma2000 including adaptive modulation of up to 8-PSK and 16-quadrature amplitude modulation (QAM), automatic repeat request (ARQ) algorithms and turbo coding. With these enhancements, 3G 1X EV-DO can transmit data in burst rates as high as 2.4 Mbps with 0.5 to 1 Mbps realistic downlink rates for individual users. The uplink design is similar to that in cdma2000. Recently, the 3G 1X EV-Data and Voice (DV) standard was finalized by the TIA and commercial equipment is currently being developed for its deployment. 3G 1X EV-DV can transmit both voice and data traffic on the same carrier with peak data throughput for the downlink being confirmed at 3.09 Mbps. As an alternative, Time Division-Synchronous CDMA (TD-SCDMA) has been developed by Siemens and the Chinese government. TD-SCDMA uses adaptive modulation of up to quadrature phase shift keying (QPSK) and 8-PSK, as well as turbo coding to obtain downlink data throughput of up to 2 Mbps. TD-SCDMA uses a 1.6 MHz time-division duplex (TDD) carrier whereas cdma2000 uses a 2  1.25 MHz frequency-division duplex (FDD) carrier (2.5 MHz total). TDD allows TD-SCDMA to use the least amount of spectrum of any 3G technologies. Table 1.2 lists the maximum data rates per user that can be achieved by various systems under ideal conditions. When the number of users increases, and if all the users share the same carrier, the data rate per user will decrease. One of the objectives of 3G systems is to provide access “anywhere, any time.” However, cellular networks can only cover a limited area due to high infrastructure costs. For this reason, satellite systems will form an integral part of the 3G networks. Satellite will provide extended wireless coverage to remote areas and to aeronautical and maritime mobiles. The level of integration of the satellite system with the terrestrial cellular networks is under investigation. A fully integrated solution will require mobiles to be dual mode terminals to allow communications with orbiting satellite and terrestrial cellular networks. Low Earth orbit (LEO) satellites are the most likely candidates for providing worldwide coverage. Currently several LEO satellite systems are being deployed to provide global telecommunications.

Table 1.2 Network technology migration paths and their associated data rates.

Technology

Carrier width (MHz)

Duplexing

Multiplexing

Modulation

Max. data rates

End-user data rates

Analog

9.6 kbps

4.8–9.6 kbps

CDPD (1G)

19.2 kbps

about 16 kbps

GSM Circuit Switched Data (2G)

0.20

FDD

TDMA

GMSK

9.6–14.4 kbps

about 12 kbps

GPRS

0.20

FDD

TDMA

GMSK

up to 115.2 kbps (8 channels)

10–56 kbps

EDGE (2.5G)

0.20

FDD

TDMA

GMSK, 8-PSK

384 kbps

about 144 kbps

WCDMA (3G)

5.00

FDD

CDMA

QPSK

2 Mbps (stationary); 384 kbps (mobile)

50 kbps uplink ; 150–200 kbps downlink

IS-54/IS-136 TDMA Circuit Switched Data (2G)

0.03

FDD

TDMA

QPSK

14.4 kbps

about 10 kbps

EDGE (2.5G) for North American TDMA system

0.20

FDD

TDMA

GMSK, 8-PSK

64 kbps uplink (initial roll out)

Initial roll out in 2001/2002: 45–50 kbps uplink; 80–90 kbps downlink

384 kbps

2003: 45–50 kbps uplink; 150–200 kbps downlink

cdma2000 (3G) 1X

1.25

FDD

CDMA

QPSK

153 kbps

90–130 kbps (depending on the number of users and distance from BS)

3G 1X EV-DO (data only)

1.25

FDD

TD-CDMA

QPSK, 8-PSK, 16-QAM

2.4 Mbps

700 kbps

3G 1X EV-DV (data and voice)

1.25

FDD

TD-CDMA

QPSK, 8-PSK, 16-QAM

3–5 Mbps

>1 Mbps

TD-SCDMA

1.60

TDD

TD-CDMA

QPSK, 8-PSK

2 Mbps

1.333 Mbps

Note: FDD  Frequency Division Duplex; TDD  Time Division Duplex; PSK  Phase Shift Keying; QPSK  Quadrature Phase Shift Keying; GMSK  Gaussian Minimum Shift Keying; QAM  Quadrature Amplitude Modulation

14

1.5

1

An Overview of Wireless Systems

Wireless 4G Systems

4G networks (see Chapter 23) can be defined as wireless ad hoc peer-to-peer networking with high usability and global roaming, distributed computing, personalization, and multimedia support. 4G networks will use distributed architecture and end-to-end Internet Protocol (IP). Every device will be both a transceiver and a router for other devices in the network eliminating the spoke-and-hub architecture weakness of 3G cellular systems. Network coverage/capacity will dynamically change to accommodate changing user patterns. Users will automatically move away from congested routes to allow the network to dynamically and automatically self-balance. Recently, several wireless broadband technologies [20] have emerged to achieve high data rates and quality of service. Navini Networks developed a wireless broadband system based on TD-SCDMA. The system, named Ripwave, uses beamforming to allow multiple subscribers in different parts of a sector to simultaneously use the majority of the spectrum bandwidth. Beamforming allows the spectrum to be effectively reused in dense environments without having to use excessive sectors. The Ripwave system varies between QPSK, 16 and 64-QAM, which allows the system to burst up to 9.6 Mbps using a single 1.6 MHz TDD carrier. Due to TDD and 64-QAM modulation the Ripwave system is extremely spectrally efficient. Currently Ripwave is being tried by several telecom operators in the United States. The Ripwave Customer Premise Equipment is about the size of a cable modem and has a self-contained antenna. Recently, PC cards for laptops have become available allowing greater portability for the user. Flarion Technologies is promoting their proprietary Flash-orthogonal frequency-division multiple (OFDM) as a high-speed wireless broadband solution. Flash-OFDM uses frequency hopping spread spectrum (FHSS) to limit interference and allows a reuse pattern of one in an OFDM access environment. Flarion’s Flash-OFDM system uses 1.25 MHz FDD carriers with QPSK and 16-QAM modulation. Peak speeds can burst up to 3.2 Mbps with sustained rates leveling off at 1.6 Mbps on the downlink. Flarion has not implemented an antenna enhancement technology that may further improve data rates. BeamReach is a wireless broadband technology based on OFDM and beamforming. It uses TDD duplexed 1.25 MHz paired carriers. Spread spectrum is used to reduce interference over the 2.5 MHz carriers allowing a frequency reuse of one. Individual users can expect downlink rates of 1.5, 1.2, and 0.8 Mbps using 32-QAM, 16-QAM, and 8-PSK modulation respectively. The aggregate network bandwidth is claimed to be 88 Mbps in 10 MHz of spectrum or 220 Mbps in 24 MHz of spectrum, which equates to a high spectral efficiency of 9 bps/Hz. It should be noted that the system uses either 4 or 6 sectors and these claims are based on those sectoring schemes. For any technology with a reuse number of 1 to achieve 9 bps/Hz per cell with 4 or 6 sectors, the efficiency in each sector would need to be a reasonable 2.3 or 1.5 bps/Hz, respectively.

1.6

Future Wireless Networks

15

IPWireless is the broadband technology based upon UMTS. It uses either 5 or 10 MHz TDD carriers and QPSK modulation. The theoretical peak transmission speeds for a 10 MHz deployment is 6 Mbps downlink and 3 Mbps uplink. The IPWireless system only uses QPSK modulation and no advanced antenna technologies. With the inclusion of advanced antenna technologies and the development of High Speed Downlink Packet Access (HSDPA), IPWireless has significant potential. SOMA networks has also developed a wireless broadband technology based on UMTS. Like UMTS, SOMA’s technology uses 5 MHz FHSS carriers. Peak throughput is claimed to be as high as 12 Mbps, making SOMA one of the faster wireless broadband technologies. Table 1.3 compares the wireless broadband technologies and their lowest order modulation data throughput.

1.6

Future Wireless Networks

As mobile networks evolve to offer both circuit and packet-switched services, users will be connected permanently (always on) via their personal terminal of choice to the network. With the development of intelligence in core network (CN), both voice and broadband multimedia traffic will be directed to their intended destination with reduced latency and delays. Transmission speeds will be increased and there will be far more efficient use of network bandwidth and resources. As the number of IP-based mobile applications grows, 3G systems will offer the most flexible access technology because it allows for mobile, office, and residential use in a wide range of public and nonpublic networks. The 3G systems will support both IP and non-IP traffic in a variety of modes, including packet, circuit-switched, and virtual circuit, and will thus benefit directly from the development and extension of IP standards for mobile communications. New developments will allow parameters like quality of service (QoS), data rate, and bit error rate (BER) — vital for mobile operation — to be set by the operator and/or service provider. Wireless systems beyond 3G (e.g., 4G) will consist of a layered combination of different access technologies: • Cellular systems (e.g., existing 2G and 3G systems for wide area mobility) • WLANs for dedicated indoor applications (such as IEEE 802.11a, 802.11b,

802.11g) • Worldwide interoperability for microwave access (WiMAX) (IEEE 802.16)

for metropolitan areas • WPANs for short-range and low mobility applications around a room in the

office or at home (such as Bluetooth) These access systems will be connected via a common IP-based core network that will also handle interworking between the different systems. The core network will enable inter- and intra-access handoff.

Table 1.3 Non-line of sight (LOS) wireless broadband technologies.

Developer

Technology

Multiplexing

Duplexing

Carrier (MHz)

Modulation

System DL Peak (Mbps)

System DL LOM (Mbps)

Avg. DL efficiency (bps/Hz)

Navini

TD-SCDMA

TD-CDMA

TDD

1.6

QPSK to 64-QAM

8.0

2.0

2.5

IPwireless

TD-WCDMA

TD-CDMA

TDD

5.0

QPSK

3.0

3.0

1.2

Flarion

Flash-OFDM

FHSS OFDM

FDD

1.25

QPSK, 16-QAM

3.2

1.6

1.28

SOMA

UMTS

CDMA

FDD

5.0

QPSK, 16-QAM

12.0

6.0

1.2

Beam Reach

AB-OFDM

DSSS-OFDM

TDD

2.5

8-PSK, 16-, 32-QAM

3.33

2.0

1.6

TDD carriers need one carrier for Tx and Rx, thus efficiency is doubled; BeamReach system throughput includes 6 sectors, thus it was divided by six; LOM — sustained system throughput estimated using lowest order modulation

1.7

Standardization Activities for Cellular Systems

17

The peak data rates of 3G systems are around 10 times more than 2G/2.5G systems. The fourth-generation systems may be expected to provide a data rate 10 times higher than 3G systems. User data rates of 2 Mbps for vehicular and 20 Mbps for indoor applications are expected. The fourth-generation systems will also meet the requirements of next generation Internet through compliance with IPv6, Mobile IP, QoS control, and so on.

1.7

Standardization Activities for Cellular Systems

The standardization activities for PCS in North America were carried out by the joint technical committee (JTC) on wireless access, consisting of appropriate groups from within the T1 committee, a unit of the Alliance for Telecommunications Industry Solutions (ATIS), and the engineering committee TR46, a unit of the Telecommunications Industry Association (TIA). The JTC was formed in November of 1992, and its first assignment was to develop a set of criteria for PCS air interfaces. The JTC established seven technical ad hoc groups (TAGs) in March of 1994, one for each of the selected air interface proposals. The TAGs drafted the specifications document for the respective air interface technologies and conducted validation and verification to ensure consistency with the criteria established by the JTC. This was followed by balloting on each of the standards. After the balloting process, four of the proposed standards were adopted as ANSI standards: IS-136-based PCS, IS-95-based PCS, GSM 1900 (a derivative of GSM) and Personal Access Communication System (PACS). Two of the proposed standards — composite CDMA/TDMA and Oki’s wide band CDMA — were adopted as trial use standards by ATIS and interim standards by TIA. The Personal Wireless Telecommunications-Enhanced (PWT-E) standard was moved from JTC to TR 46.1 which, after a ballot process, was adopted in March of 1996. Table 1.4 provides a comparison of seven technologies using a set of parameters which include access methods, duplex methods, bandwidth per channel, throughput per channel, maximum power output per subscriber unit, vocoder, and minimum and maximum cell ranges. The 3G systems were standardized under the umbrella of the International Telecommunications Union (ITU). The main proposals to the ITU International Mobile Telecommunications-2000 (IMT-2000) are: ETSI UMTS Terrestrial Radio Access (UTRA), ARIB WCDMA, TIA cdma2000, and TD-SCDMA. These thirdgeneration systems will provide the necessary quality for multimedia communications. The IMT-2000 requirements are: 384 kbps for full area coverage (144 kbps for fast-moving vehicles between 120 km per hour and 500 km per hour), and 2 Mbps for local coverage. It is, therefore, important to use packet-switched data service to dynamically allocate and release resources based on the current needs of each user. The ETSI agreed on a WCDMA solution using FDD mode. In Japan, a WCDMA solution was adopted for both TDD and FDD modes. In Korea, two

Table 1.4 Technical characteristics of North American PCS standards. TAG-1

TAG-2

TAG-3

TAG-4

TAG-5

TAG-6

TAG-7

Standard

Composite CDMA/ TDMA

IS-95 based PCS

PACS

IS-136 based PCS

GSM 1900

PWT-E

Oki’s Wide band CDMA

Access

CDMA/ TDMA/ FDMA

CDMA

TDMA

TDMA

TDMA

TDMA

CDMA

Duplex Method

TDD

FDD

FDD

FDD

FDD

TDD

FDD

Frequency Reuse

3

1

16  1

73

7  1, 3  3

Portable selected

1

Bandwidth/channel

2.5/5 MHz

1.25 MHz

300 kHz

30 kHz

200 kHz

1 kHz

5, 10, 15 MHz

Throughput/ channel (kb/s)

8

8.55/13.3

32

13

32

32

200 mW

200 mW

600 mW

500 mW

500 mW

VCELP/ ACELP

ADPCM

ADPCM

8

Max. Power/subscriber unit

600 mW

0.5–2.0 W

Vocoder

PHS HCA

CELP

ADPCM

Max. Cell Range (km)

10.0

50.0

1.6

20.0

35.0

0.15

5.0

Min. Cell Range (km)

0.1

0.05

0–1

0.5

0.5

0.01

0.05

RPE-LTE ACELP

1.8

Summary

19

different types of CDMA solutions were proposed — one based on WCDMA similar to that of Europe and Japan and the other similar to the cdma2000 proposed in North America. A number of groups working on similar technologies pooled their resources. This led to the creation of two standards groups — the third-generation partnership project (3GPP) and 3GPP2. 3GPP works on UMTS, which is based on WCDMA and 3GPP2 works on cdma2000. The IEEE standard committee 802.11 is responsible for the WLAN standard. There are two IEEE standards committees that are involved in certification of wireless broadband technologies. The 802.16x committee focuses on the wireless metropolitan area network (WMAN) using CDMA and OFDM technologies. 802.16x allows for portability and data rates above 1 Mbps. The newly formed IEEE 802.20 committee, evolved from the 802.16e committee, focuses on mobile wide area network (MWAN). Several key performance requirements include megabit data rates and mobile handoff at speeds of up to 250 km per hour. The Worldwide Interoperability for Microwave Access (WiMAX) Forum is a nonprofit organization formed by equipment and component suppliers to promote the adoption of IEEE 802.16-compliant equipment by operators of broadband wireless access systems. The organization is working to facilitate the deployment of broadband wireless networks based on IEEE 802.16 standards by helping to ensure the compatibility and interoperability of broadband wireless access equipment.

1.8

Summary

In this chapter, we presented the scope of wireless networks and gave an overview. We briefly discussed 1G, 2G/2.5G, and 3G cellular systems. The advantage of wireless data networking is apparent. Wireless data network users are not confined to the locations of “wired” data jacks, and enjoy connectivity that is less restrictive and therefore well suited to meet the needs of today’s mobile users. Wireless network deployment in three service classifications — wireless personal access network (WPAN), wireless local area network (WLAN), and wireless wide area network (WWAN) — was discussed. Today, the core technology behind the wireless service in each of these service classifications is unique and, more important, not an inherently integrated seamless networking strategy. As an example, a user of a personal digital assistant (PDA), such as a PALM (XXX) connecting to the Internet via a WWAN service provider will not be able to directly connect to a WLAN service. Simply stated, they are different services, with different hardware requirements, and have fundamentally different service limitations. In the future, wireless networks have to evolve to provide interoperability of WPAN, WLAN, and WWAN systems.

20

1

An Overview of Wireless Systems

Problems 1.1 Name the wireless access techniques used in 1G, 2G, and 3G wireless systems.

1.2 What are the three classes of wireless data networking? 1.3 Define the roles of WPAN technology in wireless data networking. 1.4 List the main features of 3G systems. 1.5 What is the role of GPRS in enhancing 2G GSM systems? 1.6 Show how CDMA IS-95 systems are moving to provide 3G services. 1.7 Show how 2G GSM systems are moving to achieve 3G services. 1.8 What are the data rate requirements for 3G systems? 1.9 Define IPWireless technology. 1.10 What are the goals of 4G systems?

References 1. Balston, D. M., and Macario, R. C. V. Cellular Radio Systems. Norwood, MA: Artech House, 1993. 2. Balston, D. M. The Pan-European Cellular Technology. IEE Conference Publication, 1988. 3. Cai, J., and Goodman, D. J. General packet radio service in GSM. IEEE Communications Magazine, vol. 35, no. 10, October 1997, pp. 121–131. 4. Crow, B. P., Widjaja, I., Kim, L. G., and Sakai, P. T. IEEE 802.11 Wireless Local Area Networks. IEEE Communications Magazine, September 1998, pp. 116–126. 5. Dasilva, J. S., Ikonomou, D., and Erben, H. European R&D programs on third-generation mobile communications systems. IEEE Personal Communications, vol. 4, no. 1, February 1997, pp. 46–52. 6. Dinan, E. H., and Jabbari, B. Spreading codes for direct sequence CDMA and wide band CDMA cellular networks. IEEE Communications Magazine, September 1998. 7. The European path towards UMTS. IEEE Personal Communications, Special Issue. February 1995. 8. Garg, V. K., and Wilkes, J. E. Wireless and Personal Communications Systems. Upper Saddle River, NJ: Prentice Hall, 1996. 9. Garg, V. K., and Wilkes, J. E. Principles and Applications of GSM. Upper Saddle River, NJ: Prentice Hall, 1998. 10. Garg, V. K. CDMA IS-95 and cdma2000. Upper Saddle River, NJ: Prentice Hall, 2000. 11. 3GPPweb Third generation partnership project website: http://www.3gpp.org. 12. 3GPP2web Third generation partnership project-2 website: http://www.3gpp2.org.

References

21

13. Marley, N. GSM and PCN Systems and Equipment. JRC Conference, Harrogate, 1991. 14. Mouly, M., and Pautet, M.-B. The GSM System for Mobile Communications. Palaiseau, France, 1992. 15. Nakajma, N. Future mobile communications systems in Japan. Wireless Personal Communications, vol. 17, no. 2–3, June 2001, pp. 209–223. 16. Nikula, E., Toshala, A., Dahlman, E., Girard, L., and Klein, A. FRAMES multiple access for UMTS and IMT-2000. IEEE Personal Communications Magazine, April 1998, pp. 16–24. 17. Negus, K., Stephens, A., and Lansford, J. Home RF: Wireless networking for the connected home. IEEE Personal Communications, February 2000, pp. 20–27. 18. Pirhonen, R., Rautava, T., and Pentinen, J. TDMA convergence for packet data services. IEEE Personal Communications Magazine, vol. 6, no. 3, June 1999, pp. 68–73. 19. Rapeli, J. UMTS: Targets, system concepts, and standardization in a global framework. IEEE Personal Communications, February 1995. 20. Shafi, M., Ogose, S., and Hattori T. (editors). Wireless Communication in the 21st Century. Wiley-Interscience, 2002. 21. Young, W. R. Advanced mobile phone services — Introduction, background and objectives. Bell Systems Technical Journal, vol. 58, 1979, pp. 1–14.

This page intentionally left blank

CHAPTER 2 Teletraffic Engineering 2.1

Introduction

There are many telephone customers, much larger than the number of available trunks, but not every customer makes or receives a telephone call at the same time. The number of trunks connecting the Mobile Switching Center (MSC) A with another MSC B are the number of voice pairs used in the connection. One of the most important steps in telecommunication engineering is to determine the number of trunks required on a route or a connection between MSCs. To dimension a route correctly, we must have some idea of its usage, that is, how many subscribers are expected to talk at one time over the route. The usage of a transmission route or an MSC brings us into the realm of teletraffic engineering. The call capacity of an MSC is expressed as the maximum number of originating plus incoming (O  I) calls that can be handled during a high-traffic hour while meeting the dial-tone delay requirements. The call capacity of an MSC depends on the subscriber call mix, feature mix, and equipment configuration. Typically, processor capacity limits the MSC call capacity under specific conditions, but the switching networking, peripheral equipment, trunk terminations, or directory numbers can limit MSC call capacity as well. The call volume offered to an MSC depends on geographical area, class of service mix, and time of day. An MSC is required to process calls while serving a representative supplementary feature (such as call waiting, call forwarding, etc.). The call capacity of an MSC is needed to plan, engineer, and administer its use. In this chapter we provide definitions of the terms used in teletraffic engineering. We include several numerical examples for calculating traffic. In the latter part of the chapter we focus on the Erlang and Poisson blocking formulas used to calculate the grade of service (GoS) and show their applications in teletraffic engineering. The readers interested in derivation of the blocking formulas should refer to queuing books [2,4,5]. Typical call mixes for systems in different environments are also given. The material in this chapter applies to both wireline and wireless systems.

2.2

Service Level

Service level for telecommunication traffic can be divided into two main areas: the delay in receiving a dial tone (for a mobile system, radio signaling delays 23

24

2

Teletraffic Engineering

contribute to dial-tone delay) and the probability of service denial. Dial-tone delay is the maximum amount of time a subscriber must wait to hear a dialtone after removing the handset from the hook. Dial-tone delay has the following characteristics: • A large number of users compete for a small number of servers (dial-tone

connections, dial-tone generators). • An assumption that the user will wait until a server is available. Service denial, or the probability that the service trunk will not be available, is similar to dial-tone delay, with several additional characteristics: • A large number of users compete for a small number of trunks. • An assumption that no delay will be encountered. The user is either given

access to a trunk or is advised by a busy signal or a recording that none are available. • The user may frequently reinstate the call attempt after receiving a busy signal. For both cases, the basic measure of performance is either the probability that service delay will exceed some specified value or the probability that the call will be denied or blocked (the blocking probability). In a system that drops calls when serving trunks are not available (a loss system), the blocking probability represents the performance measure.

2.3

Traffic Usage

A traffic path is a communication channel, time slot, frequency band, line, trunk, switch, or circuit over which individual communications take place in sequence. Traffic usage is defined by two parameters, calling rate and call holding time [6]. Calling rate, or the number of times a route or traffic path is used per unit time; more properly defined, the call intensity (i.e., calls per hour) per traffic path during busy hour. Call holding time, or the average duration of occupancy of a traffic path by a call. The carried traffic is the volume of traffic actually carried by a switch, and offered traffic is the volume of traffic offered to a switch. The offered load is the sum of the carried load and overflow (traffic that cannot be handled by the switch). Offered load  carried load  overflow

(2.1)

2.4

Traffic Measurement Units

25

No. of Calls (K)

110

10 0

6

9

12

15

18

21

24

Time of Day (Hour) Figure 2.1

Example of voice traffic variation hour by hour.

Figure 2.1 shows a typical hour-by-hour voice traffic variation for an MSC. We notice that the busiest period — the busy hour (BH) is between 10 A.M. and 11 A.M. We define the busy hour as the span of time (not necessarily a clock hour) that has the highest average traffic load for the business day throughout the busy season. The peak hour is defined as the clock hour with highest traffic load for a single day. Since traffic also varies from month to month, we define the average busy season (ABS) as the three months (not necessarily consecutive) with the highest average BH traffic load per access line. Telephone systems are not engineered for maximum peak loads, but for some typical BH load. The blocking probability is defined as the average ratio of blocked calls to total calls and is referred to as the GoS. Example 2.1 If there are 380 seizures (lines connected for service) and 10 blocked calls (lost calls) during the BH, what is the GoS? Solution Blocking Probability  (Number of lost calls)/(Total number of offered calls) 10 380  10

1 39

GoS    

2.4

Traffic Measurement Units

Traffic is measured in either Erlangs, percentage of occupancy, centrum (100) call seconds (CCS), or peg count [1].

26

2

Teletraffic Engineering

Traffic intensity is the average number of calls simultaneously in progress during a particular period of time. It is measured either in units of Erlangs or CCS. An average of one call in progress during an hour represents a traffic intensity of 1 Erlang; thus 1 Erlang equals 1  3600 call seconds (36 CCS). The Erlang is a dimensionless number. Traffic intensity can be obtained as: (the sum of circuit holding time) (the duration of monitoring period)

Traffic Intensity  

(2.2)

Nc



ti



Nct I      nct i1

T

T

(2.3)

where: I  traffic intensity T  duration of monitoring period ti  the holding time of the ith individual call Nc  the total number of calls in monitoring period t  average holding time nc  number of calls per unit time Percentage of occupancy. Percentage of time a server is busy. Peg count. The number of attempts to use a piece of equipment. The usage (U), peg count (PC) per time period, overflow (O) per period, and average holding time t are related as:  U  (PC  O) p t

(2.4)

Example 2.2 In a switching office an equipment component with an average holding time of 5 seconds has a peg count of 450 for a one-hour period. Assuming that there was no overflow (i.e., the system handled all calls), how much usage in call-seconds, CCS, and Erlangs has accumulated on the piece of the equipment? Solution 5 U  (450  0)    0.625 Erlangs 3600

0.625 Erlangs  36 CCS/Erlangs  22.5 CCS  2250 call-seconds

2.4

Traffic Measurement Units

27

Example 2.3 If the carried load for a component is 3000 CCS at 5% blocking, what is the offered load? Solution 3000 Offered load    3158 CCS (1  0.05)

Overflow  (Offered load)  (Carried load)  3158  3000  158 CCS

Example 2.4 In a wireless network each subscriber generates two calls per hour on the average and a typical call holding time is 120 seconds. What is the traffic intensity? Solution 2  120 I  0.0667 Erlangs  2.4 CCS 3600

Example 2.5 In order to determine voice traffic on a line, we collected the following data during a period of 90 minutes (refer to Table 2.1). Calculate the traffic intensity in Erlangs and CCS.

Table 2.1 Traffic data used to estimate traffic intensity. Call no.

Duration of call (s)

1

60

2

74

3

80

4

90

5

92

6

70

7

96

8

48

9

64

10

126

28

2

Teletraffic Engineering

Solution 10  6.667 calls/hour Call arrival rate   1.5

Average call holding time: (60  74  80  90  92  70  96  48  64  126) t    80 seconds per call 10

6.667  80 I  0.148 Erlangs  5.33 CCS 3600

Example 2.6 We record data in Table 2.2 by observing the activity of a single customer line during an eight-hour period from 9:00 A.M. to 5:00 P.M. Find the traffic intensity during the eight-hour period, and during busy hour (which occurs between 4:00 P.M. and 5:00 P.M). Solution 11  1.375 calls/hour Call arrival rate   8

Total call minutes  3  10  7  10  5  5  1  5  15  34  5  100 minutes Table 2.2 Traffic on customer line between 9:00 A.M. and 5:00 P.M. Call no.

Call started

Call ended

Call duration (min.)

1

9:15

9:18

3.0

2

9:31

9:41

10.0

3

10:17

10:24

7.0

4

10:24

10:34

10.0

5

10:37

10:42

5.0

6

10:55

11:00

5.0

7

12:01

12:02

1.0

8

2:09

2:14

5.0

9

3:15

3:30

15.0

10

4:01

4:35

34.0

11

4:38

4:43

5.0

2.4

Traffic Measurement Units

29

If statistical stationary condition is assumed, then the calling rate and average holding time could be converted to a one-hour period, with the same result holding for the offered traffic load. That is: 100 1  0.1515 hours/call t    11

60

The traffic intensity is: I  1.375  0.1515  0.208 Erlangs  7.5 CCS

The BH is between 4:00 P.M. and 5:00 P.M. Since there are only two calls between this period, we can write: Call arrival rate  2 calls/hour The average call holding time during BH: (34  5) t    19.5 minutes/call  0.325 hours/call 2

The traffic load during BH is: I  2  0.325  0.65 Erlangs  23.4 CCS

Example 2.7 The average mobile user has 500 minutes of use per month; 90% of traffic occurs during work days (i.e., only 10% of traffic occurs on weekends). There are 20 work days per month. Assuming that in a given day, 10% of traffic occurs during the BH, determine the traffic per subscriber per BH in Erlangs. Solution Average busy hour usage per subscriber  (minutes of use/month)  (fraction during work day)  (percentage in busy hour)/(work days/month) Average busy hour usage per subscriber  500  0.9  0.1/20  2.25 minutes of use per subscriber per busy hour. 2.25 Traffic per subscriber    0.0375 Erlangs 60

Example 2.8 If the mean holding time in Example 2.7 is 100 seconds, find the average number of busy hour call attempts (BHCAs).

30

2

Teletraffic Engineering

Solution BHCAs  (traffic in Erlangs)  3600/(mean holding time in seconds) 0.0375  3600    1.35 100

2.5

Call Capacity

Call capacity is defined with respect to a view of the mobile switching center (MSC). In general there are two approaches to view the system — global view and component view [7]. Global view. The entire MSC is considered to be a single unit. Each request to the MSC for service is counted as an attempt. This approach is applicable to central processors involved in call processing. In the global view, the call volume of interest is expressed as the sum of the originating and incoming (O  I) calls. 1. Originating call (O) includes the following: • Partial dial calls — calls with partial time-outs and abandons • Intra-office calls — all calls that originate from and terminate to the same

switch • Outgoing calls — all calls that originate from a line on the switch but terminate to a line on a different switch 2. Incoming call (I) includes the following: • Incoming-terminating calls — all calls that terminate to a line on the

switch but originate from a different switch • Tandem calls — trunk-to-trunk calls that are routed through the switch • Direct inward dialing (DID) — calls to a PBX system Component view. The component of interest is considered a subsystem. Each request to the component for service is counted as an attempt. This view applies to principal processors involved in call processing. In the component view, the call volume of interest is expressed as the sum of the originating  terminating (O  T) half-calls 1. Originating half-calls. One originating half-call is for each originating call, because two peripheral equipment connections are required for a completed call. If the component serves both lines and trunks, incoming and outgoing half-calls are added to the total half-call volume. 2. Terminating half-calls. One terminating half-call is for each incomingterminating call and each interoffice call.

2.5

Call Capacity

Call Attempts

31

Originating

Outgoing

Intra office

Tandem

Terminating

Figure 2.2

Incoming

Call types in a telephone switch.

It should be noted that false starts and permanent signals (ineffective attempts) are not counted as calls in either case. However, partial dial attempts, attempts receiving busy treatment, and attempts not answered are all considered calls. The primary determinant of a stored program control system’s call handling capacity is the maximum number of calls per hour that the processor can handle while still meeting service criteria. Central processors have high day busy hour (O  I) call capacities, whereas peripheral processors have high day busy hour (O  T) call capacities. We define the following call types (see Figure 2.2): • Originating call (O). A call placed by a subscriber of the office • Terminating call (T). A call received by a subscriber of the office • Outgoing call. A call going out of the office • Incoming call (I). A call coming into the office from another MSC • Intra-office call. A call that originates from and terminates to the same MSC • Tandem calls. Calls that come in on a trunk from another MSC and go out

over a trunk to a different MSC • (O  T). The measure of traffic load on the line side of the MSC both from within the office and from other offices • (O  I). The measure of incoming trunk-circuit traffic load and traffic load on the MSC Bell Communication Research (Bellcore) [1] suggests the use of standard call mixes given in Tables 2.3 and 2.4 for calculating call capacity of central processors (in terms of (O  I) calls) and peripheral processors (in terms of (O  T) half-calls), respectively. These call mixes do not represent the best and worst case scenarios for call capacity, but they do represent the average or typical case for each environment. These call mixes reflect the differences found in the following traffic environments: Metropolitan (metro). Class 5 office in a major metropolitan area with a high proportion of business office traffic. Single system city. Class 4 or 5 office serving a medium-sized town with a number of outlying community dial offices homing into it.

32

2

Teletraffic Engineering

Table 2.3 Standard call mixes for central processors (percentage of traffic). Traffic environments Type

Metro

SSC

Suburban

10

12

18

1

2

2

Abandon

3

4

5

Time-out

1

1

1

Answered

12

24

17

No Answer

2

2

3

Busy

2

4

4

Answered

34

17

28

No Answer

5

3

4

Busy

5

3

5

Answered

27

18

24

No Answer

5

2

5

Busy

4

3

4

Tandem

0

19

0

Ineffective False Start Permanent Signal Partial Dial

Intra office

Outgoing

Incoming-Terminating

Suburban. Class 5 office residing in the suburbs of a major city. Rural. Class 4 office residing in a rural area with a low population density and almost no business traffic. Table 2.5 shows the typical traffic intensity for metro, suburban, and rural environments in the United States. Table 2.6 provides an example of a typical traffic distribution for a mobile application in the U.S. metro environment.

2.6

Definitions of Terms

The following terms are used for mobile systems. Number of calls attempted. The total number of calls attempted (also called number of bids) is the best measure of unconstrained customer demand.

2.5

Definitions of Terms

33

Table 2.4 Standard half-call mixes for peripheral processors controlling subscriber lines (percentage of total traffic). Traffic environments Type

Metro

SSC

Suburban

Ineffective False Start

8

11

14

Permanent Signal

1

2

2

Abandon

2

3

4

Time-out

1

1

1

Answered

40

37

36

No Answer

6

5

6

Busy

8

6

7

Answered

34

38

33

No Answer

6

4

6

Busy

5

6

7

Partial Dial

Originating

Terminating

Table 2.5 Traffic intensity for (O ⴙ I). (O ⴙ I) calls/line

Environment Metro

3.5–4.0

Suburban

2.0–2.5

Rural

1.2–1.5

Table 2.6 Typical traffic distribution in U.S. metro environment for mobile. Application Mobile

Traffic Type

Distribution

Mobile to land

60%

Mobile to mobile

10%

Land to mobile

30%

34

2

Teletraffic Engineering

Number of calls completed. The number of calls completed in a network sense (i.e., calls reaching ringing tone or being answered), when compared with number of calls attempted, provides a measure of the state of network congestion. Grade of Service (GoS). (No. of busy hour call attempts)  (No. of busy hour call completed)/(No. of busy hour call attempts). The number of answered calls is lower than the number of attempted calls, since some calls are bound to encounter either a “subscriber-busy” state or a “ring tone/no reply” condition. The ratio between the number of successful calls (answered) and the total number of call attempts (seizure) is called answer-seizure ratio (ASR), whereas answer-busy ratio (ABR) is the number of successful calls (answered) and the total number of busy calls. These ratios represent the probability that a user will be able to complete a call on any given attempt. ABR  (No. of calls answered)/(No. of busy calls) ASR  (No. of calls answered)/(No. of calls attempted)

ABR and ASR are measured over relatively short periods of time (5 to 15 minutes). Both are good indicators of instantaneous network congestion. Lower value of ABR or ASR indicates higher network congestion. However, higher ABR or ASR value does not mean lower network congestion since calls may remain unanswered due to other reasons. Quality of Service (QoS). QoS has different meanings for different people. For a service provider QoS indicates how satisfied a customer is with the service. As an example, we might find that about half of the time a customer dials a number the call goes awry or the caller does not get a dial tone or is unable to hear what is being said by the calling party. All these have an impact on QoS. QoS is an important factor in many areas of the telecommunications business. Several factors affect QoS. These are: • Transmission quality (level, crosstalk, echo, etc.) • Dial-tone delay, and post dial delay • Grade of service • Fault incidence and service deficiency • Adaptation of the system to the subscribers

When QoS requirements are set, one of the two following approaches can be used: • Design for the maximum permitted impairments in the most unfavorable

condition

2.6

Definitions of Terms

35

• Design for a certain range of impairments occurring as a result of a chance

combination of elements, such as the majority of a subscriber’s opinion being favorable (also known as statistical design methodology) Both approaches have weaknesses. The first may require unnecessary high performance to satisfy those rare unfavorable cases. Some subscribers may be very unhappy with the second approach, which will be considerably more influenced by variability of plant performance. Such variability is observed in areas of signaling and transmission and should be taken into consideration while planning technical areas. The North American Region follows the statistical design approach. In Europe the maximum design impairment method is used, although it is often modified, where, for instance, certain standards will be satisfied for only 95% of the situations. Example 2.9 We consider a wireless network with following data: • Total population: 200,000 • Subscriber penetration: 25% • Average call holding time for mobile-to-land and land-to-mobile subscribers: • • • •

100 seconds Average calls/hour for mobile-to-land and land-to-mobile subscribers: 3 calls/hour Average call holding time for mobile-to-mobile subscribers: 80 seconds Average calls/hour for mobile-to-mobile subscribers: 4 calls/hour Traffic distribution is: Mobile-to-land: 50% Land-to-mobile: 40%

Mobile-to-mobile: 10% Calculate total traffic in Erlangs. If each MSC can handle 1800 Erlangs traffic, how many MSCs are required to handle the total traffic? Solution Traffic generated by a subscriber (mobile-to-land or land-to-mobile): 3  100 a(ML)/(LM)    0.0833 Erlangs 3600

Traffic generated by a subscriber (mobile-to-mobile): 4  80 aMM    0.0889 Erlangs 3600

36

2

Teletraffic Engineering

No. of wireless subscribers: 200,000  0.25  50,000 Total Traffic: 45,000  0.0833  5,000  0.0889  4194.5 Erlangs 4194.5  2.33  3 No. of MSCs required:  1800

2.7

Data Collection

Traffic data is collected periodically to access performance of the mobile switching center. The data is used in overall administration, engineering, and maintenance of an office. Traffic measurement reports can also be provided on the customer’s local exchange access line or trunk group from the serving central office. Reports are available on a one-week basis, consisting of seven consecutive days or a monthly report that contains a minimum of four consecutive weeks of data. The reports disclose minutes, attempts, overflow, etc. The traffic measurement report is intended to assist customers in designing and administering communications or business activities associated with telephone service. The following data is collected on an MSC [8]: Peg Count: One peg count for each of the following categories — call attempt, trunk-group seizure attempt, test made for dial tone speed, call queued Overflow: Overflow for each attempt collected for universal tone decoders, trunk groups, etc. Traffic Usage: Measured for trunks, decoders, etc. Customer Usage: Customer usage  Total usage  Maintenance usage

2.8

Office Engineering Considerations

The following steps are often taken in typical office engineering of a wireline or wireless office. 1. MSCs are engineered and administered based on the traffic load during the average busy hour of the busy season. 2. The busy hour is used for the overall administration, engineering, and maintenance of an office. 3. The component busy hour is used to establish trends, make projections, set capacities, and derive future requirements. 4. Dial-tone speed delay is recorded whenever a test call does not receive a dial tone within 3 seconds. 5. Terminating blockage is recorded whenever a terminating call is unable to complete because of a lack of an available path to the called line. 6. Trunk-group busy hour is the time-consistent hour during which maximum trunk-group load occurs. Trunk-group busy hour data is used to provide an adequate trunk base to meet service requirements.

2.8

Office Engineering Considerations

37

7. Traffic data is collected for one or two weeks by half-hour during all parts of the day that may produce high loads (e.g., 8 A.M. to 11 P.M.). 8. Five days of the week with the heaviest load are determined; this is the business week of the office. 9. The hour (on the clock hour or on the half-hour) with the highest total load for the business week is determined; this is the office busy hour. 10. Traffic data collected for the busy hour for the months likely to be parts of the year that may produce high loads. 11. The three months, not necessarily consecutive, having the highest busy hour load are determined; this is the busy season. 12. The average load for the busy hour for the busy season’s business day is (Average Busy Season per Busy Hour) (ABS/BH) 13. The following approximate relations can be used to estimate the design traffic: (O  T) call: (HD)/(ABS)  1.4  1.5 (O  I) call: (HD)/(ABS)  1.6  1.7 High day (HD) origination attempts per call  1.45

Example 2.10 Calculate the ABS/BH calling rate and CCS for a switch located in a large metropolitan area. The switch carries 100,000 lines. The distribution of the lines on the switch is as follows: • Residential lines: 15,000 • Business lines: 80,000 • PBX, WATS, and Foreign Exchange (FX) lines: 5000

The ABS/BH call rates for residential, business, and high-usage customers are 2, 3, and 10 calls per line, respectively. The average call holding times for these customers are 140, 160, and 200 seconds, respectively. Assuming that the HD/ABS for the switch is equal to 1.5, calculate the design call capacity for the switch and design Erlangs. Solution Percentage of residential lines: 15,000/100,000  0.15  15% Percentage of business lines: 80,000/100,000  0.80  80% Percentage of high usage lines: 5,000/100,000  0.05  5% CCS per residential line: 2  140/100  2.8 CCS per business line: 3  160/100  4.8

38

2

Teletraffic Engineering

CCS per high-usage line: 10  200/100  20 Calling rate: 2  0.15  3  0.8  10  0.05  3.2 calls per line CCS rate  2.8  0.15  4.8  0.8  20  0.05  5.26 CCS per line 5.26 Average call holding time per line for the switch:   100  164 seconds 3.2 ABS/BH calls: 3.2  100,000  320,000 5.26  100,000 36

ABS/BH usage:   14,611 Erlangs Design call capacity based on HD: 1.5  ABS/BH  1.5  320,000  480,000 calls Design Erlangs based on HD: 1.5  14,611  21,917

2.9

Traffic Types

We can classify traffic sources as either infinite or finite. For the infinite traffic sources, the probability of call arrival is constant and does not depend on the occupancy of the system. It also implies an infinite number of call arrivals, each with a small call holding time. When the number of sources offering traffic to a group of trunks or circuits is comparatively small in comparison to the number of circuits, the traffic sources are referred to as finite. We can also conclude that with a finite number of traffic sources the arrival rate is proportional to the number of sources that are not already engaged in sending a call. The probability distributions of traffic can be classified as smooth, rough, and random. Each traffic distribution can be defined by , the variance-to-mean ratio (VMR) given as   2



(2.5)

where: 2  variance  mean  is less than 1 for smooth traffic. For rough traffic,  is greater than 1. When  is equal to 1, the traffic distribution is random. The Poisson distribution function is an example of random traffic in which VMR  1. Rough traffic tends to be peakier than random or smooth traffic. For a given GoS more circuits are required for rough traffic because of greater spread of the traffic distribution curve. Smooth traffic behaves like random traffic that has been filtered. The filter is the telephone exchange. The telephone exchange looking out at its subscribers sees call arrivals as random traffic, assuming that the exchange has not been designed for more traffic (over dimensioned). Smooth traffic is the traffic on the telephone exchange outlets. The filtering or limiting of the peakiness is done by call blockage

2.10

Blocking Formulas

39

during busy hour. The blocked traffic is actually overflowed to an alternative route. Smooth traffic is characterized by the Bernoulli distribution (refer to Equation 2.6). If we assume subscribers make calls independent of each other and that each has a probability, p, of being engaged in a conversation, then if n subscribers are examined, the probability that x subscribers will be engaged in a conversation is given as: n

B(x)  C x px(1  p)n  x;

0 x n

(2.6)

where: Mean  np Variance  np(1  p) C nx means the number of ways that x entities can be taken n at a time. The Poisson probability distribution function can be derived from the Bernoulli distribution. If we assume that the number of subscribers, n, is very large and the calling rate per line, h, is low such that mean (nh  m) remains constant; allowing n to increase to infinity, then x

m em p(x)   x!

(2.7)

where: x  0, 1, 2, . . ., n m  nh (mean) n  number of subscribers h  calling rate per line

2.10

Blocking Formulas

A call is termed lost or blocked when it cannot be completed because all the connecting equipment is busy, even though the line that the caller wishes to reach may be idle. The probability of blocking is an important parameter in teletraffic engineering. In conventional teletraffic engineering three models are used for handling or dispensing lost calls. • Blocked Call Held (BCH) • Blocked Call Cleared (BCC) • Blocked Call Delayed (BCD)

The BCH concept assumes that the user will immediately reattempt the call on receipt of a congestion signal and will continue to redial. The user hopes to seize connection equipment or a trunk as soon as equipment is available. In the BCH concept, lost calls are held or waiting at the calling user’s telephone. The principal traffic formula in North America is based on the BCH concept.

40

2

Teletraffic Engineering

The BCC concept is primarily used in Europe, Asia, and Africa. In this case, the user hangs up and waits for some interval before reattempting the call if the user hears the congestion tone on the first attempt. The Erlang B formula is based on this criterion. The BCD concept assumes that the user is automatically put in a queue and is served when the connection equipment becomes available. The method by which a waiting call is selected from the pool of waiting calls is based on the queue discipline (such as first-come first-served, first-come-last-served etc.). In the queuing system the GoS is defined as the probability of delay. The Erlang C formula is based on this concept. There are several blocking formulas to determine the number of circuits (or trunks) required on a route based on busy hour traffic load. The factors that are used include: call arrivals and holding-time distributions; number of traffic sources; availability; and handling of lost calls. These factors help in determining which formulas to use given a particular set of circumstances. Erlang B loss formula has been widely used outside the United States. Loss implies the probability of blockage at the switch due to either congestion or all trunks being busy. This is expressed as the GoS (GB) or probability of finding N channels busy. In the United States the Poisson formula based on the BCH concept is used to determine number of trunks required on a route. This formula is also called the Molina formula.

2.10.1 Erlang B Formula The Erlang B formula [2] is expressed as GoS or probability of finding N channels busy. The assumptions in the Erlang B formula are: • Traffic originates from an infinite number of traffic sources independently. • Lost calls are cleared assuming a zero holding time. • Number of trunks or service channels is limited. • Full availability exists. • Inter-arrival times of call requests are independent of each other. • The probability of a user occupying a channel (called service time) is based

on an exponential distribution. • Traffic requests are represented by a Poisson distribution implying exponentially distributed call inter-arrival times. N

A ⁄A! GB  B(N, A)   N



i0

(Ai

⁄ i!)

(2.8)

2.10

Blocking Formulas

41

where: N  number of serving channels A  offered load, and B(N, A)  blocking probability

2.10.2 Poisson’s Formula The Poisson formula is used for designing trunks on a route for a given GoS [3,9]. It is used in the United States. The assumptions in Poisson’s formula are: • Traffic originates from an infinite number of independent sources • Traffic density per traffic source is equal • Lost calls are held • A limited number of trunks or service channels exist

 Ai! iN

pb  eA

i

(2.9)

where: pb  probability of blocking A  offered load, and N  number of trunks or service channels

2.10.3 Erlang C Formula The Erlang C [2] formula assumes that a queue is formed to hold all requested calls that cannot be served immediately. Customers who find all N servers busy join a queue and wait as long as necessary to receive service. This means that the blocked customers are delayed. No server remains idle if a customer is waiting. The assumptions in the Erlang C formula are: • Traffic originates from an infinite number of traffic sources independently. • Lost calls are delayed. • Number of trunks or service channels is limited. • The probability of a user occupying a channel (called service time) is based

on an exponential distribution. • Calls are served in the order of arrival. AN ⁄ [N!(1  A ⁄ N)]

C(N, A)  N 1

 Ai!  i0 i

AN N!(1  A ⁄ N)



(2.10)

42

2

Teletraffic Engineering

where: N  number of service channels A  offered load, and C(N, A)  blocking probability The Erlang B formula holds even when the load is greater than number of servers (A N) because, unlike the BCD model in which all calls are eventually served, the BCC model allows calls to be lost when all servers are busy. Therefore, the BCC system never becomes unstable.

2.10.4 Comparison of Erlang B and Poisson’s Formulas A comparison between the Erlang B and Poisson’s blocking formulas shows that Poisson’s formula results in higher blocking than that obtained by the Erlang B formula for a given traffic load. For the Erlang B system, the offered load can be divided into the load lost and the carried traffic load A* (i.e., amount of load serviced by system). A*  A[1  B(N, A)]

(2.11)

where: A  offered load N  number of servers The carried traffic equals that portion of offered traffic load A that is not lost, and AB[N, A] is the lost traffic. In a BCD system, no calls are lost. Thus, the carried load is equal to the offered load. The efficiency, , of a BCC system is defined as the load carried per server (i.e., A*/N).

2.10.5 Binomial Formula Since the binomial formula is also used in teletraffic engineering occasionally, we include the information here. More details can be found in reference [3]. The assumptions used in the binomial formula are: • Traffic originates from a finite number of traffic sources independently. • Traffic density per traffic source is equal. • Lost calls are held in the system in a queue.  D s1 pb   s s 

where: D  expected traffic density pb  blocking probability

s1

D   s N 1    s  D

iN

i

(2.12)

2.11

Summary

43

N  number of channels in the group of channels, and s  number of sources in group sources

Example 2.11 The maximum calls per hour in a mobile cell equals 4000 and the average call holding time is 160 seconds. If the GoS is 2%, find the offered load A. How many service channels are required to handle the load? Solution 4000  160 A  166.67 Erlangs (offered load) 3600

Using the Erlang B table in Appendix A; N  182 channels giving 168.3 Erlangs at 2% blocking.

Example 2.12 How many mobile subscribers can be supported with 50 service channels at 2% GoS? Assume the average call holding time equals 120 seconds and the average busy hour call per subscriber is 1.2 calls per hour. Solution From the Erlang B table in Appendix A, for 50 channels at 2% blocking, the offered load  40.26 Erlangs. The carried load will be: 40.26  (1  0.02)  39.45 Erlangs 1.2  120  0.04 Erlangs Average traffic per user   3600

39.45 No. of users    986 0.04

2.11

Summary

In this chapter we discussed the basic principles of teletraffic engineering that are used to engineer and administer a wireline or wireless switch. We also presented examples to determine the design call capacity of a switching system. We discussed Erlang B and Poisson’s blocking formulas for calculating blocking probabilities or the GoS of the system. We concluded the chapter by presenting the Erlang C and binomial formulas that apply to a queued system. Several numerical examples were given to illustrate the applications of various blocking formulas.

44

2

Teletraffic Engineering

Problems 2.1 Define pegcount, busy hour (BH), busy hour call attempts (BHCAs), and grade of service (GoS).

2.2 State the assumptions of the Erlang B formula. 2.3 Define “blocked call clear” and “blocked call held” systems. 2.4 What are the traffic measurements for a switching system? 2.5 What is the average call holding time in seconds of a group of circuits that has accumulated 0.4445 Erlang of usage in an hour, based on 330 call attempts with 10 calls overflowing (i.e., calls not served)? 2.6 A trunk accumulated 0.75 Erlang of usage while 9 calls were carried in an hour with no overflow. What is the average holding time per call in seconds? 2.7 A switching system is designed to support 80,000 subscribers during the BH. Each subscriber generates an average of 1.8 calls/hour during the BH with an average call holding time of 100 seconds. Calculate the offered traffic load to the switching system. 2.8 If there are 60 radio channels in a cell to handle all the calls and the average call holding time is 120 seconds, how many calls are handled in this cell with a GoS of 2%? 2.9 Consider an urban area in which average mobile subscriber has 600 minutes of use (MoU) per month. Eighty percent of traffic occurs during workdays (i.e., only 20% of traffic occurs on weekends). There are 20 workdays per month. Assuming that in a given day, 10% traffic occurs during busy hour, what is the traffic per subscriber in Erlangs? 2.10 Determine the number of subscribers that can be supported by a cell with 63 radio channels. Assume each subscriber generates an average of 2.9 calls per hour with an average call holding time of 110 seconds. Also determine the traffic generated by each subscriber in CCS. Assume GoS  2%. 2.11 Estimate the number of subscribers that can be supported by a cell with 400 radio channels. Assume each subscriber generates an average of 2.5 calls per hour with a call holding time of 120 seconds. Assume 2% GoS. 2.12 Consider a wireless network with the following data: • Total population: 300,000 • Subscriber penetration: 40% • Average call holding time for mobile-to-land and land-to-mobile calls: 100 seconds

References

45

• Average call holding time for mobile-to-mobile calls: 80 seconds • Average calls per hour for mobile-to-land and land-to-mobile: 3 • Average calls per hour for mobile-to-mobile: 4 • Traffic distribution: Mobile-to-land  50%; Land-to-mobile  40%;

and Mobile-to-mobile  10% a. Calculate the total traffic in Erlangs, and b. If each switch can support 3000 Erlangs, how many switches are required in the network?

References 1. Bellcore. “LATA Switching Systems Generic Requirements — Teletraffic Capacity and Environment,” Technical Reference, TR-TSY-000517, Issue 3, March 1989. 2. Copper, R. B. Introduction to Queuing Theory. New York: North-Holland, 1981. 3. Freeman, R. L. Telecommunication System Engineering. New York: John Wiley & Sons, 1989. 4. Kleinrock, L. Queuing System, Vol. 1. New York: John Wiley & Sons, 1975. 5. Kleinrock, L. Queuing System, Vol. 2. New York: John Wiley & Sons, 1976. 6. Rappaport, T. S. Wireless Communications. Upper Saddle River, NJ: Prentice Hall, 1996. 7. Rapp, Y. “Planning of Junction Network in Multiexchange Areas.” Ericsson Technics 20(1), 1964, pp. 77–130. 8. Rey, R. F. Technical Editor, Engineering and Operations in Bell System. Murray Hill, NJ: AT&T Bell Labs, 1984. 9. Sharma, R. L., et al. Network Systems. New York: Van Nostrand Reinhold Co., 1982.

This page intentionally left blank

CHAPTER 3 Radio Propagation and Propagation Path-Loss Models 3.1

Introduction

Exponential growth of mobile communications has increased interest in many topics in radio propagation. Much effort is now devoted to refine radio propagation path-loss models for urban, suburban, and other environments together with substantiation by field data. Radio propagation in urban areas is quite complex because it often consists of reflected and diffracted waves produced by multipath propagation. Radio propagation in open areas free from obstacles is the simplest to treat, but, in general, propagation over the earth and the water invokes at least one reflected wave. For closed areas such as indoors, tunnels, and underground passages, no established models have been developed as yet, since the environment has a complicated structure. However, when the environmental structure is random, the Rayleigh model used for urban area propagation may be applied. When the propagation path is on line of sight, as in tunnel and underground passages, the environment may be treated either by the Rician model or waveguide theory. Direct wave models may be used for propagation in a corridor. In general, radio wave propagation consists of three main attributes: reflection, diffraction and scattering (see Figure 3.1) [2]. Reflection occurs when radio wave propagating in one medium impinges upon another medium with different electromagnetic properties. The amplitude and phase of the reflected wave are strongly related to the medium’s instrinsic impedance, incident angle, and electric field polarization. Part of the radio wave energy may be absorbed or propagated through the reflecting medium, resulting in a reflected wave that is attenuated. Diffraction is a phenomenon by which propagating radio waves bend or deviate in the neighborhood of obstacles. Diffraction results from the propagation of wavelets into a shadowy region caused by obstructions such as walls, buildings, mountains, and so on. Scattering occurs when a radio signal hits a rough surface or an object having a size much smaller than or on the order of the signal wavelength. This causes the

47

48

3

Radio Propagation and Propagation Path-Loss Models

Building

Building Receiver

Receiver (a) Reflection of the radio wave Transmitter

Transmitter (b) Diffraction of the radio wave Scattering location

Receiver Transmitter

Figure 3.1

(c) Scattering of the radio wave

Reflection, diffraction and scattering of radio wave.

signal energy to spread out in all directions. Scattering can be viewed at the receiver as another radio wave source. Typical scattering objects are furniture, lamp posts, street signs, and foliage. In this chapter, our focus is to characterize the radio channel and identify those parameters which distort the information-carrying signal (i.e., base band signal) as it penetrates the propagation medium. The several empirical models used for calculating path-loss are also discussed.

3.2

Free-Space Attenuation

The most simple wave propagation case is that of a direct wave propagation in free space [5,7]. In this special case of line-of-sight (LOS) propagation there are no obstructions due to the earth’s surface or other obstacles (see Figure 3.2). We consider radiation from an isotropic antenna. This type of antenna is completely omni-directional, radiating uniformly in all directions. While there is no such thing as a purely isotopic antenna in practice, it is a useful theoretical concept.

3.2

Free-Space Attenuation

49

Radio Horizon Transmitting

Receiving

Antenna

Antenna

Earth

Figure 3.2

Line-of-sight propagation.

The received power, Pr, at the receiving antenna (mobile station), located at a distance, d, from the transmitter (base station) is given for free space propagation [10] as:  G G Pr  Pt  b m

 4d 

2

(3.1)

If other losses (not related to propagation) are also present, we can rewrite Equation 3.1 as:





Pr   4d Pt



2

G G L0

G G LpL0

b m b m ·  

(3.2)

where: Pr  received power Pt  transmitted power c c · 2   wave length     c f

c  carrier frequency in rad./sec c  velocity of electromagnetic waves in the free space (3  108 m/s) Gb  gain of the transmitting (base station) antenna Gm  gain of the receiving (mobile) antenna d  antenna separation distance between transmitter and receiver (i.e., base station and mobile station) L0  other losses expressed as a relative attenuation factor





4d 2  free space path loss, often expressed as an attenuation in Lp   

decibels (dB)





4d  20 log  (dB) 

50

3

Radio Propagation and Propagation Path-Loss Models

We can express Lp (dB) in free space as: (Lp)free  32.44  20 log f  20 log d dB

(3.3)

where: f  carrier frequency in MHz d  separation distance in km ( 1 km) It should be noted that the free-space attenuation increases by 6 dB whenever the length of the path is doubled. Similarly, as frequency is doubled, free-space attenuation also increases by 6 dB.

3.3

Attenuation over Reflecting Surface

Free-space propagation is encountered only in rare cases such as satellite-to-satellite paths. In typical terrestrial paths, the signal is partially blocked and attenuated due to urban clutter, trees, and other obstacles. Multipath propagation also occurs due to reflection from the ground. Signals reflected from the ground are fundamentally no different than any other reflected signal. However, they are considered separately because they contribute to the received signal power at virtually all terrestrial receiving locations. The energy radiated from a transmitting antenna may reach the receiving antenna over several possible propagation paths. For many wireless applications in the 50 to 2000 MHz range, two components of the space wave are of primary concern: energy received by means of the direct wave, which travels a direct path from the transmitter to the receiver, and a ground-reflected wave, which arrives at the receiver after being reflected from the surface of the earth. This is commonly referred to as the two-ray model. Other propagation paths such as sky and surface waves are often neglected. We assume that the base station and mobile station antenna heights, hb and hm, are much smaller compared to their separation, d, and the reflecting earth surface is flat (see Figure 3.3 for direct and reflected waves). The received power at the antenna located at a distance, d, from the transmitter, including other losses, L0, is given as [7,10] (see Appendix B for derivation):

 h dh 

b m Pr   2

2

(GbGm)  Pt L0

(3.4)

where: d

b hm » d and d  4h  

Note that under the assumed conditions the received signal level depends only on the transmitted power, antenna heights, and separation distance; there is

3.3

Attenuation over Reflecting Surface

51

no frequency dependence. Furthermore, the total attentuation increases by 12 dB when the separation distance is doubled. The expression for the effects of ground reflections from a flat (or plane)  2

 earth provides results that are approximately correct for   (dd dr)

8

(see Figure 3.3). The results are not valid for 8. When 8, the attentuation factor will be: 

Agr  1  2 2 cos

(3.5)

where:   reflection coefficient of ground (assumed to be 1) Example 3.1 With hb  100 ft, and hm  5 ft, and a frequency of 881.52 MHz (  1.116 ft), calculate signal attenuation at a distance equal to 5000 ft. Assume antenna gains are 8 dB and 0 dB for the base station and mobile station, respectively. What are the free-space and reflected surface attenuations? Assume the earth surface to be flat. Solution 



4h h 

b m 4  100  5 d     1792 ft, d d

1.116

Gb  8 dB  100.8  6.3;

Gm  0 dB  1.0

Transmitting Antenna

Direct path: 13 Reflected path: 12  23

1 

Direct Wave (dd ) 

hb

Receiving Antenna Ground Reflected Wave (dr )

3 2 Ground d

Figure 3.3

Geometry for direct and ground reflected waves.

hm

52

3

Radio Propagation and Propagation Path-Loss Models

Free-space attenuation: Pt Pr



4d   

2

1  · GbGm

4  5000  1.116 

2

1 ·  87 dB 6.3  1

Attenuation on reflecting surface: P Pr

 hdh  2

t   

b m

2

2

 100  5 

2

1  5000 1 ·  ·   86 dB GbGm

6.3  1

Example 3.2 Consider a base station transmitting to a mobile station in free space. The following parameters relate to this communication system: • Distance between base station and mobile station: 8000 m • Transmitter frequency: 1.5 GHz (  0.2 m) • Base station transmitting power, Pt  10 W(10 dBW) • Total system losses: 8 dB • Mobile receiver noise figure Nf  5 dB • Mobile receiver antenna temperature  290 K • Mobile receiver bandwidth Bw  1.25 MHz • Antenna gains are 8 dB and 0 dB for the base station and mobile station,

respectively. • Antenna height at the base station and mobile station are 30 m and 3 m, respectively. Calculate the received signal power at the mobile receiver antenna and signal-to-noise ratio (SNR) of the received signal. Solution 30  3 Free space path loss  20 log   117 dB 2

 8000 

Pr  117  8  10  0 8  107 dBW  77 dBm Te  290(3.162 1)  627 K Pn  1.38  10 23 (627  290)(1.25  106)  1.58  10 14 W  138 dBW Pr

 SNR    107 ( 138)  31 dB Pn

3.4

Effect of Earth’s Curvature

3.4

53

Effect of Earth’s Curvature

We assumed a flat earth in Section 3.3. In reality, the earth is curved, preventing LOS propagation to great distances. Thus, Equations 3.1 and 3.2 are only approximately correct for distances less than the distance d to the radio horizon (see Figure 3.4). The distance d to the radio horizon can be determined for a terrestrial transmitter as follows. We consider a circle representing the earth [12]. From Figure 3.4 we write d2  R2e  (Re  hb)2 d2  2Rehb  h2b

where: Re  radius of earth hb  antenna height Since 2Rehb

» h2b 

d ≈ 2Rehb

(3.6)

As an example with an antenna height of 600 m and earth’s radius of about 6400 km, the distance to the horizon is about 88 km. On average the refractive index of earth’s atmosphere is such that the earth’s radius appears to be about 33% more than the actual radius. Thus, the effective radius of the earth is often assumed to be 8500 km. The curvature of the earth further affects the propagation of the space

Curved Path Distance d (radio horizon) hb

Re

hb

d

Straight Path Distance d

90 Re Fictitious Horizon Actual Horizon

Earth’s Center (radius Re) Figure 3.4

Geometry for a spherical earth.

Earth’s Center (radius 4/3 Re)

54

3

Radio Propagation and Propagation Path-Loss Models

wave since the ground-reflected wave is reflected from a curved surface. Therefore, the energy on a curved surface diverges more than it does from a flat surface and the ground-reflected wave reaching the receiver is weaker than for a flat earth. The divergence factor D that describes this effect is less than unity and is given as: 1 D   



(3.7)

2

1   3



Re(hb  hm)

 2

d hbhm

 0.5  h

hb

hm

m

hb

  



where: hm  mobile antenna height D in Equation 3.7 ranges from unity for a small value of d and approaches zero as d approaches the distance to the radio horizon. It can be combined with the ground reflection coefficient so that the attenuation due to ground reflections becomes 

Agr  1  (D)2 2Dcos( )

(3.8)

where: Agr  attenuation factor due to ground reflections   reflection coefficient of the earth surface D  divergence factor In Equation 3.8 the reflection coefficient, , has been modified to account for the divergence factor, D. The effect of the divergence factor is to reduce the effective reflection coefficient of the earth. The equation for calculating the received power in dBm is given as: Pr  Pt  Gb  Gm  20 log (Afs)  20 log (Agr) dBm

(3.9)

where: Afs  [/(4d)]2, free space attenuation

3.5

Radio Wave Propagation

Radio waves propagate [7] through space as travelling electromagnetic waves. The energy of signals exists in the form of electrical and magnetic fields. Both electrical and magnetic fields vary sinusoidally with time. The two fields always exist together because a change in electrical field generates a magnetic field and a

3.5

Radio Wave Propagation

55

change in magnetic field develops an electrical field. Thus there is continuous flow of energy from one field to the other. Radio waves arrive at a mobile receiver from different directions with different time delays. They combine via vector addition at the receiver antenna to give a resultant signal with a large or small amplitude depending upon whether the incoming waves combine to strengthen each other or cancel each other. As a result, a receiver at one location may experience a signal strength several tens of decibels (dB) different from a similar receiver located only a short distance away. As a mobile receiver moves from one location to another, the phase relationship between the various incoming waves also changes. Thus, there are substantial amplitude and phase fluctuations, and the signal is subjected to fading. Figure 3.5 illustrates the fading characteristics of a mobile radio signal. A steady decrease in the received signal power at a separation distance, d, of several kilometers (or miles) occurs. This is the signal attenuation. Attenuation is proportional to the second power of distance in the free space, but can vary to the fourth or fifth power in built-up areas because of reflections and obstacles. When we focus on a distance of a couple of kilometers, we observe that signal power fluctuates around a mean value and the fluctuations have a somewhat longer period. This is referred to as long-term or slow fading. When we concentrate and examine the signal power over a few hundred meters, we find that signal power fluctuates more rapidly. These rapid fluctuations are caused by a local multipath. The phenomenon giving rise to these rapid fluctuations is referred to as short-term or fast fading. The slow fading is caused by movement over distances large enough to produce gross variations in the overall path between the base station and the mobile station. Fast fading occurs usually over distances of about half a wavelength. For VHF and Short-Term Fading r(t)

Signal

Strength dB

Long-Term Fading m(t)



 Time

Figure 3.5

Mobile radio signal fading representation.

56

3

Radio Propagation and Propagation Path-Loss Models

UHF, a vehicle travelling at 30 miles per hour (mph) can encounter several fast fades in a second. Therefore, the mobile radio signal (see Figure 3.5) contains a short-term fast fading signal superimposed on a long-term slow fading signal (which remains constant over a small area but varies slowly as the mobile receiver moves). To separate out fast fading from slow fading, the magnitude of the received signal is averaged over a distance on the order of 10 m, and the result is referred to as the small-area average or sector average [1]. The rapid fluctuation in fast fading is a result of small movements of the transmitter, receiver, and surrounding objects. Because fast fading is random, its statistical properties are used to determine system performance. For locations that are heavily shadowed by surrounding buildings, it is typically found that a Rayleigh distribution approximates the probability density function (PDF) [8,9,14,18]. For locations where there is one path making a dominant contribution to the received signal, such as when the base station is visible to the mobile station (typically in the indoor environment), the distribution function is typically found to be that of a Rician distribution [13,14]. Because of shadowing by buildings and other objects, the average within individual small areas also varies from one small area to the next in an apparently random manner, referred to as the shadow effect. Shadow effect is often called lognormal fading because its distribution is represented by lognormal distribution. In a moving vehicle, slow fading is observed over a longer time scale than fast fading. Slow fading is the average of received signal power over large transmitter and receiver separation distances. A local mean is computed by averaging signal power over 5 to 40 wavelengths (), or separation distance between 40 to 80 fades (where a signal crosses a certain level). Using a reference distance, d0, which is the signal attenuation at a standard distance from the antenna (usually d0  1 m is used for indoor environment and d0  1 km for outdoor environment) the received power, Pr, at distance d is given as: Pr  P0(d0 /d)

(3.10)

where:   path loss exponent, varying from 2 (free-space) to 5 (urban environment) P0  power at reference distance d0 Pr  received power that is proportional to d  The received power under non-line-of-sight propagation conditions can be written as: Pr(d)  10 log [P0(d0)]  10 log (d0 /d) (dBm)

(3.11)

The accuracy of Pr can be improved by accounting for a random shadow effect caused by obstructions such as buildings or mountains. Shadow effect is described

3.5

Radio Wave Propagation

57

by a zero-mean Gaussian random variable, X, with standard deviation,  (dB). Under ideal conditions, it is possible to estimate path loss Lp(d) from the transmitter (Pt) and receiver power (Pr) as Lp(d)  Pt Pr, but this approach ignores the fact that the signal undergoes lognormal fading, which could reduce the received power at any location. Since the fading is long term, no improvements can be expected if the mobile station moves through short distances. Allowable path loss Lp(d)  Path loss  Shadow effect X 

Lp(d)  Lp(d0)  10 log (d/d0)  X (dB)

(3.12) (3.13)

where:  Lp(d0) is known as the 1 m or 1 km loss, or insertion loss that arises due to free-space path loss and antenna inefficiencies (i.e., reference value of path loss). X is often based on measurement made over a wide range of locations and transmitter-receiver separation. An average value of 8 dB for  is often used giving X as 10.5 dB. It should also be noted that whenever relative motion exists between transmitter and receiver, there is a Doppler shift in the received signal. The maximum Doppler shift fm is given as: v fm   c/f

(3.14)

where: c  velocity of electromagnetic waves in free space v  velocity of the moving vehicle f  frequency of the carrier In the mobile radio case, the fading and Doppler shift occur as a result of the motion of the receiver through a spatially varying field. Doppler shift also results from the motion of the scattering of the radio waves (e.g., cars, trucks, vegetation). The effect of multipath propagation is to produce a received signal with an amplitude that varies quite substantially with location. At UHF and higher frequencies, the motion of the scattering also causes fading to occur even if the mobile station or handset is not in motion. The received signal s(t) can be expressed as the product of two components: the signal subjected to long-term fading, m(t), and the signal subjected to shortterm fading, r(t), as: s(t)  m(t) · r(t)

(3.15)

58

3

Radio Propagation and Propagation Path-Loss Models

Example 3.3 Calculate the received power at a distance of 3 km from the transmitter if the path-loss exponent  is 4. Assume the transmitting power of 4 W at 1800 MHz, a shadow effect of 10.5 dB, and the power at reference distance (d0  100 m) of 32 dBm. What is the allowable path loss? Solution Using Equation 3.11 and including shadow effect we get Pr(d)  10 log [P0(d0)]  10 log (d0/d)  X

 3000 

100  Pr  32  10  4  log   10.5  80.5 dBm

Allowable path loss  Pt Pr  36 ( 80.5)  116.5 dB

Example 3.4 What is the separation distance between the transmitter and the receiver with an allowable path loss of 150 dB and shadow effect of 10 dB? The path loss in dB is given as: Lp  133.2  43 log d

where: d  separation distance in km Solution Using Equation 3.13, we have 150  133.2  40 log d  10 6.8 log d    0.17 40

d  100.17  1.48 km

3.6

Characteristics of a Wireless Channel

The wireless channel is different and much more unpredictable than the wireline channel because of factors such as multipath and shadow fading, Doppler shift, and time dispersion or delay spread. These factors are all related to variability

3.6

Characteristics of a Wireless Channel

59

introduced by mobility of the user and the wide range of environmental conditions that are encountered as a result. Multipath delays occur as a transmitted signal is reflected by objects in the environment between a transmitter and a receiver. These objects can be buildings, trees, hills, or even trucks and cars (see Figure 3.6). The reflected signals arrive at the receiver with a random phase offset, since each reflected signal generally follows a different path to reach the user’s receiver, resulting in a random signal that fades as the reflections destructively or constructively superimpose on one another. This effectively cancels or adds part of signal energy for brief periods of time. The degree of fading will depend on the delay spread of the reflected signals as embodied by their relative phases, and their relative power. A mobile radio channel exhibits both time dispersion and frequency dispersion. Time dispersion is the distortion to the signal and is manifested by the spreading in time of the modulation symbols. This is caused by frequency-selective fading. A channel, which is said to be frequency selective, has many frequency components that take different times to arrive at the receiver and undergo different attenuation levels. The frequency band over which the attenuation remains constant provides a frequency Initial Transmitted Pulse

Received Pulses

3 1

Base Station Antenna 4

2

N  4

d

Figure 3.6

Multipath delay.

t

60

3

Radio Propagation and Propagation Path-Loss Models

region where all frequency components behave identically. We identify this frequency band as the coherence bandwidth of the channel. Time dispersion occurs when the channel is band-limited or when the coherence bandwidth of the channel is smaller than the modulation bandwidth. The time dispersion leads to inter-symbol-interference (ISI), where the energy from one symbol spills over into another symbol, thereby increasing the bit-error-rate (BER). In many instances, the fading due to multipath delay will be frequency selective, randomly affecting only a portion of the overall channel bandwidth at a given time. In the case of frequency selective fading, the delay spread exceeds the symbol duration. On the other hand, when there is no dispersion and delay spread is less than the symbol duration, the fading will be flat, thereby affecting all frequencies in the signal equally. Flat fading can lead to deep fades of more than 30 to 40 dB. Doppler shift is the random changes in a channel introduced as a result of a mobile user’s mobility. Doppler spread has the effect of shifting or spreading the frequency components of a signal. This is described in terms of frequency dispersion. Like the coherence bandwidth, coherence time is defined as the time over which the channel can be assumed to be constant. The coherence time of the channel is the inverse of the Doppler spread. It is the measure of the speed at which channel characteristics change. This determines the rate at which fading occurs. When the channel changes are higher than the modulated symbol rate, fast fading occurs. Slow fading occurs when the channel changes are slower than the symbol rate.

3.6.1

Multipath Delay Spread, Coherence Bandwidth, and Coherence Time As discussed earlier, the multipath delay spread is the time dispersion characteristic of the channel. Each multipath component is typically associated with different time delays and attenuation, the shortest of which is the LOS path. We denote the rms delay spread in multipath delay by d and the maximum spread in frequency due to Doppler shift, fm. We use the coherence bandwidth, which is a range of frequencies over which two frequency components have a strong potential for amplitude correlation, to define whether the channel fading is flat or frequency selective. The coherence bandwidth (Bc) between two frequency envelopes is given as [16]: 1 Bc ≈ 

2d

(3.16)

Frequency components of a signal separated by more than Bc will fade independently. A channel is a frequency-selective channel if Bc  Bw, where Bw is the signal bandwidth. Frequency selective distortion occurs whenever a signal’s spectral components are not all affected equally by the channel. In order to avoid channel-induced ISI distortion, the channel is required to be flat fading by ensuring

3.6

Characteristics of a Wireless Channel

61

that Bc Bw. Thus, the channel coherence bandwidth sets an upper limit on the transmission rate that can be used without incorporating an equalizer in the receiver. The coherence time, Tc, describes the expected time duration over which the impulse response of the channel stays relatively invariant or correlated. The coherence time is approximately inversely proportional to Doppler spread [16,17] 1 Tc ≈ 

2fm

(3.17)

where: v fm    maximum Doppler spread  v  velocity of moving vehicle   wavelength  c/f f  frequency of carrier c  speed of electromagnetic wave in free space (3  108 m/s) A rule of thumb for the coherence time value is Tc  0.423/fm. If the transmitted symbol interval, Ts , exceeds Tc , then the channel will change during the symbol interval and symbol distortion will occur. In such cases, a matched filter is impossible without equalization and correlator losses occur. A Rayleigh fading signal may change amplitude significantly in the interval Tc. If the signal symbol interval Ts » Tc , the channel changes or fades rapidly compared to the symbol rate. This case is called fast fading and frequency dispersion occurs, causing distortion. If Ts « Tc the channel does not change during the symbol interval. This case is called slow fading. Thus, to avoid signal distortion caused by fast fading, the channel must be made to exhibit slow fading by ensuring that the signaling rate exceeds the channel fading rate Ts  Tc. Example 3.5 Assuming the speed of a vehicle is equal to 60 mph (88 ft/sec), carrier frequency, fc  860 MHz, and rms delay spread d  2 sec, calculate coherence time and coherence bandwidth. At a coded symbol rate of 19.2 kbps (IS-95) what kind of symbol distortion will be experienced? What type of fading will be experienced by the IS-95 channel? Solution v  60 mph (88 ft/sec)  108   c  9.84 6  1.1442 ft f

860  10

v 88 Maximum Doppler shift  fm     77 Hz 

1.1442

1  1 Tc     0.0021 seconds 2fm

2  77

62

3

Radio Propagation and Propagation Path-Loss Models

6

10 Ts    52 s 19,200

The symbol interval is much smaller compared to the channel coherence time. Symbol distortion is, therefore, minimal. In this case fading is slow. 1  1 Coherence bandwidth  Bc ≈   79.56 kHz  6 2d

2  2  10

This shows that IS-95 is a wide band system in this multipath situation and experiences selective fading only over 6.5% (79.56/1228.8  0.0648 ~ 6.5%) of its bandwidth (Bw  1228.8 kHz).

3.7

Signal Fading Statistics

As discussed earlier, the rapid variations (fast fading) in signal power caused by local multipaths are represented by Rayleigh distribution. The long-term variations in the mean level are denoted by lognormal distribution. With a LOS propagation path, the Rician distribution is often used for fast fading. Thus, the fading characteristics of a mobile radio signal are described by the following statistical distributions (see Figures 3.7 and 3.8).

K   dB Rayleigh Distribution

K  6 dB

p (r)

Rician Distribution

Received signal envelope voltage r (volts) Figure 3.7

Rayleigh and Rician distribution.

3.7

Signal Fading Statistics

63

0.4

0.35

probability density p (z)

0.3

0.25

0.2

0.15

0.1

0.05

0 4

Figure 3.8

3

2

1

0 mean

1

2

3

4

z = log[(S-Sm)/s]

Lognormal distribution.

• Rician Distribution • Rayleigh Distribution • Lognormal Distribution

3.7.1 Rician Distribution When there is a dominant stationary (nonfading) signal component present, such as a LOS propagation path, the small-scale fading envelope distribution is Rician. The Rician distribution has a probability density function (PDF) given by: 2  A2 r 2

r p(r)   e 2 



2



 

Ar I0  2

for A  0,

r0

(3.18)

where: A  peak amplitude of the dominant signal and I0  (. . .) modified Bessel Function of the first kind and zero order r2/2  instantaneous power   standard deviation of the local power

64

3

Radio Propagation and Propagation Path-Loss Models

The Rician distribution is often described in terms of a parameter K, known as the Rician factor and is expressed as: 2

A dB K  10 log  2

(3.19)

2

As A → 0, K → ∞ dB and as the dominant path decreases in amplitude, the Rician distribution degenerates to a Rayleigh distribution (see Figure 3.7).

3.7.2 Rayleigh Distribution The Rayleigh distribution is used to describe the statistical time-varying nature of the received envelope of a flat fading signal, or the envelope of an individual multipath component. The Rayleigh distribution is given as: 2

r p(r)   e 2 

r 

 22

0 r ∞

(3.20)

where:   rms value of the received signal r2/2  instantaneous power 2  local average power of the received signal before envelope detection Instead of the distribution of the received envelope, we can also describe a Rayleigh fading signal in terms of the distribution function of its received normalized power. Let   (r2/2)/2, the instantaneous received power divided by the mean received power. Then d  r/2 p dr and since p(r)dr must be equal to p()d, we get 2 2  r/2 e  r /2 

p()  [p(r)dr]/[(r/2)dr]    e ,  r/2 

0  ∞

(3.21)

Equation 3.21 represents a simple exponential density function. One can rightfully say that a flat fading signal is exponentially fading in power.

3.7.3 Lognormal Distribution Lognormal distribution describes the random shadowing effects which occur over a large number of measurement locations which have the same transmitter and receiver separation, but have different levels of clutter on the propagation path. The signal, s(t), typically follows the Rayleigh distribution but its mean square value or its local mean power is lognormal in dBm with variance equal to 2s . Typically the standard deviation, s equals 8 to 10 dB.

3.8

Level Crossing Rate and Average Fade Duration

65

The lognormal distribution is given by (see Figure 3.8): 1 p(S)    e  2s



(S Sm)2  22s



(3.22)

where: Sm  mean value of S in dBm s  standard deviation of S in dB S  10 log s in dBm s  signal power in mW

3.8

Level Crossing Rate and Average Fade Duration

Doppler spread can be used to calculate the level crossing rate (LCR), or the average number of times a signal crosses a certain level. The combination of LCR and average fade duration, which is the average time over which the signal below a certain level can be used to design the signaling format and a diversity scheme for cellular systems. Let   R /Rrms be the value of the specified amplitude level, R, normalized to local rms amplitude of the fading envelope Rrms, based on the Rayleigh distribution of the received signal envelope. LCR and average fade duration are given by [1]: 

LCR  NR  2 · fm · e 2 (crossings per sec) 2

e 1 Average fade duration   ·  (ms)  · fm  2

(3.23) (3.24)

where: fm  maximum Doppler shift Example 3.6 Consider a flat Rayleigh fading channel to determine the number of fades per second for   1 and average fade duration, when the maximum Doppler frequency is 20 Hz. What is the maximum velocity of the mobile if the carrier frequency is 900 MHz? Solution 



NR   2 · fm · e   2  20  1  e 2

1

 18.44 fades per second

e 1 1 Average fade duration   ·   0.0073 s 120  2

8

3  10 v  fm · c  20   6  6.66 m/s  24 km/hour f

900  10

66

3.9

3

Radio Propagation and Propagation Path-Loss Models

Propagation Path-Loss Models

Propagation path-loss models [20] play an important role in the design of cellular systems to specify key system parameters such as transmission power, frequency, antenna heights, and so on. Several models have been proposed for cellular systems operating in different environments (indoor, outdoor, urban, suburban, rural). Some of these models were derived in a statistical manner based on field measurements and others were developed analytically based on diffraction effects. Each model uses specific parameters to achieve reasonable prediction accuracy. The long distance prediction models intended for macrocell systems use base station and mobile station antenna heights and frequency. On the other hand, the prediction models for short distance path-loss estimation use building heights, street width, street orientation, and so on. These models are used for microcell systems. When the cell size is quite small (in the range of 10 to 100 m), deterministic models based on ray tracing methods are used. Thus, it is essential to select a proper path-loss model for design of the mobile system in the given environment. Propagation models are used to determine the number of cell sites required to provide coverage for the network. Initial network design typically is based on coverage. Later growth is engineered for capacity. Some systems may need to start with wide area coverage and high capacity and therefore may start at a later stage of growth. The coverage requirement along with the traffic requirement relies on the propagation model to determine the traffic distribution, and will offload from an existing cell site to new cell sites as part of a capacity relief program. The propagation model helps to determine where the cell sites should be placed to achieve an optimal location in the network. If the propagation model used is not effective in placing cell sites correctly, the probability of incorrectly deploying a cell site in the network is high. The performance of the network is affected by the propagation model chosen because it is used for interference predictions. As an example, if the propagation model is inaccurate by 6 dB (provided S/I  17 dB is the design requirement), then the signal-to-interference ratio, S/I, could be 23 dB or 11 dB. Based on traffic conditions, designing for a high S/I could negatively affect financial feasibility. On the other hand, designing for a low S/I would degrade the quality of service. The propagation model is also used in other system performance aspects including handoff optimization, power level adjustments, and antenna placements. Although no propagation model can account for all variations experienced in real life, it is essential that one should use several models for determining the path losses in the network. Each of the propagation models being used in the industry has pros and cons. It is through a better understanding of the limitations of each of the models that a good RF engineering design can be achieved in a network. We discuss two widely used empirical models: Okumura/Hata and COST 231 models. The Okumura/Hata model has been used extensively both in Europe and

3.9

Propagation Path-Loss Models

67

North America for cellular systems. The COST 231 model has been recommended by the European Telecommunication Standard Institute (ETSI) for use in Personal Communication Network/Personal Communication System (PCN/PCS). In addition, we also present the empirical models proposed by International Mobile Telecommunication-2000 (IMT-2000) for the indoor office environment, outdoor to indoor pedestrian environment, and vehicular environment.

3.9.1 Okumura/Hata Model Okumura analyzed path-loss characteristics based on a large amount of experimental data collected around Tokyo, Japan [6,11]. He selected propagation path conditions and obtained the average path-loss curves under flat urban areas. Then he applied several correction factors for other propagation conditions, such as: • Antenna height and carrier frequency • Suburban, quasi-open space, open space, or hilly terrain areas • Diffraction loss due to mountains • Sea or lake areas • Road slope

Hata derived empirical formulas for the median path loss (L50) to fit Okumura curves. Hata’s equations are classified into three models: 1. Typical Urban L50(urban)  69.55  26.16 log fc  (44.9 6.55 log hb)log d 13.82 log hb a(hm)(dB)

(3.25)

where: a(hm)  correction factor (dB) for mobile antenna height as given by: • For large cities a(hm)  8.29[log (1.54hm)]2 11 a(hm)  3.2[log (11.75hm)]2 4.97

fc 200 MHz fc  400 MHz

(3.26) (3.27)

• For small and medium-sized cities a(hm)  [1.1 log (fc) 0.7]hm [1.56 log (fc) 0.8]

(3.28)

68

3

Radio Propagation and Propagation Path-Loss Models

2. Typical Suburban

   28f   5.4  dB c

L50  L50(urban) 2 log 

2

(3.29)

3. Rural L50  L50(urban) 4.78(log fc)2  18.33 log fc 40.94 dB

(3.30)

where: fc  carrier frequency (MHz) d  distance between base station and mobile (km) hb  base station antenna height (m) hm  mobile antenna height (m) The range of parameters for which the Hata model is valid is: 150 fc 2200 MHz 30 hb 200 m 1 hm 10 m 1 d 20 km

3.9.2 Cost 231 Model This model [19] is a combination of empirical and deterministic models for estimating the path loss in an urban area over the frequency range of 800 MHz to 2000 MHz. The model is used primarily in Europe for the GSM 1800 system. L50  Lf  Lrts  Lms dB

(3.31)

L50  Lf when Lrts  Lms 0

(3.32)

or

where: Lf  free space loss (dB) Lrts  roof top to street diffraction and scatter loss (dB) Lms  multiscreen loss (dB) Free space loss is given as: Lf  32.4  20 log d  20 log fc dB

(3.33)

The roof top to street diffraction and scatter loss is given as: Lrts  16.9 10 log W  10 log fc  20 log hm  L0 dB

(3.34)

3.9

Propagation Path-Loss Models

where: W  street width (m)

hm  hr hm m L0  10  0.354 L0  2.5  0.075( 35) dB L0  4 0.114( 55) dB

69

0  35° 35°  55° 55°  90°

where:   incident angle relative to the street The multiscreen (multiscatter) loss is given as: Lms  Lbsh  ka  kdlog d  kf log fc 9 log b

(3.35)

where: b  distance between building along radio path (m) d  separation between transmitter and receiver (km) Lbsh  18 log (11  hb)

hb  hr

Lbsh  0

hb  hr

where: hb  hb hr, hr  average building height (m) ka  54

hb hr

ka  54 0.8hb

d  500m;

h b hr

ka  54 0.8 hb(d/500)

d  500m;

h b hr

Note: Both Lbsh and ka increase path loss with lower base station antenna heights. kd  18

hb  hr 15 h

hm

kd  18 b

h b  hr

kf  4  0.7(fc / 925 1) for mid-size city and suburban area with moderate tree density

 925 fc



kf  4  1.5  1 for metropolitan area The range of parameters for which the COST 231 model is valid is: 800 fc 2000 MHz 4 hb 50 m

70

3

Radio Propagation and Propagation Path-Loss Models

1 hm 3 m 0.02 d 5 km The following default values may be used in the model: b  20–50 m W  b/2   90° Roof  3 m for pitched roof and 0 m for flat roof, and hr  3 (number of floors)  roof Example 3.7 Using the Okumura and COST 231 models, calculate the L50 path loss for a PCS system in an urban area at 1, 2, 3, 4 and 5 km distance (see Table 3.1). Assume hb  30 m, hm  2 m, and carrier frequency fc  900 MHz. Use the following data for the COST 231 model: W  15 m,

b  30 m,

  90°, hr  30 m

COST 231 Model L50  Lf  Lrts  Lms Lf  32.4  20 log d  20 log fc  32.4  20 log d  20 log 900 dB Lf  91.48  20 log d dB Lrts  16.9 10 log W  10 log fc  20 log hm  L0

hm  hr hm  30 2  28 m Table 3.1 Summary of path losses from COST 231 model. d (km)

Lf (dB)

Lrts (dB)

1

91.49

29.82

Lms (dB)

9.72

L50 (dB)

131.03

2

97.50

29.82

15.14

142.46

3

101.03

29.82

18.31

149.16

4

103.55

29.82

20.56

153.91

5

105.47

29.82

22.30

157.59

Note: The table applies to this example only.

3.9

Propagation Path-Loss Models

71

L0  4 0.114( 55)  4 0.114(90 55)  0 Lrts  16.9 10 log 15  10 log 900  20 log 28  0  29.82 dB Lms  Lbsh  ka  kd log d  kf log fc 9 log b ka  54 0.8hb  54 0.8  30  30

hb  hb hr  30 30  0 m Lbsh  18 log 11  0  18.75 dB 15 h

hm

15  0 kd  18 b  18   18

 925



28

fc 900 kf  4  0.7  1  4  0.7  1  3.98 (for mid-sized city)

 925



Lms  18.75  30  18 log d  3.98 log 900 9 log 30  9.72  18 log d dB

Okumura/Hata Model L50  69.55  26.16 log fc  (44.9 6.55hb)log d 13.82 log hb a(hm) dB a(hm)  (1.1 log fc 0.7)hm (1.56 log fc 0.8)  (1.1 log 1800 0.7)(2) (1.56 log 900 0.8)  1.29 dB L50  69.55  26.16 log 900  (44.9 6.55 log 30)log d 13.82 log 30 1.29 dB  125.13  35.23 log d dB (refer to Table 3.2)

The results from the two models are given in Figure 3.9. Note that the calculated path loss with the COST 231 model is higher than the value obtained by the Okumura/Hata model. Table 3.2 Summary of path losses from Okumura model. d (km)

L50 (dB)

1

125.13

2

135.74

3

141.94

4

142.34

5

145.76

72

3

Radio Propagation and Propagation Path-Loss Models

160

COST 231 model

155

Hata model

Path loss in [dB]

150

145

140

135

130

125

1

Figure 3.9

1.5

2

2.5 3 3.5 Distance from transmitter in [km]

4

4.5

5

Comparison of COST 231 and Hata-Okumura models.

3.9.3 IMT-2000 Models The operating environments are identified by appropriate subsets consisting of indoor office environments, outdoor to indoor and pedestrian environments, and vehicular (moving vehicle) environments. For narrowband technologies (such as FDMA and TDMA), delay spread is characterized by its rms value alone. However, for wide band technologies (such as CDMA), the strength and relative time delay of the many signal components become important. In addition, for some technologies (e.g., those using power control) the path-loss models must include the coupling between all co-channel propagation links to provide accurate predictions. Also, in some cases, the shadow effect temporal variations of the environment must be modeled. The key parameters of the IMT-2000 propagation models are: • Delay spread, its structure, and its statistical variation • Geometrical path loss rule (e.g., d , 2  5) • Shadow fading margin • Multipath fading characteristics (e.g., Doppler spectrum, Rician vs. Rayleigh

for envelope of channels) • Operating radio frequency

3.9

Propagation Path-Loss Models

73

Indoor Office Environment This environment is characterized by small cells and low transmit powers. Both base stations and pedestrian users are located indoors. RMS delay spread ranges from around 35 nsec to 460 nsec. The path loss rule varies due to scatter and attenuation by walls, floors, and metallic structures such as partition and filing cabinets. These objects also produce shadowing effects. A lognormal shadowing with a standard deviation of 12 dB can be expected. Fading characteristic ranges from Rician to Rayleigh with Doppler frequency offsets are determined by walking speeds. Path-loss model for this environment is: L50  37  30 log d  18.3 · n  n  2 / n  1  0.46  dB

(3.36)

where: d  separation between transmitter and receiver (m) n  number of floors in the path Outdoor to Indoor and Pedestrian Environment This environment is characterized by small cells and low transmit power. Base stations with low antenna heights are located outdoors. Pedestrian users are located on streets and inside buildings. Coverage into buildings in high power systems is included in the vehicular environment. RMS delay spread varies from 100 to 1800 nsec. A geometric path-loss rule of d 4 is applicable. If the path is a lineof-sight on a canyon-like street, the path loss follows a rule of d 2, where there is Fresnel zone clearance. For the region with longer Fresnel zone clearance, a path loss rule of d 4 is appropriate, but a range of up to d 6 may be encountered due to trees and other obstructions along the path. Lognormal shadow fading with a standard deviation of 10 dB is reasonable for outdoors and 12 dB for indoors. Average building penetration loss of 18 dB with a standard deviation of 10 dB is appropriate. Rayleigh and/or Rician fading rates are generally set by walking speeds, but faster fading due to reflections from moving vehicles may occur sometimes. The following path-loss model has been suggested for this environment: L50  40 log d  30 log fc  49 dB

(3.37)

This model is valid for non-line-of-sight (NLOS) cases only and describes the worst-case propagation. Lognormal shadow fading with a standard deviation equal to 10 dB is assumed. The average building penetration loss is 18 dB with a standard deviation of 10 dB. Vehicular Environment This environment consists of larger cells and higher transmit power. RMS delay spread from 4 microseconds to about 12 microseconds on elevated roads in hilly or mountainous terrain may occur. A geometric path-loss rule of d 4 and lognormal

74

3

Radio Propagation and Propagation Path-Loss Models

shadow fading with a standard deviation of 10 dB are used in the urban and suburban areas. Building penetration loss averages 18 dB with a 10 dB standard deviation. In rural areas with flat terrain the path loss is lower than that of urban and suburban areas. In mountainous terrain, if path blockages are avoided by selecting base station locations, the path-loss rule is closer to d 2. Rayleigh fading rates are determined by vehicle speeds. Lower fading rates are appropriate for applications using stationary terminals. The following path-loss model is used in this environment: L50  40(1 4  10 2 hb)log d 18 log ( hb)  21 log fc  80 dB (3.38)

where:

hb  base station antenna height measured from average roof top level (m) Delay Spread A majority of the time rms delay spreads are relatively small, but occasionally, there are “worst case” multipath characteristics that lead to much larger rms delay spreads. Measurements in outdoor environments show that rms delay spread can vary over an order of magnitude within the same environment. Delay spreads can have a major impact on the system performance. To accurately evaluate the relative performance of radio transmission technologies, it is important to model the variability of delay spread as well as the “worst case” locations where delay spread is relatively large. For each environment IMT-2000 defines three multipath channels: low delay spread, median delay spread, and high delay spread. Channel “A” represents the low delay spread case that occurs frequently; channel “B” corresponds to the median delay spread case that also occurs frequently; and channel “C” is the high delay spread case that occurs only rarely. Table 3.3 provides the rms values of delay spread for each channel and for each environment. Table 3.3 Rms delay spread (IMT-2000). Channel “A”

Channel “B”

Channel “C”

Environment

rms (ns)

% Occurrence

rms (ns)

% Occurrence

Indoor office

35

50

100

45

460

5

Outdoor to indoor and pedestrian

100

40

750

55

1800

5

Vehicular (high antenna)

400

40

4000

55

12,000

5

rms (ns) % Occurrence

3.10

Indoor Path-Loss Models

3.10

75

Indoor Path-Loss Models

Picocells cover part of a building and span from 30 to 100 meters [13,15]. They are used for WLANs and PCSs operating in the indoor environment. The path-loss model for a picocell is given as: 

Lp  Lp(d0)  10 log d  Lf(n)  X dB

(3.39)

where:  Lp(d0)  reference path loss at the first meter (dB)   path-loss exponent d  distance between transmitter and receiver (m) X  shadowing effect (dB) Lf (n)  signal attenuation through n floors Indoor-radio measurements at 900 MHz and 1.7 GHz values of Lf per floor are  10 dB and 16 dB, respectively. Table 3.4 lists the values of Lp(d0), Lf (n), , and X. Partition dependent losses for signal attenuation at 2.4 GHz are given in Table 3.5. 

Table 3.4 Values of Lp(d0), , Lf (n) and X␴ in Equation 3.39. Environment

Residential



Office

Commercial

Lp(d0) (dB)

38

38

38



2.8

3.0

2.2

Lf (n) (dB)

4n

15  4(n  1)

6  3(n  1)

X (dB)

8

10

10

Table 3.5 Partition dependent losses at 2.4 GHz. Signal attenuation through

Loss (dB)

Window in brick wall

2

Metal frame, glass wall in building

6

Office wall

6

Metal door in office wall

6

Cinder wall

4

Metal door in brick wall Brick wall next to metal door

12.4 3

76

3

Radio Propagation and Propagation Path-Loss Models

Femtocellular systems span from a few meters to a few tens of meters. They exist in individual residences and use low-power devices using Bluetooth chips or Home RF equipment. The data rate is around 1 Mbps. Femtocellular systems use carrier frequencies in the unlicensed bands at 2.4 and 5 GHz. Table 3.6 lists the values of Lp(d0) and  for LOS and NLOS conditions. Example 3.8 In a WLAN the minimum SNR required is 12 dB for an office environment. The background noise at the operational frequency is 115 dBm. If the mobile terminal transmit power is 100 mW, what is the coverage radius of an access point if there are three floors between the mobile transmitter and the access point? Solution • Transmit power of mobile terminal  10 log 100  20 dBm • Receiver sensitivity  background noise  minimum SNR  115  12  103 dB • Maximum allowable path loss  transmit power receiver sensitivity  20 ( 103)  123 dB  • Lp(d0)  38 dB, Lf (n)  15  4(3 1)  23 dB,   3, and X  10 dB (from Table 3.4) Maximum allowable path loss  123  38  23  10  30 log d d  54 m

3.11

Fade Margin

As we discussed earlier, the local mean signal strength in a given area at a fixed radius, R, from a particular base station antenna is lognormally distributed [7]. The local mean (i.e., the average signal strength) in dB is expressed by a normal random variable with a mean Sm (measured in dBm) and standard deviation s (dB). If Smin is the receiver sensitivity, we determine the fraction of the locations (at d  R) wherein a mobile would experience a received signal above the receiver sensitivity. The receiver sensitivity is the value that provides an acceptable signal under Rayleigh fading conditions. The probability distribution function for a lognormally distributed random variable is: 1 [(S Sm) /(2s )] p(S)    e 2

s 2

2

(3.40)

3.11

Fade Margin

77

Table 3.6 Values of A and  for femtocellular systems.

Environment

Center frequency (GHz)

Indoor office

2.4

Meeting room

5.1

Suburban residence

5.2



Lp(d0)(dB)



LOS

41.5

2.0

NLOS

37.7

3.3

LOS

46.6

2.22

NLOS

61.6

2.22

47

2 to 3

Scenario

LOS (same floor) NLOS (same floor)

4 to 5

NLOS & room in higher floor directly above Tx

4 to 6

NLOS & room in higher floor not directly above Tx

6 to 7

The probability for signal strength exceeding receiver sensitivity PSmin(R) is given as PSmin(R)  P[S  Smin] 





Smin



S

S

m 1 1 erf min p(S)dS     

2

2

s2



(3.41)

Note: See Appendix D for erf, erfc and Q functions. Example 3.9 If the mean signal strength and receiver sensitivity are 100 dBm and 110 dBm, respectively and the standard deviation is 10 dB, calculate the probability for exceeding signal strength beyond the receiver sensitivity. Solution 1 1 erf 110  100  0.5  0.5 erf(0.707)  0.84 PSmin(R)      2

2



102



78

3

Radio Propagation and Propagation Path-Loss Models

Next, we determine the fraction of the coverage within an area in which the received signal strength from a radiating base station antenna exceeds Smin. We define the fraction of the useful service area Fu as that area, within an area for which the signal strength received by a mobile antenna exceeds Smin. If PSmin is the probability that the received signal exceeds Smin in an incremental area dA, then 1 P dA Fu   Smin 2

(3.42)

R

Using the power law we express mean signal strength Sm as d Sm  10log 

(3.43)

R

where accounts for the transmitter effective radiated power (ERP), receiver antenna gain, feed line losses, etc. Substituting Equation 3.43 into 3.41 we get:



Smin  10log(d ⁄ R)

1 1 erf PSmin      2

2

s2





(3.44)



Let a  (Smin )/(s 2 ) and b  (10log (d/R))/(s 2 ) Substituting Equation 3.44 into 3.42, we get

 x{erf[a  b log (x/R)]}dx R

1 1 Fu   2 2

R

(3.45)

0

Let t  a  b log (x/R), then



1 1 e( 2a)/b ∴Fu    2

b

a



e(2t)/berf(t)dt

(3.46)

or





1 1 erf(a)  e(1 2ab)/b2 1 erf 1 ab Fu    2



b



(3.47)

If we choose such that Sm  Smin at d  R, then a  0 and 1 2

Fu   1 e1/b [1 erf(1/b)]

2

Figure 3.10 shows the relation in terms of the parameter s .

(3.48)

3.12

Link Margin

79

FRACTION OF TOTAL AREA WITH SIGNAL ABOVE THRESHOLD, Fu

1.0

PSmin (R ) 5 0.95 0.9 0.85

0.9

0.8 0.75 0.7

0.8

0.65 0.6 0.7

0.6

0.5

Figure 3.10

0.55

σm 5 STANDARD DEVIATION dB; PATH LOSS VARIES AS 1/r r, PSmin (R) COVERAGE PROBABILITY ON AREA BOUNDARY (d 5 R)

0

1

2

3

0.5

4 ss / g

5

6

7

8

Fraction of total area with average power above threshold.

(After: W. C. Jakes, Jr., (editor), Microwave Mobile Communications. New York: John Wiley & Sons, 1974, p. 127)

3.12

Link Margin

To consider the losses incurred in transmitting a signal from point A to point B, we start by adding all the gains and losses in the link to estimate the total overall link performance margin [4]. The receiver power Pr is given as: PtGtGr

Pr   Lp

(3.49)

where: Pt is the transmitter power Gt and Gr are the gains of transmitter and receiver Lp is the path loss between the transmitter and receiver. In addition, there are also the effects due to receiver thermal noise, which is generated due to random noise inherent within a receiver’s electronics. This

80

3

Radio Propagation and Propagation Path-Loss Models

increases with temperature. We account for this thermal noise effect with the following: N  kTBw

(3.50)

where: k  Boltzmann’s constant (1.38  10 23 W/Kelvin-Hz) T  temperature in Kelvin Bw  receiver bandwidth(Hz). Spectral noise density, N0, is the ratio of thermal noise to receiver bandwidth N0  N/Bw  kT

(3.51)

Finally, there is an effect on signal-to-noise (SNR) ratio due to the quality of the components used in the receiver’s amplifiers, local oscillators (LOs), mixers, etc. The most basic description of a component’s quality is its noise figure, Nf, which is the ratio of the SNR at the input of the device versus the SNR at its output. The overall composite effect of several amplifiers’ noise figures is cumulative, and can be obtained as: Nf,total  Nf1  (Nf2 1)/G1  (Nf3 1)/(G1G2)  . . .

(3.52)

where: Nfk is the noise figure in stage k Gk  gain of the kth stage. By combining all the factors, we can develop a relation that allows us to calculate the overall link margin PtGtGrAg

M  

N 

Eb Nf, totalTkLpLfL0Fmargin R 

0 reqd

where: Ag  gain of receiver amplifier in dB R  data rate in dB Fmargin  fade margin in dB Tk  noise temperature in Kelvin (Eb / N0)reqd  required value in dB Lp  path losses in dB Lf  antenna feed line loss in dB L0  other losses in dB

(3.53)

3.13

Summary

81

Expressing Equation 3.53 in dB, we obtain M  Pt  Gt  Gr  Ag Nf, total Tk Lp Lf L0 Fmargin R (Eb /N0)reqd dB

(3.54)

Example 3.10 Given a flat rural environment with a path loss of 140 dB, a frequency of 900 MHz, 8 dB transmit antenna gain and 0 dB receive antenna gain, data rate of 9.6 kbps, 12 dB in antenna feed line loss, 20 dB in other losses, a fade margin of 8 dB, a required Eb /N0 of 10 dB, receiver amplifier gain of 24 dB, noise figure total of 6 dB, and a noise temperature of 290 K, find the total transmit power required of the transmitter in watts for a link margin of 8 dB. k  10 log (1.38  10 23)  228.6 dBW

Lp  140 dB; Ag  24 dB; Nf  6 dB; Fmargin  8 dB; Gt  8 dB; Gr  0 dB; L0  20 dB; Lfeed  12 dB; T  24.6 dB; R  39.8 dB; (Eb /N0)reqd  10 dB; and M  8 dB From Equation (3.54) Pt  M Gt Gr Ag  Nf, total  T  k  Lp  Lf  L0  Fmargin  R  (Eb /N0)reqd Pt  8 8 0 24  6  (24.6 228.6)  140  12  20  8  39.8  10  7.8 dBW ∴Pt  100.78 ≈ 6 W

3.13

Summary

In this chapter we discussed propagation and multipath characteristics of a radio channel. The concepts of delay spread that causes channel dispersion and intersymbol interference were also presented. Since the mathematical modeling of the propagation of radio waves in a real world environment is complicated, empirical models were developed by several authors. We presented these empirical and semi-empirical models used for calculating the path losses in urban, suburban, and rural environments and compared the results obtained with each model. Doppler spread, coherence bandwidth, and time dispersion were also discussed. The

82

3

Radio Propagation and Propagation Path-Loss Models

forward error correcting algorithms [3] for improving radio channel performances are given in Chapter 8.

Problems 3.1 Define slow and fast fading. 3.2 What is a frequency selective channel? 3.3 Define receiver sensitivity. 3.4 A vehicle travels at a speed of 30 m/s and uses a carrier frequency of 1 GHz. What is the maximum Doppler shift? What is the approximate fade duration? 3.5 A mobile station traveling at 30 km per hour receives a flat Rayleigh fading signal at 800 MHz. Determine the number of fades per second above the rms level. What is the average duration of fade below the rms level? What is the average duration of fade at a level 20 dB below the rms level? 3.6 Find the received power for the link from a synchronous satellite to a terrestrial antenna. Use the following data: height  60,000 km; satellite transmit power  4 W; transmit antenna gain  18 dBi; receive antenna gain  50 dBi; and transmit frequency  12 GHz. 3.7 Determine the SNR for the spacecraft that uses a transmitter power of 16 W at a frequency of 2.4 GHz. The transmitter and receiver antenna gain are 28 dBi and 60 dBi, respectively. The distance from the spacecraft to ground is 2  1010 m, the effective noise temperature of antenna plus receiver is 14 degrees Kelvin, and a bit rate of 120 kbps. Assume the bandwidth of the system to be half of the bit rate, 60 kHz. 3.8

A base station transmits a power of 10 W into a feeder cable with a loss of cable 10 dB. The transmit antenna has a gain of 12 dBi in the direction of the mobile receiver with a gain of 0 dBi and feeder loss of 2 dB. The mobile receiver has a sensitivity of 104 dBm. (a) Determine the effective isotropic radiated power, and (b) maximum acceptable path loss.

3.9 A receiver in a digital mobile communication system has a noise bandwidth of 200 kHz and requires that its input signal-to-noise ratio should be at least 10 dB when the input signal is 104 dBm. (a) What is the maximum permitted value of the receiver noise figure, and (b) What is the equivalent input noise temperature? 3.10 Calculate the maximum range of the communication system in Problem 8, assuming a mobile antenna height (hm) of 1.5 m, a base station antenna height (hb) of 30 m, a frequency equal to 900 MHz and propagation that

References

83

takes place over a plane earth. Assume base station and mobile station antenna gains to be 12 dBi and 0 dBi, respectively. How will this range change if the base station antenna height is doubled?

3.11 A mobile station traveling at a speed of 60 km/h transmits at 900 MHz. If it receives or transmits data at a rate of 64 kbps, is the channel fading slow or fast? 3.12 The power received at a mobile station is lognormal with a standard deviation of 8 dB. Calculate the outage probability assuming the average received power is 96 dBm and the threshold power is 100 dBm. 3.13 Determine the minimum signal power for an acceptable voice quality at the base station receiver of a GSM system (bandwidth 200 kHz, data rate 271 kbps). Assume the following data: Receiver noise figure  5 dB, Boltzmann’s constant  1.38  10 23 Joules/K, mobile radiated power  30 dBm, transmitter cable losses  3 dB, base station antenna gain  16 dBi, mobile antenna gain  0 dBi, fade margin  10.5 dB, and required Eb /N0  13.5 dB. What is the maximum allowable path loss? What is the maximum cell radius in an urban area where a 1 km intercept is 108 dB and the path-loss exponent is 4.2? 3.14 Develop a MATLAB program and obtain a curve for maximum path loss versus cell radius. Test your program using the following data: base station transmit power  10 W, base station cable loss  10 dB, base station antenna gain  8 dBi, base station antenna height  15 m, mobile station antenna gain  0 dBi, mobile station antenna height  1 m, body and matching loss  6 dB, receiver noise bandwidth  200 kHz, receiver noise figure  7 dB, noise density  174 dBm/Hz, required SNR  9 dB, building penetration loss  12 dB, and fade margin  10 dB. 3.15 In the Bluetooth device with NLOS, S /N required is 10 dB in an indoor office environment. The background noise at the operating frequency is 80 dBm. If the transmit power of the device is 20 dBm, what is its coverage?

References 1. Bertoni, H. L. Radio Propagation for Modern Wireless Systems. Upper Saddle River, NJ: Prentice Hall, 2000. 2. Clarke, R. H. “A Statistical Theory of Mobile Radio Reception.” Bell System Technical Journal 47 (July–August 1968): 957–1000. 3. Forney, G. D. “The Viterbi Algorithm.” Proceedings of IEEE, vol. 61, no. 3, March 1978, pp. 268–278. 4. Garg, V. K., and Wilkes, J. E. Wireless and Personal Communications Systems. Upper Saddle River, NJ: Prentice Hall, 1996.

84

3

Radio Propagation and Propagation Path-Loss Models

5. Hanzo, L., and Stefanov, J. “The Pan-European Digital Cellular Mobile Radio System — Known as GSM.” Mobile Radio Communications. R. Steele, Ed., Chapter 8, London: Prentech Press, 1992. 6. Hata, M. “Empirical Formula for Propagation Loss in Land Mobile Radio Services.” IEEE Transactions Vehicular Technology, vol. 29, no. 3, 1980. 7. Jakes, W. C., Ed., Microwave Mobile Communications. New York: John Wiley, 1974. 8. Lecours, M., Chouinard, I. Y., Delisle, G. Y., and Roy, J. “Statistical Modeling of the Received Signal Envelope in a Mobile Radio Channel.” IEEE Transactions Vehicular Technology, vol. VT-37, 1988, pp. 204–212. 9. Lee, William, C. Y. Mobile Communications Engineering. New York: John Wiley, 1989. 10. Mark, J. W., and Zhuang, W. Wireless Communications and Networking. Upper Saddle River, NJ: Prentice Hall, 2003. 11. Okumura, Y., et al. “Field Strength and Its Variability in VHF and UHF Land Mobile Radio Service.” Review Electronic Communication Lab 16, no. 9–10, 1968. 12. Parsons, D. The Mobile Radio Propagation Channel. Chichester, West Sussex, England: John Wiley, 1996. 13. Rappaport, T. S. Wireless Communications. Upper Saddle River, NJ: Prentice Hall, 1996. 14. Sampei, Seiichi. Applications of Digital Wireless Technologies to Global Wireless Communications. Upper Saddle River, NJ: Prentice Hall, 1997. 15. Seidel, S. Y., and Rappaport, T. S. “914 MHz Path Loss Prediction Models for Indoor Wireless Communications in Multifloor Buildings.” IEEE Transactions, Antenna & Propagation, 40(2), February 1992. 16. Sklar, B. “Rayleigh Fading Channels in Mobile Digital Communication Systems Part I: Characterization.” IEEE Communications Magazine, vol. 35, no. 9, September 1997, pp. 136–146. 17. Sklar, B. “Rayleigh Fading Channels in Mobile Digital Communication Systems Part II: Mitigation.” IEEE Communications Magazine, vol. 35, no. 9, September 1997, pp. 148–155. 18. Turin, G. L., et al. “A Statistical Model of Urban Multipath Propagation.” IEEE Transactions Vehicular Technology, vol. VT-21, 1972, pp. 1–9. 19. Walfisch, J., and Bertoni, H. L. “A Theoretical Model of UHF Propagation in Urban Environment.” IEEE Transactions, Antennas & Propagation, AP-36:1788–1796, October 1988. 20. Weissberger, M. A. “An Initial Critical Summary of Models for Predicting the Attenuations of Radio Waves.” ESD-TR-81–101, Electromagnetic Compatibility Analysis Center, Annapolis, MD, July 1982.

CHAPTER 4 An Overview of Digital Communication and Transmission 4.1

Introduction

The basic part of any digital communication system is the communication channel. This is the physical medium that carries information bearing signals from the source of the information to the sink. In a radio system the communication channel is the propagation of radio waves in free space (see Figure 4.1). As discussed in Chapter 3, radio waves in free space are subjected to fading. In nearly all communication systems some equipment is required to convert the information-bearing signal into a suitable form for transmission over the communication channel and then back into a form that is comprehensible to the end-user. This equipment is the transmitter and receiver. The receiver does not only perform the inverse translation to the transmitter, but it also has to overcome the distortions and disturbances (see Chapter 3) that occur over the communication channel. Thus, it is often more difficult to design the receiver than the transmitter. Speech coding, forward-error-correcting (FEC) coding, bit-interleaving, diversity, equalization, and modulation play important roles in a communication system, particularly in a radio system (see Chapters 7, 8, and 9). The transmitter for a radio system consists of antenna, RF section, encoder, and modulator. An antenna converts the electrical signal into a radio wave propagating in free space. The RF section of the transmitter generates a signal of sufficient power at the required frequency. It typically consists of a power amplifier, a local oscillator, and an up-converter. However, generally the RF section only amplifies and frequency-converts a signal (see Figure 4.2). At the input of the transmitter the user interface interacts and converts the information into a suitable digital data stream. The information source can be analog (such as speech) or discrete (such as data). Analog information is converted into digital information through the use of sampling and quantization. Sampling, quantization, and encoding techniques are called formatting and source coding. The source encoder and modulator bridge the gap between the digital data and electrical signal required at the input to the RF section. The encoder converts the data stream into a form that is more resistant to the degradations introduced

85

86

4

An Overview of Digital Communication and Transmission

Sink

Source Channel Transmitter

Figure 4.1

Receiver

Overview of a communication system.

Antenna

Input (Binary)

Encoder

Modulator

RF Section

Power amplifier Local oscillator Up-converter Figure 4.2

Structure of the transmitter for a radio system.

by the communication channel (see Chapter 3). The modulator takes a sine wave at the required carrier frequency and modifies the signal characteristics. We may regard the encoder/modulator as a single subsystem that maps data presented to it by the user interface onto a modulated RF carrier for subsequent processing, amplification, and transmission by the RF section. The demodulator/decoder takes the received RF signal and performs the inverse mapping back to the data stream for onward transmission. The encoder/modulator determines the bandwidth occupied by the transmitted signal. The demodulator/decoder determines the quality of the resulting service, in terms of bit-error-rate (BER), availability and delay. These subsystems also determine the robustness of the communication system to channel impairments due to the RF subsystem (such as phase noise and nonlinearity) and RF channel (such as multipath dispersion and fading, discussed in Chapter 3). Therefore, the correct choice of the coding/modulation scheme is vital for efficient operation of the whole system. In addition to coding/modulation, other means for improving signal quality in a wireless system include the speech coding, bitinterleaving, equalization, and diversity [4,6,9,10,14,16]. The basic objective of this chapter is to familiarize the readers who are not exposed to digital communications. The chapter may be omitted by the students who have been exposed to digital communications course(s). Details of techniques

4.3

Messages, Characters, and Symbols

87

Antenna

Digital Texual Analog

Sampling

Quantization Formatting

Encoding

Pulse Code Modulation

Transmit

Antenna

Receive

Demodulation /Detection

Formatting Low Pass Decoding Filter

Analog Texual Digital

Figure 4.3

Formatting, transmission, and reception of baseband signals.

such as speech/channel coding, bit-interleaving, modulation, diversity, and the antenna used for improving signal quality are discussed in Chapters 7, 8, and 9. This chapter focuses only on the performance parameters of coding and the modulation scheme. It outlines the OSI layers, types of data services, and discusses briefly transmission media to familiarize the readers.

4.2

Baseband Systems

Source information may contain either analog, textual, or digital data. Formatting involves sampling, quantization, and encoding. It is used to make the message compatible with digital processing. Transmit formatting transforms source information into digital symbols. When data compression is used in addition to formatting, the process is referred to as source coding. Figure 4.3 shows a functional diagram that primarily focuses on the formatting and transmission of baseband (information bearing) signals. The receiver with a detector followed by a signal decoder performs two main functions: (1) does reverse operations performed in the transmitter, and (2) minimizes the effect of channel noise for the transmitted symbol.

4.3

Messages, Characters, and Symbols

During digital transmission the characters are first encoded into a sequence of bits, called a bit stream or baseband signal. Groups of b bits form a finite symbol set or word M ( 2b) of such symbols [14,17]. A system using a symbol set size of M is referred to as an M-ary system. The value of b or M is an important initial choice in the design of any digital communication system. For b  1, the system is called a binary system, the size of symbol set M is 2, and the modulator uses two different waveforms to represent the binary “1” and the binary “0” (see Figure 4.4). In this case, the symbol rate and the bit rate are the same. For b  2, the system is called

88

4

An Overview of Digital Communication and Transmission

1 (⫺1, 1)

(⫺1, ⫺1)

(1, 1)

(1, ⫺1)

⫺1 Figure 4.4

Binary and quatenary systems.

quatenary or 4-ary (M  4) system. At each symbol time, the modulator uses one of the four different waveforms that represents the symbol (see Figure 4.4).

4.4

Sampling Process

The analog information is transformed into a digital format. The process starts with sampling the waveform to produce a discrete pulse-amplitude-modulated waveform (see Figure 4.5) [6,16,17]. The sampling process is usually described in time domain. This operation is basic to digital signal processing and digital communication. Using the sampling process, we convert the analog signal into corresponding sequences of samples that are usually spaced uniformly in time. The sampling process can be implemented in several ways, the most popular being the sample-and-hold operation. In this operation, a switch and storage mechanism (such as a transistor and a capacitor, or shutter and a filmstrip) form a sequence of samples of the continuous input waveform. The output of the sampling process is called pulse amplitude modulation (PAM) (see Section 4.6) because the successive output intervals can be described as a sequence of pulses with amplitudes derived from the input waveform samples. The analog waveform can be retrieved from a PAM waveform by simple low-pass filtering provided we choose the sampling rate properly. The ideal form of sampling is called instantaneous sampling. We sample the signal s(t) instantaneously at a uniform rate, fs, once every Ts (1/fs) seconds. Thus, we obtain: 

s(t) 



n  

s(nTs)(t  nTs)

(4.1)

where: s(t) is the ideal sampled signal (t  nTs) is the delta function positioned at time t  nTs A delta function is closely approximated by a rectangular pulse of duration t and amplitude s(nTs)/t; the smaller t the better the approximation.

4.4

Sampling Process

89

s(t)

s(t)

Ts

t

0

Figure 4.5

0

Sampling process.

We determine the Fourier transform of the ideal sampled signal s(t) by using the duality property of the Fourier transform and the fact that a delta function is an even function [6] as: s(t) ⇔ fs





m  

S(f  mfs)

(4.2)

where: S(f ) is the Fourier transform of the original signal s(t) and fs is the sampling rate. Equation 4.2 states that the process of uniformly sampling a continuoustime signal of finite energy results in a periodic spectrum with a period equal to the sampling rate. Taking the Fourier transform of both sides of Equation 4.1 and noting that the Fourier transform of the delta function (t  nTs) is equal to ej2nfTs, we get 

 s(nTs)ej2nfT

S(f )

s

(4.3)

n  

Equation 4.3 is called the discrete-time Fourier transform. It is the complex Fourier series representation of the periodic frequency function S(f ), with the sequence of samples {s(nTs)} defining the coefficients of the expansion. From Equation 4.2, we see that the Fourier transform of s(t) may also be expressed as: 

S(f )  fsS(f )  fs



S(f  mfs)

(4.4)

m  , m ⬆ 0

Next, we consider a continuous-time signal s(t) of finite energy and infinite duration. The signal is strictly band-limited with no frequency component

90

4

An Overview of Digital Communication and Transmission

higher than Bw Hz. This implies that the Fourier transform S(f ) of the signal s(t) has the property that S(f ) is zero for |f | Bw. If we choose the sampling period Ts  1/(2Bw), then the corresponding spectrum is given as: 

S(f ) 



n  

B 

jnf 

n e s

 2Bw

w

(4.5)

Using the following two conditions: 1. S(f )  0 for |f | Bw and 2. fs  2Bw We find from Equation 4.4 1 2Bw

S(f ) S(f), Bw f Bw

(4.6)

Substituting Equation 4.5 into Equation 4.6, we can write 

1  S(f ) 

2Bw

n  

 B , Bw f Bw

jnf 

n e  s

2B  w

w

(4.7)

Thus, if the sample value s(n/2Bw) of a signal s(t) is specified for all n, then the Fourier transform S(f) of the signal is uniquely determined by using the discretetime Fourier transform of Equation 4.7. Because s(t) is related to S(f) by inverse Fourier transform, it follows that the signal s(t) is itself uniquely determined by the sample values s(n/2Bw) for  n . In other words, the sequence {s(n/2Bw)} has all the information contained in s(t). We state the sampling theorem for band-limited signals of finite energy in two parts, which apply to the transmitter and receiver of a pulse modulation system, respectively. 1. A band-limited signal of finite energy with no frequency components higher than Bw Hz is completely described by specifying the values of signals at instants of time separated by 1/(2Bw) seconds. 2. A band-limited signal of finite energy with no frequency components higher than Bw Hz, may be completely recovered from a knowledge of its samples taken at the rate of 2Bw samples per second. This is known as the uniform sampling theorem. The sampling rate of 2Bw samples per second for a signal bandwidth Bw Hz, is called the Nyquist rate and 1/(2Bw) second is called the Nyquist interval [8]. We discussed the sampling theorem by assuming that signal s(t) is strictly band limited. In practice, however, an information-bearing signal is not strictly

4.4

Sampling Process

91

band limited, with the result that some degree of undersampling is encountered. Consequently, some aliasing is produced by the sampling process. Aliasing refers to the phenomenon of a high-frequency component in the spectrum of the signal seemingly taking on the identity of a lower frequency in the spectrum of its sampled version.

4.4.1 Aliasing Figure 4.6 shows the part of the spectrum that is aliased due to undersampling [13]. The aliased spectral components represent ambiguous data that can be retrieved only under special conditions. In general, the ambiguity is not resolved and ambiguous data appears in the frequency band between (fs  fm) and fm, where fm is the maximum frequency and fs is the sampling rate. In Figure 4.7 we use the higher sampling rate fs to eliminate the aliasing by separating the spectral replicas. Figures 4.8 and 4.9 show two ways to eliminate

fs  fm

fm

fm

fs

fs  fm

fs /2

Figure 4.6

Sampled signal spectrum.

fs /2

fm Figure 4.7

fs  fm

fs

Higher sampling rate to eliminate aliasing.

f

92

4

An Overview of Digital Communication and Transmission

fs /2

f 9m Figure 4.8

fs 2 f 9m

f

fs

Pre-filtering to eliminate aliasing.

fm

fm

fs  fm

fm

fs

fs  fm

fs /2 Figure 4.9

Post-filter to eliminate aliasing portion of the spectrum.

aliasing using anti-aliasing filters. The analog signal is prefiltered so that the new maximum frequency fm is less than or equal to fs /2. Thus there are no aliasing components seen in Figure 4.8, since fs 2fm . Eliminating aliasing terms prior to sampling is good engineering practice. When the signal structure is well known, the aliased terms can be eliminated after sampling with a linear pass filter (LPF) operating on the sampled data. In this case the aliased components are removed by post-filtering after sampling. The filter cutoff frequency fm removes the aliased components; fm needs to be less than (fs  fm). It should be noted that filtering techniques for eliminating the aliased portion of the spectrum will result in a loss of some of the signal information. For this reason, the sample rate, cutoff bandwidth, and filter type selected for a particular signal bandwidth are all interrelated.

4.4

Sampling Process

93

Realizable filters require a non-zero bandwidth for the transition between the passband and the required out-of-band attenuation. This is called the transition bandwidth. To minimize the system sample rate, we want the anti-aliasing filter to have a small transition bandwidth. Filter complexity and cost increase sharply with a narrower transition bandwidth, so a trade-off is required between the cost of a small transition bandwidth and the costs of the higher sampling rate, which are those of more storage and higher transition rates. In many systems the answer is to make the transition bandwidth 10 to 20% of the signal bandwidth. If we account for a 20% transition bandwidth of the anti-aliasing filter, we have an engineering version of the Nyquist sampling rate: fs 2.2fm. Example 4.1 What should be the sampling rates to produce a high-quality digitalization of a 20-kHz bandwidth music signal? Solution A sampling rate, of 44 ksps should be used. The sampling rate for a compact disc digital audio player  44.1 ksps and for a studio-quality audio player  48 ksps are used.

4.4.2 Quantization In Figure 4.10, each pulse is expressed as a level from a finite number of predetermined levels; each level can be represented by a symbol from a finite alphabet. The pulses in Figure 4.10 are called quantized samples. When the sample values are quantized to a finite set, this format can interface with a digital system. After quantization, the analog waveform can still be recovered, but not precisely; improved reconstruction fidelity of the analog waveform can be achieved by increasing the number of quantization levels (requiring increased system bandwidth). s(t)

T t Ts

Figure 4.10

Flat-top quantization.

94

4

An Overview of Digital Communication and Transmission

4.4.3 Sources of Error The quantization process introduces an error defined as the difference between the input and output signal. The error is referred to as quantization noise. Other sources of error are due to sampling and channel condition [2,13]. Quantization Noise The distortion inherent in quantization is a round-off or truncation error. The process of encoding the pulse amplitude modulation (PAM) waveform into a quantized waveform involves discarding some of the original analog information. This distortion is called quantization noise; the amount of such noise is inversely proportional to the number of quantization levels used in the process. Quantizer Saturation The quantizer allocates L levels to the task of approximating the continuous range of inputs with a finite set of outputs. The range of inputs for which the difference between the input and output is small is called the operating range of the converter. If the input exceeds this range, the difference between the input and output becomes large and we say that the converter is operating in saturation mode. Saturation errors are more objectionable than quantizing noise. Generally, saturation is avoided by use of automatic gain control, which effectively extends the operating range of the converter. Timing Jitter If there is a slight jitter in the position of the sample, the sampling is no longer uniform. The effect of the jitter is equivalent to frequency modulation of the baseband signal. If the jitter is random, a low-level wide band spectral contribution is induced whose properties are very close to those of the quantizing noise. Timing jitter can be controlled with good power supply isolation and stable clock reference. Channel Noise Thermal noise, interference from other users, and interference from circuit switching transients can cause errors in detecting the pulses carrying the digitized samples. Channel-induced errors can degrade the reconstructed signal quality quite significantly. The rapid degradation of the output signal quality with channelinduced errors is called a threshold effect. Inter-Symbol Interference A bandwidth-limited channel spreads a pulse waveform passing through it. When the channel bandwidth is much greater than the pulse bandwidth, the spreading of the pulse will be slight. When the channel bandwidth is close to the signal bandwidth, the spreading will exceed a symbol duration and cause signal pulses to overlap. This overlapping is called inter-symbol interference (ISI) (see Chapter 3).

4.4

Sampling Process

95

ISI causes system degradation (higher bit-error-rate); it is a particularly insidious form of interference because raising the signal power to overcome interference will not improve the error performance.

4.4.4 Uniform Quantization We consider a uniform quantization process as shown in Figure 4.11. With the quantizer input having zero mean, and the quantizer being symmetric, the quantizer output and quantization error will also have zero mean. Since the mean of quantization error e is zero, its variance will be: 2 



q2

q2

e2p(e)de 



q/2

q2 12

1 de   Average quantization noise power e2

q q/2

where: p(e)  probability of error  1/q ( for uniform distribution of quantization error) 2  variance of the average quantization noise power L  number of quantization levels.



V

 2 2

L 2

pp Signal power  (Vp)2 

 q

2

L2q2 4



(4.8)

where: Vpp  peak-to-peak voltage range 2

 p Signal power SNR 



2

V

Average quantization noise power



Vp Vp  q/2 Vp  q/2 q (volts) L levels

Vpp  peak-topeak voltage range

VP Figure 4.11

Uniform quantization.

96

4

An Overview of Digital Communication and Transmission

(L2q2) ⁄ 4

 3L  N q 

q2/12

S 

2

(4.9)

With L  2R, R  bits per sample, we obtain 2

2  2Vp

q  2 

R 12

2

1 (V )2 22R  /12 

p 3

Let P denote the average power of the message signal s(t), then P  3P (22R)  NS q 



2 (Vp)2 

(4.10)

The output signal-to-noise ratio (SNR) of the quantizer increases exponentially with increasing number of bits per sample, R. An increase in R requires a proportionate increase in the channel bandwidth. In the limit as L → , the signal approaches the PAM format (with no quantization error) and signal-to-quantization noise ratio is infinite. In other words, with an infinite number of quantization levels, there is no quantization error. Example 4.2 We consider a full-load sinusodial modulating signal of amplitude A, which uses all representation levels provided. If the average signal power corresponds to a load of 1 Ohm, determine the SNR for L  32, 64, 128, and 256. Solution 2

A P

2

1 A222R 2 

3

(SNR)q 

A2

2

1 A222R

3





3 2R 

(2 )  1.8  6R dB 2

L  2R

R (bits)

(SNR)q(dB)

32

5

31.8

64

6

37.8

128

7

43.8

256

8

49.8

4.5

Voice Communication

4.5

97

Voice Communication

For most voice communication, very low speech volumes predominate: 50% of the time, the voltage characterizing detected speech energy is less than 1/4 of the rms value of the voltage. Large amplitude values are relatively rare: only 15% of the time does the voltage exceed the rms value. Uniform quantization would be wasteful for speech signals; many of the quantizing steps would rarely be used. In a system that uses equally spaced quantization levels, the quantization noise is the same for all signal magnitudes because noise depends on the step size of quantization. The uniform quantization results in poor signal to quantization noise ratios for low-amplitude signals. Nonuniform quantization can provide better quantization of the weak signals and coarse quantization of the strong signals. Thus, in the case of nonuniform quantization, quantization noise can be made proportional to signal magnitude. The effect is to improve overall SNR by reducing the noise for predominant weak signals, at the expense of an increase in noise for rarely occurring signals. The nonuniform form of quantization is used to make the SNR a constant for all signals within the input range. Nonuniform quantization is achieved by first distorting the original signal with a logarithmic compression characteristic, and then using a uniform quantizer. For small magnitude signals the compression characteristic has a much steeper slope than large magnitude signals. Thus, a given signal change at small magnitudes will carry the uniform quantizer through more steps than the same change at large magnitudes. The compression characteristic effectively changes the distribution of the input signal magnitude. By compression, the low amplitudes are scaled up while the high amplitudes are scaled down. After compression, the distorted signal is used as input to the uniform quantizer. Thus, we achieve nonuniform quantization. There are two compression algorithms commonly used, the -law and the A-law [6,10]. We use a device called expander at the receiver with characteristics complementary to the compression. It is used so that the overall transmission is not distorted. The whole process (compression and expansion) is called companding [7]. The  - and A-law compression characteristics as given below are used. • -Law compression characteristic used in North America is given as: log 1  |in| 

|out| 

log{1  }

(4.11)

where: in and out are the normalized input and output voltages, respectively   a positive constant   0 represents uniform quantization,   255 is used in North America

98

4

An Overview of Digital Communication and Transmission

• A-Law compression characteristic used in Europe and most of the rest of

the world is given as: A|in|

|out| 

1  logA

1  logA|in|

|out| 

1  logA

1 0  |in| 

A

1 A

 |in|  1

(4.12a)

(4.12b)

where A is the positive constant. A  87.6 is the standard value used in Europe. The case of uniform quantization corresponds to A  1.

4.6

Pulse Amplitude Modulation (PAM)

Pulse amplitude modulation [6] is a process that represents a continuous analog signal with a series of discrete analog pulses in which the amplitude of the information signal at a given time is coded as a binary number. PAM is now rarely used, having been largely superseded by pulse code modulation (PCM). Two operations involved in the generation of the PAM signal are: 1. Instantaneous sampling of the message signal s(t) every Ts seconds, where fs  1/Ts is selected according to the sampling theorem. 2. Lengthening the duration of each sample obtained to some constant value T. These operations are jointly referred to as sample and hold. One important reason for intentionally lengthening the duration of each sample is to avoid the use of an excessive channel bandwidth, since bandwidth is inversely proportional to pulse duration. The Fourier transform of the rectangular pulse h(t) is given as (see Figure 4.12): S(f )  Tsinc ( f T)ej2fT

(4.13)

Using flat-top sampling of an analog signal with a sample-and-hold circuit such that the sample has the same amplitude for its whole duration introduces amplitude distortion as well as a delay. This effect is similar to the variation in transmission frequency that is caused by the finite size of the scanning aperture in television. The distortion caused by the use of PAM to transmit an analog signal is called the aperture effect. The distortion may be corrected by use of an equalizer (see Figure 4.13). The equalizer decreases the in-band loss of the reconstruction filter as the frequency increases in such a manner to compensate for the aperture effect. The amount of equalization required in practice is usually small. For T/Ts  0.1, the amplitude distortion is less than 0.5%, in which case the need for equalization may be omitted altogether.

4.6

Pulse Amplitude Modulation (PAM)

99

T h(t) S(f)

1.0

Spectrum Magnitude

0

3/T

T

1/T

0

1/T

3/T

Pulse arg [S(f)]

 Spectrum Phase

3/T

Figure 4.12

1/T

1/T

3/T

Rectangular pulse and its spectrum.

PAM signal s(t)

Figure 4.13

Filter (LPF)

Equalizer

Message Signal m(t)

An equalizer application.

Example 4.3 Twenty-four voice signals are sampled uniformly and then time-division multiplexed. The sampling operation is flat-top samples with 1 s duration. The multiplexing operation includes provision for synchronization by adding an extra pulse of sufficient amplitude and also 1 s duration. The highest frequency components of each voice signal is 3.4 kHz. a. Assuming a sampling rate of 8 kHz, calculate the spacing between successive pulses of the multiplexed signal. b. Repeat your calculations using the engineering version of Nyquist rate sampling.

100

4

An Overview of Digital Communication and Transmission

Solution 1  125 s a. Ts 

8000

25 channels (24 voice channels  1 synchronization channel), time allocated for each channel  125/25  5 s. Since the pulse duration is 1 s, the time between pulses is (5  1)  4 s b. Nyquist rate  7.48 kHz (2.2  3.4) 1  134 s Ts 

7480

134 Tc 

 5.36 s 25

The time between pulses is (5.36  1)  4.36 s.

4.7

Pulse Code Modulation

Pulse code modulation (PCM) [13] is a digital scheme for transmitting analog data. It converts an analog signal into digital form. Using PCM, it is possible to digitize all forms of analog data, including full-motion video, voice, music, telemetry, etc. To obtain a PCM signal from an analog signal at the source (transmitter) of a communications circuit, the analog signal is sampled at regular time intervals. The sampling rate is several times the maximum frequency of the analog signal. The instantaneous amplitude of the analog signal at each sample is rounded off to the nearest of several specific, predetermined levels (quantization). The number of levels is always a power of 2. The output of a pulse code modulator is a series of binary numbers, each represented by some power of 2 bits. At the destination of the communications circuit, the pulse code modulator converts the binary numbers back into pulses having the same quantum levels as those in the modulator. These pulses are further processed to restore the original analog waveform. When pulse modulation is applied to a binary symbol, the resulting binary wave form is called a pulse code modulation waveform. When pulse modulation is applied to a nonbinary symbol, the resulting waveform is called M-ary pulse modulation waveform. Each analog sample is transmitted into a PCM word consisting of groups of b bits. The PCM word size can be described by the number of quantization levels that are used for each sample. The choice of the number of quantization levels, or bits per sample, depends on the magnitude of quantization distortion that one is willing to tolerate with the PCM format. In North America and Japan, PCM samples the analog waveform 8000 times per second and converts each sample into an 8-bit number, resulting in a 64 kbps data stream. The sample rate is twice the 4 kHz bandwidth required for a toll-quality voice conversion.

4.7

Pulse Code Modulation

101

Differential (or Delta) PCM (DPCM) encodes the PCM values as differences between the current and the previous value. For audio, this type of encoding reduces the number of bits required per sample by about 25% compared to PCM. Adaptive Differential PCM (ADPCM) is a variant of DPCM that varies the size of the quantization step to allow further reduction of the required bandwidth for a given signal-to-noise ratio. Example 4.4 The information in an analog waveform with a maximum frequency of fm  3 kHz is transmitted over an M-ary PCM system, where the number of pulse levels is M  25  32. The quantization distortion is specified not to exceed 1% of the peak-to-peak analog signal. a. What is the minimum number of bits/sample or bits/PCM word that should be used? b. What is the minimum sampling rate, and what is the resulting transmission rate? c. What is the PCM pulse or symbol transmission rate? Solution |e|  pVpp

where: p  the fraction of the peak-to-peak analog voltage, Vpp |e|  the magnitude of quantization distortion error specified as a fraction p of the peak-to-peak analog voltage Vpp Vpp

|emax| 

2L





Vpp

 pVpp

2L



1 2R  L

2p

where R is the number of bits required to represent quantization levels L 1 2R

 50 2  0.01

 (R 5.64)

use R  6

fs  2fm  6000 samples per second

102

4

An Overview of Digital Communication and Transmission

fs  6  6000  36 kbps M  2b  32 b  5 bits per symbol 36,000 5

Rs   7200 symbols/sec.

4.8

Shannon Limit

For a Gaussian channel (with additive white Gaussian noise [AWGN]) a theoretical limit to the maximum data rate (Rb) that can be transmitted in a channel with a given bandwidth (Bw) is given by the Shannon theorem as [11,12]:



S C  Bwlog2 1 



C Bw

N

Eb N0



Rb Bw

 log2 1  

(4.14)



(4.15)

where: Eb/N0  signal-to-noise density ratio At channel capacity Rb  C Let C/Bw   (spectral efficiency of the channel), then Equation 4.15 can be written as: 1 E  b 2

 N

(4.16)

0

where: C  channel capacity (bits per second) S  signal strength N  noise power N0  noise density In the limit, as  → 0, we get Eb N0



1 (2  loge2) loge2 0.6931 1.592 dB  2

  lim →0 →0

lim

(4.17)

Figure 4.14 shows  versus Eb/N0 for the Gaussian channel. The Shannon capacity bound states that there exists a coding/modulation scheme that achieves, over the AWGN channel, an arbitrarily low probability of bit-error, Pb provided that Rb C(SNR)

(4.18)

4.9

Modulation

103

Capacity Boundary C=R Spectral efficiency of channel 

Region for which R < C

8 Bandwidth Limited Region

Shannon Limit 4 2 1.6 1

1

6

12

Eb /N0 (dB)

18 Power Limited Region

1/2 1/4

Figure 4.14 ␩ versus Eb/N0.

4.9

Modulation

Baseband signals are generated at low rates, therefore these signals are modulated onto a radio frequency carrier for transmission. Baseband signal s(t) is complex, and can be represented mathematically as s(t)  a(t)ej(t)

(4.19)

where: a(t) is the amplitude and (t) is the phase Assuming a sampling rate the same as the Nyquist rate, the low-pass reconstruction filter extends from fm to fm (see Figure 4.15). The maximum frequency fm of s(t) is an approximate measure of its bandwidth. The Fourier transform of s(t) is given by S(f ) 







s(t)ej2ft dt

(4.20)

A functional block diagram of a generic modulation procedure for signal s(t) is given in Figure 4.16. t

x(t)  Real [s(t)Ac e j2f c]  Ac a(t)cos[2fc t  (t)]

(4.21)

x(t)  Ac a(t)cos(t)cos(2fc t)  Ac a(t)sin(t)sin(2fc t)

(4.22)

104

4

An Overview of Digital Communication and Transmission

S(f )

fm

Figure 4.15

f fm

Power spectral density (PSD).

s(t)  a(t)ej2f(t) Ac  amplitude of carrier fc  carrier frequency

Figure 4.16

0

Modulator

x(t)  Real {s(t)Acej2fct}

Carrier Acej2fct

Functional block diagram of a generic modulator.

The modulation can be classified as linear modulation or nonlinear modulation. A modulation process is linear when a(t)cos(t) and a(t)sin(t) are linearly related to the message information signal. Examples of linear modulation are amplitude modulation, where the modulating signal affects only the amplitude of the modulated signal (i.e., (t) is constant for any t), and phase modulation, where the modulating signal affects only the phase of the modulated signal (i.e., when (t) is a constant over each signaling (symbol) interval and a(t) is constant for any t). The modulation process is nonlinear when the modulating signal s(t) affects the frequency of the modulated signal. The definition of a nonlinear system is that superposition does not apply. The modulation process is nonlinear whether or not the amplitude of the modulating signal is a function of time. We consider a frequency modulation process and let a(t)  a for any t. Then, the nonlinear modulated signal is x(t)  aAc cos[2fct  (t)], where (t) is the integral of a frequency function. Selection of modulation and demodulation schemes is based on spectral efficiency, power efficiency, and fading immunity. During the late 1970s and early 1980s, constant envelope modulation schemes were used for cellular systems to achieve a high-power efficient terminal with a C class amplifier. As a result, Gaussian minimum shift keying (GMSK) (see Chapter 9) is the widely used modulation scheme in the GSM and DECT systems. In the mid-1980s, when cellular systems’ capacity became a serious problem, developments of linear modulations with two bits per second per Hz (bps/Hz) transmission capability were initiated. To apply a linear modulation in a wireless communication system, we need high spectral efficiency as well as high-power efficiency

4.10

Performance Parameters of Coding and Modulation Scheme

105

at the same time. /4-quadrature phase shift keying (QPSK) (see Chapter 9) was used in the Japanese and North American digital cellular and personal systems. More details of modulation schemes used in various wireless systems are given in Chapter 9.

4.10

Performance Parameters of Coding and Modulation Scheme

The most important parameter of a coding and modulation scheme is the bandwidth requirement, which is determined by the spectrum of the modulated signal usually presented as a plot of power spectral density (PSD) against frequency (see Figure 4.17) [1]. Ideally, the PSD should be zero outside the band occupied. However, in practice this can never be achieved, and the spectrum extends to infinity beyond the band. This is either because of the inherent characteristics of the modulation scheme, or because of the practical implementation of filters. Hence, we must define the bandwidth, Bw, such that the signal power falling outside the band is below a specified threshold. In practice, this threshold is determined by the tolerance of the system to adjacent channel interference. The coding and modulation selection should be based on the following factors: • bit error rate probability, Pb • bandwidth efficiency, 

Frequency (f )

PSD

Nominal Carrier Frequency

Bandwidth: Bw

Figure 4.17

PSD versus frequency.

106

4

An Overview of Digital Communication and Transmission

• signal-to-noise density ratio, Eb/N0 (Eb is the energy per bit and N0 is the

noise density) • complexity of transmitter/receiver The bandwidth efficiency (or spectrum efficiency), , of a coding and modulation scheme determines the bandwidth requirement. This is defined as the information bit rate, Rb, per unit bandwidth occupied, and is measured in bits/sec/Hz (bps/Hz). R Bw

  b

(4.23)

An ideal coding and modulation system should provide a small Pb with a high bandwidth efficiency () and a low signal-to-noise density ratio (Eb/N0). The information rate, Rb, is related to the number of waveforms, M, used by the modulator and the duration of these waveforms, Ts log2M Ts

Rb 

(4.24)

The average power used by the modulator is P  Es/Ts, where Es is the average energy of the modulator signals. Each signal carries a log2 M information bit. Eblog2M Ts

 P   EbRb

(4.25)

The signal-to-noise ratio (SNR) is the ratio between the average signal power and the average noise power over the signal bandwidth

N  B 

P  Eb  Rb SNR 



N0Bw

(4.26)

w

0

Equation 4.26 shows that SNR is the product of (Eb/N0) and (Rb/Bw), the bandwidth (or spectral) efficiency of a modulation scheme. Table 4.1 lists the bandwidth efficiencies of several 2G wireless systems. Table 4.1 Bandwidth efficiencies of wireless systems. Wireless System

Rb (kb/s)

Bw (kHz)

Rb/Bw (bps/Hz)

CT2

72

100

0.72

GSM

270.8

200

1.354

IS-54

48.6

30

1.62

PDC

42.0

25

1.68

IS-95-CDMA

1228

1230

0.998

4.10

Performance Parameters of Coding and Modulation Scheme

107

For high SNR, the error probability can be closely approximated by a

complementary error function erfc[dmin/(2 N0 )], where dmin is the minimum Euclidean distance between any two elements of the modulator signal set [1]. It is related to the power efficiency () of a modulation scheme. The parameter  expresses how efficiently a modulation scheme uses the available signal energy to generate minimum Euclidean distance.  is defined as: d2 4Eb

min 

(4.27)

Equation 4.27 provides an approximation to error probability that is asymptotically tight for large SNR. M  1 erfc Pb

2



E  

N 

(4.28)

b

0

Thus, for high signal-to-noise ratios, a modulation scheme is better if its power efficiency is greater. Table 4.2 lists bandwidth and power efficiency of several M-ary modulation schemes. The BER performance of a coding and modulation scheme is an important parameter. The shape of Eb/N0 versus BER (Pb) curves depends on the channel and the modulation scheme. Ideally, the BER of the service offered to the user should be zero, but it is not possible in practice. Hence, we must specify a BER to obtain the required Eb/N0. There is an inherent trade-off between  and the Eb/N0 requirement of a coding and modulation scheme. The greater the bandwidth efficiency, the greater the required Eb/N0 is likely to be. Hence, it is usually possible to increase the capacity of a communication system for a given bandwidth allocation by increasing the signal power.

Table 4.2 Bandwidth and power efficiency of M-ary modulation scheme [1]. Modulation

Rb/Bw



Pulse Amplitude Modulation (PAM)

2 log2 M

3 log2 M/(M 2ⴚ1)

Phase Shift Keying (PSK)

log2 M

sin2(␲/M ) log2 M

Quadrature Amplitude Modulation (QAM)

log2 M

3 log2 M/(2Mⴚ1)

Frequency Shift Keying (FSK)

2 log2 M/M

1/2 log2 M

108

4.11

4

An Overview of Digital Communication and Transmission

Power Limited and Bandwidth-Limited Channel

The objective of a general communication system design is to use transmitted power and channel bandwidth as efficiently as possible. In most communication channels, one resource may be more important than another. We may therefore classify communication channels as power-limited or bandwidthlimited. For a power-limited channel, the desired system performance is obtained with the smallest possible power. One choice is the use of standard error control codes [5] which increase the power efficiency by adding extra bits to the transmitted symbol sequence. This procedure requires the modulator to operate at a higher data rate and, hence, requires a larger bandwidth. In a bandwidth-limited channel, increased efficiency in both power and frequency utilization is obtained by selecting an integrated coding and modulation solution, where higher-order modulation schemes (e.g., 8-PSK instead of 4-PSK; see Chapter 9) are combined with low-complexity coding schemes. Trellis codes [9] for bandwidth-limited channels result from the treatment of modulation and coding as a combined entity rather than as two separate operations. The trellis-coded-modulation (TCM) solution combines the choice of a modulation scheme with that of a convolutional code, while the receiver, instead of performing demodulation and decoding in two separate steps, combines the two operations into one. Table 4.3 summarizes some of the energy savings (coding gains) in dB that can be obtained by doubling the constellation size and using TCM. These are considered for coded 8-PSK (relative to uncoded 4-PSK) and for coded 16-QAM (relative to uncoded 8-PSK). These gains can be achieved only for high signal-to-noise ratios, and decrease as the signal-to-noise ratios decrease. The complexity of decoder/demodulator is proportional to the number of states.

Table 4.3 Coding gain of TCM [1]. Coding gain (dB) (8-PSK)

Coding gain (dB) (16-QAM)

4

3.0

4.4

No. of states

8

3.6

5.3

16

4.1

6.1

32

4.6

6.1

64

4.8

6.8

128

5.0

7.4

256

5.4

7.4

4.12

Nyquist Bandwidth

4.12

109

Nyquist Bandwidth

As discussed earlier, Nyquist showed that the theoretical minimum bandwidth (Nyquist bandwidth) required for the baseband transmission of Rs symbols per second without inter-symbol interference (ISI) is Rs/2 Hz. In practice, the Nyquist minimum bandwidth is increased by about 10 to 20% because of the constraints of real filters. Thus, typical baseband digital communication throughput is reduced from the ideal two symbols per Hz to a range of about 1.8 to 1.6 symbols per Hz. Assuming Nyquist filtering at baseband, the required bandwidth is related to the symbol rate by (1  )  R (1  ) Bw 

s Ts

(4.29)

where:   roll-off factor (may vary between 0 and 1). This allows a trade-off between bandwidth and desirable time-domain properties of the signal. Very low values are difficult to realize in practice because they require very rapid roll-off in the filter response and because of the ringing in time-domain response. From the set of M symbols, the coding system assigns to each symbol b bits, where M  2b. The number of bits per symbol will be: b  log2M Rb  bRs R b

R log2M

b Rs  b 

(4.30)

For signaling at a fixed symbol rate as b is increased, the data rate Rb is increased. In the case of M-PSK, increasing b results in an increased bandwidth efficiency . In other words, with the same system bandwidth, we can transmit M-PSK signals at an increased data rate and hence at an increased . The effective duration Tb of each bit in terms of the symbol duration, Ts, or the symbol rate, Rs, is T b

T log2M

s 1  s Tb 



Rb

Rb  log2M  l 



bits/s/Hz Bw

Bw Ts

Bw Tb

(4.31)

From Equation 4.31, we can observe that a digital communication system becomes more bandwidth efficient as its BwTb product is decreased. Thus, signals with small BwTb product are often used with a bandwidth-limited system. For example the Global System of Mobile (GSM) communication uses GMSK modulation

110

4

An Overview of Digital Communication and Transmission

with a BwTb product equal to 0.3, where Bw is the 3 dB bandwidth of a Gaussian filter (see Chapter 9). For uncoded bandwidth-limited systems, the objective is to maximize the transmitted information rate within the allowable bandwidth, at the expense of a signal-to-noise density ratio (Eb/N0) (while maintaining a specified value of BER). For the case of power-limited systems, power is scarce but system bandwidth is available. The following trade-offs are possible: (1) improved BER at the expense of bandwidth for a fixed signal-to-noise density ratio (Eb/N0); or (2) reduction in signal-to-noise density ratio (Eb/N0) at the expense of bandwidth for a fixed BER. A natural modulation choice for a power-limited system is M-FSK (see Chapter 9). Using Equations 4.29 and 4.31, we get log2M Rb

 bit/s/Hz (1  )

Bw

(4.32)

The M-PSK modulation is a bandwidth-efficient scheme. As M increases in value, Rb/Bw also increases. Thus M-PSK modulation can achieve improved bandwidth efficiency at the cost of increased signal to noise density ratio (Eb/N0). For M-FSK, the Nyquist minimum bandwidth is given as M 1   

Bw   MRs 1    TS

(4.33)

From Equations 4.32 and 4.30, we get Rb

log2M



Bw M 1   

(4.34)

For simplicity, the modulation choice is limited to constant-envelope types — either M-PSK or noncoherent orthogonal M-FSK. Thus, for an uncoded system, if the channel is bandwidth limited, M-PSK is selected, and if the channel is power limited, M-FSK is selected. Note that, when we consider error-correction coding, the selection of a modulation scheme is not so simple because there may exist coding schemes which can provide power-bandwidth trade-off more effectively than would be possible using any M-ary modulation scheme. Example 4.5 In a digital communication system, the received power to noise-power spectral density (S/N0) is 53 dB-Hz, the required data rate is 9.6 kbps and the available bandwidth is 4.8 kHz. The required BER performance Pb is 105. What is the design choice of the modulation scheme if no error-correction coding is used?

4.12

Nyquist Bandwidth

111

Solution Since the required data rate of 9.6 kbps is more than the available bandwidth of 4.8 kHz, the channel is bandwidth-limited. Eb

S (dB  Hz)  R(dB)

(dB) 

N N0

0

E N0

 b  53  10 log9600  20.78 dB Try the 8-PSK modulation scheme Ps  (log2M)  Pb  (log28)  105  3  105 Es N0

Eb N0

 (log28)   3  20.78  62.34

  NE   sin 8    2Q(4.28)  2.2  10

Ps(8)  2Q 2 s

5



0

3  105 (see Chapter 9)

Therefore, use 8-PSK modulation. Example 4.6 In a digital communication system, the signal-to-noise spectral density ratio is 48 dB-Hz, the available bandwidth is equal to 45 kHz, and the data rate is 9.6 kbps. The required BER performance, Pb, is 105. What is the design choice of the modulation scheme without an error-correction coding? Solution In this case, the channel is not bandwidth limited since the available bandwidth of 45 kHz is more than adequate to support the required data rate of 9.6 kbps. Eb

S (dB  Hz)  R(dB)

(dB) 

N N0

0

E N0

 b  48  10 log 9600  6.61 dB We try the 16-FSK modulation scheme E N0

E N0

 s  (log216)  b  4  6.61  26.44 Es



15 13.22 M  1 e 2N0 

Ps (M) 

e  1.36  105 (see Chapter 9) 2

2

112

4

An Overview of Digital Communication and Transmission

For orthogonal signaling  1 ខ Pb  65,535  105  2  105 (see Chapter 9) Ps  2

M

2M – 1

32,768

We can meet the given specifications for this power-limited channel with a 16-FSK modulation scheme without any error-correction coding.

4.13

OSI Model

Coding and modulation are the primary tasks of the OSI physical layer, which adapts the information transmitted to it from the higher layers into a form that can be transmitted over the physical medium. Some functions of coding and modulation, such as error control [15] and multiple access, appear higher up the OSI stack. Table 4.4 summarizes network-related functions performed by the first three OSI layers.

4.13.1 OSI Upper Layers The role of the OSI upper layers (Transport, Session, Presentation, and Application) are summarized as follows [3,14]: The transport layer segments the messages into packets of acceptable sizes and performs the reassembly at the destination. It may multiplex many low-rate transmissions onto one virtual circuit or divide a high-rate transmission into parallel virtual circuits. The transport layer controls transmission errors and requests retransmissions of packets corrupted by transmission errors. In addition, the flow may be controlled by some mechanism to prevent one host from sending data faster than the destination host can handle. The session layer sets up the call and takes care of the authentication of the user and of billing. The session layer supervises the synchronization (packet Table 4.4 Functions of first four OSI layers. OSI layers

Function

Network Layer

Routing

Data Link Layer • Logical Link Control (LLC)

Error Control

• Media Access Control (MAC)

Multiple Access protocols

Physical Layer

Modulation, Forward Error Correction (FEC), Encryption, Equalization, Synchronization

Physical Medium

Radio Propagating Mechanism

4.14

Data Communication Services

113

numbering) and the recovery in case of failures. The session layer closes the session at the end of transmission. The presentation layer asks the session layer to set up a call. It specifies the destination’s name and type of transmission (e.g., datagram, high priority). The presentation layer translates between the local syntax used by the application process and transfer syntax. The presentation layer also performs the required encryption and data compression. The application layer provides information transfer services for user application programs. The user interacts with the application layer through a user interface. The application layer is composed of Specific Application Service Elements (SASEs) that use the services of Common Application Service Elements (CASEs). A CASE establishes are association between SASEs and may include an Association Control Service Element (ACSE), a Remote Operation Service Element (ROSE), and the Commitment Concurrency and Recovery (CCR) element.

4.14

Data Communication Services

End-to-end communication services are classified as either synchronous (sync) or asynchronous (async) [2]. A sync communication service delivers a bit stream with a fixed delay and a given bit error rate. The voice communication is an example of the sync communication service. The sync delivery of a 64 kbps voice bit stream can be implemented by dividing the bit stream into packets that are received with random delays and are stored in a buffer to hold the bits until they are delivered. This implementation of a sync transmission service is called packetized-voice. In packetized-voice, a buffer is used to absorb the random fluctuations in the packet transmission delays. Another implementation of the sync transmission of the bit stream is to use a dedicated coaxial cable that propagates the bits one after the other, all with the same delay. In an async communication service, the bit stream to be transferred is divided into packets. The packets are received by the destination with varying delays, and a fraction of them may not be received correctly at the destination. An async communication service is evaluated by its QoS. The QoS deals with parameters, such as the packet error rate, delay, throughput, reliability, and security of the communication. There are two classes of async communication services: connection-oriented and connectionless. A connection-oriented communication service delivers the packet in sequence, i.e., in the correct order and confirms the delivery. Depending on the QoS requirements, the delivery may be guaranteed to be error-free. Thus, connection-oriented service looks from end-to-end like a dedicated link, which may be noiseless or noisy. A connectionless communication service delivers the packets individually. The packets can be delivered out of order, and some may contain errors and others may be lost. Some connectionless services provide an acknowledgment (ACK) of the correctly delivered packets. The

114

4

An Overview of Digital Communication and Transmission

connectionless services are similar to the mail service provided by the post office: letters may be delivered out of order; normal mail delivery does not guarantee or acknowledge the delivery. Another class of connection-oriented communication service is also used in some applications. It is called expedited data, and corresponds to a potentially faster delivery of certain packets, usually by making them jump to the head of the queues of packets that are waiting to be transmitted. Communication services are implemented by transporting bits over the network. One essential objective of the bit transport is connectivity, in which one network user should be able to exchange information with many other users. It should be possible to route the bits of one user to any one of a large number of other users. The property to vary the path followed by the bits is called switching. Three basic methods used for switching bits in communication networks are: • circuit-switching • virtual-circuit packet switching • datagram packet switching

In circuit-switching, the switch connects transmission paths to establish a circuit between transmitter and receiver. Circuit-switching is quite suitable for continuous data transmission services. It is a sync service. A packet-switched network uses another scheme. The nodes of the network, packet-switching nodes, play a role similar to that of switches in a circuit-switched network. Packet-switched networks can use two different methods for selecting the path followed by the packets: virtual-circuit (VC) and datagram. In the VC transport, the different packets that are part of the same information transfer are sent along the same path. The packets follow one another as if they were using a dedicated circuit even though they may be interleaved with other packet streams. Some implementations of VC perform an error control on each link between successive nodes. Thus, not only are the packets delivered in sequence by each node to the next node along the path, but they are also transmitted without errors. This is implemented by each node checking the correctness of the packets it receives and asking the previous node along the path to retransmit incorrect packets. VC packet-switching does not need a buffer at the destination. VC packet-switching is a sync service. Since multiple virtual circuits may exist between the source-destination pair, routing cannot be done on the basis of the source-destination address only. Data packets must carry an indication of VC identification as well. Routing is done on the basis of explicit route number and destination address. An explicit routing table at each node associates an appropriate outgoing transmission group with the destination address and explicit route number. By changing the explicit route number for a given destination, a new path will be followed. This introduces alternative route capability. If a link or node along the path becomes inoperative, any session using that path can be reestablished on an explicit route by bypassing the failed element. Explicit routes can also be assigned on the basis of type of

4.15

Multiplexing

115

traffic, type of physical media along the path (satellite or terrestrial, for example), or other criteria. Routes could also be listed on the basis of cost, the smallest-cost route being assigned first, then the next-smallest-cost route, and so forth. In datagram packet-switching, the bits are grouped as packets. Each packet is labeled with the address of its destination. The packets are routed independently of one another and arrive at destination out-of-sequence. The datagram packet switching requires buffers at the source and destination. In datagram packet switching networks, each network node keeps a complete (global) topological database that is updated regularly as topological changes occur. Generally, the routing philosophy of datagram networks is to route packets (datagram) along paths of minimum time delay. The datagram is an async service.

4.15

Multiplexing

Multiplexing [2] refers to a variety of techniques that are used to make an efficient use of transmission facility. In many cases, the capacity of transmission facility exceeds the requirements for the transfer of data between two devices. That capacity can be shared among multiple transmitters by multiplexing a number of signals onto the same medium. In this case, the actual transmission path is called a circuit or link, and the portion of capacity devoted to each pair of transmitter/receivers is called a channel. There are three types of multiplexing techniques: frequency-division multiplexing (FDM), synchronous time-division multiplexing (TDM), and improved synchronous TDM. FDM is the most widespread. TDM is commonly used for multiplexing digitized voice streams. The third type improves the efficiency of synchronous TDM. It is known by various names: • Statistical TDM • Asynchronous TDM • Intelligent TDM

Figure 4.18 shows the concept of multiplexing. We have m inputs to a multiplexer. The multiplexer is connected by a single data link to a demultiplexer. m inputs

MUX

Figure 4.18

Multiplexing.

m outputs

One link, m channels

DEMUX

116

4

An Overview of Digital Communication and Transmission

The link carries m separate channels of data. The multiplexer combines data from m input lines and transmits over a higher-capacity data link. The demultiplexer accepts the combined data stream, separates the data according to channel, and delivers them to the appropriate output lines.

4.16

Transmission Media

The transmission medium [2] is the physical path between the transmitter and receiver in a communication system. The characteristics and quality of transmission depend on the nature of the signal and the nature of the medium. Two general ranges of frequencies are of interest. Microwave frequencies cover a range of about 2 to 40 GHz. At these frequencies, highly directional beams are possible. Microwave is quite suitable for point-to-point transmission. Signals in the frequency range 30 MHz to 1 GHz are referred to as radio waves. Omnidirectional transmission is used and signals at these frequencies are often employed for broadcast applications. The following are the commonly used transmission mediums used in wired and wireless communications: • Twisted pair • Coaxial cable • Fiber optics • Terrestrial microwave • Radio • Satellite

A twisted pair contains two insulated copper wires arranged in a regular spiral pattern. A wire pair acts as a single communication link. Typically, a number of these pairs are bundled together into a cable by wrapping them in a tough protective sheath. Over longer distances, cables may contain hundreds of pairs. The twisting of the individual pairs minimizes electromagnetic interference between the pairs. The wires in a pair have a thickness of 0.016 to 0.036 in. The twisted pair is by far the most common medium for both analog and digital data transmission. It is the backbone of the telephone system as well as the workhorse for intra-building communications. A coaxial cable is simply a transmission line consisting of a pair made up of an inner conductor surrounded by a grounded outer conductor, which is held in a concentric configuration by a dielectric. Systems have been designed to use coaxial cable as a transmission medium with a capability of transmitting an FDM configuration from 120 to 10,800 voice channels. Community antenna television (CATV) systems use single cable for transmitting a bandwidth of the order of 300 MHz. One of the advantages of coaxial cable systems is reduced noise accumulation when compared to radio links. A coaxial cable system is attractive for television transmission or other video applications.

4.16

Transmission Media

117

Fiber optics as a transmission medium has a comparatively unlimited bandwidth. It has excellent attenuation properties — as low as 0.25 dB/km. A major advantage of fiber optics compared to coaxial cable is that no equalization is needed. Also repeaters separation of the order of 10 to 1000 times that of coaxial cable for equal transmission bandwidths can be used. Other advantages of fiber optics are: • Electromagnetic immunity • Ground loop elimination • Security • Small size and light weight • Expansion capabilities require change out of electronics only, in most

cases. Fiber optics uses three wavelength bands: around 800, 1300, and 1600 nm or near-visible infrared. Fiber optics has analog transmission applications, particularly for video/TV. The primary use of terrestrial microwaves is in long-haul telecommunications service as an alternative to coaxial cable for transmitting television and voice. Microwaves can support high data rates over long distances. The microwave requires far fewer amplifiers or repeaters than coaxial cable for the same distance, but requires line-of-sight transmission. One of the potential uses for terrestrial microwaves is to provide digital data transmission in small regions (radius 10 km). Another common use of microwaves is for short point-to-point links between buildings. It can also be used as a data link between local area networks. The principal difference between radio and microwaves is that radio is omnidirectional and microwaves are focused. Radio does not require dish-shaped antennas, and the antennas need not be rigidly mounted to precise alignment. Radio covers VHF and some UHF bands: 30 MHz to 1 GHz. A primary source of impairment for radio waves is multipath interference. Reflection from land, water, and natural or human-made objects can create multiple paths between antennas. This effect is frequently evident when TV reception displays multiple images as a plane passes by. Satellite communication is nothing more than a radio link communication using one or two RF repeaters located at great distances from terminal earth stations. If the range from an earth antenna to a satellite is the same as the satellite altitude, the round-trip delay is around 500 ms, which is more than that encountered in conventional terrestrial systems. Thus, one major problem is propagation time and the resulting echo on telephone circuits. To reply for packet transmission systems affects the delay and requires careful selection of telephone signaling systems, otherwise the call setup time may become excessive. The equatorial orbit is filling with geostationary satellites and RF interference from one satellite system to another is increasing. The most desirable frequency bands for commercial satellite communication are in the spectrum 1000–10,000 MHz.

118

4

4.17

An Overview of Digital Communication and Transmission

Transmission Impairments

The nominal voice channel occupies the band 0–4 kHz. The CCITT (ITU) voice channel occupies the band 300–3400 Hz. There are four other parameters that are used to characterize the voice channel: 1. 2. 3. 4.

Attenuation distortion Phase distortion Level (signal power level) Noise and SNR

4.17.1 Attenuation Distortion A communication channel suffers various types of distortion, i.e., the output signal from the channel is distorted in some manner such that it is not an exact replica of the input. Attenuation distortion can be avoided if all frequencies within the passband are subjected to exactly the same loss (or gain). One type of distortion is referred to as attenuation distortion and is the result of imperfect amplitude-frequency response. 4.17.2 Phase Distortion A voice channel may be regarded as a band-pass filter. A signal takes a finite time to pass through a filter. This time is a function of the velocity of propagation, which varies with the medium involved. The velocity also tends to vary with frequency because of the electrical characteristics associated with it. The finite time a signal takes to pass through the total extension of a voice channel or any network is called delay. Absolute delay is the delay a signal experiences while passing through the channel at a reference frequency. The propagation time is different for different frequencies with the wave front of one frequency arriving prior to the wave front of another in the passband. Thus, the phase is shifted or distorted. If the phase shift is uniform with respect to frequency, a modulated signal will not be distorted, but if the phase shift is nonlinear with respect to frequency, the output signal is distorted compared to the input. 4.17.3 Level System levels are used to engineer a communication system. A zero test-level point is established. A zero test-level is the point at which the test-tone level should be 0 dBm. From zero test-level other points may be shown using dBr (decibel reference). The dBm and dBr are related as: dBm  dBm0  dBr

(4.35)

For example, a value of 30 dBm at a 20 dBr point corresponds to a reference level of 10 dBm0. A 10 dBm0 signal introduced at 0 dBr point (zero test-level) has an absolute signal level of 10 dBm.

4.17

Transmission Impairments

4.17.4

119

Noise and SNR

Noise Noise can be put into the following categories: • Thermal noise • Inter modulation (IM) noise • Crosstalk • Impulse noise

Thermal noise occurs in all transmission media and all communication equipment. It occurs due to random electron motion and is characterized by a uniform distribution of energy over the frequency spectrum with a Gaussian distribution of levels. The white noise refers to the average uniform spectral distribution of energy with respect to frequency. Thermal noise is directly proportional to bandwidth (Bw) and temperature (T). The amount of thermal noise to be found in 1 Hz of bandwidth in an actual device is given as: N0  kT (W/Hz)

(4.36)

where: k  Boltzmann’s constant (1.3803  1023 J/K) T  absolute temperature (K) At room temperature T  290 K N0  1.3803  290  1023  204 dBW/Hz  174 dBm/Hz

The noise power at temperature T with bandwidth Bw is given as: N  kTBw N  198.6 dBW  10 log T  10 log Bw

(4.37)

Inter modulation (IM) noise is the result of the presence of IM products. If two signals with frequencies f1 and f2 are passed through a nonlinear device or medium, the result will be IM products that are spurious frequency components. These components may be present either inside or outside the band of interest for the device. IM noise may result from a number of causes: • Improper level setting. If the level of input to a device is too high, the device

is driven into its nonlinear operating region • Improper alignment causing a device to function nonlinearly

120

4

An Overview of Digital Communication and Transmission

• Nonlinear envelope • Device malfunction

Crosstalk refers to unwanted coupling between signal paths. Crosstalk is caused by (1) electrical coupling between transmission media, (2) poor control of frequency response, and (3) the nonlinear performance in an analog multiplex system. There are two types of crosstalk: • Intelligible — where at least four words are intelligible to the listener from

extraneous conversations in a seven-second period (for voice applications) • Unintelligible — crosstalk resulting from any other form of disturbing effects of one channel on another. Impulse noise is noncontinuous and consists of irregular pulses or noise spikes of short duration and of relatively high amplitude. These spikes are often called “hits.” Impulse-noise degrades voice telephony only marginally, if at all; however, it may seriously degrade error rate on data or other digital circuits. SNR SNR is expressed in decibels (dB) — the amount by which a signal level exceeds the noise within a specified bandwidth. S N

(dB)  (Signal level)(dBm)  (Noise level)(dBm)

(4.38)

The following are the suggested SNR values for various services: • Voice: 40 dB* • Video: 45 dB* • Data ~15 dB, based on specified BER and modulation type.

Note: The values marked by * are based on customer input.

4.18

Summary

In this chapter we presented the essentials of digital communications. This chapter may be skipped by those readers who have been exposed to digital communications. The chapter outlines the process of converting an analog signal to a digital signal. The chapter presented the modulation and error correcting schemes that are used in wireless communications systems. It also discussed data communication services and defined circuit-switched and packet-switched services. Finally, it looks at the most common types of transmission media, and the behavior of signals on those media.

References

121

The readers who wish to learn more about digital communication theory should refer to several excellent books listed in the reference section (for example [6] and [7]). More details about modulation schemes used in wireless communications can be found in Chapter 9.

Problems 4.1 Define linear and nonlinear modulation schemes. 4.2 What are the power limited and bandwidth limited channels? 4.3 Define the Nyquist bandwidth. 4.4 A digital signaling system is required to operate at 9.6 kbps. If a signal element encodes a 4-bit word, what is the minimum bandwidth of the channel? 4.5 Given a channel with intended capacity of 20 Mbps, the bandwidth of the channel is 3 MHz. What signal-to-noise ratio is required to achieve this capacity? 4.6 The receiver in a communication system has a received signal power of 134 dBm, a received noise power spectral density of 174 dBm/Hz, and a bandwidth of 2000 Hz. What is the maximum rate of error-free information for the system? 4.7 Find the minimum required bandwidth for the baseband transmission of an 8-level PAM pulse sequence having a data rate, Rb, of 9600 bps. The system characteristics consist of a raised-cosine spectrum with a 50% excess bandwidth. 4.8 Find the BER probability, Pe, of an M-PSK system for 14.4 kbps signal, amplitude is 15 mv, and noise density, N0 is 109 W/Hz. 4.9 A noncoherent orthogonal M-FSK system carries 3 bits per symbol. The system is designed for Eb /N0  6 dB. What is the maximum bit error rate probability? Find the bandwidth efficiency of the system. 4.10 A noncoherent orthogonal M-FSK system carries 4 bits per symbol. The system is designed to have a maximum probability of symbol-error of 106. What is the required Eb/N0? What is the bandwidth efficiency?

References 1. Burr, Alister. Modulation and Coding for Wireless Communications. Upper Saddle River, NJ: Prentice Hall, 2001. 2. Freeman, R. L. Telecommunication System Engineering. New York: John Wiley and Sons, 1989.

122

4

An Overview of Digital Communication and Transmission

3. Garg, V. K., and Wilkes, J. E. Principles & Applications of GSM. Upper Saddle River, NJ: Prentice Hall, 1999. 4. Glover, I. A., and Grant, P. M. Digital Communications. Upper Saddle River, NJ: Prentice Hall, 1998. 5. Hamming, R.W. Error detecting and correcting codes. Bell System Technical Journal, 29:147–160, 1950. 6. Haykin, S. Communication Systems. New York: John Wiley and Sons, 2001. 7. Lathi, B. P. Modern Digital and Analog Communications Systems. 2nd Edition. Holt, Reinhart and Winston, 1989. 8. Nyquist, H. Certain topics in telegraph transmission theory. AIEE Transactions, 47: 617– 644, 1928. 9. Proakis, J. G. Digital Communications. 3rd Edition. McGraw-Hill, 1995. 10. Rappaport, T. S. Wireless Communications: Principles and Practice, 2nd Edition. Upper Saddle River, NJ: Prentice Hall, 2000. 11. Shannon, C. E. A mathematical theory of communication, Part 1. Bell System Technical Journal, 27:379, 1948. 12. Shannon, C. E. A mathematical theory of communication, Part 2. Bell System Technical Journal, 27:623, 1948. 13. Skalar, B. Digital Communication — Fundamental and Applications. Englewood Cliffs, NJ: Prentice Hall, 1988. 14. Stallings, W. Data and Computer Communications. New York: Macmillan Publishing Company, 1988. 15. Sweeney, P. Error Control Coding: An Introduction. Prentice Hall, 1991. 16. Taub, H., and Schilling, D. L. Principles of Communications Systems. New York: McGraw Hill, 1999. 17. Ziemer, R. E., and Peterson, R. L. Introduction to Digital Communication. New York: Macmillan Publishing Company, 1992.

CHAPTER 5 Fundamentals of Cellular Communications 5.1

Introduction

In this chapter, we present the concept of a cellular system and discuss the fundamentals of cellular communications. We develop a relationship between reuse ratio (q) and cluster size or reuse factor (N) for hexagonal cell geometry, as well as study cochannel interference for omnidirectional and sectorized cells. We also discuss cell splitting and segmentation procedures used in cellular systems.

5.2

Cellular Systems

Most commercial radio and television systems are designed to cover as much area as possible. These systems typically operate at maximum power and with the tallest antennas allowed by the Federal Communications Commission (FCC). The frequency used by the transmitter cannot be reused again until there is enough geographical separation so that one station does not interfere significantly with another station assigned to that frequency. There may even be a large region between two transmitters using the same frequency where neither signal is received. The cellular system takes the opposite approach [1,3,4,5,9,11–14]. It seeks to make an efficient use of available channels by employing low-power transmitters to allow frequency reuse at much smaller distances (see Figure 5.1). Maximizing the number of times each channel may be reused in a given geographic area is the key to an efficient cellular system design. Cellular systems are designed to operate with groups of low-power radios spread out over the geographical service area. Each group of radios serve mobile stations located near them. The area served by each group of radios is called a cell. Each cell has an appropriate number of low-power radios to communicate within the cell itself. The power transmitted by the cell is chosen to be large enough to communicate with mobile stations located near the edge of the cell. The radius of each cell may be chosen to be perhaps 28 km (about 16 miles) in a start-up system with relatively few subscribers, down to less than 2 km (about 1 mile) for a mature system requiring considerable frequency reuse.

123

124

5

Fundamentals of Cellular Communications

1 Same frequencies are used in these cells (first-tier)

1

N  i2  j2  i ? j N  22  12  2.1  7

2 7

1

3 1

6

First-tier interfacing cell

4

1

5 1 1 Cluster of 7 cells (N  7)

Figure 5.1

Cell arrangement with reuse factor.

As the traffic grows, new cells and channels are added to the system. If an irregular cell pattern is selected, it would lead to an inefficient use of the spectrum due to its inability to reuse frequencies because of cochannel interference. In addition, it would also result in an uneconomical deployment of equipment, requiring relocation from one cell site to another. Therefore, a great deal of engineering effort would be required to readjust the transmission, switching, and control resources every time the system goes through its development phase. The use of a regular cell pattern in a cellular system design eliminates all these difficulties. In reality, cell coverage is an irregularly shaped circle. The exact coverage of the cell depends on the terrain and many other factors. For design purposes and as a first-order approximation, we assume that the coverage areas are regular polygons. For example, for omnidirectional antennas with constant signal power, each cell site coverage area would be circular. To achieve full coverage without dead spots, a series of regular polygons are required for cell sites. Any regular polygon such as an equilateral triangle, a square, or a hexagon can be used for cell design. The hexagon is used for two reasons: a hexagonal layout requires fewer cells and, therefore, fewer transmitter sites, and a hexagonal cell layout is less expensive

5.3

Hexagonal Cell Geometry

125

compared to square and triangular cells. In practice, after the polygons are drawn on a map of the coverage area, radial lines are drawn and the signal-to-noise ratio (SNR) calculated for various directions using the propagation models (discussed in Chapter 3), or using appropriate computer programs [2,6–8]. For the remainder of this chapter, we assume regular polygons for coverage areas even though in practice that is only an approximation.

5.3

Hexagonal Cell Geometry

We use the u-v axes to calculate the distance D between points C1 and C2 (see Figure 5.2). The u-v axes are chosen so that u-axis passes through the centers of the hexagons. C1 and C2 are the centers of the hexagonal cells with coordinates (u1,v1) and (u2,v2) [11,12]. 2

2

2 1/2

D   (u2u1) (cos30)  [(v2v1)  (u2u1)(sin30)]  2

2

(5.1)

  (u2u1)  (v2v1)  (v2v1) (u2u1) 

1/2

(5.2)

v,y

v (u 2,v2) C2 u

C1 2R Cos 308

(u 1,v1)

308

x

Figure 5.2

Coordinate system.

126

5

Fundamentals of Cellular Communications

If we assume (u1,v1)  (0,0), or the origin of the coordinate system is the center of a hexagonal cell, and restrict (u2,v2) to be a positive integer valued (i,j) then normalized distance Dnorm from Equation 5.2 can be written as: Dnorm  [i2  j2  ij]1/2

(5.3)

The normalized distance between two adjacent cells is unity for (i  1, j  0) or (i  0, j  1). The actual center-to-center distance (D) between two  adjacent hexagonal cells is 2Rcos30 or  3 R where R is the center-to-vertex distance. We assume the size of all the cells is roughly the same. As long as the cell size is fixed, and each cell transmits the same power, cochannel interference will be independent of the transmitted power of each cell. The cochannel interference is a function of q where q  D/R. But D (distance between any two cells) is a function of N I and S/I. N I is the number of cochannel-interfering cells in the first tier (see Figure 5.3, Note: N I is 6; the number of cochannel-interfering cells in the second tier will be 12) and S/I is the received signal-to-interference ratio at the desired mobile receiver. We neglect the effects of cochannel-interfering cells in the second, third, and higher tiers because their contributions are much smaller compared to the first tier (1% of the total interference is caused by cells in the second and higher tiers). Assuming that the first cell is centered at the origin (u  0, v  0), the distance between any two cells will be: 

D  (Dnorm)( 3 R)

(5.4a)

Using Equation 5.3 cochannel separation (D) (Figure 5.4) will be: D2  3R2(i2  j2  ij)

(5.4b)

Since the area of a hexagon is proportional to the square of the distance between the center and vertex, the area enclosed by the large hexagon is: Alarge  k[3R2(i2  j2  ij)]

(5.5)

where k is a constant Similarly the area enclosed by the small hexagon is given as: Asmall  k(R2)

(5.6)

5.3

Hexagonal Cell Geometry

127

First-tier

cell #1 Interfering Cell

Cell Site-to-Mobile Interference (Downlink) Mobile-to-Cell Site Interfaces (Uplink) Figure 5.3

Cochannel interference with omnidirectional cell site.

Comparing Equation 5.5 and Equation 5.6, we can write Alarge D2 2 2   3(i  j  ij)   R2 Asmall

(5.7)

From symmetry, we can see that the large hexagonal encloses the center cluster of N cells plus one-third of the number of cells associated with six other peripheral hexagons. Thus, the total number of cells enclosed in the large hexagon is equal to N  6(N/3)  3N. Since the area is proportional to the number of cells, Alarge  3N and Asmall  1. A

large   3N

Asmall

(5.8)

128

5

Fundamentals of Cellular Communications

First-Tier D N57

1 Interfering Cell

R

D

Figure 5.4

Six effective interfering cells in tier 1 of cell 1.

Substituting Equation 5.8 into Equation 5.7 we get: N  i2  j2  ij 2

D 2  3N R



D   q   3N R

where: q  reuse ratio (refer to Figure 5.5) N  cluster size (see Figure 5.1) or reuse factor Table 5.1 lists the values of q for different values of N.

(5.9) (5.10) (5.11)

5.3

Hexagonal Cell Geometry

129

2

5

1 4

D 3N  q   R

R 7

3

D 2

6

1

5 4 N7

3

7

6

Figure 5.5

Relationship between q and N.

Equation 5.11 is important because it affects the traffic-carrying capacity of a cellular system and the cochannel interference. By reducing q, the number of cells per cluster is reduced. If total RF channels are constant, then the number of channels per cell is increased, thereby increasing the system capacity. On the other hand, cochannel interference is increased with small q. The reverse is true when q is increased — an increase in q reduces cochannel interference and also the traffic capacity of the cellular system. Table 5.1 shows that a 2-cell, 5-cell, etc., reuse pattern does not exist. However, the basic assumption in Equation 5.9 is that all six first-tier interferers are located at the same distance from the desired cell. Asymmetrical reuse arrangements, where interferers are located at various distances, do allow for 2-cell, 5-cell, etc., reuse.

130

5

Fundamentals of Cellular Communications

Table 5.1 Cochannel reuse ratio (q) vs. frequency reuse pattern (N ). i

j

N

1

0

1

q ⴝ D/R 1.73

1

1

3

3.00

2

0

4

3.46

2

1

7

4.58

3

0

9

5.20

2

2

12

6.00

3

1

13

6.24

4

0

16

6.93

3

2

19

7.55

4

1

21

7.94

4

2

28

9.17

Example 5.1 We consider a cellular system in which total available voice channels to handle the traffic are 960. The area of each cell is 6 km2 and the total coverage area of the system is 2000 km2. Calculate (a) the system capacity if the cluster size, N (reuse factor), is 4 and (b) the system capacity if the cluster size is 7. How many times would a cluster of size 4 have to be replicated to cover the entire cellular area? Does decreasing the reuse factor N increase the system capacity? Explain.

Solution • Total available channels  960 • Cell area  6 km2 • Total coverage area  2000 km2

N4 • Area of a cluster with reuse N  4: 4  6  24 km2 • Number of clusters for covering total area with N equals 4  2000/24

 83.33 ⵑ 83 • Number of channels per cell  960/4  240 • System capacity  83  960  79, 680 channels N7

5.4

Cochannel Interference Ratio

131

• Area of cluster with reuse N  7: 7  6  42 km2 • Number of clusters for covering total area with N equals 7  2000/42

 47.62 ⵑ 48

• Number of channels per cell  960/7  137.15 ⵑ 137 • System capacity  48  960  46,080 channels

It is evident when we decrease the value of N from 7 to 4, we increase the system capacity from 46,080 to 79,680 channels. Thus, decreasing the reuse factor (N) increases the system capacity.

5.4

Cochannel Interference Ratio

The S/I ratio at the desired mobile receiver is given as [5,10]: S I

S

 N I

(5.12)

 Ik

k1

where: Ik  the interference due to the kth interferer NI  the number of interfering cells in the first tier. In a fully equipped hexagonal-shaped cellular system, there are always six cochannel-interfering cells in the first tier (i.e., NI  6, see Figure 5.1). Most of the cochannel interference results from the first tier. Contribution from second and higher tiers amounts to less than 1% of the total interference and, therefore, it is ignored. Cochannel interference can be experienced both at the cell site and the mobile stations in the center cell. In a small cell system, interference will be the dominating factor and thermal noise can be neglected. Thus the S/I ratio can be given as: S

1

 6 I Dk   R k1

(5.13)



where: 2  5  the propagation path loss, and  depends upon the terrain environment (see Chapter 3) Dk  the distance between mobile and kth interfering cell R  the cell radius

132

5

Fundamentals of Cellular Communications

If we assume Dk is the same for the six interfering cells for simplification, or D  Dk, then Equation 5.13 becomes: S I



1 6(q)

q     6

(5.14)

where: q  D/R, reuse ratio

 I

 q  6 S

1/

(5.15)

Substituting for q in Equation 5.15 from Equation 5.11 we get: 3  I

1 6 S N 

2/

(5.16)

Example 5.2 Consider the advanced mobile phone system in which an S/I ratio of 18 dB is required for the accepted voice quality. What should be the reuse factor for the system? Assume   4. What will be the reuse factor of the Global System of Mobile (GSM) system in which an S/I of 12 dB is required? Solution Using Equation 5.16, we get 1 [6(101.8)]2/4  6.486 7 NAMP   3

and 1 6 101.2 NGSM  

2/4

3

 3.251 4

Example 5.3 Consider a cellular system with 395 total allocated voice channel frequencies. If the traffic is uniform with an average call holding time of 120 seconds and the call blocking during the system busy hour is 2%, calculate: 1. The number of calls per cell site per hour (i.e., call capacity of cell) 2. Mean S/I ratio for cell reuse factor equal to 4, 7, and 12.

5.4

Cochannel Interference Ratio

133

Assume omnidirectional antennas with six interferers in the first tier and a slope for path loss of 40 dB/decade (  4).

Solution For a reuse factor N  4, the number of voice channels per cell site  395/4  99. Using the Erlang-B traffic table (see Appendix A) for 99 channels with 2% blocking, we find a traffic load of 87 Erlangs. The carried load will be (1  0.02)  87  85.26 Erlangs. Ncall  120   85.26 3600

Ncall  2558 calls/hour/cell

Using Equation 5.16, we get 3  I

1 6 S 4 

2/4

 S  24 (13.8 dB) I

The results for N  7 and N  12 are given in Table 5.2. It is evident from the results that, by increasing the reuse factor from N  4 to N  12, the mean S/I ratio is improved from 13.8 to 23.3 dB. However, the call capacity of cell (i.e., calls per hour per cell) is reduced from 2558 to 724 calls per hour.

Table 5.2 Cell reuse factor vs. mean S/I ratio and call capacity of cell. N

Voice channels per cell

Calls per hour per cell (Ncall)

Mean S/I (dB)

4

99

2558

13.8

7

56

1349

18.7

12

33

724

23.3

134

5

Fundamentals of Cellular Communications

Example 5.4 Consider a GSM system with a one-way spectrum of 12.5 MHz and channel spacing of 200 kHz. There are three control channels per cell, and the reuse factor is 4. Assuming an omnidirectional antenna with six interferers in the first tier and a slope of path loss of 40 dB/decade, calculate the number of calls per hour per cell site with 2% blocking during the system busy hour and an average call holding time of 120 seconds. The GSM uses eight voice channels per RF channel. Solution 12.5  10  8  No. of voice channels per cell site    3  122 3 6

200  10  4

Using the Erlang-B traffic table for 122 channels with 2% blocking, we find a traffic load of 110 Erlangs. The carried traffic load will be (1  0.02)  110  107.8  120

N

call   107.8

3600

Ncall  3234 calls/hour/cell

Using Equation 5.16 we calculate the mean S/I ratio as 13.8 dB.

5.5

Cellular System Design in Worst-Case Scenario with an Omnidirectional Antenna

In the previous section we showed that with a seven-reuse pattern (N  7), the value of q  4.6 (see Table 5.1) is adequate for a normal interference condition [5]. Let us reexamine the seven-cell reuse pattern and consider the worst case in which the mobile unit is located at the cell boundary as shown in Figure 5.6. In this situation, the mobile unit receives the weakest signal from its own cell and is subjected to strong interference from all the interfering cells in the first tier. The distances from the six interfering cells are given in Figure 5.6. The S/I ratio can be given as: S I

R  2D  2(D  R)

     

2(DR)

(5.17)

Using the path-loss exponent   4 and D/R  q, we rewrite Equation 5.17 as: S

1

   I 2 q1 4  2q4  2 q1 4

(5.18)

5.5

Cellular System Design in Worst-Case Scenario

135

D

DR

DR

DR DR

D

Figure 5.6

Worst-case scenario for cochannel interference.

where: q  4.6 for a normal seven-cell reuse pattern (N  7). Substituting q  4.6 in Equation 5.18, we get S/I  54.3 or 17.3 dB. For a conservative estimate, if we use the shortest distance (DR) then S

1

1

  28 or 14.47 dB  I 6 q  1 4 6 3.6 4

(5.19)

In a real situation, because of imperfect cell-site locations and the rolling nature of the terrain configuration, the S/I ratio is often less than 17.3 dB. It could be 14 dB

136

5

Fundamentals of Cellular Communications

or lower. Such conditions may occur in heavy traffic. Therefore, the cellular system should be designed around the S/I ratio of worst case. If we consider the worst case for a seven-cell reuse (N  7) pattern, we conclude that q  4.6 is not enough in an omnidirectional cell system. In an omnidirectional cell system, N  9 (q  5.2) or N  12 (q  6.0) cell reuse pattern would be a better choice. These cell reuse patterns would provide the S/I ratio of 19.78 and 22.54 dB, respectively.

5.6

Cochannel Interference Reduction

In the case of increased call traffic, the frequency spectrum should be used efficiently. We should avoid increasing the number of cells N in a frequency reuse pattern. As N increases, the number of frequency channels assigned to a cell is reduced, thereby decreasing the call capacity of the cell [7]. Instead of increasing N, we use either a directional antenna arrangement (sectorization) to reduce cochannel interference or perform cell splitting to subdivide a congested cell into smaller cells. In case of a sectorized cell, each cell is divided into three or six sectors and uses three or six directional antennas at the base station to reduce the number of cochannel interferers (refer to Figures 5.7 and 5.8). Each

Two Interferers in First Ring per Sector

1208

1208 1208

Cell Site-to-Mobile Interference (Downlink) Mobile-to-Cell Site Interference (Uplink)

Figure 5.7

Cochannel interference with 120ⴗ sectorized cells.

5.7

Directional Antennas in Seven-Cell Reuse Pattern

137

One Interferer in First-Tier per Sector 608 608

608

608

608 608

Cell Site-to-Mobile Interference (Downlink) Mobile-to-Cell Site Interference (Uplink)

Figure 5.8

Cochannel interference with 60° sectorized cells.

sector is assigned a set of channels (frequencies) (either 1/3 or 1/6 of the frequencies of the omnidirectional cell).

5.7 Directional Antennas in Seven-Cell Reuse Pattern 5.7.1 Three-Sector Case We consider the worst case in which the mobile unit is at position M (see Figure 5.9). In this situation, the mobile receives the weakest signal from its own cell and fairly strong interference from two interfering cells 1 and 2. Because of the use of directional antennas, the number of interfering cells is reduced from six to two. At point M, the distances between the mobile unit and the two interfering cells are D and (D  0.7R), respectively. The S/I ratio in the worst case with   4 will be:    4 4

S I

D

R4  (D  0.7R)

(5.20)

S I

q

1  (q  0.7)

(5.21)

   4 4

138

5

2

Fundamentals of Cellular Communications

Interfering sector

D  0.7R

R

Mobile unit M D

1

Figure 5.9

Interfering sector

Worst-case interference with 120ⴗ sectorized cells.

Using q  4.6 in Equation 5.21 for N  7, we get S/I  285 or 24.5 dB. The S/I for a mobile unit served by a cell site with 120 directional antenna exceeds 18 dB in the worst case. It is evident from Equation 5.21 that the use of a directional antenna is helpful in reducing cochannel interference. In a real situation, under heavy traffic, the S/I could be up to 6 dB weaker than in Equation 5.21 due to irregular terrain configurations and imperfect site locations. The resulting 18.5 dB S/I is still adequate.

5.7.2 Six-Sector Case In this case the cell is divided into six sectors by using a 60 beam width directional antenna. In this case, only one interference can occur. The worst case S/I ratio with   4 will be (see Figure 5.10): S I

R4 (D  0.7R)

 (q  0.7)4    4

(5.22)

For q  4.6 (N  7), Equation 5.22 gives S/I  789 or 29 dB. This indicates a further reduction in cochannel interference. Using the argument that was used for the three-sector case and subtracting 6 dB from 29 dB, the remaining 23 dB is still more than adequate. Under heavy traffic, the 60 sector configuration can be used to reduce cochannel interference. However, with the six-sector configuration, the call capacity of the cell is decreased.

5.7

Directional Antennas in Seven-Cell Reuse Pattern

1

139

Interfering sector

D  0.7R

R M Mobile unit

Figure 5.10

Worst-case interference with 60° sectorized cells.

Example 5.5 We consider a cellular system with 395 total allocated voice channels of 30 kHz each. The total available bandwidth in each direction is 12.5 MHz. The traffic is uniform with the average call holding time of 120 seconds, and call blocking during the system busy hour is 2%. Calculate: a. The calls per hour per cell site b. The mean S/I ratio c. The spectral efficiency in Erlang/km2/MHz For a cell reuse factor N  4, 7, and 12, respectively, and for omnidirectional (see Example 5.3) 120 and 60 antenna systems, calculate the call capacity. Assume that there are 10 mobiles/km2 with each mobile generating traffic of 0.02 Erlangs. The slope of path loss is   40 dB per decade. The mean S/I ratio is given as: 

Mean S  log 3N 10log m I

where:   slope of path loss (dB/decade) m  number of interferers in the first tier m  6 for omnidirectional system m  2 for 120 sectorized system m  1 for 60 sectorized system

140

5

Fundamentals of Cellular Communications

Solution We consider only the first tier interferers and neglect the effects of cochannel interference from the second and other higher tiers. The traffic carried per cell site  V  t  Ac  V  t  2.6R2 Erlangs where: V  number of mobiles per km2 t  traffic in Erlangs per mobile Ac  area of hexagonal cell (i.e., Ac  2.6R2) The traffic carried per cell site  10  0.02  2.6R2  0.52R2 Erlangs Traffic per cell  Nc Bw  A

The spectral efficiency   Erlangs/km2/MHz where: Nc  number of cells in the system (i.e., A/2.6R2) and A  area of the system Traffic per cell 2.6R  Bw

The spectral efficiency   Erlangs/km2/MHz 2 We will demonstrate the procedure for calculating results in one row in Table 5.3; the remaining calculations can be performed without any difficulty. 120° Sectorized Cell a. N  7 b. Number of voice channels per sector  395/(7  3) 19 c. Offered traffic load per sector from Erlang-B tables  12.3 Erlangs d. Offered traffic load per cell site  3  12.3  36.9 Erlangs e. Carried traffic load per cell site  (1  0.02)  36.9  36.2 Erlangs 36.2  3600 No. of calls per hour per cell site    1086 120



 0.52

36.2 R    8.3 km 36.2 Spectral efficiency    0.016 Erlang/km2/MHz 2 2.6  12.5  8.3



Mean S/I  40log21 10log2  26.44  3.01  23.43 dB

5.8

Cell Splitting

141

Table 5.3 Omni vs. sectorized cellular system performance.

System Omni

120 Sector

60 Sector

Channels per sector

Offered load/ cell (E)

Carried load/ cell (E)

Calls/ hour/ cell (cell capacity)

4

99

87.0

85.3

2558

12.8

0.016

13.8

N

Required cell radius R (km)

Spectral efficiency E/km2/ MHz

Mean S/I (dB)

7

56

45.9

45.0

1349

9.3

0.016

18.7

12

33

24.6

24.1

724

6.8

0.016

23.3

4

33

73.8

72.3

2169

11.8

0.016

18.6

7

19

36.9

36.2

1086

8.3

0.016

23.4

12

11

17.5

17.2

516

5.8

0.016

28.1

3

22

84.2

82.6

2477

12.6

0.016

19.1

4

17

64.2

62.9

1887

11.0

0.016

21.6

7

9

26.0

25.5

765

7.0

0.016

26.4

12

6

13.7

13.4

402

5.1

0.016

31.1

From the results in Table 5.3, we can draw the following conclusions: 1. Sectorization reduces cochannel interference and improves the mean S/I ratio for a given cell reuse factor. However, it reduces cell capacity since the channel resource is distributed more thinly among various sectors. 2. Since a sectorized cellular system has fewer cochannel interferers, it is possible to reduce the cluster size. 3. An omnidirectional cellular system requires a cluster size of 7, while a 120 sectorized system requires a cluster size of 4 and a 60 sectorized system requires a cluster size of 3 for a desired mean S/I ratio of approximately 18 dB.

5.8

Cell Splitting

After a cellular system has been in operation, traffic will grow in the system and will require that additional channels be made available. The cell with heavy traffic is split into smaller cells (see Figure 5.11) [10]. This is done in such a way that cell areas, or the individual component coverage areas of the cellular system, are

142

5

Fundamentals of Cellular Communications

2 1

5 4

1

3

7 6

3

Growing by splitting cell numbered 4 into cells of small size

2 (3)

1 (6)

5 (1) 1

(4) (7) 3

(2) (5)

7

6

Figure 5.11

Radius of big cell:R Radius of small cell:R/2

3

Cell splitting.

further divided to yield yet more cell areas. The splitting of cell areas by adding new cells provides for an increasing amount of channel reuse and, hence, increasing system capacity. When cell splitting occurs, the designer should minimize changes in the system. The value of the reuse factor, N, is important with respect to the type of split. The following splitting patterns can be used for various values of N: • For N  3, use 4:1 cell splitting. • For N  4, use 3:1 cell splitting.

5.8

Cell Splitting

143

• For N  7, use 3:1 or 4:1 cell splitting. • For N  9, use 4:1 cell splitting.

The 4:1 cell splitting works as follows. When the new cell is located on the border between two existing cell areas, the new cell will cover an area with a radius of one-half that of the larger cell areas from which it was split. Thus, the area covered by the new cell is one-fourth of the area of the old cell; hence, this is called 4:1 splitting. For example, if the old cell had a radius of 8 km, the new cell would have a radius of 4 km. The 3:1 cell splitting works as follows. When the new cell is located on the corner between the cell areas covered by three existing cells, the radius of the cell  area covered by the new cell would be 1/3 of the radius of the cell areas covered by the existing cell sites. Thus, the new coverage area is only one-third of the larger cell’s coverage area; hence, this is called 3:1 splitting. Decreasing cell radii imply that cell boundaries will be crossed more often. This will result in more handoffs per call and a higher processing load per subscriber. Simple calculations show that a reduction in a cell radius by a factor of four will produce about a tenfold increase in the handoff rate per subscriber. Since the call processing load tends to increase geometrically with the increase in the number of subscribers, with cell splitting the handoff rate will increase exponentially. Therefore, it is desirable to perform a cost-benefit analysis to compare the overall cost of cell splitting versus other available alternatives to handle increased traffic load. The large cell with radius R is split into cells with radius R/2. Let d be the separation between the transmitter and receiver, and d0 be the distance from the transmitter to a close-in reference point. P0 is the power received at the close-in reference point. The average received power, Pr , is given by: 

d

d Pr  P0 

(5.23)

0

Expressing Equation 5.23 in decibels, we have:

d

d Pr (dBm)  P0(dBm)  10 log 

(5.24)

0

Pt1 and Pt2 are the transmit power of the large cell (radius R) base station and small cell (radius R/2) base station, respectively. The received power, Pr at the  large cell boundary is proportional to Pt1R , and Pr at the small cell boundary is  proportional to Pt2(R/2) . On the basis of equal power, we get 

Pt1R



R  Pt2  2

(5.25)

144

5

Fundamentals of Cellular Communications

or Pt1   4 Pt2

(5.26)

For   4, Pt1/Pt2  16 (12 dB). Thus, with cell splitting, where the radius of the new cell is one-half of that of the old cell, we achieve a 12 dB reduction in the transmit power.

5.9

Adjacent Channel Interference (ACI)

Signals which are adjacent in frequency to the desired signal cause adjacent channel interference [7,8]. ACI is brought about primarily because of imperfect receiver filters which allow nearby frequencies to move into the pass band, and nonlinearity of the amplifiers. The ACI can be reduced by: (1) using modulation schemes which have low out-of-band radiation; (2) carefully designing the bandpass filter at the receiver front end; and (c) assigning adjacent channels to different cells in order to keep the frequency separation between each channel in a given cell as large as possible. The effects of ACI can also be reduced using advanced signal processing techniques that employ equalizers.

5.10

Segmentation

Sometimes engineers have to add an additional cell at less than the reuse distance without using a complete cell splitting process. This method might be used to fill in a coverage gap in the system. This can result in cochannel interference. The most straightforward method to avoid an increase in cochannel interference is simply not to reuse them. Segmentation divides a channel group into segments of mutually exclusive channel frequencies. Then, by assigning different segments to particular cell sites, cochannel interference between these cell sites can be avoided. The disadvantage of segmentation is that the capacity of the segmented cells is lower than the unsegmented cell. When a cellular system is growing, there may be cells of different radii in the same region of the coverage area. This also can result in cochannel interference. By dividing the radii at the cell site into two separate serving groups, one for the larger (overlaid) cell and other for the smaller (underlaid) cell, the interference can be minimized (see Figure 5.12). The radii for the primary serving group serve the underlaid cell, and the radii of the secondary serving group are used to serve mobiles in the overlaid cell areas. As traffic in the smaller cells grows, more and more channels are removed from the secondary serving group and assigned to the primary group until the secondary group (and its larger cell) disappears.

5.11

Summary

145

Underlaid Cell R1

R2

N1

dunder N1

Overlaid Cell

dover

N2 5 N 1 N1

N1

Note: Total channels N are divided between two segments as N1 and N2 Figure 5.12

5.11

Cell segmentation.

Summary

In this chapter, we developed a relationship between the reuse ratio (q) and cell cluster size (N) for the hexagonal geometry. Cochannel interference ratios for the omnidirectional and sectorized cell were derived. A numerical example was given to demonstrate that, for a given cluster size, sectorization yields a higher S/I ratio, but reduces spectral efficiency. However, it is possible to achieve a higher spectral efficiency by reducing the cluster size in a sectorized system without lowering the S/I ratio below the minimum requirement. We concluded the chapter by discussing cell splitting, adjacent channel interference, and segmentation.

146

5

Fundamentals of Cellular Communications

Problems 5.1 Why is cell splitting needed? Define 4:1 and 3:1 cell splitting. 5.2 Explain segmentation in a cellular system. Why is it required? 5.3 What is adjacent channel interference? How can it be minimized? 5.4 The S/I ratio was calculated by neglecting the interference from cells other than in the first tier. Calculate the amount of interference from the second tier of cells. Is it reasonable to neglect this interference? 5.5 Consider a GSM system with a one-way spectrum of 12.5 MHz and channel spacing of 200 kHz. There are three control channels per cell and the reuse factor is 4. Assuming an omnidirectional antenna with six interferers in the first tier and slope for path loss of 45 dB/decade, calculate the number of calls per cell site per hour with 2% blocking during system busy hour and an average call holding time of 120 seconds. What is the S/I ratio? 5.6 Repeat Problem 5.5 with 3-sector and 6-sector antenna systems. Discuss your results with an omnidirectional antenna and provide comments. 5.7 Compare the spectral efficiency of the GSM system with the total access communication system (TACS) using the following data: • GSM channel spacing: 200 kHz • TACS channel spacing: 25 kHz • The required S/I for the GSM: 12 dB • The required S/I for the TACS: 16 dB • The total available one-way spectrum: 25 MHz • The number of control channels for GSM per cell: 3 • The number of control channels for TACS per cell: 6 • The reuse factor for GSM: 4 • The reuse factor for TACS: 7 • Total coverage area: 10,000 km2

5.8 A signal-to-interference ratio of 16 dB is required for a satisfactory forward link performance of a cellular system. What is the frequency reuse factor and cluster size that should be used for maximum capacity if the path loss is 40 dB/decade? Assume that there are six cochannel interferers in the first tier and all of them are the same distance from the mobile. Neglect interference from higher tiers. 5.9 A suburb has an area of 1500 square miles and is covered by a cellular system that uses a seven reuse pattern. Each cell has a radius of four miles and the city is allocated 50 MHz of spectrum of a full duplex channel

References

147

bandwidth of 60 kHz. Assuming a GoS of 2% for an Erlang-B system and a traffic load per user of 0.03 Erlangs, calculate: • The number of cells in the service area • The number of channels per cell • The maximum carried traffic • Traffic intensity of each cell • Total number of users that can be served for 2% GoS • Number of mobiles per channel

References 1. AT&T Technical Center, “Cellular System Design and Performance Engineering I,” CC 1400, version 1.12, 1993. 2. Chen, G. K. Effects of Sectorization on Spectrum Efficiency of Cellular Radio Systems. IEEE Transactions on Vehicular Technology, 41, no. 3 (August 1992): 217–225. 3. Dersh, U., and Braun, W. A Physical Mobile Radio Channel Model. Proceedings of IEEE Vehicular Technology Conference. May 1991, 289–294. 4. French, R. C. The effects of Fading and Shadowing on Channel Reuse in Mobile Radio. IEEE Transactions on Vehicular Technology, 28 (August 1979). 5. Lee, W. C. Y. Elements of Cellular Radio System. IEEE Transactions on Vehicular Technology, 35 (May 1986): 48–56. 6. Lee, W. C. Y. Spectrum Efficiency and Digital Cellular. Presented at 38th IEEE Vehicular Technology Conference, Philadelphia, June 1988. 7. Lee, W. C. Y., Mobile Cellular Telecommunications System. New York: McGraw-Hill, 1989. 8. Lee, W. C. Y. Spectrum Efficiency in Cellular. IEEE Transactions on Vehicular Technology, 38 (May 1989): 69–75. 9. MacDonald, V. H. The Cellular Concept. Bell System Technical Journal, 58, no. 1 (January 1979): 15–41. 10. Mark, J. W., and Weihua, Zhuang. Wireless Communications and Networking. Upper Saddle River, NJ: Prentice Hall, 2003. 11. Mehrotra, A. Cellular Radio Analog and Digital System. Boston: Artech House, 1994. 12. Rappaport, T. S. Wireless Communication. Upper Saddle River, NJ: Prentice Hall, 2002. 13. Whitehead, J. F. Cellular System Design: An Emerging Engineering Discipline. IEEE Communications Society Magazine, 24, no. 2 (February 1986): 8–15. 14. Young, W. R. Advanced Mobile Phone Service: Introduction, Background, and Objectives. Bell System Technical Journal, 58, no. 1 (January 1979): 1–14.

This page intentionally left blank

CHAPTER 6 Multiple Access Techniques 6.1

Introduction

In Chapter 5 we discussed that cellular systems divide a geographic region into cells where mobile units in each cell communicate with the cell’s base station. The goal in the design of a cellular system is to be able to handle as many calls as possible in a given bandwidth with the specified blocking probability (reliability). Multiplexing deals with the division of the resources to create multiple channels. Multiplexing can create channels in frequency, time, etc., and the corresponding terms are then frequency division multiplexing (FDM), time division multiplexing (TDM), etc. [1,3]. Since the amount of spectrum available is limited, we need to find ways to allow multiple users to share the available spectrum simultaneously. Shared access is used to implement a multiple access scheme when access by many users to a channel is required [13,14,15]. For example, one can create multiple channels using TDM, but each of these channels can be accessed by a group of users using the ALOHA multiple access scheme [8,9]. The multiple access schemes can be either reservation-based or random. Multiple access schemes allow many users to share the radio spectrum. Sharing the bandwidth efficiently among users is one of the main objectives of multiple access schemes [16,17]. The variability of wireless channels presents both challenges and opportunities in designing multiple access communications systems. Multiple access strategy has an impact on robustness and interference levels generated in other cells. Therefore, multiple access schemes are designed to maintain orthogonality and reduce interference effects [10]. Multiple access schemes can be classified as reservation-based multiple access (e.g., FDMA, TDMA, CDMA) [4,5] and random multiple access (e.g., ALOHA, CSMA) (see Figure 6.1) [9,23]. If data traffic is continuous and a small transmission delay is required (for example in voice communication) reservationbased multiple access is used. The family of reservation-based multiple access includes frequency division multiple access (FDMA), time division multiple access (TDMA), and code division multiple access (CDMA) [6,7,12,21,22]. In many wireless systems for voice communication, the control channel is based on random multiple access and the communication channel is based on FDMA, TDMA, or CDMA. The reservation-based multiple access technique has a disadvantage in that once the channel is assigned, it remains idle if the user has nothing to transmit, 149

150

6

Multiple Access Techniques

Multiple Access Techniques

Reservationbased

FDMA

TDMA

Random

CDMA

Random with reservation

Random

ALOHA

CSMA

ISMA

Reservation ALOHA

PRMA

ISMA : Idle Signal Casting Multiple Access PRMA : Packet Reservation Multiple Access

Figure 6.1

Multiple access schemes.

while other users may have data waiting to be transmitted. This problem is critical when data generation is random and has a high peak-rate to average-rate ratio. In this situation, random multiple access is more efficient, because a communication channel is shared by many users and users transmit their data in a random or partially coordinated fashion. ALOHA and carrier sense multiple access (CSMA) are examples of random multiple access [8]. If the data arrives in a random manner, and the data length is large, then random multiple access combined with a reservation protocol will perform better than both random- and reservationbased schemes. We first focus on the reservation-based multiple access schemes including narrowband channelized and wideband nonchannelized systems for wireless communications. We discuss access technologies — FDMA, TDMA, and CDMA. We examine FDMA and TDMA from a capacity, performance, and spectral efficiency viewpoint. As networks have evolved, the demand for higher capacities has encouraged researchers and system designers to examine access schemes that are even more spectrally efficient than TDMA. Therefore, we also examine the CDMA system. Work in standards bodies around the world indicates that the 3G/4G wireless systems are evolving to wideband CDMA (WCDMA) to achieve high efficiencies and high access data rates. The later part of the chapter is devoted to the discussion of random multiple access schemes.

6.2

Narrowband Channelized Systems

Traditional architectures for analog and digital wireless systems are channelized [6,11]. In a channelized system, the total spectrum is divided into a large number of

6.2

Narrowband Channelized Systems

151

relatively narrow radio channels that are defined by carrier frequency. Each radio channel consists of a pair of frequencies. The frequency used for transmission from the base station to the mobile station is called the forward channel (downlink channel) and the frequency used for transmission from the mobile station to the base station is called the reverse channel (uplink channel). A user is assigned both frequencies for the duration of the call. The forward and reverse channels are assigned widely separated frequencies to keep the interference between transmission and reception to a minimum. A narrowband channelized system demands precise control of output frequencies for an individual transmitter. In this case, the transmission by a given mobile station occurs within the specified narrow bandwidth to avoid interference with adjacent channels. The tightness of bandwidth limitations plays a dominant role in the evaluation and selection of modulation technique. It also influences the design of transmitter and receiver elements, particularly the filters which can greatly affect the cost of a mobile station. A critical issue with regulators and operators around the world is how efficiently the radio spectrum is being used. Regulatory bodies want to encourage competition for cellular services. Thus for a given availability of bandwidth, more operators can be licensed. For a particular operator, a more efficient technology can support more users within the assigned spectrum and thus increase profits. When we examine efficiencies of various technologies, we find that each system has made different trade-offs in determining the optimum method for access. Some of the parameters that are used in the trade-off are bandwidth per user, guard bands between channels, frequency reuse among different cells in the system, the signal-to-noise and signal-to-interference ratio, the methods of channel and speech coding, and the complexity of the system. The first-generation analog cellular systems showed signs of capacity saturation in major urban areas, even with a modest total user population. A major capacity increase was needed to meet future demand. Several digital techniques were deployed to solve the capacity problem of analog cellular systems. There are two basic types of systems whereby a fixed spectrum resource is partitioned and shared among different users [13,16]. The channels are created by dividing the total system bandwidth into frequency channels through the use of FDM and then further dividing each frequency channel into time channels through the use of TDM. Most systems use a combination of FDMA and TDMA.

6.2.1

Frequency Division Duplex (FDD) and Time Division Duplex (TDD) System Many cellular systems (such as AMP, GSM, DAMP, etc.) use frequency division duplex (FDD) in which the transmitter and receiver operate simultaneously on different frequencies. Separation is provided between the downlink and uplink channels to avoid the transmitter causing self interference to its receiver. Other

152

6

Multiple Access Techniques

precautions are also needed to prevent self interference, such as the use of two antennas, or alternatively one antenna with a duplexer (a special design of RF filters protecting the receiver from the transmit frequency). A duplexer adds weight, size, and cost to a radio transceiver and can limit the minimum size of a subscriber unit. A cellular system can be designed to use one frequency band by using time division duplex (TDD). In TDD a bidirectional flow of information is achieved using the simplex-type scheme by automatically alternating in time the direction of transmission on a single frequency. At best TDD can only provide a quasisimultaneous bidirectional flow, since one direction must be off while the other is using the frequency. However, with a high enough transmission rate on the channel, the off time is not noticeable during conversations, and with a digital speech system, the only effect is a very short delay. The amount of spectrum required for both FDD and TDD is the same. The difference lies in the use of two bands of spectrum separated by the required bandwidth for FDD, whereas TDD requires only one band of frequencies but twice the bandwidth. It may be easier to find a single band of unassigned frequencies than finding two bands separated by the required bandwidth. With TDD systems, the transmit time slot and the receiver time slot of the subscriber unit occur at different times. With the use of a simple RF switch in the subscriber unit, the antenna can be connected to the transmitter when a transmit burst is required (thus disconnecting the receiver from the antenna) and to the receiver for the incoming signal. The RF switch thus performs the function of the duplexer, but is less complex, smaller in size, and less costly. TDD uses a burst mode scheme like TDMA and therefore also does not require a duplexer. Since the bandwidth of a TDD channel is twice that of a transmitter and receiver in an FDD system, RF filters in all the transmitters and receivers for TDD systems must be designed to cover twice the bandwidth of FDD system filters.

6.2.2 Frequency Division Multiple Access The FDMA is the simplest scheme used to provide multiple access. It separates different users by assigning a different carrier frequency (see Figure 6.2). Multiple users are isolated using bandpass filters. In FDMA, signals from various users are assigned different frequencies, just as in an analog system. Frequency guard bands are provided between adjacent signal spectra to minimize crosstalk between adjacent channels. The advantages and disadvantages of FDMA with respect to TDMA or CDMA are: Advantages 1. Capacity can be increased by reducing the information bit rate and using an efficient digital speech coding scheme (See Chapter 8) [20].

6.2

Narrowband Channelized Systems

153

2. Technological advances required for implementation are simple. A system can be configured so that improvements in terms of a lower bit rate speech coding could be easily incorporated. 3. Hardware simplicity, because multiple users are isolated by employing simple bandpass filters. Disadvantages 1. The system architecture based on FDMA was implemented in firstgeneration analog systems such as advanced mobile phone system (AMPS) or total access communication system (TACS). The improvement in capacity depends on operation at a reduced signal-to-interference (S/I) ratio. But the narrowband digital approach gives only limited advantages in this regard so that modest capacity improvements could be expected from the allocated spectrum. 2. The maximum bit-rate per channel is fixed and small, inhibiting the flexibility in bit-rate capability that may be a requirement for computer file transfer in some applications in the future.

Frequency 1

Circuit Guard

Frequency 2

Circuit

Downlink Path Frequency n

Circuit

Frequency 1

Circuit

Frequency 2

Circuit

Frequency Domain

Uplink Path

Frequency n

Figure 6.2

FDMA/FDD channel architecture.

Circuit

154

6

Multiple Access Techniques

3. Inefficient use of spectrum, in FDMA if a channel is not in use, it remains idle and cannot be used to enhance the system capacity. 4. Crosstalk arising from adjacent channel interference is produced by nonlinear effects.

6.2.3 Time Division Multiple Access In a TDMA system, each user uses the whole channel bandwidth for a fraction of time (see Figure 6.3) compared to an FDMA system where a single user occupies the channel bandwidth for the entire duration (see Figure 6.2) [2]. In a TDMA system, time is divided into equal time intervals, called slots. User data is transmitted in the slots. Several slots make up a frame. Guard times are used between each user’s transmission to minimize crosstalk between channels (see Figure 6.4). Each user is assigned a frequency and a time slot to transmit data. The data is transmitted via a radio-carrier from a base station to several active mobiles in the downlink. In the reverse direction (uplink), transmission from mobiles to base stations is time-sequenced and synchronized on a common frequency for TDMA. The preamble carries the address and synchronization information that both the base station and mobile stations use for identification. Slot 1

Slot 2

...

Slot m

Frequency 1

Circuit

Circuit

Circuit

Frequency 2

Circuit

Circuit

Circuit

Downlink Path Frequency n

Circuit

Circuit

Circuit

Frequency 1

Circuit

Circuit

Circuit

Frequency 2

Circuit

Circuit

Circuit

Frequency Domain

Uplink Path

Frequency n

Figure 6.3

Circuit

TDMA / FDD channel architecture.

Circuit

Circuit

6.2

Narrowband Channelized Systems

155

One frame

Preamble

Slot 1

Preamble (Trail + Synch.)

Figure 6.4

Slot 2

Information message

Slot m

Trail bits

Guard bit (time)

TDMA frame.

In a TDMA system, the user can use multiple slots to support a wide range of bit rates by selecting the lowest multiplexing rate or multiple of it. This enables supporting a variety of voice coding techniques at different bit rates with different voice qualities. Similarly, data communications customers could use the same kinds of schemes, choosing and paying for the digital data rate as required. This would allow customers to request and pay for a bandwidth on demand. Depending on the data rate used and the number of slots per frame, a TDMA system can use the entire bandwidth of the system or can employ an FDD scheme. The resultant multiplexing is a mixture of frequency division and time division. The entire frequency band is divided into a number of duplex channels (about 350 to 400 kHz). These channels are deployed in a frequency-reuse pattern, in which radio-port frequencies are assigned using an autonomous adaptive frequency assignment algorithm. Each channel is configured in a TDM mode for the downlink direction and a TDMA mode for the uplink direction. The advantages and disadvantages of TDMA are: Advantages 1. TDMA permits a flexible bit rate, not only for multiples of the basic single channel rate but also submultiples for low bit rate broadcast-type traffic. 2. TDMA offers the opportunity for frame-by-frame monitoring of signal strength/bit error rates to enable either mobiles or base stations to initiate and execute handoffs. 3. TDMA, when used exclusively and not with FDMA, utilizes bandwidth more efficiently because no frequency guard band is required between channels.

156

6

Multiple Access Techniques

4. TDMA transmits each signal with sufficient guard time between time slots to accommodate time inaccuracies because of clock instability, delay spread, transmission delay because of propagation distance, and the tails of signal pulse because of transient responses. Disadvantages 1. For mobiles and particularly for hand-sets, TDMA on the uplink demands high peak power in transmit mode, that shortens battery life. 2. TDMA requires a substantial amount of signal processing for matched filtering and correlation detection for synchronizing with a time slot. 3. TDMA requires synchronization. If the time slot synchronization is lost, the channels may collide with each other. 4. One complicating feature in a TDMA system is that the propagation time for a signal from a mobile station to a base station varies with its distance to the base station.

6.3

Spectral Efficiency

An efficient use of the spectrum is the most desirable feature of a mobile communications system. To realize this, a number of techniques have been proposed or already implemented. Some of these techniques used to improve spectral efficiency are reducing the channel bandwidth, information compression (low-rate speech coding), variable bit rate codec (see Chapter 8), improved channel assignment algorithms (dynamic channel assignment), and so on [11,17,19]. Spectral efficiency of a mobile communications system shows how efficiently the spectrum is used by the system. Spectral efficiency of a mobile communications system depends on the choice of a multiple access scheme. The measure of spectral efficiency enables one to estimate the capacity of a mobile communications system. The overall spectral efficiency of a mobile communications system can be estimated by knowing the modulation and the multiple access spectral efficiencies separately [16].

6.3.1 Spectral Efficiency of Modulation The spectral efficiency with respect to modulation is defined as [16]: Number of Channels Available in the System) m  (Total  (Bandwidth)(Total Coverage Area)

m 

Bw Nc  Bc N 

Bw  Nc  Ac

(6.1a)

(6.1b)

6.3

Spectral Efficiency

157

1 m   Channels/MHz/km2 Bc  N  Ac

(6.1c)

where: m  modulation efficiency (channels/MHz/km2) Bw  bandwidth of the system (MHz) Bc  channel spacing (MHz) Nc  total number of cells in the covered area N  frequency reuse factor of the system (or cluster size) Ac  area covered by a cell (km2) Equation 6.1c indicates that the spectral efficiency of modulation does not depend on the bandwidth of the system. It only depends on the channel spacing, the cell area, and the frequency reuse factor, N. By reducing the channel spacing, the spectral efficiency of modulation for the system is increased, provided the cell area (Ac) and reuse factor (N) remain unchanged. If a modulation scheme can be designed to reduce N then more channels are available in a cell and efficiency is improved. Another definition of spectral efficiency of modulation is Erlangs/MHz/km2 Maximum Total Traffic Carried by System (System Bandwidth)(Total Coverage Area)

m  





B /B N m   w c Total Traffic Carried by  Channels

Bw Ac

(6.2a)

(6.2b)

By introducing the trunking efficiency factor, t in Equation 6.2b (1, it is a function of the blocking probability and number of available channels per cell), the total traffic carried by the system is given as:





B /B N m   w c t 

Bw Ac



t m  

BcNAc

(6.2c) (6.2d)

where: t is a function of the blocking probability and the total number of available



B /B N

w c channels per cell 



158

6

Multiple Access Techniques

Based on Equation 6.2d we can conclude: 1. The voice quality will depend on the frequency reuse factor, N, which is a function of the signal-to-interference (S/I) ratio of the modulation scheme used in the mobile communications system (see Chapter 5). 2. The relationship between system bandwidth, Bw, and the amount of traffic carried by the system is nonlinear, i.e., for a given percentage increase in Bw, the increase in the traffic carried by the system is more than the increase in Bw. 3. From the average traffic per user (Erlang/user) during the busy hour and Erlang/MHz/km2, the capacity of the system in terms of users/km2/MHz can be obtained. 4. The spectral efficiency of modulation depends on the blocking probability. Example 6.1 In the GSM800 digital channelized cellular system, the one-way bandwidth of the system is 12.5 MHz. The RF channel spacing is 200 kHz. Eight users share each RF channel and three channels per cell are used for control channels. Calculate the spectral efficiency of modulation (for a dense metropolitan area with small cells) using the following parameters: • Area of a cell  8 km2 • Total coverage area  4000 km2 • Average number of calls per user during the busy hour  1.2 • Average holding time of a call  100 seconds • Call blocking probability  2% • Frequency reuse factor  4

Solution 12.5  1000  62 Number of 200 kHz RF channels   200

Number of traffic channels  62  8  496 Number of signaling channels per cell  3 496 Number of traffic channels per cell   3  121 4000 Number of cells    500

4

8

With 2% blocking for an omnidirectional case, the total traffic carried by 121 channels (using Erlang-B tables)  108.4 (1.0  0.02)  106.2 Erlangs/cell or 13.28 Erlangs/km2

6.3

Spectral Efficiency

159

106.2  3600 Number of calls per hour per cell    3823, calls/hour/km2  3823 8

100

  477.9 calls/hour/km2

3823 Max. number of users/cell/hour    3186, users/hour/channel  3186  1.2

121

26.33 per cell)  no. of cells 106.2  500 m  (Erlangs     1.06 Erlangs/MHz/km2 Bw  Coverage Area

12.5  4000

6.3.2 Multiple Access Spectral Efficiency Multiple access spectral efficiency is defined as the ratio of the total time or frequency dedicated for traffic transmission to the total time or frequency available to the system. Thus, the multiple access spectral efficiency is a dimensionless number with an upper limit of unity. In FDMA, users share the radio spectrum in the frequency domain. In FDMA, the multiple access efficiency is reduced because of guard bands between channels and also because of signaling channels. In TDMA, the efficiency is reduced because of guard time and synchronization sequence. FDMA Spectral Efficiency For FDMA, multiple access spectral efficiency is given as: BcNT

a   1 B w

(6.3)

where: a  multiple access spectral efficiency NT  total number of traffic channels in the covered area Bc  channel spacing Bw  system bandwidth Example 6.2 In a first-generation AMP system where there are 395 channels of 30 kHz each in a bandwidth of 12.5 MHz, what is the multiple access spectral efficiency for FDMA? Solution 30  395 12.5  1000

a    0.948

160

6

Multiple Access Techniques

TDMA Spectral Efficiency TDMA can operate as wideband or narrowband. In the wideband TDMA, the entire spectrum is used by each individual user. For the wideband TDMA, multiple access spectral efficiency is given as: Mt Tf

a  

(6.4)

where:  duration of a time slot that carries data Tf  frame duration Mt  number of time slots per frame In Equation 6.4 it is assumed that the total available bandwidth is shared by all users. For the narrowband TDMA schemes, the total band is divided into a number of sub-bands, each using the TDMA technique. For the narrowband TDMA system, frequency domain efficiency is not unity as the individual user channel does not use the whole frequency band available to the system. The multiple access spectral efficiency of the narrowband TDMA system is given as:



 Mt 

a   Tf

  BBN   u u 

(6.5)

w

where: Bu  bandwidth of an individual user during his or her time slot Nu  number of users sharing the same time slot in the system, but having access to different frequency sub-bands

6.3.3 Overall Spectral Efficiency of FDMA and TDMA Systems The overall spectral efficiency, , of a mobile communications system is obtained by considering both the modulation and multiple access spectral efficiencies    m a

(6.6)

Example 6.3 In the North American Narrowband TDMA cellular system, the one-way bandwidth of the system is 12.5 MHz. The channel spacing is 30 kHz and the total number of voice channels in the system is 395. The frame duration is 40 ms, with six time slots per frame. The system has an individual user data rate of 16.2 kbps in which the speech with error protection has a rate of 13 kbps. Calculate the multiple access spectral efficiency of the TDMA system.

6.3

Spectral Efficiency

161

Solution

 16.2   6 

13 40 The time slot duration that carries data:     5.35 ms

Tf  40 ms, Mt  6, Nu  395, Bu  30 kHz, and Bw  12.5 MHz 5.35  6 30  395 a     0.76 40

12500

The overhead portion of the frame  1.0  0.76  24% Capacity and Frame Efficiency of a TDMA System Cell Capacity The cell capacity is defined as the maximum number of users that can be supported simultaneously in each cell. The capacity of a TDMA system is given by [16]: b

B RN

w Nu    f

(6.7)

where: Nu  number of channels (mobile users) per cell b  bandwidth efficiency factor (1.0)

 bit efficiency ( 2 bit/symbol for QPSK,  1 bit/symbol for GMSK as used in GSM) f  voice activity factor (equal to one for TDMA) Bw  one-way bandwidth of the system R  information (bit rate plus overhead) per user N  frequency reuse factor Nu  R

Spectral efficiency    bit/sec/Hz B w

(6.8)

Example 6.4 Calculate the capacity and spectral efficiency of a TDMA system using the following parameters: bandwidth efficiency factor b  0.9, bit efficiency (with QPSK)

 2, voice activity factor f  1.0, one-way system bandwidth Bw  12.5 MHz, information bit rate R  16.2 kbps, and frequency reuse factor N  19. Solution 12.5  106 16.2  10  19

0.9  2 Nu     3 1.0

N  73.1 (say 73 mobile users per cell)

162

6

Multiple Access Techniques

73  16.2 Spectral efficiency     0.094 bit/sec/Hz 12.5  1000

Efficiency of a TDMA Frame We refer to Figure 6.4 that shows a TDMA frame. The number of overhead bits per frame is: b0  Nrbr Ntbp (Nt Nr)bg

(6.9)

where: Nr number of reference bursts per frame Nt  number of traffic bursts (slots) per frame br  number of overhead bits per reference burst bp  number of overhead bits per preamble per slot bg  number of equivalent bits in each guard time interval The total number of bits per frame is: bT  Tf  Rrf

(6.10a)

where: Tf  frame duration Rrf  bit rate of the RF channel Frame efficiency   (1  b0 /bT)  100%

(6.10b)

It is desirable to maintain the efficiency of the frame as high as possible. The number of bits per data channel (user) per frame is bc  RTf, where R  bit rate of each channel (user). No. of channels/frame

(Total Data Bits)/(frame) (Bits per Channel)/(frame)

NCF   RrfTf

NCF   RTf

Rrf

NCF   R

Equation 6.11b indicates the number of time slots per frame.

(6.11a)

(6.11b)

6.4

Wideband Systems

163

Example 6.5 Consider the GSM TDMA system with the following parameters: Nr  2 Nt  24 frames of 120 ms each with eight time slots per frame br  148 bits in each of 8 time slots bp  34 bits in each of 8 time slots bg  8.25 bits in each of 8 time slots Tf  120 ms Rrf  270.8333333 kbps R  22.8 kbps Calculate the frame efficiency and the number of channels per frame. Solution b0  2  (8  148) 24  (8  34) 8  8.25  10,612 bits per frame bT  120  103  270.8333333  103  32,500 bits per frame





10612  1  100  67.35% 32500

0.6735  270.8333333 22.8

Number of channels/frame    8 The last calculation, with an answer of 8 channels, confirms that our calculation of efficiency is correct.

6.4

Wideband Systems

In wideband systems, the entire system bandwidth is made available to each user, and is many times larger than the bandwidth required to transmit information. Such systems are known as spread spectrum (SS) systems. There are two fundamental types of spread spectrum systems: (1) direct sequence spread spectrum (DSSS) and (2) frequency hopping spread spectrum (FHSS) [3,26]. In a DSSS system, the bandwidth of the baseband information carrying signals from a different user is spread by different codes with a bandwidth much larger than that of the baseband signals (see Chapter 11 for details). The spreading codes used for different users are orthogonal or nearly orthogonal to each other. In the DSSS, the spectrum of the transmitted signal is much wider than the spectrum associated with the information rate. At the receiver, the same code is used for despreading to recover the baseband signal from the target user while suppressing the transmissions from all other users (see Figure 6.5). One of the advantages of the DSSS system is that the transmission bandwidth exceeds the coherence bandwidth (see Chapter 3). The received signal, after despreading (see Chapter 11 for details), resolves into multiple signals with different time delays. A Rake receiver (see Chapter 11) can be used to recover the multiple time

164

6

Multiple Access Techniques

Code c (t ) signal s (t )

BC

BS

After spreading s (t ) c (t ) BC ω

After modulation s (t ) c (t ) cos ω τ

ω

BC

frequency

BC

Modulator spreading s (t )

modulation s( t ) c( t )

s (t ) c (t ) cos ω τ cos ωt

c (t ) Demodulator

despreading

demodulation s(t ) c (t ) cos ω τ

LPF

s( t ) c( t )

cos ω τ

dt

s (t )

c (t )

LPF: Low-pass filter

Figure 6.5

Direct sequence spread spectrum.

delayed signals and combine them into one signal to provide a time diversity with a lower frequency of deep fades. Thus, the DSSS system provides an inherent robustness against mobile-channel degradations. Another potential benefit of a DSSS system is the greater resistance to interference effects in a frequency reuse situation. Also, there may be no hard limit on the number of mobile users who can simultaneously gain access. The capacity of a DSSS system depends upon the desired value of Eb/I0 instead of resources (frequencies or time slots) as in FDMA or TDMA systems. Frequency hopping (FH) is the periodic changing of the frequency or the frequency set associated with transmission (see Figure 6.6). If the modulation is M-ary frequency-shift-keying (MFSK) (see Chapter 9 for details), two or more frequencies are in the set that change at each hop. For other modulations, a single center or carrier frequency is changed at each hop. An FH signal may be considered a sequence of modulated pulses with pseudorandom carrier frequencies. The set of possible carrier frequencies is called the hop set. Hopping occurs over a frequency band that includes a number of frequency channels. The bandwidth of a frequency channel is called the instantaneous bandwidth (BI). The bandwidth of the frequency band over which the hopping occurs is called the total hopping bandwidth (BH). The time duration between hops is called the hop duration or hopping period (TH).

6.4

Wideband Systems

165

Frequency

fn fn 1 fn 2 f3 f2 f1 t 0 Tc Figure 6.6

Tc

2

Frequency hopping spread spectrum system.

Frequency hopping can be classified as fast or slow. Fast frequency hopping occurs if there is frequency hop for each transmitted symbol. Thus, fast frequency hopping implies that the hopping rate equals or exceeds the information symbol rate. Slow frequency hopping occurs if two or more symbols are transmitted in the time interval between frequency hops. Frequency hopping allows communicators to hop out of frequency channels with interference or to hop out of fades. To exploit this capability, error-correcting codes, appropriate interleaving, and disjoint frequency channels are nearly always used. A frequency synthesizer is required for frequency hopping systems to convert a stable reference frequency into the various frequency of hop set. Frequency hopping communicators do not often operate in isolation. Instead, they are usually elements of a network of frequency hopping systems that create mutual multiple-access interference. This network is called a frequency-hopping multiple-access (FHMA) network. If the hoppers of an FHMA network all use the same M frequency channels, but coordinate their frequency transitions and their hopping sequence, then the multipleaccess interference for a lightly loaded system can be greatly reduced compared to a non-hopped system. For the number of hopped signals (Mh) less than the number of channels (Nc), a coordinated hopping pattern can eliminate interference. As the number of hopped signals increases beyond Nc, then the interference will increase in proportion to the ratio of the number of signals to the number of channels. In the absence of fading or multipath interference, since there is no interference suppression system in frequency hopping, for a high channel loading the performance of a frequency hopping system is no better than a non-hopped system. Frequency hopping systems are best for light channel loadings in the presence of conventional non-hopped systems.

166

6

Multiple Access Techniques

When fading or multipath interference is present, the frequency hopping system has better error performance than a non-hopped system. If the transmitter hops to a channel in a fade, the errors are limited in duration since the system will shortly hop to a new frequency where the fade may not be as deep. Network coordination for frequency hopping systems are simpler to implement than that for DSSS systems because the timing alignments must be within a fraction of a hop duration, rather than a fraction of a sequence chip (narrow pulse). In general, frequency hopping systems reject interference by trying to avoid it, whereas DSSS systems reject interference by spreading it. The interleaving and error-correcting codes that are effective with frequency hopping systems are also effective with DSSS systems. The major problems with frequency hopping systems with increasing hopping rates are the cost of the frequency synthesizer increases and its reliability decreases, and synchronization becomes more difficult. In theory, a wideband system can be overlaid on existing, fully loaded, narrowband channelized systems (as an example, the IS-95 CDMA system overlays on existing AMPS [FDMA]). Thus, it may be possible to create a wideband network right on top of the narrowband cellular system using the same spectrum.

6.5

Comparisons of FDMA, TDMA, and DS-CDMA

The DSSS approach is the basis to implementation of the direct sequence code division multiple access (DS-CDMA) technique introduced by Qualcom. The DSCDMA has been used in commercial applications of mobile communications. The primary advantage of DS-CDMA is its ability to tolerate a fair amount of interfering signals compared to FDMA and TDMA that typically cannot tolerate any such interference(Figure 6.7). As a result of the interference tolerance of CDMA, the problems of frequency band assignment and adjacent cell interference are greatly simplified. Also, flexibility in system design and deployment are significantly improved since interference to others is not a problem. On the other hand, FDMA and TDMA radios must be carefully assigned a frequency or time slot to assure that there is no interference with other similar radios. Therefore, sophisticated filtering and guard band protection is needed with FDMA and TDMA technologies. With DS-CDMA, adjacent microcells share the same frequencies whereas with FDMA/TDMA it is not feasible for adjacent microcells to share the same frequencies because of interference. In both FDMA and TDMA systems, a time-consuming frequency planning task is required whenever a network changes, whereas no such frequency planning is needed for a CDMA network since each cell uses the same frequencies. Capacity improvements with DS-CDMA also result from voice activity patterns during two-way conversations, (i.e., times when a party is not talking) that cannot be cost-effectively exploited in FDMA or TDMA systems. DS-CDMA radios can, therefore, accommodate more mobile users than FDMA/TDMA radios

6.5

Comparisons of FDMA, TDMA, and DS-CDMA

Time

167

Time FDMA TDMA User 1 User 3 User 2 User 1 1 2 3 4

Frequency

Frequency

Time DS-CDMA

Frequency

Figure 6.7

Comparison of multiple access methods.

on the same bandwidth. Further capacity gains for FDMA, TDMA, and CDMA can also result from antenna technology advancement by using directional antennas that allow the microcell area to be divided into sectors. Table 6.1 provides a summary of access technologies used for various wireless systems. Table 6.1 Access technologies for wireless system.

System

Access technology

Mode of operation

Frame rate (kbps)

North American IS-54 (Dual Mode)

TDMA/FDD FDMA/FDD

Digital/ Analog FM

48.6 —

North American IS-95 (Dual Mode)

DS-CDMA/FDD FDMA/FDD

Digital/ Analog FM

1228.8 —

North American IS-136

TDMA/FDD

Digital

48.6

GSM (used all over world)

TDMA/FDD

Digital

270.833

European CT-2 Cordless

FDMA/TDD

Digital

72.0

DECT Cordless

TDMA/TDD

Digital

1152.0

168

6.6

6

Multiple Access Techniques

Capacity of a DS-CDMA System

The capacity of a DS-CDMA system depends on the processing gain, Gp (a ratio of spreading bandwidth, Bw, and information rate, R), the bit energy-to-interference ratio, Eb/I0, the voice duty cycle, vf, the DS-CDMA omnidirectional frequency reuse efficiency, f, and the number of sectors, G, in the cell-site antenna. The received signal power at the cell from a mobile is S  R  Eb. The signal-to-interference ratio is S  R  Eb   Bw I0 I



(6.12)

where: Eb  energy per bit I0  interference density In a cell with Nu mobile transmitters, the number of effective interferers is Nu  1 because each mobile is an interferer to all other mobiles. This is valid regardless of how the mobiles are distributed within the cell since automatic power control (APC) is used in the mobiles. The APC operates such that the received power at the cell from each mobile is the same as for every other mobile in the cell, regardless of the distance from the center of the cell. APC conserves battery power in the mobiles, minimizes interference to other users, and helps overcome fading. In a hexagonal cell structure, because of interference from each tier, the S/I ratio is given as (see Chapter 5): S I

1 (Nu  1)  [1 6  k1 12  k2 18  k3 . . . ]

  

(6.13)

where: Nu  number of mobile users in the band, Bw ki, i  1, 2, 3, . . .  the interference contribution from all terminals in individual cells in tiers 1, 2, 3, etc., relative to the interference from the center cell. This loss contribution is a function of both the path loss to the center cell and the power reduction because of power control to an interfering mobile’s own cell center. If we define a frequency reuse efficiency, f, as in Equation 6.14a, then Eb /I0 is given by Equation 6.15. 1 [1 6  k1 12  k2 18  k3 . . .]

f  

(6.14a)

6.6

Capacity of a DS-CDMA System

169

f S 

(6.14b)

(Nu  1)

I

f Bw Eb    R I0 (Nu  1)

(6.15)

This equation does not include the effect of background thermal and spurious noise (i.e., ) in the spreading bandwidth Bw. Including this as an additive degradation term in the denominator results in a bit energy-to-interference ratio of: f Bw Eb      R I0 (Nu  1) /S

(6.16)

Note that from Equation 6.16 the capacity of the DS-CDMA system is reduced by /S which is the ratio of background thermal plus spurious noise to power level. For a fixed Gp  Bw /R, one way to increase the capacity of the DS-CDMA system is to reduce the required Eb /I0, which depends upon the modulation and coding scheme. By using a powerful coding scheme, the Eb /I0 ratio can be reduced, but this increases system complexity. Also, it is not possible to reduce the Eb/I0, ratio indefinitely. The only other way to increase the system capacity is to reduce the interference. Two approaches are used: one is based on the natural behavior of human speech and the other is based on the application of the sectorized antennas. From experimental studies it has been found that typically in a full duplex 2-way voice conversation, the duty cycle of each voice is, on the average, less than 40%. Thus, for the remaining period of time the interference induced by the speaker can be eliminated. Since the channel is shared among all the users, noise induced in the desired channel is reduced due to the silent interval of other interfering channels. It is not cost-effective to exploit the voice activity in the FDMA or TDMA system because of the time delay associated with reassigning the channel resource during the speech pauses. If we define vf as the voice activity factor (1), then Equation 6.16 can be written as: f Bw Eb 1   vf  R   (Nu  1) /S I0

f

      I

0 Bw (Nu  1)    vf     Eb S R

(6.17a)

(6.17b)

170

6

Multiple Access Techniques

The equation to determine the capacity of a DS-CDMA system should also include additional parameters to reflect the bandwidth efficiency factor, the capacity degradation factor due to imperfect power control, and the number of sectors in the cell-site antenna. Equation 6.17b is augmented by these additional factors to provide the following equation for DS-CDMA capacity at one cell:  c 

f b d Bw Nu    1   vf

R  (Eb/I0)

S

(6.18a)

Equation 6.18a can be rewritten as Equation 6.18b by neglecting the last two terms.  c  vf

f b d Bw Nu   

R  (Eb /I0)

(6.18b)

where: f  frequency reuse efficiency 1 b  bandwidth efficiency factor 1 cd  capacity degradation factor to account for imperfect APC 1 vf  voice activity factor 1 Bw  one-way bandwidth of the system R  information bit rate plus overhead Eb  energy per bit of the desired signal Eb /I0  desired energy-to-interference ratio (dependent on quality of service)   efficiency of sector-antenna in cell ( G, number of sectors in the cell-site antenna) For digital voice transmission, Eb /I0 is the required value for a bit error rate (BER) of about 103 or better, and f depends on the quality of the diversity. Under the most optimistic assumption, f 0.5. The voice activity factor, vf is usually assumed to be less than or equal to 0.6. Eb /I0 for a BER of 103 can be as high as 63 (18 dB) if no coding is used and as low as 5 (7 dB) for a system using a powerful coding scheme. The capacity degradation factor, cd will depend on the implementation but will always be less than 1. Example 6.6 Calculate the capacity and spectral efficiency of the DS-CDMA system with an omnidirectional cell using the following data: • bandwidth efficiency b  0.9 • frequency reuse efficiency f  0.45

6.7

Comparison of DS-CDMA vs. TDMA System Capacity

171

• capacity degradation factor cd  0.8 • voice activity factor vf  0.4 • information bit rate R  16.2 kbps • Eb /I0  7 dB • one-way system bandwidth Bw  12.5 MHz

Neglect other sources of interference. Solution Eb /I0  5.02 (7 dB) 12.5  10 0.45  0.9  0.8  1   Nu   3 6

0.4

16.2  10  5.02

Nu  124.5 (say 125) 125  16.2  0.162 bits/sec/Hz The spectral efficiency,    3 12.5  10

In these calculations, an omnidirectional antenna is assumed. If a three sector antenna (i.e., G  3) is used at a cell site with   2.6, the capacity will be increased to 325 mobile users per cell, and spectral efficiency will be 0.421 bits/sec/Hz.

6.7

Comparison of DS-CDMA vs. TDMA System Capacity

Using Equations 6.7 and 6.18b with f  1 (no voice activity) for TDMA and   1.0 (omnidirectional cell) for DS-CDMA the ratio of the cell capacity for the DS-CDMA and TDMA systems is given as: cdNf NCDMA 1  1  RTDMA     

f NTDMA RCDMA Eb /I0 cdma

(6.19)

Example 6.7 Using the data given in Examples 6.4 and 6.6, compare the capacity of the DS-CDMA and TDMA omnidirectional cell. Solution NCDMA 1  1  16.2  1.703 0.8  19  0.45        NTDMA 0.4 2 16.2 5.02

172

6.8

6

Multiple Access Techniques

Frequency Hopping Spread Spectrum with M-ary Frequency Shift Keying

The FHSS system uses M-ary frequency shift keying modulation (MFSK) and involves the hopping of the carrier frequency in a random manner. It uses MFSK, in which b  log2M information bits determine which one of M frequencies is to be used [19]. The portion of the M-ary signal set is shifted pseudo-randomly by the frequency synthesizer over a hopping bandwidth, Bss. A typical block diagram is shown in Figure 6.8. In a conventional MFSK system, the data symbol is modulated on a carrier whose frequency is pseudo-randomly determined. The frequency synthesizer produces a transmission tone based on simultaneous dictates of the pseudonoise (PN) code (see Chapter 11) and the data. At each frequency hop time a PN generator feeds the frequency synthesizer a frequency word (a sequence of L chips), which dictates one of 2L symbol-set positions. The FH bandwidth, Bss, and the minimum frequency spacing between consecutive hop positions, f, dictate the minimum number of chips required in the frequency word. Example 6.8 A hopping bandwidth, Bss, of 600 MHz and a frequency step size, f, of 400 Hz are used. What is the minimum number of PN chips that are required for each frequency word? Solution Bss  106  600 Number of tones contained in Bss     1.5  106 f

400

Minimum number of chips required  L log2(1.5  106) M  20 chips

Transmitter Data

MFSK Modulator

FH Modulator

Receiver

Channel 

FH Demodulator

Interference PN Generator

Figure 6.8 Frequency hopping using MFSK.

PN Generator

MFSK Demodulator

Data

6.9

6.9

Orthogonal Frequency Division Multiplexing (OFDM)

173

Orthogonal Frequency Division Multiplexing (OFDM)

In this section we briefly introduce OFDM. For more details readers should refer to [19]. OFDM uses three transmission principles, multirate, multisymbol, and multicarrier. OFDM is similar to frequency division multiplexing (FDM). OFDM distributes the data over a large number of carriers that are spaced apart at precise frequencies. The spacing provides the orthogonality in this technique, which prevents the demodulator from seeing frequencies other than their own. Multiple Input, Multiple Output-OFDM (MIMO-OFDM) uses multiple antennas to transmit and receive radio signals. MIMO-OFDM allows service providers to deploy a broadband wireless access system that has non-line-of-sight (NLOS) functionality. MIMO-OFDM takes advantage of the multipath properties of the environment using base station antennas that do not have LOS. The MIMO-OFDM system uses multiple antennas to simultaneously transmit data in small pieces to the receiver, which can process the data flow and put it back together. This process, called spatial multiplexing, proportionally boosts the data transmission speed by a factor equal to the number of transmitting antennas. In addition, since all data is transmitted both in the same frequency band and with separate spatial signatures, this technique utilizes spectrum efficiently. VOFDM (vector OFDM) uses the concept of MIMO technology. We consider a data stream operating at R bps and an available bandwidth of Nf centered at fc. The entire bandwidth could be used to transmit a data stream, in which case the bit duration would be 1/R. By splitting the data stream into N substreams using a serial-to-parallel converter, each substream has a data rate of R/N and is transmitted on a separate subcarrier, with spacing between adjacent subcarriers of f (see Figure 6.9). The bit duration is N/R. The advantage of OFDM is that on a multiple channel the multipath is reduced relative to the symbol interval by a ratio of 1/N and thus imposes less distortion in each modulated symbol. OFDM overcomes inter-symbol interference (ISI) in a multipath environment. ISI has a greater impact at higher data rates because the distance between bits or symbols is smaller. With OFDM, the data rate is reduced by a factor of N, which increases the symbol duration by a factor of N. Thus, if the symbol duration is Ts for the source stream, the duration of OFDM signals is NTs. This significantly reduces the effect of ISI. As a design criterion, N is selected so that NTs is significantly greater than rms (rms delay spread) of the channel. With the use of OFDM, it may not be necessary to deploy an equalizer. OFDM is an ideal solution for broadband communications, because increasing the data rate is simply a matter of increasing the number of subcarriers. To avoid overlap between consecutive symbols, a time guard is enforced between the transmissions of two OFDM pulses that will reduce the effective data rate. Also, some subcarriers are devoted to synchronization of signal, and some are reserved for redundancy.

174

6

Multiple Access Techniques

R/N Modulator

fc  (N  1)f/2 R/N Modulator

R

Serial-toparallel converter

fc  f/2 R/N Modulator

fc  f/2 R/N Modulator

R  Input data rate R/N  Input to each sub-channel

fc  (N  1)f/2

Figure 6.9 Orthogonal frequency division multiplexing (OFDM).

The most important feature of OFDM is the orthogonal relationship between the subcarrier signals. Orthogonality allows the OFDM subcarriers to overlap each other without interference. OFDM uses FH to create a spread spectrum system. FH has several advantages over DSSS, for example, no near-far problem, easier synchronization, less complex receivers, and so on. In the OFDM the input information sequence is first converted into parallel data sequences and each serial/parallel converter output is multiplied with spreading code. Data from all subcarriers is modulated in baseband by inverse fast Fourier transform (IFFT) and converted back into serial data. The guard interval is inserted between symbols to avoid ISI caused by multipath fading and finally the signal is transmitted after RF up-conversion. At the receiver, after down-conversion, the m-subcarrier component corresponding to the received data is first coherently detected with FFT and then multiplied with gain to combine the energy of the received signal scattered in the frequency domain (see Figure 6.10). Wireless Local Area Networks (WLAN) development is ongoing for wireless point-to-point and point-to-multipoint configurations using OFDM technology.

6.10

Multicarrier DS-CDMA (MC-DS-CDMA)

Data

Serial To Parallel

Modulation

175

Parallel To Serial

Guard Interval

Demodulation

Parallel To Serial

IFFT

Data

OFDM Transmitter Data

Guard Interval Removal

Serial To Parallel

FFT

Data

OFDM Receiver Figure 6.10 IEEE 802.11 a Transmit and Receive OFDM.

In a supplement to the IEEE 802.11 standard, the IEEE 802.11 working group published IEEE 802.11a, which outlines the use of OFDM in the 5.8 GHz band. The basic principal of operation is to divide a high-speed binary signal to be transmitted into a number of lower data rate subcarriers. There are 48 data subcarriers and 4 pilot subcarriers for a total of 52 subcarriers. Each lower data rate bit stream is used to modulate a separate subcarrier from one of the channels in the 5 GHz band. Prior to transmission the data is encoded using convolutional code (see Chapter 8) of rate, R  1/2 and bit interleaved for the desired data rate. Each bit is then mapped into a complex number according to the modulation type and subdivided in 48 data subcarriers and 4 pilot subcarriers. The subcarriers are combined using an IFFT and transmitted. At the receiver, the carrier is converted back to a multicarrier lower data rate form using FFT. The lower data subcarriers are combined to form a high rate data unit.

6.10

Multicarrier DS-CDMA (MC-DS-CDMA)

Future wireless systems such as a fourth-generation (4G) system will need flexibility to provide subscribers with a variety of services such as voice, data, images, and video. Because these services have widely differing data rates and traffic profiles, future generation systems will have to accommodate a wide variety of data rates. DS-CDMA has proven very successful for large-scale cellular voice systems, but there are concerns whether DS-CDMA will be well-suited to non-voice traffic. The DS-CDMA system suffers inter-symbol interference (ISI) and multi-user interference (MUI) caused by multipath propagation, leading to a high loss of performance. With OFDM, the time dispersive channel is seen in the frequency domain as a set of parallel independent flat subchannels and can be equalized at a low complexity. There are potential benefits to combining OFDM and DS-CDMA. Basically the frequency-selective channel is first equalized in the frequency domain using the

176

6

Multiple Access Techniques

OFDM modulation technique. DS-CDMA is applied on top of the equalized channel, keeping the orthogonality properties of spreading codes. The combination of OFDM and DS-CDMA is used in MC-DS-CDMA. MC-DS-CDMA [4,5,12,25] marries the best of the OFDM and DS-CDMA world and, consequently, it can ensure good performance in severe multipath conditions. MC-DS-CDMA can achieve very large average throughput. To further enhance the spectral efficiency of the system, some form of adaptive modulation can be used. Basically, three main designs exist in the literature, namely, MC-CDMA, MC-DS-CDMA, and multitone (MT)-CDMA. In MC-CDMA, the spreading code is applied across a number of orthogonal subcarriers in the frequency domain. In MC-DS-CDMA, the data stream is first divided into a number of substreams. Each substream is spread in time through a spreading code and then transmitted over one of a set of orthogonal subcarriers. In MT-CDMA the system undergoes similar operations as MC-DS-CDMA except that the different subcarriers are not orthogonal after spreading. This allows higher spectral efficiencies and longer spreading codes; however, different substreams interfere with one other. The MCDS-CDMA transmitter spreads the original data stream over different orthogonal subcarriers using a given spreading code in the frequency domain.

6.11

Random Access Methods

So far we have discussed the reservation-based schemes, now we focus on random-access schemes [8]. When each user has a steady flow of information to transmit (for example, a data file transfer or a facsimile transmission), reservationbased access methods are useful as they make an efficient use of communication resources. However, when the information to be transmitted is bursty in nature, the reservation-based access methods result in wasting communication resources. Furthermore, in a cellular system where subscribers are charged based on a channel connection time, the reservation-based access methods may be too expensive to transmit short messages. Random-access protocols provide flexible and efficient methods for managing a channel access to transmit short messages. The randomaccess methods give freedom for each user to gain access to the network whenever the user has information to send. Because of this freedom, these schemes result in contention among users accessing the network. Contention may cause collisions and may require retransmission of the information. The commonly used random-access protocols are pure ALOHA, slotted-ALOHA, and CSMA/CD. In the following section we briefly describe details of each of these protocols and provide the necessary throughput expressions.

6.11.1 Pure ALOHA In the pure ALOHA [18,23] scheme, each user transmits information whenever the user has information to send. A user sends information in packets. After

6.11

Random Access Methods

177

sending a packet, the user waits a length of time equal to the round-trip delay for an acknowledgment (ACK) of the packet from the receiver. If no ACK is received, the packet is assumed to be lost in a collision and it is retransmitted with a randomly selected delay to avoid repeated collisions.* The normalized throughput S (average new packet arrival rate divided by the maximum packet throughput) of the pure ALOHA protocol is given as: S  Ge2G

(6.20)

where G  normalized offered traffic load From Equation 6.20 it should be noted that the maximum throughput occurs at traffic load G  50% and is S  1/2e. This is about 0.184. Thus, the best channel utilization with the pure ALOHA protocol is only 18.4%.

6.11.2 Slotted ALOHA In the slotted-ALOHA [23] system, the transmission time is divided into time slots. Each time slot is made exactly equal to packet transmission time. Users are synchronized to the time slots, so that whenever a user has a packet to send, the packet is held and transmitted in the next time slot. With the synchronized time slots scheme, the interval of a possible collision for any packet is reduced to one packet time from two packet times, as in the pure ALOHA scheme. The normalized throughput S for the slotted-ALOHA protocol is given as: S  GeG

(6.21)

where G  normalized offered traffic load The maximum throughput for the slotted ALOHA occurs at G  1.0 (Equation 6.21) and it is equal to 1/e or about 0.368. This implies that at the maximum throughput, 36.8% of the time slots carry successfully transmitted packets. The best channel utilization with the slotted ALOHA protocol is 36.8% — twice the pure ALOHA protocol.

*It should be noted that the protocol on CDMA access channels as implemented in TIA IS-95-A is based upon the pure ALOHA approach. The mobile station randomizes its attempt for sending a message on the access channel and may retry if an acknowledgment is not received from the base station. For further details, one should reference Section 6.6.3.1.1.1 of TIA IS-95-A.

178

6

Multiple Access Techniques

6.11.3 Carrier Sense Multiple Access (CSMA) The carrier sense multiple access (CSMA) [8,18] protocols have been widely used in both wired and wireless LANs. These protocols provide enhancements over the pure and slotted ALOHA protocols. The enhancements are achieved through the use of the additional capability at each user station to sense the transmissions of other user stations. The carrier sense information is used to minimize the length of collision intervals. For carrier sensing to be effective, propagation delays must be less than packet transmission times. Two general classes of CSMA protocols are nonpersistent and p-persistent. • Nonpersistent CSMA: A user station does not sense the channel continu-

ously while it is busy. Instead, after sensing the busy condition, it waits for a randomly selected interval of time before sensing again. The algorithm works as follows: if the channel is found to be idle, the packet is transmitted; or if the channel is sensed busy, the user station backs off to reschedule the packet to a later time. After backing off, the channel is sensed again, and the algorithm is repeated again. • p-persistent CSMA: The slot length is typically selected to be the maximum

propagation delay. When a station has information to transmit, it senses the channel. If the channel is found to be idle, it transmits with probability p. With probability q  1  p, the user station postpones its action to the next slot, where it senses the channel again. If that slot is idle, the station transmits with probability p or postpones again with probability q. The procedure is repeated until either the frame has been transmitted or the channel is found to be busy. If the station initially senses the channel to be busy, it simply waits one slot and applies the above procedure. • 1-persistent CSMA: 1-persistent CSMA is the simplest form of the p-persistent

CSMA. It signifies the transmission strategy, which is to transmit with probability 1 as soon as the channel becomes idle. After sending the packet, the user station waits for an ACK, and if it is not received within a specified amount of time, the user station waits for a random amount of time, and then resumes listening to the channel. When the channel is again found to be idle, the packet is retransmitted immediately. For more details, the reader should refer to [18]. The throughput expressions for the CSMA protocols are: • Unslotted nonpersistent CSMA aG

Ge S   aG G(1 2a) e

(6.22)

6.11

Random Access Methods

179

• Slotted nonpersistent CSMA aG

aGe S   aG 1e

a

(6.23)

• Unslotted 1-persistent CSMA G[1 G aG(1 G (aG)/2)]eG(1 2a) G(1 2a)  (1  e ) (1 aG)e

S   aG G(1 a)

(6.24)

• Slotted 1-persistent CSMA GeG(1 a)[1 a  eaG] (1 a)(1  e ) ae

S   aG G(1 a)

(6.25)

where: S  normalized throughput G  normalized offered traffic load a  /Tp  maximum propagation delay Tp  packet transmission time Example 6.9 We consider a WLAN installation in which the maximum propagation delay is 0.4 sec. The WLAN operates at a data rate of 10 Mbps, and packets have 400 bits. Calculate the normalized throughput with: (1) an unslotted nonpersistent, (2) a slotted persistent, and (3) a slotted 1-persistent CSMA protocol. Solution 400 Tp    40 s 10

 0.4  0.01 a   Tp

40

 106  10  106 G  40   1 400

180

6

Multiple Access Techniques

• Slotted nonpersistent:  1  e0.01  0.495 S  0.01  0.01 1e

0.01

• Unslotted nonpersistent: 0.01

1e S   0.01  0.493 (1 0.02) e

• Slotted 1-persistent: e1.01(1 0.01  e0.01) (1 0.01)(1  e ) 0.01e

S   0.01 1.01  0.531

6.11.4 Carrier Sense Multiple Access with Collision Detection A considerable performance improvement in the basic CSMA protocols can be achieved by means of the carrier sense multiple access with collision detection (CSMA/CD) technique. The CSMA/CD protocols are essentially the same as those for CSMA with addition of the collision-detection feature. Similar to CSMA protocols, there are nonpersistent, 1-persistent, and p-persistent CSMA/CD protocols. More details about CSMA/CD protocols can be found in [27]. When a CSMA/CD station senses that a collision has occurred, it immediately stops transmitting its packets and sends a brief jamming signal to notify all stations of this collision. Collisions are detected by monitoring the analog waveform directly from the channel. When signals from two or more stations are present simultaneously, the composite waveform is distorted from that of a single station. This is manifested in the form of larger than normal voltage amplitude on the cable. In the Ethernet the collision is recognized by the transmitting station, which goes into a retransmission phase based on an exponential random backoff algorithm. The normalized throughput for unslotted nonpersistent and slotted nonpersistent CSMA/CD is given as: Unslotted nonpersistent CSMA/CD aG

Ge S   aG aG aG aG Ge

bG(1  e

) 2aG(1  e

where b  jamming signal length

) (2  e

)

(6.26)

6.11

Random Access Methods

181

Slotted nonpersistent CSMA/CD aG

aGe S   aG aG aG aG aG aGe

b(1  e

 aGe

) a(2  e

 aGe

)

(6.27)

While these collision detection mechanisms are a good idea on a wired local area network (LAN), they cannot be used on a wireless local area network (WLAN) environment for two main reasons: • Implementing a collision detection mechanism would require the implemen-

tation of a full duplex radio capable of transmitting and receiving at the same time — an approach that would increase the cost significantly. • In a wireless environment we cannot assume that all stations hear each other (which is the basic assumption of the collision detection scheme), and the fact that a station wants to transmit and senses the medium as free does not necessarily mean that the medium is free around the receiver area.

6.11.5

Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) IEEE 802.11 uses a protocol known as carrier sense multiple access with collision avoidance (CSMA/CA) or distributed coordination function (DCF). CSMA/CA attempts to avoid collisions by using explicit packet acknowledgment (ACK), which means an ACK packet is sent by the receiving station to confirm that the data packet arrived intact. The CSMA/CA protocol works as follows. A station wishing to transmit senses the medium, if the medium is busy (i.e., some other station is transmitting) then the station defers its transmission to a later time. If no activity is detected, the station waits an additional, randomly selected period of time and then transmits if the medium is still free. If the packet is received intact, the receiving station issues an ACK frame that, once successfully received by the sender, completes the process. If the ACK frame is not detected by the sending station, either because the original data packet was not received intact or the ACK was not received intact, a collision is assumed to have occurred and the data packet is transmitted again after waiting another random amount of time. The CSMA/CA provides a way to share access over the medium. This explicit ACK mechanism also handles interference and other radio-related problems very effectively. However, it does add some overhead to 802.11 that 802.3 does not have, so that an 802.11 WLAN will always have slower performance than the equivalent Ethernet LAN (802.3). The CSMA/CA protocol is very effective when the medium is not heavily loaded since it allows stations to transmit with minimum delay. But there is always a chance of stations simultaneously sensing the medium as being free

182

6

Multiple Access Techniques

and transmitting at the same time, causing a collision. These collisions must be identified, so that the media access control (MAC) layer can retransmit the packet by itself and not by the upper layers, which would cause significant delay. In particular, the hidden node and exposed node problems should be addressed by MAC. Both of them give rise to many performance problems including throughput degradtion, unfair throughput distribution, and throughput instability (see Chapter 18 for details). The IEEE 802.11 uses a collision avoidance (CA) mechanism together with a positive ACK. The MAC layer of a station wishing to transmit senses the medium. If the medium is free for a specified time (called distributed inter-frame space (DIFS)), then the station is able to transmit the packet; if the medium is busy (or becomes busy during the DIFS interval) the station defers using the exponential backoff algorithm. This scheme implies that, except in cases of very high network congestion, no packets will be lost, because retransmission occurs each time a packet is not acknowledged. This entails that all packets sent will reach their destination in sequence. The IEEE 802.11 MAC layer provides cyclic redundancy check (CRC) checksum and packet fragmentation. Each packet has a CRC checksum calculated and attached to ensure that the data was not corrupted in transmit. Packet fragmentation is used to segment large packets into smaller units when sent over the medium. This is useful in very congested environments or when interference is a factor, since large packets have a better chance of being corrupted. This technique reduces the need for retransmission in many cases and improves overall wireless network performance. The MAC layer is responsible for reassembling fragments received, rendering the process transparent to higher-level protocols. The following are some of the reasons it is preferable to use smaller packets in a WLAN environment. • Due to higher BER of a radio link, the probability of a packet getting cor-

rupted increases with packet size. • In case of packet corruption (either due to collision or interference), the smaller the packet, the less overhead it needs to retransmit. A simple stop-and-wait algorithm is used at the MAC sublayer. In this mechanism the transmitting station is not allowed to transmit a new fragment until one of the following happens: • Receives an ACK for the fragment, or • Decides that the fragment was retransmitted too many times and drops the

whole frame.

6.11

Random Access Methods

183

Exponential backoff scheme is used to resolve contention problems among different stations wishing to transmit data at the same time. When a station goes into the backoff state, it waits an additional, randomly selected number of time slots known as a contention window (in 802.11b a slot has a 20 s duration and the random number must be greater than 0 and smaller than a maximum value referred to as a contention window (CW)). During the wait, the station continues sensing the medium to check whether it remains free or if another transmission begins. At the end of its window, if the medium is still free the station can send its frame. If during the window another station begins transmitting data, the backoff counter is frozen and counting down starts again as the channel returns to the idle state. There is a problem related to the CW dimension. With a small CW, if many stations attempt to transmit data at the same time it is possible that some of them may have the same backoff interval. This means that there will continuously be collisions, with serious effects on network performance. On the other hand, with a large CW, if a few stations wish to transmit data they will likely have long backoff delays resulting in degradation of network performance. The solution is to use an exponentially growing CW size. It starts from a small value (CWmin  31) and doubles after each collision, until it reaches the maximum value CWmax (CWmax  1023). In 802.11 the backoff algorithm is executed in three cases: 1. When the station senses the medium busy before the first transmission of a packet 2. Before each retransmission 3. After a successful transmission This is necessary to avoid a single host wanting to transmit a large quantity of data and occupying the channel for too long, denying access to all other stations. The backoff mechanism is not used when the station decides to transmit

DIFS

DIFS PIFS

Contention Window

SIFS Busy Medium

Backoff Window

Next Frame

Slot-time Defer Access

Figure 6.11 CSMA /CA in IEEE 802.11b.

Select slot and decrement backoff as long as medium is idle

184

6

Multiple Access Techniques

a new packet after an idle period and the medium has been free for more than a distributed inter-frame space (DIFS) (see Figure 6.11). To support time-bounded services, the IEEE 802.11 standard defines the point coordinate function (PCF) to let stations have priority access to the wireless medium, coordinated by a station called point coordinate (PC). The PCF has higher priority than DIFS, because it may start transmissions after a shorter duration than DIFS; this time space is called PCF inter frame space (PIFS), which is 25 s for IEEE 802.11 and larger than SIFS. D s The transmission time for a data frame   PLCP   R

where: PLCP  the time required to transmit the physical layer convergence protocol (PLCP) D  the frame size R  the channel bit rate D SIFS A s CSMA/CA packet transmission time  BO DIFS 2PLCP   R

R

where: A  the ACK frame size BO  the backoff time DIFS  the distributed inter-frame space SIFS  the short inter-frame space PIFS  the point coordinate interframe space

6.12

Idle Signal Casting Multiple Access

In the CSMA scheme, each terminal must be able to detect the transmissions of all other terminals. However, not all packets transmitted from different terminals can be sensed, or terminals may be hidden from each other by buildings or some other obstacles. This is known as the hidden terminal problem, which severely degrades the throughput of the CSMA. The idle signal casting multiple access (ISMA) system transmits an idle/busy signal from the base station to indicate the presence or absence of another terminal’s transmission. The ISMA and CSMA are basically the same. In the CSMA, each terminal must listen to all other terminals, whereas in the ISMA, each terminal is informed from the base station of the other terminals’ transmission. Similar to CSMAs, there are nonpersistent ISMAs and 1-persistent ISMAs.

6.13

Packet Reservation Multiple Access

Packet reservation multiple access (PRMA) allows a variety of information sources to share the same communication channel and obtains a statistical multiplexing

6.14

Error Control Schemes for Link Layer

185

effect. In PRMA, time is divided into frames, each of which consists of a fixed number of time slots. For voice terminals, voice activity detection is adopted. The voice signal comprises a sequence of talk spurts. At the beginning of a talk spurt, the terminal transmits the first packet based on slotted ALOHA. Once the packet is transmitted successfully, that terminal is allowed to use the same time slots in the succeeding frames (reservation is made). The reservation is kept until the end of the talk spurt. The status “reserved” or “unreserved” of each slot is broadcast from the base station.

6.14

Error Control Schemes for Link Layer

Error control schemes for the link layer are used to improve the performance of mobile communication systems [8]. Several automatic repeat request (ARQ) schemes are used. At the physical layer of wireless mobile communication systems, error detection and correction techniques such as forward error correction (FEC) schemes are used. For some of the data services, higher layer protocols use ARQ schemes to enable retransmission of any data frames in which an error is detected. The ARQ schemes are classified as follows [23]: • Stop and Wait: The sender transmits the first packet numbered 0 after

storing a copy of that packet. The sender then waits for an ACK numbered 0, (ACK0) of that packet. If the ACK0 does not arrive before a time-out, the sender transmits another copy of the first packet. If the ACK0 arrives before a time-out, the sender discards the copy of the first packet and is ready to transmit the next packet, which it numbers 1. The sender repeats the previous steps, with numbers 0 and 1 interchanged. The advantages of the Stop and Wait protocol are its simplicity and its small buffer requirements. The sender needs to keep only the copy of the packet that it last transmitted, and the receiver does not need to buffer packets at the data link layer. The main disadvantage of the Stop and Wait protocol is that it does not use the communication link very efficiently. The total time taken to transmit a packet and to prepare for transmitting the next one is T  Tp 2Tprop 2Tproc Ta

where: T  total time for transmission time Tp  transmission time for a packet Tprop  propagation time of a packet or an ACK Tproc  processing time for a packet or an ACK Ta  transmission time for an ACK

(6.28)

186

6

Multiple Access Techniques

The protocol efficiency without any error is: Tp

(0)  

(6.29)

T

If p is the probability that a packet or its ACK is corrupted by transmission errors, and a successful transmission of a packet and its ACK takes T seconds and occurs with probability 1  p, the protocol efficiency for full duplex (FD) is given as: (1  p)Tp

FD  

(1  p)T pTp

(6.30)

• Selective Repeat Protocol (SRP): In case of the SRP, only the selected pack-

ets are retransmitted. The data link layer in the receiver delivers exactly one copy of every packet in the correct order. The data link layer in the receiver may get the packets in the wrong order from the physical layer. This occurs, for example, when transmission errors corrupt the first packet and not the second one. The second packet arrives correctly at the receiver before the first. The data link layer in the receiver uses a buffer to store the packets that arrive out of order. Once the data link layer in the receiver has a consecutive group of packets in its buffer, it can deliver them to the network layer. The sender also uses a buffer to store copies of the unacknowledged packets. The number of the packets which can be held in the sender/receiver buffer is a design parameter. Let W  the number of packets which the sender and receiver buffers can each hold and SRP  number of packets in modulo 2W. The protocol efficiency without any error and with a packet error probability of p is given as:



WTp

(0)  min , 1 T



(6.31)

For very large W, the protocol efficiency is (p)  1  p

(6.32)

where: WTp  time-out 2 p(W  1) 2 p(3W  1)

(p)  

(6.33)

6.14

Error Control Schemes for Link Layer

187

SRP is very efficient, but it requires buffering packets at both the sender and the receiver. • Go-Back-N (GBN): The Go-Back-N protocol allows the sender to have mul-

tiple unacknowledged packets without the receiver having to store packets. This is done by not allowing the receiver to accept packets that are out of order. When a time-out timer expires for a packet, the transmitter resends that packet and all subsequent packets. The Go-Back-N protocol improves on the efficiency of the Stop and Wait protocol, but is less efficient than SRP. The protocol efficiency for full duplex is given as: 1 FD  

(6.34)

p

1  p

1 W

• Window-control Operation Based on Reception Memory (WORM) ARQ:

In digital cellular systems, bursty errors occur by multipath fading, shadowing, and handoffs. The typical bit-error rate fluctuates from 101 to 106. Therefore, the conventional ARQ schemes do not operate well in a digital cellular system. WORM ARQ has been suggested for control of dynamic error characteristics. It is a hybrid scheme that combines SRP GBN protocol. GBN protocol is chosen in the severe error condition whereas SRP is selected in the normal error condition. • Variable Window and Frame Size GBN and SRP [24]: Since wireless systems have bursty error characteristics, the error control schemes should have a dynamic adaptation to a bursty channel environment. The SRP and GBN with variable window and frame size have been proposed to improve error control in wireless systems. Table 6.2 provides the window and frame size for different BER. If the error rate increases, the window and frame size are decreased. In the case of the error rate being small, the window and frame size are increased. The optimum threshold values of BER, window and frame size were obtained through computer simulation.

Table 6.2 Bit-error rate versus window and size. Bit-error rate (BER) BER  104 104

 BER 

103

103

 BER 

102

102

 BER

Window size (W)

Frame size (bits)

32

172

8

80

4

40

2

16

188

6

Multiple Access Techniques

Example 6.10 We consider a WLAN in which the maximum propagation delay is 4 sec. The WLAN operates at a data rate of 10 Mbps. The data and ACK packets are of 400 and 20 bits, respectively. The processing time for a data or ACK packet is 1 sec. If the probability p that a data packet or its ACK can be corrupted during transmission is 0.01, find the data link protocol efficiency with (1) Stop and Wait protocol — full duplex, (2) SRP with window size W  8, and (3) Go-Back-N protocol with window size W  8. Solution 400 Tp    40 s 10

20 Ta    2 s 10

Tprop  4 s Tproc  1 s T  40 2  4 2  1 2  52 s

Stop and Wait: (1  0.01)  40 (1  0.01)  52 0.01  40

    0.763

SRP: 2 0.01(8  1) 2 0.01(24  1)

    0.954

GBN: 1     0.925

 1  0.01 

0.01 1 8 

6.15

Summary

The chapter described the access technologies used in wireless communications including reservation-based multiple access and random multiple access. FDMA,

Problems

189

TDMA, and CDMA technologies were discussed and their advantages and disadvantages were listed. Illustrated examples were given to show calculations for determining the capacity of TDMA and CDMA systems. Brief descriptions of the FDD, TDD, TDM/TDMA, and TDM/TDMA/FDD approaches were also given. Since packet networks are an important part of wireless networks, we briefly stated the characteristics of the access methods in common use and defined their throughput equations. The common packet protocols such as ALOHA, slotted ALOHA, and Carrier Sense Multiple Access (CSMA) were discussed. We also presented the methods used to control errors for data link protocols.

Problems 6.1 In a proposed TDMA cellular system, the one-way bandwidth of the system

6.2

6.3

6.4

6.5

6.6

is 40 MHz. The channel spacing is 30 kHz and total voice channels in the system are 1333. The frame duration is 40 ms divided equally between six time slots. The system has an individual user data rate of 16.2 kbps in which the speech with error protection has a rate of 13 kbps. Calculate the efficiency of the TDMA system.What is the efficiency of the system with 20, 60, 80 and 100 MHz? Recompute the capacity of the GSM system in Example 5.1 when a sectorized system is used. With sectorization, there are 12 channel sets of 39 channels each with three sets assigned at each cell, one for each sector. In the IS-54 (TDMA/FDD), the frame duration is 40 ms. The frame contains six time slots. The transmit bit rate is 48.6 kbps. Each time slot carries 260 bits of user information. The total number of 30 kHz voice channels available is 395 and the total system bandwidth is 12.5 MHz. Calculate the access efficiency of the system. Calculate the capacity and spectral efficiency () of the IS-54 system using the following parameters: b  0.96,  2 (i.e., /4-DQPSK), voice activity factor vf  1.0, information bit rate  19.5 kbps, frequency reuse factor  7 and system bandwidth  12.5 MHz. Calculate the cell capacity and spectral efficiency of a GSM system using the following data: (1) bandwidth efficiency factor  1, (2) bit efficiency (with GMSK modulation)  1, (3) voice activity factor  1, (4) one-way bandwidth of the system  10 MHz, (4) information bit rate per frame  270.83 kbps, (5) number of users per frame  8, and (6) frequency reuse factor  4. Consider a CDMA system that uses QPSK modulation and convolutional coding. The system has a bandwidth of 1.25 MHz and transmits data at 9.6 kbps. Find the number of users that can be supported by the system and bandwidth efficiency. Assume a three-sector antenna system with an effective gain of 2.6, power control efficiency  90%, and frequency reuse efficiency of 66.67%. A bit-error rate of 103 is required.

190

6

Multiple Access Techniques

6.7 A QPSK/DSSS WLAN is designed to transmit in the 902- to 928-MHz ISM band. The symbol transmission rate is 0.25 Megasymbols per second. An orthogonal code with eight symbols is used. A bit-error rate of 105 is required. How many users can be supported by the WLAN? A threesector antenna with gain  2.6 is used. Assume frequency reuse efficiency of 66.67% and power control efficiency of 90%.What is the bandwidth efficiency? 6.8 A WLAN accommodates 50 stations running the same application. The transmission rate per station is 2 Mbps and the stations use slotted ALOHA protocol. The total traffic produced by the stations is assumed to form a Poisson process. What is the maximum throughput in Erlangs? What is the maximum throughput in Mbps? What is the maximum throughput in Mbps for each station? 6.9 Consider a WLAN installation in which maximum propagation delay is 0.5 s. The WLAN operates at a data rate of 12 Mbps, and each packet is 600 bits. Calculate the throughput with (1) an unslotted nonpersistent, (2) a slotted persistent, and (3) a slotted 1-persistent CSMA protocol. 6.10 Consider a WLAN in which the maximum propagation delay is 5 s. The WLAN operates at a data rate of 12 Mbps. The data and ACK packet are 600 and 24 bits, respectively. The processing time for data or ACK packet is 2 s. If the probability p that a data packet or its ACK can be corrupted during transmission is 1%, find the data link protocol efficiency with (1) Stop-and-Wait protocol, full duplex, (2) SRP with window size W  12, and (3) Go-Back-N protocol with window size W  12.

References 1. Bates, R. J. Wireless Network Communication. New York: McGraw-Hill, Inc., 1994. 2. Bell Communications Research. Generic Framework Criteria for version 1.0 Wireless Access Communication System (WACS), FA-NWT-001318. Piscataway, NJ, June 1992. 3. Calhoun, G. Digital Cellular Radio. Boston: Artech House, 1988. 4. Chen, Q., Sousa, E. S., and Pasupathy, S. Performance of a coded multi-carrier DS-CDMA in multipath fading channels. Kluwer Journal of Wireless Personal Communications, vol. 2, 1995, pp. 167–183. 5. Fazel, K. Performance of CDMA/OFDM for mobile communication system. IEEE Proceedings of ICUPC, vol. 2, October 1993, pp. 975–979. 6. Garg, V. K., and Wilkes, J. E. Wireless and Personal Communications. Upper Saddle River, NJ, Prentice Hall, 1996. 7. Gilhousen, S., et al. Increased Capacity Using CDMA for Mobile Satellite Communication. IEEE Journal on Selected Areas in Communications, vol. 8, no. 4, May 1990.

References

191

8. Hammond, J. L., and O’Reilly, J. P. Performance Analysis of Local Computer Networks. Reading, MA: Addison-Wesley Publishing Company, 1986. 9. Habab, I. M., Kavehrad, M., and Sundberg, C.-E. W. ALOHA with Capture Over Slow and Fast Fading Radio Channels with Coding and Diversity. IEEE Journal of Selected Areas of Communications, vol. 6, 1988, pp. 79–88. 10. Jacobs, I. M., et al. Comparison of CDMA and FDMA for the MobileStar System. Proceedings of Mobile Satellite Conference, Pasadena, CA, May 3–5, 1988, pp. 283–290. 11. Jellamy, J. C., Digital Telephony. Second Edition. New York: John Wiley & Sons, Inc., 1990. 12. Kondo, S., and Milstein, L. Performance of multi-carrier DS-CDMA systems. IEEE Transactions on Communications, vol. 44, no. 2, February 1996, pp. 238–246. 13. Lee, W. C. Y. Spectrum efficiency in cellular. IEEE Transactions Vehicular Technology, vol. 38, May 1989, pp. 69–75. 14. Lee, W. C. Y. Overview of cellular CDMA. IEEE Transactions Vehicular Technology, vol. 40, no. 2, May 1991, pp. 291–302. 15. Lee, W. C. Y. Mobile Communications Design Fundamentals. Second Edition. New York: John Wiley & Sons, Inc., 1993. 16. Mehrotra, A. Cellular Radio — Analog & Digital Systems. Boston: Artech House, 1994 ISBN 0-89006-731-7. 17. Parsons, D., and Gardiner, J. G. Mobile Communication Systems. New York: Halsted Press, 1988. 18. Pahlavan, K., and Levesque, A. H. Wireless Information Networks. New York: John Wiley and Sons, Inc., 1995. 19. Skalar, B. Digital Communications — Fundamentals and Applications. Englewood Cliffs: Prentice Hall, 1988. 20. Torrieri, J. Principles of Secure Communication Systems. Second Edition. Boston: Artech House, 1992. 21. Viterbi, A. J. When not to spread spectrum-A sequel. IEEE Communications Magazine, vol. 23, April 1985, pp. 12–17. 22. Viterbi, A. J., and Padovani, Roberto. Implications of mobile cellular CDMA. IEEE Communication Magazine, vol. 30, no. 12, 1992, pp. 38–41. 23. Walrand, J. Communications Networks: A First Course. Homewood, IL: Irwin, 1991. 24. Woo, Ill and Cho, Dong-Ho. A Study on the Performance Improvements of Error Control Schemes in Digital Cellular DS/CDMA Systems. IEICE Transactions Communications, vol. E77-B, no. 7, July 1994. 25. Yee, N., Linnartz, J -P., and Feltweis, G. Multi-carrier CDMA in indoor wireless radio networks. IEEE Proceedings of PIMRC, vol. 1, September 1993, pp. 109–113. 26. Ziemer, E., and Peterson, R. L. Introduction to Digital Communications. New York: Macmillan Publishing Company, 1992. 27. Keiser, G. E. Local Area Networks. New York: McGraw-Hill Publishing Company, 1989.

This page intentionally left blank

CHAPTER 7 Architecture of a Wireless Wide-Area Network (WWAN) 7.1

Introduction

A wireless network does not operate in isolation; it uses the services of public switched telephone networks (PSTNs) to make or receive calls from wireline users. A number of functions is required to support the services and facilities in a wireless wide-area network (WWAN). The basic subsystems of the WWAN are: radio station subsystem (RSS), networking and switching subsystem (NSS), and operational and maintenance subsystem (OMSS) (see Figure 7.1). The radio subsystem is responsible for providing and managing transmission paths between the user equipment and the NSS. This includes management of the radio interface between the user equipment and the rest of the WWAN system. The NSS has the responsibility of managing communications and connecting user equipment to the relevant networks or other users. The NSS is not in direct

WWAN

RSS

NSS

OMSS

UE

External Network

Service Provider

User RSS: Radio Station Subsystem NSS: Network and Switching Subsystem OMSS: Operational and Maintenance Subsystem UE: User Equipment

Figure 7.1

Model of a WWAN system.

193

194

7

Architecture of a Wireless Wide-Area Network (WWAN)

contact with the user equipment, nor is the radio subsystem in direct contact with external networks. The user equipment, radio subsystem, and NSS form the operational part of the WWAN system. The OMSS provides the means for a service provider to control them. Figure 7.1 shows the model for the WWAN system. In the WWAN, interaction between the subsystems can be grouped into two main parts [3]: • Operational part: External Networks ⇔ NSS ⇔ RSS ⇔ UE ⇔ User • Control and maintenace part: OMSS ⇔ Service Provider

The operational part provides transmission paths and establishes them. The control and maintenance parts interact with the traffic-handling activity of the operational part by monitoring and modifying it to maintain or improve functions. In this chapter, we present the architecture of a WWAN and discuss subsystem entities and their roles. We also look into the frame and channel structure and point out the role of different logical channels used in the GSM. We then describe how information is processed in the GSM. We conclude the chapter with a brief description of services available in the GSM900. More detailed discussion of WWANs will be given in Chapter 15.

7.2

WWAN Subsystem Entities

Figure 7.2 shows the functional entities of a WWAN and their logical interaction. A brief description of these functional entities is provided below [1,10,11].

7.2.1 User Equipment The user equipment (UE) consists of the physical equipment used by the subscriber to access a WWAN for offered telecommunication services. Functionally, the UE includes a mobile terminal and, depending on the services it can support, various terminal equipment and combinations of terminal equipment and terminal adaptor (TA) functions (TA acts as a gateway between the terminal equipment and mobile terminal) (see Figure 7.3). Various types of mobile stations, such as a vehicle-mounted station, portable station, or handheld station, are used [7]. Basically, a mobile station can be divided into two parts. The first part contains the hardware and software to support radio and man-machine interface functions and is available at retail stores to buy or rent. The second part contains terminal/ user-specific data in the form of a smart card (subscriber identity module (SIM) card), which can effectively be considered a sort of logical terminal. The SIM card plugs into the first part of the mobile station and remains in it for the duration of use. Without the SIM card, the mobile station is not associated with any user and cannot make or receive calls (except possibly an emergency call if the network allows). The SIM card is issued by the mobile service provider after

7.2

WWAN Subsystem Entities

195

G OMC

VLR

D

B Abis UE

A

BTS

C MSC

BSC

HLR F

Um

RSS

E Other MSC

Other Network

EIR

A, B, C, . . . ,G Interfaces Abis: Interface between BTS and BSC AuC: Authentication Center BSC: Base Station Controller BTS: Base Transceiver Station EIR: Equipment Identity Register HLR: Home Location Register

MSC: Mobile Switching Center OMC: Operations and Maintenance Center RSS: Radio Station Subsystem UE: User Equipment (MS) VLR: Visitor Location Register

Signaling links Data + Signaling links

Figure 7.2

WWAN reference model.

MT0

MT1

TE1

TE2

TA

MT1 S

TE3

MT2 R Um Mobile Station (MS)

MT: Mobile Terminal TE: Terminal Equipment TA: Terminal Adaptor Um: Air interface R, S: Interfaces

Figure 7.3

Other MSC’s VLR

Functional model of a mobile station.

AuC

196

7

Architecture of a Wireless Wide-Area Network (WWAN)

subscription. This type of SIM-card mobility is analogous to terminal mobility, but it also provides a personal-mobility-like service within the WWAN. A mobile station has a number of identities, including the international mobile subscriber identity (IMSI), the temporary mobile subscriber identity (TMSI), the international mobile equipment identity (IMEI), and integrated services of digital network (ISDN) number. The IMSI is embodied in the SIM. SIM contains all subscriber-related information stored on the user’s side of the radio interface. • International Mobile Subscriber Identity: The IMSI is assigned to a mobile

station at subscription time. It uniquely identifies a given mobile station. IMSI is transmitted over the radio interface only if necessary. IMSI contains 15 digits and has: 1. 2. 3. 4.

Mobile Country Code (MCC) — 3 digits (home country) Mobile Network Code (MNC) — 2 digits (home network) Mobile Subscriber Identification Number (MSIN) National Mobile Subscriber Identity Number (NMSI)

• Temporary Mobile Subscriber Identity: The TMSI is assigned to a mobile

station by the visitor location register (VLR). The TMSI uniquely identifies a mobile station within the area controlled by a given visitor location register. A maximum of 32 bits can be used for TMSI. • International Mobile Equipment Identity: The IMEI uniquely identifies the mobile station equipment. It is assigned by the equipment manufacturer. The IMEI contains 15 digits and carries: 1. 2. 3. 4.

Type of Approval Code — 6 digits Final Assembly Code — 2 digits Serial Number — 6 digits Spare — 1 digit

The Subscriber Identity Module (SIM) contains: IMSI, authentication key (Ki), subscriber information, access control class, cipher key (Kc), TMSI, information about additional services, location area identity (LAI) and forbidden networks list.

7.2.2 Radio Station Subsystem The radio station subsystem (RSS) is the physical equipment that provides radio coverage to prescribed geographical areas, known as cells. It contains equipment required to communicate with the user equipment. Functionally, an RSS consists of a control function performed by the base station controller (BSC) and a transmitting/receiving function carried out by the base station transceiver (BTS) system. The BTS is the radio transmission/receiving equipment and covers a cell. An RSS can serve several cells and can have multiple base station transceivers.

7.2

WWAN Subsystem Entities

197

The base station transceiver contains the transcoder rate adapter unit (TRAU). In the GSM TRAU, the speech encoding and decoding is carried out, as well as the rate adaptation function for data. In certain situations TRAU is located between the base station controller (BSC) and the mobile switching center (MSC) to gain an advantage of a more-compressed transmission between the BTS and the TRAU. Interface between the BTS and BSC is Abis. The interface between the user equipment and radio station subsystem is air interface (Um).

7.2.3 Network and Switching Subsystem The NSS includes the main switching functions of the WWAN, databases required for the subscribers, and mobility management. Its main role is to manage the communications between the WWAN and other network users [8]. Within the NSS, the switching functions are performed by the mobile switching center (MSC). Subscriber information relevant to provisioning of service is kept in the home location register (HLR). The other database in the NSS is the visitor location register (VLR), which maintains data required for mobility management. The MSC performs the necessary switching functions required for the user equipment located in an associated geographical area, called an MSC area. The MSC monitors the mobility of its subscribers and manages necessary resources required to handle and update the location registration procedures and to carry out the handoff functions. The MSC is involved in the interworking functions to communicate with other networks such as PSTN and ISDN. The interworking functions of the MSC depend upon the type of the network to which it is connected and the type of service to be performed. The call routing, call control, and echo control functions are also performed by the MSC. The HLR is the functional unit used for management of mobile subscribers. The number of home location registers in a network varies with the characteristics of the network. Two types of information are stored in the HLR: subscriber information and part of the mobile information to allow incoming calls to be routed to the MSC for the particular mobile. Any administrative action (such as changes in service profile, etc.) by a service provider on subscriber data is carried out in the HLR. The HLR stores IMSI, MS ISDN number, VLR address, and subscriber data (e.g., supplementary services, etc.). The VLR is linked to one or more MSCs. The VLR is a functional unit that stores subscriber information, such as location area, when the subscriber is located in the area covered by the VLR. When a roaming user enters an MSC area, the MSC informs the associated VLR about the UE; the UE goes through a registration procedure that includes the following steps: • The VLR recognizes that the UE is from another network. • If roaming is allowed, the VLR finds the UE’s HLR in its home network.

198

7

Architecture of a Wireless Wide-Area Network (WWAN)

• The VLR constructs a global title from the IMSI to allow signaling from the

VLR to the UE’s HLR via the PSTN/ISDN networks. • The VLR generates a mobile subscriber roaming number (MSRN) that is used to route incoming calls to the UE. • The MSRN is sent to the UE’s HLR. The information included in the VLR is: 1. 2. 3. 4. 5. 6. 7. 8.

MSRN TMSI LAI where the UE has been registered data related to supplementary services MS ISDN number IMSI HLR address or global title (GT) local UE identity, if used

The NSS contains more than MSCs, HLRs, and VLRs. In order to set up a call, the call is first routed to a gateway switch, referred to as the gateway MSC (GMSC). The GMSC is responsible for collecting the location information and routing the call to the MSC through which the subscriber can obtain service at that instant (i.e., the visited MSC). The GMSC first finds the right HLR from the directory number of the subscriber and interrogates it to obtain user information. The GMSC has an interface with external networks for which it provides a gateway function. It also has an interface with the signaling system 7 (SS7) network for interworking with other NSS entities.

7.2.4 Operation and Maintenance Subsystem (OMSS) The OMSS is responsible for handling system security based on the validation of identities of various telecommunications entities. These functions are performed in the authentication center (AuC) and equipment identity register (EIR). The AuC is accessed by the HLR to determine whether a UE will be granted service. The EIR provides UE information used by the MSC. The EIR maintains a list of legitimate, fraudulent, or faulty UEs. The OMSS is also in charge of remote operation and maintenance of the network. Functions are monitored and controlled in the OMSS. The OMSS may have one or more network management centers (NMCs) to centralize network control. The operations and maintenance center (OMC) is the functional entity through which a service provider monitors and controls the system. The OMC provides a single point for maintenance personnel to maintain the entire system. One OMC can serve multiple MSCs.

7.3

Logical Channels

199

7.2.5 Interworking and Interfaces Necessary interfaces are required to achieve an optimum interworking between different entities of the network. The use of the SS7 between the MSC and VLR and between the MSC and HLR allows transmission of both call control signals and other information. The corresponding signaling capabilities are supported by the mobile application part (MAP) in SS7 defined in the particular network standards. The interface labels on the reference model (Figure 7.2 — A, Abis, B, C, D, E, F, G, and Um) corresponds to interfaces between network nodes. Each interface is specified in network standards along with their corresponding procedures. For example the GSM recommendations in the 09 series cover interworking procedures between a public land mobile network (PLMN) and other networks [9].

7.3

Logical Channels

A WWAN uses a variety of channels in which the information is carried. In GSM, these channels are separated into physical channels and logical channels [10,11]. The physical channels are determined by the time slots, whereas the logical channels are determined by the information carried within the physical channels. It can be further summarized by saying that several recurring time slots on a carrier constitute a physical channel. These are then used by different logical channels to transfer information. These channels may either be used for user data (payload) or signaling to enable the system to operate. In GSM logical channels (see Figure 7.4) are used to carry traffic and control information. There are three types of logical channels: traffic channels (TCHs), control channels (CCHs), and the cell broadcast channel (CBCH). The traffic channels are used to transmit user information (speech or data). The control channels are used to transmit control and signaling information. The cell broadcast channel is used to broadcast user information from a service center to the mobile stations listening to a given cell area. It is a unidirectional (downlink-only, base station to mobile station), point-to-multipoint channel used for a short-information message service. Some special constraints are imposed on the design of the CBCH because of the requirement that this channel can be listened in parallel with the broadcast control channel (BCCH) information and the paging messages. The control channels consist of broadcast channel (BCH), common control channel (CCCH), and dedicated control channel (DCCH). The broadcast channels are used for such functions as correcting mobile frequencies, frame synchronization, control channel structure, and so on. They are point-to-multipoint, downlink-only channels. These channels consist of BCCH, frequency correction channel (FCCH), and synchronization channel (SCH). The BCCH is used to send cell identities, organization information about common control channels, cell service available, and so on. The frequency correction channel is used to transmit frequency correction data bursts that contain the set of

200

7

Architecture of a Wireless Wide-Area Network (WWAN)

Logical Channels

CCH

TCH

TCH/F

CBCH

TCH/H

BCH

FCCH

CCCH

SCH

DCCH

BCCH

PCH

AGCH

RACH

ACCH

SACCH

SACCH/TF

SACCH/TH

SACCH/C4

SDCCH

FACCH

SACCH/C8

FACCH/F

SDCCH/4

FACCH/H

ACCH: Associated Control Channel AGCH: Access Grant Channel BCH: Broadcast Channel BCCH: Broadcast Control Channel CBCH: Cell Broadcast Channel CCH: Control Channel CCCH: Common Control Channel DCCH: Dedicated Control Channel FACCH: Fast Associated Control Channel associated with full rate (F) or half rate (H) traffic channel FCCH: Frequency Correction Channel PCH: Paging Channel RACH: Random Access Channel SACCH: Slow Associated Control Channel associated with full rate (TF) or half rate (TH) traffic channel SCH: Synchronization Channel SDCCH: Stand-alone Dedicated Control Channel TCH: Traffic Channel TCH/F: Traffic Channel/Full TCH/H: Traffic Channel/Half

Figure 7.4

GSM logical channel.

SDCCH/8

7.4

Channel and Frame Structure

201

all “0.” This gives a constant frequency shift of the RF carrier that can be used by the mobile station for frequency correction. The synchronization channel (SCH) is used for time synchronization of the mobile stations. The data on this channel include frame number as well as the base station identity code (BSIC) required by mobile stations when measuring base station signal strength. The CCCHs include paging channel (PCH), access grant channel (AGCH) and random access channel (RACH). The CCCHs are point-to-multipoint downlink-only channels that are used for paging and access. The PCHs are used to page mobile stations. The mobile stations need to listen for paging during certain times. The AGCHs are downlink-only channels used to assign mobiles to stand-alone dedicated control channels (SDCCHs) for initial assignment. The RACHs are uplink-only channels used by mobile stations for transmitting their requests for dedicated connections to the network. There are two types of DCCHs: SDCCH and associated control channel (ACCH). The SDCCHs are bidirectional, point-to-point channels that are used for service request, subscriber authentication, ciphering initiation, equipment validation, and assignment to a traffic channel (TCH). The ACCHs are bidirectional, point-to-point channels that are associated with a given TCH and SDCCH. These channels are used to send out-of-band signaling and control data between the mobile station and the base station. Examples of their use are to send signal strength measurements from a mobile station to the base station or to send transmission timing information from the base station to the mobile station. The associated control channels are further divided as slow associated control channels (SACCHs) and fast associated control channels (FACCHs). Figure 7.4 shows all GSM logical channels. Table 7.1 lists briefly the role of each logical channel. For more details, refer to the GSM 04.03, 05.01, and 05.2 series recommendations [4,5].

7.4

Channel and Frame Structure

We discuss channel and frame structure with reference to the GSM900 system that uses FDMA/TDMA (see Chapter 6) and a bandwidth of 25 MHz. The frequency band used for uplink (i.e., from the mobile station to the base station) is 890 to 915 MHz, whereas for the downlink (i.e., from the base station to the mobile station) is 935 to 960 MHz. The GSM900 has 124 RF channels, each with a bandwidth of 200 kHz. For a given RF channel the uplink (ULf) and downlink (DLf) frequency can be obtained from Equations 7.1 and 7.2, respectively: ULf  890.2  0.20(N  1) MHz

(7.1)

DLf  935.2  0.20(N  1) MHz

(7.2)

202

7

Architecture of a Wireless Wide-Area Network (WWAN)

Table 7.1 Role of logical channels in GSM. Role of the logical channel

Logical channel TCH/F

Full rate traffic channel to carry payload at 22.8 kbps

TCH/H

Half rate traffic channel to carry payload at 11.4 kbps

BCCH

Broadcast network information

SCH

Synchronization of the mobile stations

FCCH

Used for frequency correction

AGCH

Acknowledges channel request from mobile and allocates an SDCCH

FACCH

For time critical signaling over TCH (e.g., for handoff signaling); traffic burst is stolen for a full signaling burst

SACCH

TCH in-band signaling, e.g., for link monitoring

SDCCH

For signaling exchange, e.g., during call setup, registration/location updates

FACCHs

FACCH for SDCCH. The SDCCH burst is stolen for a full signaling burst.

SACCHs

SDCCH in-band signaling, e.g., for link monitoring

where: N  1, 2,. . ., 124 When the mobile station is assigned a channel, an RF channel and a time slot are also assigned. RF channels are assigned in frequency pairs — one for the uplink and the other for the downlink. Each pair of RF channels supports up to eight simultaneous voice calls (see Figure 7.5). Thus, the GSM can support up to 992 simultaneous users with a full-rate speech coder. This number is doubled to 1984 users with the half-rate speech coder. In GSM, the RF carriers are divided in time using a TDMA scheme. The fundamental unit of time is called a burst (time slot) period and it lasts for 0.577 ms. Eight of these burst periods are grouped into a TDMA frame. The frame lasts for 4.615 ms and it forms the basic unit for the definition of logical channels. One physical channel is one burst period allocated in each TDMA frame. There are different types of frames that are transmitted to carry different data. The frames are organized into multiframes, superframes, and hyperframes to provide overall synchronization. The GSM multiframe is 120 ms. It contains 26 frames of 8 bursts. The structure of a GSM hyperframe, superframe, multiframe, frame, and burst is given in Figure 7.6. A burst carries 156.25 bits. The same format is used for the uplink and downlink transmission with various burst types as shown in Figure 7.7. In a normal burst, two user information groups of 58 bits account for most of the transmission time in a time slot (57 bits carry user data, while the H bit is used to distinguish speech from other transmissions). Twenty-six training (T) bits are

7.6

Speech Processing

203

Time Domain TS0

TS1

TS2

TS4

TS5

TS6

TS7

TS8

f1

Ch # 1

Ch # 2

Ch # 3

Ch # 4

Ch # 5

Ch # 6

Ch # 7

Ch # 8

f2

Ch # 1

Ch # 2

Ch # 5

Ch # 6

Ch # 7

Ch # 8

TS: Time-slot

Frequency Domain

f124

Ch # 1

Figure 7.5

Ch # 2

Ch # 5

Ch # 6

Ch # 7

Ch # 8

GSM FDMA / TDMA structure.

used in the middle of the time slot. The time slot starts and ends with three tail bits. The time slot also contains 8.25 guard (G) bits.

7.5

Basic Signal Characteristics

The GSM system operates using frequency division duplex (FDD) and as a result, paired frequency bands are required for the up- and downlink transmissions. The frequency separation is dependent upon the band in use (for GSM900, it is 45 MHz). The RF carrier is modulated using Gaussian minimum shift keying (GMSK) (see Chapter 9 for details). The GMSK occupies a relatively narrow bandwidth, and it has a constant power level. The data transported by the RF carrier serves up to eight different users under the basic system. Even though the full data rate on the carrier is about 270.83 kbps, some of this supports the management overhead, and therefore the data rate allocated to each time slot is only 22.8 kbps. In addition to error correction, the problem of interference, fading and the like must be overcome. This means that the available data rate for transporting the digitally encoded speech is 13 kbps for the basic vocoder.

7.6

Speech Processing

Two major steps are taken in transmitting and receiving information over a digital radio link: information processing and modulation processing. Information processing deals with the preparation of the basic information signals so that they are

204

7

Architecture of a Wireless Wide-Area Network (WWAN)

Tail

Data

H

T

H

Data

Tail

G

3

57

1

26

1

57

3

8.25

156.25 bits

GSM Time Slot (Normal Burst) (0.576 ms)

1250 bits

1

0

2

3

4

5

6

7

GSM Frame (4.62 ms)

0

1

2

23

24

25

GSM Multiframe (120 ms)

0

1

2

48

49

50

GSM Superframe (6.12 second)

0

1

2

2045

2046

2047

GSM Hyperframe (3.48 hours) G : Guard bits T : Training bits H : Bit used to distinguish speech from other transmission

Figure 7.6 Physical structure for GSM hyperframe, superframe, multiframe, frame, and time slot.

protected and converted into a form that the radio link can handle. Information processing involves transcoding, channel coding, bit-interleaving, encrypting, and multiplexing. Modulation processing involves the physical preparation of the signal to carry information on an RF carrier [6]. Each digital radio link process in the transmitting path has its peer in the receiving path (see Figure 7.8). The delay equalization process in the receiving path is required to compensate for the spread in time delays resulting from the multipath propagation (see Chapter 3). It can be part of the demodulation process. However, it should be emphasized that it is required when delay spreads are significant compared to the information bit duration. In the GSM the transmitted bit duration is about 37 sec, and a delay spread of about 6 to 8 sec is quite common. The delay spread problem becomes critical and is difficult to solve with the higher transmission bit rate.

7.6

Speech Processing

156.25 bits

205

Tail

Data

H

T

H

Data

Tail

G

3

57

1

26

1

57

3

8.25

Tail

G

3

8.25

Tail

G

3

8.25

Tail

G

3

8.25

GSM Normal Burst (0.576 ms)

Tail 156.25 bits

3

Fixed Bits (142) GSM Frequency Correction Burst

Tail 3

Data (39)

Training (64)

Data (39)

GSM Synchronization Burst

Tail 3

Mixed Bits (58)

Training (26)

Mixed Bits (58)

GSM Dummy Burst

Tail 3

Tail Synch. Seq. (41)

Data (36)

3

Guard (G) (68.25)

GSM Access Burst

Figure 7.7

GSM different time slot structures.

In the GSM, the analog speech from the mobile station is passed through a low-pass filter to remove the high-frequency contents from the speech. The speech is sampled at the rate of 8000 samples per second, uniformly quantized (see Chapter 4) to 213 (8192) levels and coded using 13 bits per sample. This results in a digital information stream at a rate of 104 kbps. At the base station, the speech signal is digital (64 kbps), which is first transcoded from the A-law (see Chapter 4) 8-bit samples into 13-bit samples corresponding to a linear representation of amplitudes. This results in a digital information stream at a rate of 104 kbps.

206

7

Transmitting Path

Base Station

Architecture of a Wireless Wide-Area Network (WWAN)

Speech Coding

Speech Decoding

Channel Coding

Channel Decoding

Bit Interleaving

Bit Deinterleaving

Encryption

Decryption

Burst Forming & Multiplexing

Demultiplexing

Receiving Path

Mobile Station

Delay Equalization

Modulation

Figure 7.8

Demodulation

Digital radio link process.

The 104-kbps digital signal stream is fed into the speech encoder (see Figure 7.9) which then transcodes the speech into a 13-kbps stream. The full-rate speech encoder takes a 2080 bit block from the 13-bit transcoder every 20 ms (i.e., 160 samples) and produces 36 “filter parameters” bits over the 20 ms period, 9 long-term prediction (LTP) bits every 5 ms, and 47 regular pulse excited-linear prediction coding (RPE-LPC) bits every 5 ms (refer to Table 7.2). Thus, 260 bits are generated every 20 ms. Of these 260 bits, 182 are classified as class I bits related to the excitation signal, whereas the remaining 78 are class II bits related to the parameters of linear prediction coding (LPC) and LTP filters. The class I bits

7.6

Speech Processing

Base Station Digital Speech 8-bit A-law Signal to13-bit Uniform Convertor Mobile Station Analog Speech Signal

Figure 7.9

207

13 kbps 13 bits  8,000  104 kbps

13  8,000  104 kbps Low-Pass Filter

A/D Convertor

Speech Encoder

To Channel Encoder

13 kbps Speech Encoder

To Channel Encoder

Speech processing in GSM.

Table 7.2 Details of class I and class II bits in GSM vocoder. Bits per 5 ms Linear Prediction Coding (LPC) filter

Bits per 20 ms 36

Long Term Prediction (LTP) filter

9

36

Excitation signal

47

188

Total

260

Class I

182 (class Ia  50, class Ib  132)

Class II

78

are further divided into class Ia (50 bits) and class Ib (132 bits). The class Ia bits are the most significant bits which are used to generate 3 cyclic redundancy check (CRC) bits. The 3 CRC bits along with 4 tail bits are added to the 182 class I bits before they are passed through the half-rate convolutional codec to produce twice as many bits output as there are bits input (i.e., 2  189  378). The 78 class II bits remain uncoded and are bypassed (see Figure 7.10). The total 456 bits (i.e., 378  78) are fed to the bit interleaver. Since 456 bits are generated during 20 ms, the user data rate is 456/0.02  22.8 kbps. This includes 13 kbps raw data and 9.8 kbps of parity, tail, and channel coding. Another vocoder (the enhanced full rate (EFR)) (see Chapter 8) has been added in response to the poor voice quality perceived by the users of the RPE-LPC vocoder. This new vocoder gives much better voice quality. It uses the algebraic code excitation linear prediction (ACELP) compression technology (see Chapter 8) to provide a significant improvement in voice quality compared to the original RPE-LPC. There is also a half-rate vocoder (11.4 kbps). Although this vocoder

208

7

Architecture of a Wireless Wide-Area Network (WWAN)

4 Tail bits 50 Class Ia bits

3-bit CRC

3-bits

132 Class Ib bits 160 Speech Samples (20 ms)

Figure 7.10

Coded Half-rate Class I convolution bits coder 378

Speech coder 78 Class II bits

Bit Interleaver (8 bursts)

456 bits (data rate 22.8 kbps)

Channel coding for full-rate speech in GSM.

gives more inferior voice quality compared to full-rate vocoders, it does allow for an increase in network capacity. It is used in some instances when network loading is very high to handle all the calls. Of the 456 bits, 57 at a time are interleaved with 57 other bits from an adjacent data block to form a data burst of 114 bits. At this stage, 42.25 overhead bits are added to the data burst to carry it in a time slot (see Figure 7.11). Bit interleaving is used to reduce the adverse effects of fading by preventing entire blocks of bits from being destroyed by a signal fade. Interleaved data is passed through the GMSK modulator where it is filtered by a Gaussian filter before applying it to the modulator. The modulated data passes through a duplexer switch where filtering is provided between the transmitted and the received signal. On the receiving side the signal is demodulated and deinterleaved before the error correction is applied to the recovered bits.

7.7

Power Levels in Mobile Station

A variety of power levels are allocated by the GSM standard, the lowest being only 800 mW (29 dBm). As mobiles may only transmit for one-eighth of the time, i.e., for their allocated slot, which is one of eight, the average power level is an eighth of the maximum power. To reduce the levels of transmitted power additionally and hence the levels of interference, mobiles are able to step the power down in increments of 2 dB from the maximum to a minimum of 20 mW (13 dBm). The mobile station measures the signal strength or signal quality (based on bit-error-rate), and passes the information to BTS and hence to BSC, which ultimately decides if and when the power level should be changed. A further power-saving and interference-reducing facility is the discontinuous transmission (DTx) capability (an optional feature in GSM) that is incorporated

7.8

GSM Public Land Mobile Network Services

160 samples 2,080 bits (20 ms)

209

104 kbps

Speech Coder

Speech Coder

182 Class I 78 Class II

160 samples 2,080 bits (20 ms)

13 kbps

260 bits Channel Encoding

260 bits Channel Encoding

22.8 kbps

456 bits

456 bits 1 2 3 4 5 6 7 8

57 bits Stream of bursts

1

2

ST (3)

3

4

Data (57)

5

6

H T (1) (26)

7

57 bits

1 2 3 4 5 6 7 8

8 (114)

(114)

bits

bits

Data (57)

H ST G (1) (3) (8.25)

Normal Burst

Figure 7.11

Speech coding in GSM (full rate).

within specifications. It is particularly useful because there are long pauses in speech, for example, when the person using the mobile phone is listening, and during these periods there is no need to transmit a signal. The most important element of DTx is the voice activity detector. It must distinguish between voice and noise inputs, a task that is not trivial.

7.8

GSM Public Land Mobile Network Services

GSM offers users good voice quality, call privacy, and network security. The SIM card provides the security mechanism for GSM. SIM cards are like credit cards and identify the user to the GSM network. They can be used with any GSM handset, providing phone access, ensuring delivery of appropriate services to that user, and automatically billing the subscribers’s network usage to the home network. Roaming agreements have been established between most GSM service providers in different countries, allowing subscribers to roam between networks and have access to the same services no matter where they travel. Of major importance is GSM’s potential for delivering enhanced services requiring multimedia communication: voice, image, and data. Several mobile service providers offer free voice mailboxes and phone answering services to subscribers.

210

7

Architecture of a Wireless Wide-Area Network (WWAN)

The basic telecommunication services provided by the GSM public land mobile network (PLMN) are divided into three main groups: bearer services, teleservices, and supplementary services [2]. Bearer services give the subscriber the capacity required to transmit appropriate signals between certain access points (i.e., usernetwork interfaces). Table 7.3 provides a summary of these services and compares them with service available with integrated services of digital network (ISDN). Teleservices provide the subscriber with necessary capabilities, including terminal equipment functions, to communicate with other subscribers. The GSM teleservices are: • Speech transmission — telephony, emergency call • Short message services — mobile terminating point-to-point, mobile• • • •

originating point-to-point, cell broadcast Message handling and storage services Videotex access Teletext transmission Facsimile transmission

A summary of the teleservices is given in Table 7.4. A comparison is made between the GSM and ISDN teleservices. Supplementary services modify or supplement basic telecommunications services and are offered with or in association with basic telecommunications services. Table 7.5 summarizes the supplementary services and compares them to the supplementary services available with ISDN. The GSM system was designed as a 2G cellular communications system. One of the basic aims was to provide a system that would enable greater capacity to be achieved than 1G analog systems. GSM achieved this by using a digital TDMA approach. By adopting TDMA more users could be accommodated within the Table 7.3 A comparison of bearer services supported by GSM and ISDN. Service Data Service

GSM

ISDN

yes

yes

Alternate speech/data

yes

yes

Speech followed by data

yes

yes

Clear 3.1-kHz audio

yes

yes

Unrestricted digital information (UDI)

yes

yes

Packet Assembler/Disassembler (PAD)

yes

no

3.1-kHz external to PLMN

yes

no

Others

no

yes

7.8

GSM Public Land Mobile Network Services

211

available spectrum. In addition to this, ciphering of digitally encoded speech was used to retain privacy. Table 7.4 A comparison of teleservices supported by GSM and ISDN. GSM

ISDN

Circuit switch (telephony)

yes

yes

Emergency call

yes

yes

Short message point-to-point

yes

yes

Short message cell broadcast

yes

yes

Alternate speech/facsimile group 3

yes

yes

Automatic facsimile group 3 service

yes

yes

Voice-band modem (3.1-kHz audio)

yes

yes

Messaging teleservices

yes

no

Paging teleservices

yes

no

Others

no

yes

Service

Table 7.5 A comparison of supplementary services supported by GSM and ISDN. Service

GSM

ISDN

Call Number ID Presentation

yes

yes

Call Number ID Restriction

yes

yes

Connected Number ID Presentation

yes

yes

Connected Number ID Restriction

yes

yes

Malicious Call Identification

yes

yes

Call Forwarding Unrestricted

yes

yes

Call Forwarding Mobile Busy

yes

yes

Call Forwarding No Reply

yes

yes

Call Forwarding Mobile Not Reachable

yes

yes

Call Transfer

yes

yes

Call Waiting

yes

yes

Call Hold

yes

yes

Completion of Call to Busy Subscriber

yes

yes

3-Party Service

yes

yes

Conference Calling

yes

yes

(continued)

212

7

Architecture of a Wireless Wide-Area Network (WWAN)

Table 7.5 (continued) Service

GSM

ISDN

Closed User Group

yes

yes

Multiparty

yes

yes

Advice of Charge

yes

yes

Reverse Charging

yes

yes

Flexible Altering

no

yes

Mobile Access Hunting

yes

yes

Freephone

yes

no

Barring All Originating Calls

yes

yes

Barring Outgoing Calls

yes

yes

Barring Outgoing International Calls

yes

yes

Barring All International Calls

yes

yes

Barring Outgoing International Calls — except Home

yes

yes

Barring Incoming Calls when Roaming

yes

yes

Barring

yes

SMS

yes

Preferred Language Service

no

yes

Remote Feature Control

no

yes

Do Not Disturb Message Waiting Notification

Selective Call Acceptance

Barring

yes

Encryption

yes

Priority Access & Channel Assignment

no

yes

Password Call Acceptance

no

yes

Voice Privacy

Barring

yes

Subscriber PIN Access

Subscriber PIN Intercept

no

yes

Voice Mail Retrieval

no

yes

Others

no

yes

7.9

Summary

The use of GSM has grown steadily since it was first deployed in 1991. It is now the most widely used mobile phone system in the world. GSM reached the 1 billion subscriber point in 2004, and continues to grow in popularity. In this chapter, we presented an overview of GSM900 — a typical WWAN — and discussed network architecture, logical channels, frame structures, speech processing, power control, and services. We will discuss enhancements to the GSM in Chapter 15, including general packet radio services (GPRS) and universal mobile telecommunication services (UMTS), which are derived from the GSM.

References

213

Problems 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 7.10

What are the three subsystems that are used in the WWAN? Define the role of a VLR in the WWAN architecture. Define the role of an HLR in the WWAN architecture. Why are the IMSI, TMSI, and IMEI used in a WWAN? Define the three services available in the GSM. What is the role of a speech codec in the speech processing of a WWAN? What is a bit-interleaver? Why it is used in a WWAN? Why do you need an equalizer in the GSM? Why are convolution codecs used in wireless networks? Why are so many logical channels used in the GSM?

References 1. GSM Specification Series 1.02–1.06, “GSM Overview, Glossary, Abbreviation, Service Phases.” 2. GSM Specification Series 2.01–2.88, “GSM Services and Features.” 3. GSM Specification Series 3.01–3.88, “GSM PLMN Functions, Architecture, Numbering and Addressing Procedures.” 4. GSM Specification Series 4.01– 4.88, “MS-BSS Interface.” 5. GSM Specification Series 5.01–5.10, “Radio Link.” 6. GSM Specification Series 6.01–6.32, “Speech Processing.” 7. GSM Specification Series 7.01–7.03, “Terminal Adaptation.” 8. GSM Specification Series 8.01–8.60, “BSS-MSC Interface, BSC-BTS Interface.” 9. GSM Specification Series 9.01–9.11, “Network Interworking, MAP.” 10. Mouly, M., and Pautet, M. The GSM System for Mobile Communications. Palaiseau, France, 1992. 11. Garg, V. K., and Wilkes, J. E. Wireless and Personal Communications System. Upper Saddle River, NJ: Prentice Hall, 1996.

This page intentionally left blank

CHAPTER 8 Speech Coding and Channel Coding 8.1

Introduction

In Chapter 7 we introduced speech and channel coding with reference to the GSM. In this chapter we provide detailed descriptions of speech and channel coding schemes. We first discuss speech coding methods and attributes of a speech codec. (The speech codec is also called a voice codec or vocoder. It is a hardware circuit that converts the spoken word into digital code and vice versa.) These are then followed by a brief discussion of linear-prediction-based analysis-by-synthesis (LPAS) method. We also discuss the QCELP, EVRC, EFR, and AMR codecs that are used in WWAN systems. We then focus on channel coding, concentrating on three channel coding schemes — convolutional code, Reed-Solmon (R-S) block code, and turbo code. The convolutional codes have been used in direct sequence spread spectrum CDMA (IS-95) and R-S and turbo code are the proposed channel coding schemes for cdma2000 and WCDMA. To achieve reliable communication with minimum possible signal power is the main objective of a communications system engineer. Speech coding is used to save bandwidth and improve bandwidth efficiency whereas channel coding is employed to improve signal quality and reduce bit-error-rate (BER). The idea of using a channel coding scheme is to recover from errors that occur during transmission over the communication channel. The channel coding strategy is aimed at allowing the transmitter to use minimum possible signal power in accomplishing the design objective of providing a specified error rate.

8.2

Speech Coding

The speech quality of codecs operating at a fixed bit rate is largely determined by the worst-case speech segments, i.e., those that are the most difficult to code at the given rate. Variable rate coding can provide a given level of speech quality at an average bit rate that is substantially less than the bit rate that would be required by an equivalent quality fixed rate codec [1–4]. The original CDMA codec known as Qualcomm code-excited linear predictive (QCELP), IS 96A was an 8 kbps code-excited linear predictive (CELP) variable rate

215

216

8

Speech Coding and Channel Coding

codec designed for use in the 900 MHz digital cellular band [11,12]. The desire for improved voice quality spurred the TIA to begin working on a 13 kbps CELP variable rate codec to provide higher quality voice transmissions. The 13 kbps CDMA codec takes advantage of the higher data rate (14.4 kbps as compared to 9.6 kbps for the 8 kbps codec) to improve speech quality. It has produced mean opinion scores (MOS) close to toll-quality voice, the benchmark used for comparison. Unfortunately, system capacity and cell coverage are reduced by the higher data rate codec.

8.2.1 Speech Coding Methods Speech coding is the process for reducing the bit rate of digital speech representation for transmission or storage, while maintaining a speech quality that is acceptable for the application. Speech coding methods are classified as waveform coding, source coding, and hybrid coding. The following sections will explain these concepts. In Figure 8.1, the bit rate is plotted on a logarithmic axis versus speech quality classes of “poor to excellent” corresponding to the five-point MOS scale values of 1 to 5, defined by the International Telecommunications Union (ITU). It may be noted that for low complexity and low delay, a bit rate of 32 to 64 kbps is required. This suggests the use of waveform codecs. However, for low bit rate of 4 to 8 kbps, hybrid codecs should be used (see Figure 8.1). These types of codecs tend to be complex with high delay [5–8].

SPEECH QUALITY WAVEFORM CODECS

EXCELLENT HYBRID CODECS

GOOD

FAIR

POOR VOCODERS BAD 1

Figure 8.1

2

4

8 16 BIT RATE (kbits/s)

Quality of service versus bit rate.

32

64

8.2

Speech Coding

217

8.2.2 Speech Codec Attributes Speech quality as produced by a codec is a function of transmission bit rate, complexity, delay, and bandwidth. Therefore, when considering speech codecs it is essential to consider all of these attributes and their interactions. For example, low bit rate codecs tend to have more delay compared to the higher bit rate codecs. They are generally more complex to implement and often have lower speech quality than the higher bit rate codecs. The following are the speech codec attributes. Transmission Bit Rate Since the speech codec shares the communications channel with other data, the peak bit rate should be as low as possible so as not to use a disproportionate share of the channel. The codecs below 64 kbps are primarily developed to increase the capacity of circuit multiplication equipment used for narrow bandwidth links. For the most part, they are fixed bit rate codecs, meaning they operate at the same rate regardless of the input. In the variable bit rate codecs, network loading and voice activity determine the instantaneous rate assigned to a particular voice channel. Any of the fixed rate speech codecs can be combined with a voice activity detector (VAD) and made into a simple two-state variable bit rate system. The lower rate could be either zero or some low rate needed to characterize slowly changing background noise characteristics. Either way, the bandwidth of the communications channel is only used for active speech. Delay The delay of a speech codec can have a great impact on its suitability for a particular application. For a one-way delay of conversation greater than 300 ms, the conversation becomes more like a half-duplex or push-to-talk experience, rather than an ordinary conversation. The components of total system delay include: (1) frame size, look ahead, multiplexing delay, (2) processing delay for computations, and (3) transmission delay. Most low bit rate speech codecs process a frame of speech data at a time. The speech parameters are updated and transmitted for every frame. In addition, to analyze the data properly it is sometimes necessary to analyze data beyond the frame boundary; hence, before the speech can be analyzed it is necessary to buffer a frame’s worth of data. The resulting delay is referred to as algorithmic delay. This delay component cannot be reduced by changing the implementation, but all other delay components can. The second major contribution for delay comes from the time taken by the encoder to analyze the speech and the decoder to reconstruct the speech. This part of the delay is referred to as processing delay. It depends on the speed of the hardware used to implement the coder. The sum of the algorithmic and processing delays is called the one-way codec delay. The third component of delay is due to transmission. It is the time taken for an entire frame of data to be

218

8

Speech Coding and Channel Coding

transmitted from the encoder to the decoder. The total of the three delays is the oneway system delay. In addition, frame interleaving delay adds an additional frame delay to the total transmission delay. Frame interleaving is necessary to combat channel fading (discussed in Chapter 3), and is part of the channel coding process. Complexity Speech codecs are implemented on special purpose hardware, such as digital signal processing (DSP) chips. DSP attributes are the computing speed in millions of instructions per second (MIPS), random access memory (RAM), and read only memory (ROM). For a speech codec, the system designer makes a choice about how much of these resources are to be allocated to the speech codec. Speech codecs using less than 15 MIPS are considered low-complexity codecs; those requiring 30 MIPS or more are thought of as high-complexity codecs. More complexity results in higher costs and greater power usage; for portable applications, greater power usage means reduced time between battery recharges or use of larger batteries, which means more expense and weight. Quality Of all the attributes, quality has the most dimensions. In many applications there are large amounts of background noise (car noise, street noise, office noise, etc.). How well does the codec perform under these adverse conditions? What happens when there are channel errors during transmission? Are the errors detected or undetected? If undetected, the codec must perform even more robustly than when it is informed that entire frames are in error. How good does the codec sound when speech is encoded and decoded twice? All these questions must be carefully evaluated during the testing phase of a speech codec. The speech quality is often based on the five-point MOS scale as defined by the International Telecommunication Union-Technical (ITU-T).

8.2.3 Linear-Prediction-Based Analysis-by-Synthesis (LPAS) In a linear predictive codec attempt is made to synthesize certain features of the voice message (time waveform). The linear-prediction-based analysis-by-synthesis (LPAS) methods provide efficient speech coding at rates between 4 and 16 kbps. In LPAS speech codecs [10], the speech is divided into frames, each of about 20 ms length, for which the coefficients of the linear predictor (LP) are computed. The resulting LP filter predicts each sample from a set of previous samples. In LPAS codecs, the residual signal is quantized on a subframe-by-subframe basis (there are commonly 2 to 8 subframes per frame). The resulting quantized signal forms the excitation signal for the LP synthesis filter. For each subframe, a criterion is used to select the best excitation signal from a set of trial excitation signals. The criterion compares the original speech signal with trial reconstructed speech signals. Because of the synthesis implicit in the evaluation criterion, the

8.2

Speech Coding

219

method is called analysis-by-synthesis coding. Various representations of excitation have been used. For lower bit rates, the most efficient representation is achieved by using vector quantization. For each subframe, the excitation signal is selected from a multitude of vectors, which are stored in a codebook. (The codebook consists of a set of stochastic excitation signals, ensembles of zero-mean Gaussian noise.) The index of the best matching vector is transmitted. At the receiver this vector is retrieved from the same codebook. The resulting excitation signal is filtered through the LP synthesis filter to produce the reconstructed speech. Linear prediction analysis-by-synthesis codecs using a codebook approach are commonly known as code-excited linear predictive (CELP) codec. The parametric coding is used for those aspects of the speech signal which are well understood. Parametric codecs are traditionally used at low bit rates. A proper understanding of a speech signal and its perception is essential to obtaining good speech quality with a parametric codec. The waveform matching procedure is employed for those aspects of speech which are not well understood. Waveform matching constraints are relaxed for those aspects which can be replaced by parametric models without degrading the quality of the reconstructed speech. A parameter which is well understood in parametric coding is the pitch of the speech signal. Satisfactory pitch estimation procedures are available. Piecewise linear interpolation of the pitch does not degrade speech quality. Pitch period is typically determined once every 20 ms and linearly interpolated between the updates. The challenge is to generalize the LPAS method such that its matching accuracy becomes independent of the synthetic pitch-period contour used. This is done by determining a time wrap of speech signal such that its pitch-period contour matches the synthetic pitch-period contour. The time wraps are determined by comparing a multitude of time-wrapped original signals with a synthesized signal. This coding scheme is called the generalized analysis-by-synthesis method and is referred to as Relaxed CELP (RCELP). The generalization relaxes the waveformmatching constraints without affecting speech quality.

8.2.4 Waveform Coding In general, waveform codecs are designed to be independent of signal. They map the input waveform of the encoder into a facsimile-like replica of it at the output of the decoder. Coding efficiency is quite modest. The coding efficiency can be improved by exploiting some statistical signal properties, if the codec parameters are optimized for most likely categories of input signals, while still maintaining good quality for other types of signals as well. The waveform codecs are further subdivided into time domain waveform codec and frequency domain waveform codec. Time Domain Waveform Coding The well-known representation of speech signal using time domain waveform coding is the A-law (in Europe) or μ-law (in North America) companded pulse code

220

8

Speech Coding and Channel Coding

modulation (PCM) at 64 kbps (see Chapter 4). Both use nonlinear companding characteristics to give a near-constant signal-to-noise ratio (SNR) over the total input dynamic range. The ITU G.721, 32 kbps adaptive differential PCM (ADPCM) codec is an example of a time domain waveform codec. More flexible counterparts of the G.721 are the G.726 and G.727 codecs. The G.726 codec is a variable-rate arrangement for bit rates between 16 and 40 kbps. This may be advantageous in various networking applications to allow speech quality and bit rate to be adjusted on the basis of the instantaneous requirement. The G.727 codec uses core bits and enhancement bits in its bit stream to allow the network to drop the enhancement bits under restricted channel capacity conditions, while benefiting from them when the network is lightly loaded. In differential codecs a linear combination of the last few samples is used to generate an estimate of the current one, which occurs in the adaptive predictor. The resultant difference signal (i.e., the prediction residual) is computed and encoded by the adaptive quantizer with a lower number of bits than the original signal, since it has a lower variance than the incoming signal. For a sampling rate of 8000 samples per second, an 8-bit PCM sample is represented by a 4-bit ADPCM sample to give a transmission rate of 32 kbps. Time domain waveform codecs encode the speech signal as a full-band signal and map it into as close a replica of the input as possible. The difference between various coding schemes is their way of using prediction to reduce the variance of the signal to be encoded in order to reduce the number of bits necessary to represent the encoded waveform. Frequency Domain Waveform Coding In the frequency domain waveform codecs the input signal undergoes some shorttime spectral analysis. The signal is split into a number of frequency-domain subbands. The individual sub-band signals are then encoded by using different numbers of bits to fulfill the quality requirements of that band based on its prominence. The various schemes differ in their accuracies of spectral analysis and in the bit allocation principle (fixed, adaptive, semi-adaptive). Two well-known representatives of this class are sub-band coding (SBC) and adaptive transform coding (ATC).

8.2.5 Vocoders Vocoders are parametric digitizers which use certain properties of the human speech production mechanism. Human speech is produced by emitting sound pressure waves which are radiated primarily from the lips, although, with some sounds, significant energy emanates also from the nostrils, throat, etc. In human speech, the air compressed by the lungs excites the vocal cord in two typical modes. When generating voice sounds, the vocal cord vibrates and generates quasi-periodic voice sounds. In the case of lower energy unvoiced sounds, the

8.2

Speech Coding

221

vocal cord does not participate in voice production and the source acts like a noise generator. The excitation signal is then filtered through the vocal apparatus, which behaves like a spectral shaping filter. This can be described adequately by an all-pole transfer function that is constituted by the spectral shaping action of the vocal tract, lip radian characteristics, etc. In the case of vocoders, instead of producing a close replica of an input signal at the output, an appropriate set of source parameters is generated to characterize the input signal sufficiently close for a given period of time. The following steps are used in this process: 1. Speech signal is partitioned in segments of 5 to 20 ms. 2. Speech segments are subjected to spectral analysis to produce the coefficients of the all-zero analysis filter to minimize the prediction residual energy. This process is based on the computation of the speech autocorrelation coefficients and then using either matrix inversion or iterative scheme. 3. The corresponding source parameters are specified. The excitation parameters as well as filter coefficients are quantized and transmitted to the decoder to synthesize a replica of the original signal by exciting the all-pole synthesis filter. The quality of this type of scheme is predetermined by the accuracy of the source model rather than the accuracy of the quantization of the parameters. The speech quality is limited by the fidelity of the source model used. The main advantage of vocoders is their low bit rate, with the penalty of relatively low, synthetic speech quality. Vocoders can be classified into the frequency domain and time domain subclasses. However, frequency domain vocoders are generally more effective than time domain vocoders.

8.2.6 Hybrid Coding Hybrid coding is an attractive trade-off between waveform coding and vocoder, both in terms of speech quality and transmission bit rate, although generally at the price of higher complexity. They are also referred to as analysis-by-synthesis (ABS) codecs. The most recent international and regional speech coding standards belong to a class of LPAS codecs. This class of codecs includes ITU G723.1, G.728 (lowdelay (LD) CELP, 16 kbps), and G.729 and all the current digital cellular standards including: • European global system for mobile communications (GSM), full-rate, half-

rate, enhanced full-rate (EFR), and adaptive multiple rate (AMR) codec • North American full-rate, half-rate, and enhanced full-rate for time division

multiple access (TDMA) IS-136 and code division multiple access (CDMA) IS-95 systems • Japanese public digital cellular (PDC) full-rate and half-rate

222

8

Speech Coding and Channel Coding

In an LPAS coder, the decoded speech is generated by filtering the signal produced by the excitation generator through both a long-term (LT) predictor synthesis filter and a short-term (ST) predictor synthesis filter. The excitation signal is found by minimizing the mean-squared error over a block of samples. The error signal is the difference between the original and decoded signal. It is weighted by filtering it through a weighting filter. Both ST and LT predictors are adapted over time. Since the analysis procedure (encoder) includes the synthesis procedure (decoder), the description of the encoder defines the decoder. The short-term synthesis filter models the short-term correlations (spectral envelope) in the speech signal. This is an all-pole filter. The predictor coefficients are determined from the speech signal using linear prediction (LP) techniques. The coefficients of the short-term predictor are adapted in time, with rates varying from 30 to as high as 400 times per second. The long-term predictor filter models the long-term correlations (fine spectral structure) in the speech signal. Its parameters are a delay and gain coefficient. For periodic signals, delay corresponds to the pitch period (or possibly an integral number of pitch periods). The delay is random for non-periodic signals. Typically, the long-term predictor coefficients are adapted at rates varying from 100 to 200 times per second. An alternative structure for the pitch filter is the adaptive codebook. In this case, the long-term synthesis filter is replaced by a codebook that contains the previous excitation at different delays. The resulting vectors are searched, and the one that provides the best result is selected. In addition, an optimal scaling factor is determined for the selected vector. This representation simplifies the determination of the excitation for delays smaller than the length of excitation frames. Code-excited linear predictive (CELP) codecs use another approach to reduce the number of bits per sample. Both encoder and decoder store the same collection of codes (C) of possible length L in a codebook. The excitation for each frame is described completely by the index to an appropriate vector. This index is found by an exhaustive search over all possible codebook vectors, using the one that gives the smallest error between the original and decoded signals. To simplify the search it is common to use a gain-shape codebook in which the gain is searched and quantized separately. Further simplifications are obtained by populating the codebook vectors with a multipulse structure. By using only a few non-zero unit pulses in each codebook vector, efficient search procedures are derived. This partitioning of excitation space is referred to as an algebraic codebook. The excitation method is known as algebraic codebook excited linear prediction (ACELP).

8.3 Speech Codecs in European Systems 8.3.1 GSM Enhanced Full-Rate (EFR) GSM EFR is the same as a US1 vocoder with 8-PSK modulation and is a 12.2 kbps vocoder. It is used in GSM1900. The distribution of bits in a GSM EFR vocoder is given in Table 8.1.

8.3

Speech Codecs in European Systems

223

The IS-136 (TDMA) system uses the vector self-excited linear predictor (VSELP) codec. The VSELP algorithm uses a codebook with a predefined structure to reduce the number of computations. The output of the VSELP codec for IS-136 is 7.95 kbps. It produces a speech frame every 20 ms containing 159 bits. Recently, the algebraic codebook excited prediction (ACELP) codec (IS-641 vocoder) was selected to replace the VSELP codec in IS-136. This codec has an output bit rate of 7.4 kbps. Comparisons of MOS and delay for IS-641, ITU LD-CELP and GSM EFR are given in Tables 8.2 and 8.3. Table 8.1 GSM enhanced full-rate vocoder (12.2 kbps). Bits per 20 ms LP Filter Coefficients

38

Adaptive Excitation

30

Fixed or Algebraic Excitation (4 subframes/ frame each with 35 bits)

4  35  140

Gains (4 subframes/frame each with 9 bits)

4  9  36 244 or 12.2 kbps

Total

Table 8.2 Comparison of MOS (5 is the best and 1 is the lowest). Condition

Original

IS-641

LD-CELP

GSM EFR

Clean Speech

4.34

4.09

4.23

4.26

15 dB Babble

3.75

3.49

3.81

3.70

20 dB Car Noise

3.72

3.61

3.64

3.75

15 dB Office Noise

3.70

3.40

3.61

3.58

15 dB Music

3.99

3.82

3.98

3.99

Table 8.3 Comparisons of delay (ms). Delay Cause

IS-641

GSM EFR

LD-CELP

Look-ahead

5

0

0

Frame Size

20

20

0.625

Processing

16

16

0.5

0

0

19.375

Transmission

Bit Stream Buffer

26.6

6.6

6.6

Total Delay

67.6

42.6

27.1

224

8

Speech Coding and Channel Coding

Results show an advantage for ITU LD-CELP and GSM EFR when compared to IS-641 for every condition. The disadvantage of the LD-CELP is its higher bit rate. This means fewer bits are available for error protection. In weaker channel conditions, GSM EFR will have an advantage over LD-CELP due to its lower bit rate, fewer sensitive bits to protect, and faster recovery from frame erasures. IS-641 also has the same advantages, and on an even weaker channel its performance would surpass that of GSM EFR. ITU LD-CELP has a distinct advantage as far as total delay is concerned. The advantages of IS-641 over LD-CELP are in complexity and bit rate. LD-CELP is a low-delay coder and can produce better clear channel quality than IS-641 for a variety of conditions. GSM EFR has clear channel quality performance on a par with LD-CELP, has a small delay advantage compared to IS-641, and is a good candidate as an upgrade to IS-641 for strong channel conditions.

8.3.2 Adaptive Multiple Rate Codec ETSI adaptive multiple rate (AMR) speech codec design (used for 3G UMTS) incorporates multiple codecs for use in full-rate or half-rate mode that are determined by channel quality. AMR operates in eight submodes that incorporate bitexact versions of both 12.2 kbps US1/GSM EFR and 7.4 kbps IS-641 full-rate speech coders. The AMR can increase voice capacity by 150% compared to the GSM. Tables 8.4 and 8.5 provide details of AMR speech codecs. The AMR codec allows dynamic management of voice quality and error control to provide good voice quality even under adverse radio conditions. AMR is not only used in GSM, but also in EDGE and UMTS networks [15]. It is designed to work with both GSM full-rate (one user per each of eight time slots in each radio channel) and GSM half-rate (two users per time slot). AMR defines multiple

Table 8.4 ETSI AMR speech coder full-rate (22.8 kbps) submodes.

Submode

Speech codec source rate (kbps)

Channel coding rate (kbps)

1

12.2 (US1/GSM EFR)

10.6

2

10.2

12.6

3

7.95

14.85

4

7.4 (IS-641 ACELP)

15.4

5

6.7

16.1

6

5.9

16.9

7

5.15

17.65

8

4.75

18.05

8.3

Speech Codecs in European Systems

225

Table 8.5 ETSI AMR half-rate (11.4 kbps) submodes.

Submodes

Speech codec source rate (kbps)

Channel coding rate (kbps)

1

7.95

3.45

2

7.4 (IS-641 ACELP)

4.0

3

6.7

4.7

4

5.9

5.5

5

5.15

6.25

6

4.75

6.65

voice encoding rates, each with a different level of error control (see Table 8.4). The AMR codec dynamically responds to radio conditions by using the most effective mode of operation at each moment of time. Compared to the GSM EFR codec, AMR can operate under much worse radio conditions, such as with a heavily loaded network. AMR offers the following benefits: • Greater spectral efficiency, hence higher capacity from higher frequency

reuse with frequency hopping • Better voice quality throughout the cell, particularly at cell edges and deep inside buildings, and increased overall coverage • The potential of operating with toll-quality voice in half-rate mode, which reduces network costs The 12.2 kbps rate of the AMR is the same as GSM EFR codec, and the 7.4 kbps rate is the same as the IS-136 TDMA codec. The gross bit rate of the channel (one time slot) is 22.8 kbps, which is divided into voice information and error control. As an example, 7.95 kbps mode means that more than half of the bit rate (i.e., 22.8  7.95  14.85 kbps) can be allocated to channel coding. By reducing the AMR rate, resistance to errors increases further. Figure 8.2 shows the resulting effect from voice quality (MOS) versus the S/I ratio. Dynamic capability means AMR can compensate for the higher error rates arising from techniques such as tighter frequency reuse and higher fractional network loading, which inherently force the mobile to operate at lower S/I. The six submodes of AMR half-rates are given in Table 8.5. Since the gross bit rate in a half-rate channel is only 11.4 kbps, a much smaller number of bits is available for channel coding, thus requiring a better S/I. But with a better radio signal, AMR can enable half-rate operation, which translates to more users in the same number of radio channels. AMR half-rate mode is further enhanced in EDGE radio networks where more bits per time slot are available.

226

8

Speech Coding and Channel Coding

Voice quality Mode 1

AMR performance Voice Codec mode changes

Mode 2

Mode 3

S/I Channel Quality

Figure 8.2(a)

30 AMR-HR

25 Speech Quality (MOS)

20

AMR-FR

15 GSM-HR 10

GSM-EFR

5 0 5 10

Full Role Half Role 5

0

5

Figure 8.2(b) Figure 8.2

MOS versus S/I for AMR codec.

10

15

20

25

30 S/I (dB)

8.4

CELP Speech Codec

227

The AMR codec operates on speech frames of 20 ms corresponding to 160 samples at the sampling frequency of 8000 samples per second. Depending on the air interface loading and quality of speech connections, the bit rate of the AMR speech connection can be controlled by the radio access network. During high loading, such as during busy hours, it is possible to use lower AMR bit rates to offer higher capacity while providing slightly lower speech quality. Also, if the mobile is running out of the cell coverage area and using maximum transmission power, a lower AMR bit rate can be used to extend the cell coverage area. With the AMR codec it is possible to achieve a trade-off between the network’s capacity, coverage and speech quality according to the operator’s requirements. Example 8.1 The coverage gain in dB for the AMR is given as: DPDCH (kbps)  DPCCH

 DPDCH[(AMR bit rate (kbps))]  DPCCH 

10 log  dB

Assuming the power difference between the dedicated physical control channel (DPCCH) and dedicated physical data channel (DPDCH) of the WCDMA to be 3.0 dB for 12.2 kbps AMR speech, calculate the gain in the link budget in dB by reducing the AMR bit rate from 12.2 to 7.95 kbps, and by reducing the AMR bit rate from 12.2 to 4.75 kbps. Solution (a)  12.2  103/10 18.315 Coverage gain  10 log 12.2  10 log   1.15 dB  3/10

 7.95  12.2  10



 14.064 

(b)

 10.865 

12.2  12.2  103/10 18.315 Coverage gain  10 log   10 log   2.27 dB 3/10

 4.75  12.2  10

8.4



CELP Speech Codec

Code excited linear predictive (CELP) codec dynamically selects one of the four data rates every 20 ms, depending on the speech activity. The four rates are 8 kbps (full-rate), 4 kbps (half-rate), 2 kbps (quarter-rate), and 1 kbps (eighth-rate). Typically, active speech is coded at the 8 kbps rate, while silence and background noise are coded at the lower rates. MOS testing has shown that QCELP provides speech quality equivalent to that of 8 kbps VSELP, while maintaining an

228

8

Hamming Window

DC Offset Removal

Analog to Uniform PCM Converter

Analog Speech Input

Speech Coding and Channel Coding

Compute Auto Correlation Coefficients

Determine LPC Coefficients

Scale LPC Coefficients

µ-law PCM to Uniform PCM Converter

Transform LPC Coefficients into LSPs

µ-law PCM Speech Input

Convert LSP Frequencies into LSP Codes

Output LSP Codes (a)

Figure 8.3(a)

Uniform PCM

Compute Error Function

Choose L & b to Minimize Error

Output Lag (L) & Gain (b)

LSP Codes (a)

All Possible Values of L & b

Figure 8.3(b)

Uniform PCM

LSP Codes (a) Pitch Lag and gain (L & b) All Possible Values of I & G Figure 8.3(c)

Compute Error Function

Choose G & I to Minimize Error

Output Codebook (I) & Gain (G)

8.4

CELP Speech Codec

229

Synthesis Filter — Pitch & Format

Code Book

X

Codebook Index (I)

Codebook Gain (G)

Spectral Post Filter and Gain Control

Output Speech

Pitch lag (L), Pitch Gain (b), Pitch Spectral Lines (a)

Inputs to CELP Decoder Figure 8.3(d)

X

Codebook Seed (CBSEED)

Codebook Gain (G)

Synthesis Filter — Pitch & Format

Spectral Post Filter and Gain Control

Output Speech

Figure 8.3(e)

Pseudo Random Vector Generator

Pitch Spectral Lines (a)

Inputs to CELP Decoder Figure 8.3

Block diagram for CELP encoder/decoder.

Table 8.6 Bit allocation for CELP at 8 kbps (full rate). 40

LPC Pitch Code-book Parameter

10 10

10 10

10

10 10

10

10 10

10

10

average data rate under 4 kbps in a typical conversation. Figure 8.3 shows block diagrams of the encoder/decoder. The bit allocation for each data rate is given in Tables 8.6, 8.7, 8.8, and 8.9, respectively. A 10th order LPC filter is used. Its coefficients are encoded using linear spectral pair (LSP) frequencies due to the good

230

8

Speech Coding and Channel Coding

Table 8.7 Bit allocation for CELP at 4 kbps (half rate). 20

LPC Pitch

10

Code-book Parameter

10

10

10

10

10

Table 8.8 Bit allocation for CELP at 2 kbps (1/4 rate). 10

LPC Pitch

10

Code-book Parameter

10

10

Table 8.9 Bit allocation for CELP at 1 kbps (1/8 rate). LPC

10

Pitch

0

Code-book Parameter

6

quantization, interpolation, and stability properties of LSPs. LSP is a representation of digital filter coefficients in pseudo-frequency domain.

8.5

Enhanced Variable Rate Codec

The CDMA Development Group’s (CDG) enhanced variable rate codec (EVRC) takes advantage of signal processing hardware and software techniques to provide 13 kbps voice quality at 8 kbps data rate, thereby maximizing both quality and system capacity [13,14]. MOS data has shown that the 8 kbps EVRC (see Figure 8.4) compares favorably to 16 kbps LD-CELP and ADPCM, the industry standards for comparison. More important, the tests have shown that EVRC maintains superior quality over 13 kbps CELP as frame error rates (FERs) rise. One of the important and unique aspects of CDMA wireless systems is that while there are limits to the number of mobile calls that can be handled by a given carrier at a time, this capacity limit is not a fixed number. In CDMA, cell coverage depends on the way the system is designed and implemented. System capacity, voice quality, and coverage are all interrelated, enabling a service provider to trade any one off against the other two. To maximize the number of simultaneous calls that can be handled at any given time in the allocated frequency spectrum, digital wireless systems utilize

8.5

Enhanced Variable Rate Codec

231

Rate 1/8 Encoding

Sampled Speech

Signal Preprocessing

External Rate Command

Model Parameter Estimation

Packet Formatting Rate 1/2 or Rate 1 Encoding

Rate Determination

Formatted Packet

Rate Decision

Note: Less number of bits are used for pitch and more to protect against errors

Figure 8.4

Enhanced variable rate codec (EVRC).

speech compression between the mobile and base stations. A lower speech transmission rate enables a higher number of simultaneous calls that a system can handle at any given time within the allocated carrier frequency spectrum. The variable rate vocoder in CDMA uses the smallest number of bits to represent each call without sacrificing voice quality. The basis of the standard EVRC algorithm is the Relaxed Code-Excited Linear Predictive (RCELP) coding. RCELP coding is a generalization of the CELP speech coding algorithm. It is particularly well suited for variable rate operation and robustness in the CDMA environment. CELP uses 20 ms speech frames for coding and decoding. In each 20 ms time interval, the encoder processes 160 samples of speech. With variable rate coders, the encoder examines the contents of each speech frame to determine the necessary coding rate. Depending on the voice waveform (volume, pitch, rate, and so on), the coder represents speech at one of three bit rates: 8, 4, or 1 kbps. As a result, the average bit rate is less than 8 kbps. This differs from a half-rate or multirate coder, where the bit rate is determined once for each call. In addition, when no voice is detected, the vocoder drops its encoding rate and the effective bit rate goes to 1 kbps, further reducing the interference energy produced. CELP codecs use three sets of bits to represent speech: linear predictor filter coefficients, pitch parameters, and excitation waveform. For each 20 ms speech frame, the CELP algorithm examines the data and generates 10 linear prediction coding filter coefficients. With EVRC, the coefficients are represented by a vector, which is a set of the most likely usable coefficients. Increased or decreased bit precision is applied as necessary. CELP speech coders also perform long-term pitch analysis to generate a 7-bit pitch period and a 3-bit pitch gain. The analysis is based on a mathematical model of the human vocal track. At the different rates, the coder performs this

232

8

Speech Coding and Channel Coding

pitch analysis on either four 5-ms subframes, two 10-ms subframes or one 20-ms frame. The result is a variable number of bits per frame representing the pitch information. EVRC makes two pitch measurements and uses wrapped pitch delays, removing the requirement for a large number of bits and increased computations for fractional pitch delays. The excitation waveform for every frame or subframe is selected from a codebook consisting of a large number of candidate waveform vectors. The codebook vector chosen to excite the speech coder filters minimizes the weighted error between the original and synthesized speech. The technique used by EVRC to reduce the number of bits required for linear predictor coefficients and pitch synthesis enables the algebraic codebook to generate excitation. As a result, EVRC has higher voice quality. Unlike conventional CELP encoders, EVRC does not attempt to match the original speech signal exactly. Instead, EVRC matches a time-wrapped version of the residual that conforms to a simplified pitch contour. The contour is obtained by estimating the pitch delay in each frame and linearly interpolating the pitch from frame to frame. While this adds to computational complexity, the result is higher voice quality per bit transmitted. The simplified pitch representation also leaves more bits available in each packet for the stochastic excitation and channel impairment protection than would be possible if a traditional fractional pitch approach were used. The result is enhanced error performance without degraded speech quality at the small cost of added processing requirements. EVRC also enhances call quality by suppressing background noise. The IS-127 [9] standard recommends a noise suppressor algorithm, but allows system designers to define their own. This is an important factor in choosing a processing platform, making programmable DSPs a desirable choice. The EVRC algorithm offers a significant performance improvement over the IS-96A speech codec. Table 8.10 shows the performance of the CDMA Development Group’s 13 kbps (CDG-13 kbps) speech codec along with IS-96A and EVRC codecs. The CDG-13 kbps offers high voice quality but results in a decrease in channel capacity of about 40% over the 8 kbps codec (the processing gain is reduced from 128 to 85.2, resulting in a 40% reduction in system capacity). Table 8.11 provides the bit allocations by packet type. Table 8.10 Comparisons of CDG-13 kbps, IS-96A and EVRC vocoders in MOS. Frame error rate (FER)%

CDG-13 kbps

IS-96A

EVRC

0

4.00

3.29

3.95

1

3.95

3.17

3.83

2

3.88

2.77

3.66

3

3.67

2.55

3.50

8.6

Channel Coding

233

Table 8.11 Bit allocations by packet type in EVRC. Packet Type Rate 1

Field

Rate 1/2

Spectral Transition Indicator

1

LSP

28

22

Pitch Delay

7

7

Delta Delay

5

ACB Gain

9

9

FCB Shape

105

30

FCB Gain

15

12

Total Encoded Bits

Blank

8

8

Frame Energy Unused

Rate 1/8

1 171

80

Mixed Mode Bit (MM)

1

Frame Quality Indicator (CRC) (F)

12

8

Encoder Tail Bits (T)

8

8

16

8

8

Total Bits

192

96

24

8

Rate (kb/s)

9.6

4.8

1.2

0.4

LSP (Line Spectral Pair)  A representation of digital filter coefficient in a pseudo frequency domain. This representation has good quantization and interpolation properties. ACB  Adaptive Codebook FCB  Fixed Codebook

8.6

Channel Coding

High-speed digital communication systems development demands the optimization of: • Data transmission rate • Data reliability • Transmission energy • Bandwidth • System complexity • Cost

Error correcting codes help to meet the requirements cost effectively with a reduced cost. A higher error correction code performance offers more flexibility to a designer in determining the required transmission energy, bandwidth, and system

234

8

Speech Coding and Channel Coding

complexity. Signal-to-noise ratio (SNR) improvement of a communication channel depends on the error correction code used and the channel’s characteristics. Error control coding is the process of adding redundant information to a message to be transmitted that can then be used at the receiving end to detect and possibly corrects errors in the transmission. Since the redundant information adds overhead to the transmission, the type of coding must be chosen based upon how many errors the system is expected to see, and whether the capability to request retransmission of data is available. There are two basic error control coding classifications: automatic repeat request (ARQ) and forward error correction (FEC). ARQ is a detection-only type coding, in which transmission errors can be detected by the receiver but not corrected. The receiver must ask for any data received and request that detected errors be retransmitted. FEC allows not only detection of errors at the receiving end but correction of errors as well. We primarily focus on the FEC techniques and present the error correction codes (i.e., block, convolutional, and turbo) that are generally used for FEC. Reed-Solomon (RS), Viterbi (V) Convolutional, and concatenated ReedSolomon Viterbi (RSV) are the most common error correction codes implemented today. At a bit error rate (BER) of 106, these codes are at least 2.5 to 3.0 dB short of the Shannon limit in an additive white Gaussian noise (AWGN) channel. Turbo codes have been shown to perform within 1 dB of the Shannon limit at a BER of 106. Turbo codes break a complex decoding problem into simple steps, in which each step is repeated until a solution is reached.

8.6.1 Reed-Solomon (RS) Codes RS coding is a type of FEC. It has been widely used because of its relatively large error correction capability when weighed against its minimal added overhead. RS codes are also easily scaled up or down in error correction capability to match the error rates expected in a given system. It provides a robust error control method for many common types of data transfer mediums, particularly those that are one-way or noisy and sure to produce errors. In block codes a sequence of K information symbols is encoded in a block of N symbols, N  K, to be transmitted over the channel. For a data source that delivers the information bits at the rate B bps, every T seconds the encoder receives a sequence of K  BT bits which defines a message. After K information bits have entered the encoder, the encoder generates a sequence of coded symbols of length N to be transmitted over the channel. In this transmitted sequence or codeword, N must be greater or equal to K in order to guarantee a unique relationship between each codeword and each of the possible 2K messages. Such a code which maps a block of K information symbols into a block of N coded symbols is called an (N,K) block code. The code rate is r  K/N bits/symbol, N is called the block length. RS codes are an example of a block coding technique. The data stream to be transmitted is broken up into blocks and redundant data is then added to each

8.6

Channel Coding

235

Transmitting Side Data (N-R)

RS Encoder (check symbols added)

If all errors are corrected, sent and received data should be identical

Data (N-R)  Check symbols (R)  Codeword (N)

Transmission medium Receiving Side

Corrected Data

Figure 8.5

RS Decoder (errors detected/corrected)

Possibly corrupted data and check symbols (received codeword)

RS system level block diagram.

block. The size of these blocks and the amount of redundant data added to each block is either specified for a particular application or can be user-defined for a closed system. Within these blocks, the data is further subdivided into a number of symbols, which are generally from 6 to 10 bits in size. The redundant data then consists of additional symbols being added to the end of the transmission. The system-level block diagram for an RS codec is shown in Figure 8.5. The original data, which is a block consisting of N-R symbols, is run through an RS encoder and R check symbols are added to form a code word of length N. Since RS can be done on any message length and can add any number of check symbols, a particular RS code is expressed as RS (N, N-R) code. N is the total number of symbols per code word; R is the number of check symbols per code word, and N-R is the number of actual information symbols per code word. RS encoding consists of the generation of check symbols from the original data. The process is based upon finite field arithmetic. The variables to generate a particular RS code include field polynomial and generator polynomial starting roots. The field polynomial is used to determine the order of the elements in the finite field. Another system-level characteristic of RS coding is whether the implementation is systematic or nonsystematic. A systematic implementation produces a code word that contains the unaltered original input data stream in the first R symbols of the code word. In contrast, in a nonsystematic implementation, the input data stream is altered during the encoding process. Most specifications require systematic coding. The simplified schematic representation for a systematic RS encoder is shown in Figure 8.6. The input data stream is immediately clocked back out of the function into the check symbol generation circuitry. The fact that the input data stream is clocked out immediately without being altered means that the implementation is systematic. A series of finite fields adds and multiplies results in each register

236

8

Speech Coding and Channel Coding

Output Select

Check Symbol R

Check Symbol R-1

Check Symbol 1

Output Data Clock Input Data Finite field multiplier Finite field adder

Figure 8.6

Systematic RS encoder schematic representation.

(Division by the generator polynomial) (Find factors in remainder) (Evaluate polynomials to find errors)

Corrected Codeword Syndrome Calculations

Euclid Algorithm

Chien Search Error Flag

Figure 8.7

RS decoder block diagram.

containing one check symbol after the entire input stream has been entered. At that point, the output select is switched over to the check symbol registers, and the check symbols are shifted out at the end of the original message. The size of the encoder is most heavily affected by the number of check symbols required for the target RS code. The total message length, as well as the field polynomial and first root value, do not have any appreciable effect on the device performance. A typical RS decode algorithm consists of several major blocks. The first of these blocks is the syndrome calculation, where the incoming symbols are divided into the generator polynomial, which is known from the parameters of the decoder. The check symbols, which form the remainder in the encoder section, will cause the syndrome calculation to be zero in the case of no errors. If there are errors, the resulting polynomial is passed to the Euclid algorithm, where the factors of the remainder are found (see Figure 8.7). The result is evaluated for each of the incoming symbols over many iterations, and any errors are found and

8.6

Channel Coding

237

corrected. The corrected code word is the output from the decoder. If there are more errors in the code word than can be corrected by the RS code used, then the received code word is output with no changes and a flag is set, stating that the error correction has failed for that code word. The error correction capability of a given RS code is a function of the number of check bits appended to the message. In general, it may be assumed that correcting an error requires one check symbol to find the location of the error, and a second check symbol to correct the error. In general then, a given RS code can correct R/2 symbol errors, where R is the number of check symbols in the given RS code. Since RS codes are generally described as an RS (N, N-R) value, the number of errors correctable by this code is [N- (N-R)]/2. This error control capability can be enhanced by use of erasures, a technique that helps to determine the location of an error without using one of the check symbols. An RS implementation supporting erasures would then be able to correct up to R errors. Since RS codes work on symbols (most commonly equal to one 8-bit byte) as opposed to individual data bits, the number of correctable errors refers to symbol errors. This means that a symbol with all of the bits corrupted is no different than a symbol with only one of its bits corrupted, and error control capability refers to the number of corrupted symbols that can be corrected. RS codes are more suitable to correct consecutive bits. RS codes are generally combined with other coding methods such as Viterbi, which is more suited to correcting evenly distributed errors. The effective throughput of an RS decoder is a combination of the number of clock cycles required to locate and correct errors after the code word has been received and the speed at which the design can be clocked. Knowing the latency and clock speed allows the user to determine how many symbols per second may be processed by the decoder. In the RS code, there are two RS decoder choices: a high-speed decoder and a low-speed decoder. The trade-off is that the low-speed decoder is usually approximately 20% smaller in device utilization. Note that both decoders operate at the same clock rate, but the low-speed decoder has a longer latency period, resulting in a slower effective symbol rate. As the number of check symbols decreases, the complexity of the decoder decreases, resulting in a smaller design and an increase in performance. In a real-life RS coding implementation, functions that tend to reside on either side of the RS encoder or decoder are often implemented in programmable logic. One function that often resides after an RS encoder is an interleaver. The task of an interleaver is to scramble the symbols in several RS code words before transmission, effectively spreading any burst error that occurs during transmission over several code words. Spreading this burst error over several code words increases the chance of each code word being able to correct all of its induced errors.

8.6.2 Convolutional Code Convolutional codes are suitable when the information symbols to be transmitted arrive serially in long sequences rather than in blocks. In convolutional codes,

238

8

Speech Coding and Channel Coding

long sequences of information symbols are encoded continuously in serial form. To provide the extra bits required for error control, an output rate greater than the message bit is achieved by connecting two or more mod-2 adders to the shift registers (see Figure 8.8). Each message bit influences a span of n(L  1) successive output bits. The quantity n(L  1) is called the constraint length measured in terms of encoded output bits, where L is the encoder’s memory measured in terms of input message. L represents the number of shift-registers in the encoder. One at a time the symbols enter into the encoder, which has some finite memory capacity. The information symbols are sequentially shifted through an L-stage shift register, and following each shift some number, v, of coded symbols are generated and transmitted. These v coded symbols are obtained by parity checking, that is, by modulo-2 addition of the contents of various stages of the shift register according to the specific code. The length K  L  1 is called the constraint length of the code, and the code rate is r  1/v bit per transmitted symbol. A binary convolutional code of rate 1/v bits per symbol can be generated by a linear finite-state machine consisting of an L-stage shift register, v modulo-2 adders connected to some of the shift registers, and a commutator that scans the output of the modulo-2 adders. The whole system is called a convolutional encoder (see Figure 8.8). The error-correcting capability of a convolutional coding scheme increases as rate r decreases. However, the channel bandwidth and decoder complexity both increase with K. The advantage of lower code rates when using a convolutional code with coherent phase-shift-keying (PSK), is that the required Eb/N0 is

Rate (r )  1/2 Constraint length K  9

X1

X2

g0

X3

X4

c0

X5

X6

X8

X7

Shift Register

Information Bits (input) g1

c1 Code Symbols (output)

Figure 8.8

Convolutional encoder.

8.6

Channel Coding

239

g1  (1011)

k  4, r  1/2

v1 10000

11101 x0

x1

x2

g2  (1111)

v2 11011

g1  1  x 2  x 3 g2  1  x  x 2  x 3

Figure 8.9

Convolution encoder with K ⴝ 4 and r ⴝ 1/2.

decreased, permitting the transmission of higher data rates for a given amount of power, or permitting reduced power for a given data rate. Simulation studies have indicated that for a fixed constraint length, a decrease in code rate from 1/2 to 1/3 results in reduction of the required Eb /N0 of about 0.4 dB. However, the corresponding increase in decoder complexity is about 17%. For smaller values of code rate, the improvement in performance relative to the increased decoding complexity diminishes rapidly. The major drawback of the Viterbi algorithm is that while error probability decreases exponentially with constraint length, the number of code states, and consequently decoder complexity, grows exponentially with constraint length. Example 8.2 Figure 8.9 shows a convolution encoder with K  4 and r  1/2. The generator polynomials g1 and g2 are 1  x2  x3 and 1  x  x2  x3, respectively. If the input is 1 1 1 0 1 (first bit), calculate the output of the encoder. Solution Refer to Table 8.12. Figures 8.10 and 8.11 indicate the implementations of convolutional coding in GSM for downlink and uplink. Figure 8.10 shows the downlink channel coding that is used with the GSM EFR speech coder and 8-PSK modulation. The bits are combined such that every two class IA bits are used with one class II bit to form the first 89 triads. The remaining 44 triads are formed by using three class IB bits. For the uplink (Figure 8.11) the bits are combined in a somewhat similar manner. The 86 triads consist of two class IA bits and one class II bit, three triads contain two class IB bits along with one class II bit, and 35 triads consist of three class IB bits.

240

8

Speech Coding and Channel Coding

Table 8.12 Input /output state table of the encoder. X0 (ⴝx)

X1(ⴝx2)*

X2(ⴝx3)*

V1

V2

1

0

0

0

1

1

0

1

0

0

0

1

1

0

1

0

0

0

1

1

0

1

0

1

1

1

1

0

0

1

Input after each shift*

The bits in columns marked with an * are added using modulo-2 to produce V1 after each shift. The output of the encoder is 1 0 1 0 0 0 1 0 1 1.

Convolutional Encoder r  1/2, K  6

8 CRC bits Speech Coder IA 81 bits

89 bits

178 bits

2  89

Punching 244 bits

IB 74 bits

II 89 bits

Figure 8.10

2  74

148

148  16  132

132 bits

Convolutional Encoder

Channel coding (downlink) with GSM EFR speech coder.

The triads are reordered to provide additional interleaving gain, and then intra-slot interleaving is applied. There are three primary interleaving options: one-slot, two-slot, and three-slot. In the one-slot interleaving, there is no intra-slot interleaving and the current triads are simply transmitted in the current slot. In the two-slot interleaving, certain triads (from the current triad vector) are transmitted in the current slot, and the remaining triads in the next slot. Thus, the current slot contains triads from the current and previous triad vector. In the three-slot interleaving, the concept is extended to include another set of triads in each transmitted slot. Certain triads from the current triad vector are transmitted in the next slot, and the remaining triads are transmitted in the slot after the next slot. Thus, the current slot contains triads from the current triad vector, the previous triad vector, and one before the previous triad vector.

8.6

Channel Coding

241

Convolutional Encoder r  1/2, K  6

8 CRC bits Speech Coder IA 81 bits

89 bits

2  89

178

178  6  172

172 bits

Punching 244 bits

IB 74 bits

II 89 bits

Figure 8.11

2  74

148

148  37  111

111 bits

Convolutional Encoder

Channel coding (uplink) with GSM EFR speech coder.

8.6.3 Turbo Coding A turbo decoder consists of two concatenated decoders, each providing soft information (see the next section) and so-called intrinsic information. Two main classes of algorithms available for turbo codes are soft-output Viterbi algorithm (SOVA) and iterative soft-input and soft-output decoding algorithms such as the symbol-by-symbol maximum a posterior algorithm, which is more complex but yields better performance. Since the capacity of a CDMA system is strongly dependent on the required Eb /N0 value of the receiver, any improvement in the Eb/N0 value due to coding translates directly into capacity increase. The inner receiver’s (in serial arrangement) task is to provide the best estimate for the outer receiver. It deals with the following signal impairments: • The presence of multipath components • The presence of multi-user interference (both inter- and intracell) • The fading of each transmission path • The near-far effect due to the relative position of all mobiles and base

station • The time-varying nature of these impairments These issues are tackled with a combination of the following techniques: • Channel estimation and tracking • A maximum ratio combination-based (coherent) Rake receiver, to take

advantage of multipath characteristics • Multi-user detection (MUD) schemes, such as interference cancellation (IC)

or decorrelating receivers

242

8

Speech Coding and Channel Coding

• Fast power control based on SNR estimation • Antenna arrays (in the base station) to provide another form of diversity

(space diversity) Turbo codes may use serial (concatenated) and/or parallel recursive convolution codes. Recursive means the output not only depends on the input sequence but also depends on the previous output. A turbo encoder uses two convolutional encoders separated by an N-bit random interleaver or permuter, together with an optional puncturing mechanism. The interleaver is used to spread coded symbols from the other encoder or from the input. The input to an encoder can be from an input sequence and/or from a coded output from another encoder. Instead of cascading the encoders in the usual serial fashion, they can be arranged in parallel concatenation (see Figure 8.12).

Information Bits Encoder Constituent Encoder #1

Constituent Encoder #2

Interleaver

SoftDecision

Received Parity bits 1st Code Received Info bits

Parity Bits MUX Parity Bits

SoftDecision

De-interleaver

Soft1st Decision Constituent Decoder

Interleaver

SoftDecision

2nd Constituent Decoder

Final Output

Interleaver De-interleaver Received parity bits 2nd code

Figure 8.12

Decoder

Turbo encoding and decoding.

8.6

Channel Coding

243

The function of the permuter is to take each incoming block of N data bits and rearrange them in a pseudo-random fashion prior to encoding by the second encoder. Unlike the classical interleaver (e.g., block or convolution interleaver), which arranges the bits in some systematic manner, it is important that the permuter sort the bits in a manner that lacks any apparent order. It is also important that N be selected quite large (N  1000). The role of the turbo code puncture is identical to that of its convolutional code counterpart, to periodically delete selected bits to reduce coding overhead. Each decoding process yields a soft output decision for the next decoder. The key component is the soft input, soft output (SISO) decoder. To achieve the benefits of turbo code, several iterations are provided. Thus, the turbo codes are processing intensive and are applied to less delay-sensitive applications such as data. Turbo codes have been implemented in both software and hardware as a single integrated circuit. Turbo codes don’t suffer from the error at low BERs that have been attributed to other codes. Turbo codes are capable of providing high coding gain, even with high code rates. Turbo codes provide about 1.5 to 3 dB bit energy-to-noise improvement over the current R-S and RSV error correction codes. A 1.5 to 3.0 dB performance gain allows the system developer to reduce transmitter power and/or bandwidth and still maintain the same BER performance. Conversely, a 1.5 to 3 dB improvement can be used to improve overall system performance if transmitter power is not reduced. The use of turbo codes can allow the system engineer to reduce antenna size, lowering system cost. When used to provide improved BER performance, turbo codes can also lead to a clearer image transmission, improved audio, greater range, or better data integrity. A turbo code is applicable where digital data is transmitted over a noisy channel because it spreads error uniformly over the interleaved duration. Turbo codes are better than convolutional codes due to their strength in nonlinear coding/decoding and feedback property. Turbo codes can support the data rates achieved with other error correction codes while offering improved correction capability. Figure 8.13 provides a comparison of a turbo code having constraint length K  4 decoded with 4 iterations with a convolutional code of constraint length K  9 decoded with a Viterbi decoder. Figure 8.13 shows the performance of a rate 1/3 turbo code compared to the corresponding rate 1/3 convolutional code. The results indicate that the turbo code out-performs the corresponding convolutional code with the same decoding complexity for data rates larger than 9.6 kbps; the improvement in performance increases with the data rate for a fixed frame length of 20 ms. The performance improvement is due to the fact that the number of bits in the 20 ms frames increases with the data rate and the performance of a turbo code improves with the number of bits in the frame. With a large number of bits in a frame, the interleaver separating the codes can randomize the errors more effectively.

244

8

Speech Coding and Channel Coding

101 Turbo Code (9.6 kbps) Bit error rate (BER)

102 Convolutional code (9.6 kbps) 103 Turbo Code 76.8 kbps

104 105

Turbo Code 460.8 kbps

106 0

0.5

1.0

1.5 2.0 Eb/N0 (dB)

2.5

3.0

3.5

Turbo K = 4,4 iterations, 460.8 kb/s Turbo K = 4,4 iterations, 76.8 kb/s Turbo K = 4,4 iterations, 9.6 kb/s Convolutional K = 9

Figure 8.13

Performance of turbo code compared with convolutional code.

8.6.4 Soft and Hard Decision Decoding The decoding of a block code can be performed with hard or soft decision input, and the decoder may output hard or soft decision data. In hard decision decoding, each received channel bit is assigned a value of 1 or 0 at the demodulator depending whether the received noisy data is higher or lower than a threshold. The decoder then uses the redundancy added by the encoder to determine if there are errors, and, if possible, to correct the errors. The desired output of the decoder is a corrected code word. A soft decision decoder receives not only the binary value of 1 or 0, but also a confidence value associated with the given bit. If the demodulator is certain that the bit is a 1, it places very high confidence on it. If it is less certain, it places a lower confidence value. A soft input decoder can output either hard decision data or soft decision data. For example, a Viterbi decoder receives soft information from the demodulator and outputs hard decision data. The decoder can use the soft information to determine if a given bit is a solid 1 or solid 0, and it outputs this hard decision. A SISO decoder both receives soft decision data and generates soft decision output. For each bit in the code word, the SISO decoder examines the confidence of the other bits in the code word, and, using the redundancy of the code, generates an updated soft output for the given bit.

8.6

Channel Coding

245

RS error correction codes are the “standard” algorithm for FEC. RS codes are block codes that are very efficient for error correction implemented either in hardware or software. RS codes are hard decision codes. RS codes followed with concatenated Viterbi codes (RSV) offer an improvement over the stand-alone RS codes in terms of BER performance. The concept of SISO decoders has been applied to turbo codes. A turbo code feeds demodulated soft decision data into a SISO decoder. The output of this decoder is then fed into the same (or a different) SISO decoder. The output of this decoder is then fed again. This iterative process continues until a confident solution is reached. The concept of feeding the output back into the input is similar to a turbo charger of an engine, therefore the name turbo code is used. For a turbo code to be effective, the given data must be encoded with two (or more) different codes. Then, when decoding, each of the codes will modify the confidence of each bit. With each iteration, all of the codes modify the confidence of the data, so each code sees slightly different data for each iteration. Each code pushes the confidence of a given bit higher or lower, and consequently changes the hard decision value of the bits in error. Eventually, the data will settle on an arrangement in which all codes are pushing the confidence of all bits higher. The hard decision values at this point are closer to the transmitted data.

8.6.5 Bit-Interleaving and De-Interleaving Transmission errors occur randomly in bursts. Burst errors happen when the signal undergoes deep fades. The ability of FEC is limited in correcting a long string of errors, bit-interleaving is used between the encoder and modulator. At the receiver, the de-interleaver is employed between the demodulator and FEC decoder. The convolutional codecs perform poorly on bursts of errors. Bit-interleaving is used to randomize the errors so that the convolutional codec can correct them. The purpose of bit-interleaving is to avoid loss of consecutive information bits. Interleaving is performed by storing the data in a table containing rows and columns at the transmitter. The data is written in rows and transmitted in a vertical direction (according to columns). At the receiver, the data is written and read in the opposite manner. GSM blocks of full-rate speech are interleaved over 8 bursts; i.e., 456 bits of one block are divided into 8 sub-blocks, each containing 57 bits. Each subblock is carried by a different burst and in a different TDMA frame. Thus a burst contains contributions from two successive speech blocks, j and j  1. To avoid proximity relations between successive bits, bits from block j use even positions in the burst and bits from block j  1 use odd positions. This ensures sufficient redundancy within the interleaving process of the GSM signal structures to allow for one frame in the five to be lost without significant loss in voice quality. De-interleaving in the receiver is the reverse operation of interleaving. The major drawback of de-interleaving and interleaving is that they introduce delay.

246

8

Speech Coding and Channel Coding

The delay amounts to the transmission time from the first burst to the last burst in the block and is equal to 8 TDMA frames, i.e., about 37 ms. Example 8.3 Using a bit string of 0 0 0 0 0 1 1 1 0 0 0 1 0 0 0 1 (first bit) demonstrate operations of 4  4 bit interleaver/de-interleaver in converting a burst error into bit errors. Solution Input to the interleaver:



1 1 0000011100010001→ 1 0

0 0 1 0

0 0 1 0



0 0 0 write the data in rows 0

Output of the interleaver: 0 0 0 0 0 1 0 0 0 1 0 0 0 1 1 1



1 0 Input to the de-interleaver: 0 0 0 0 0 1 0 0 0 1 0 0 0 1 1 1 → 0 0

1 0 0 0

1 1 1 0

0 0 0 0



Burst error is indicated by italic bold-face (under-score). The output of the de-interleaver: 0 0 0 0 0 1 1 1 0 0 0 1 0 0 0 1

8.7

Summary

In this chapter, we presented speech codec attributes and briefly discussed different coding schemes. We provided details of the enhanced variable rate codec (EVRC) which takes advantage of a signal processing hardware and software technique to provide 13 kbps voice quality at 8 kbps data rate by maximizing both quality and system capacity. MOS data indicates that 8 kbps EVRC compares favorably to 16 kbps LD-CELP and ADPCM, the industry standards for comparison. The tests also show that EVRC maintains superior quality over 13 kbps CELP as frame error rates increase. We focused on three types of channel coding schemes: RS codes, convolutional codes, and turbo codes. The error correcting capability of a convolutional coding scheme increases as rate decreases and constraint length increases. However, the number of code states and, consequently, decoder complexity, grows with the constraint length. Turbo codes use serial and/or parallel recursive convolutional codes. They appear to approach the Shannon limit with large enough

References

247

iterations of decoding. Turbo codes spread error uniformly over the interleaved duration and, therefore, they are very useful for fading channels. Turbo codes provide an additional coding gain of 1 to 2 dB over convolutional codes.

Problems 8.1 Define speech coding methods used in digital communication. 8.2 What are the main features of a hybrid coding scheme? 8.3 List the main attributes of a speech codec. 8.4 What is a CELP codec? 8.5 List steps used in speech processing in a vocoder. 8.6 What is an AMR codec? 8.7 How is capacity improvement in a GSM system achieved with the AMR codec compared to the FER codec? 8.8 What are the information rates in EVRC codec for rate set 1 in the CDMA IS-95 B? 8.9 Discuss various forward error correcting schemes used in digital communication. Why are convolutional schemes used in wireless communications? 8.10 Calculate the output of the convolution encoder shown in Figure 8.8, if the input is 1 0 1 0 0 1 1 (first bit). 8.11 Why is bit-interleaving used in wireless communications? 8.12 Using a bit string of 1 0 1 1 0 0 0 0 0 1 1 1 0 0 0 1 0 0 0 1 0 1 0 1 0 demonstrate operations of 5  5 bit interleaver/de-interleaver in converting burst error into bit errors.

References 1. Atal, B. S. Predictive Coding of Speech Signals at Low Bit Rates. IEEE Trans., Comm. vol. 30, no. 4, 1982, pp. 600–614. 2. Atal, B. S., and Schroeder, M. R. Stochastic Coding of Speech at Very Low Bit Rate. Proc. International Conf. Comm. Amsterdam, 1984, pp. 1610–1613. 3. Bahl, L. R., Cocke, J., Jelinek, F., and Raviv, J. Optimal Decoding of Linear Codes for Minimizing Symbol Error Rate. IEEE Trans. Info. Theory. vol. IT-20, no. 2. 4. Chen, J., Cox, R., Lin, Y., Jayant, N., and Melchner, M. Coder for the CCITT 16 kb/s Speech Coder Standard. IEEE Journal of Selected Areas of Communications 6, 1988, pp. 353–363.

248

8

Speech Coding and Channel Coding

5. Furuskar, A., et al. System Performance of EDGE, a Proposal for Enhanced Data Rates in Existing Digital Cellular System. IEEE VTC 98, pp. 1284–1289. 6. Garg, V. K., and Wilkes, J. E. Principles & Applications of GSM. Upper Saddle River, NJ: Prentice Hall, 1999. 7. Hess, W. Pitch Determination of Speech Signals. Berlin: Springer Verlag, 1983. 8. Jarvinen, K., et al. GSM Enhanced Full Rate Speech Codec. IEEE GLOBECOM ’97. 9. Kleijn, W. B., Ramachandran, R. P., and Kroon, P. Generalized Analysis-by-Synthesis Coding and Its Application to Pitch Prediction. International Conf., Acoust. Speech Signal Processing. San Francisco, 1992, pp. 1337–1340. 10. Kroon, P., and Deprettere, E. F. A Class of Analysis-by-Synthesis Prediction Coders for High Quality Speech Coding at Rates Between 4.8 and 16 kbps. IEEE Journal of Selected Areas of Communications 6, 1988, pp. 353–363. 11. TIA IS-96A, “Speech Service Option Standard for Wideband Spread Spectrum Digital Cellular System.” 12. TIA IS-733, “13 kbps Speech Coder.” 13. TIA IS-127, “Enhanced Variable Rate Codec (EVRC) 8.5 kbps Speech Coder.” 14. TR45.5/98.04.03.03. “The cdma2000 ITU-R RTT Candidate Submission.” April, 1998. 15. Van Nobelen, R., et al. An Adaptive Radio Link Protocol with Enhanced Data Rates for GSM Evolution. IEEE Personal Communications. vol. 6, no. 1, February. 1999, pp. 54–64.

CHAPTER 9 Modulation Schemes 9.1

Introduction

The digital signals that are developed in the process of transmitting voice, data, video, and signaling information are generated at low data rates. These data rates, typically 1 to 50 kbps, are so low in frequency that their transmissions directly from the transmitter to the receiver would require antennas that are thousands of meters long. Furthermore, the signals from one transmitter would interfere with signals from other transmitters if they all use the same frequency band. Therefore, baseband signals (see Chapter 4) are modulated onto a radio frequency carrier for transmission from the transmitter to the receiver [2,3]. The radio environment at 8002000 MHz (used for mobile communications) is hostile. We must therefore select modulation schemes that are robust. In addition to the modulation schemes we must also choose encoding algorithms that improve the performance of the system [9,10]. We introduced modulation in Chapter 4. In this chapter we focus primarily on three modulation schemes, minimum shift keying (MSK), Gaussian minimum shift keying (GMSK), and /4-Differential Quadrature Phase Shift Keying (/4-DQPSK) that have been used in different cellular systems [8]. GMSK is the modulation scheme used in GSM900, GSM1800, GSM1900, and digital enhanced cordless telecommunications (DECT). MSK was introduced as a first step toward GMSK. Personal wireless telecommunications (PWT) and PWT-E (enhanced), the variations of DECT for the licensed and unlicensed 1900-MHz band, respectively, in North America, use /4-DQPSK. Since each of these schemes is descended from phase shift keying (PSK), we first discuss PSK and then show its relationship to others. We also examine quadrature amplitude modulation (QAM) and M-ary frequency shift keying (MFSK) schemes [11,13].

9.2

Introduction to Modulation

When we want to send signals over any distance, baseband signaling is not sufficient [6]. We must therefore modulate the signals onto an RF carrier. When we transmit the digital bit stream, we convert the bit stream into the analog signal, a(t) cos  t   . This signal has amplitude a(t), frequency   2 , and phase . We can change any of these three characteristics to formulate the modulation

249

250

9

Modulation Schemes

scheme. The basic forms of the three modulation methods used for transmitting digital signals are [11]: • Amplitude shift keying (ASK) • Frequency shift keying (FSK) • Phase shift keying (PSK)

When  and  remain unchanged, we have ASK. In this case, the transmitted carrier wave takes two amplitude values during the duration of the pulse. When a(t) and  remain unchanged, we have binary (or M-ary) FSK. In this case the carrier frequency is shifted up or down by a fixed value corresponding to a binary 1 or 0. When a(t) and  remain unchanged, we have binary (or M-ary) PSK. In binary PSK, the information is contained in the phase. During the transmission of a binary 1 the carrier phase is zero, and the carrier phase is changed to a value of  during the transmission of a binary 0. Hybrid schemes exist where two characteristics are changed with each symbol transmitted. The most common method is to fix  and change a(t) and . This method is known as quadrature amplitude modulation (QAM). Each modulation method results in a different transmitter and receiver design, different occupied bandwidth, and different error rate [4]. Since all signals have a theoretical bandwidth that is infinite, all modulation schemes must be band limited. Band limiting introduces detection errors, and the filter bandwidths must be selected to achieve an optimization between bandwidth and error rate [12]. The baseband outputs of the data transmitters are a series of binary data that cannot be sent directly over a radio link. The communications designer must choose radio signals that represent the binary data and allow the receiver to decode the data with minimal errors. For the simplest binary signaling system, we choose two signals denoted by s0(t) and s1(t) to represent the binary values of 0 and 1, respectively. Since no channel is perfect, the receiver will also have additive white Gaussian noise (AWGN), n(t). The data receiver (Figure 9.1) will then process the signal and noise through a filter, h(t), and, at the end of the signaling interval, T, make a determination of whether the transmitter sent a 0 or a 1.

S0(t) or S1(t )



Linear Filter h(t) v(t)

V

Threshold Comparison A Decision

AWGN n(t) AWGN: Additive white Gaussian noise

Figure 9.1

Receiver structure to detect binary signal with AWGN.

9.2

Introduction to Modulation

251

The energies of s0(t) and s1(t) in a T interval are assumed to be finite and denoted as E0 and E1, respectively. For simplicity we assume that the noise has a probability density function of amplitude that is Gaussian and that the noise spectral density is flat with frequency with the double-sided Power Spectral Density of N0/2, where N0 is the noise density. When s0(t) is present at the filter as an input, its output at time t T is V S0  N, with s0(t) present

(9.1)

where: S0 output signal component at t T for the input s0(t), and N the output noise component Similarly, when s1(t) is present at the filter as an input, its output at t T is V S1  N, with s1(t) present

(9.2)

Since noise is Gaussian with zero mean (implied by its constant PSD), N is also Gaussian. The variance of the noise, 2, can be determined as:

2

   H f  



(9.3)

 H f  2 df

(9.4a)

2





2 N0



0

N

0 df 2

2 N0 BN

(9.4b)

where: H(f ) is the transfer function of filter, and BN





0

 H f  2 df is the noise-equivalent bandwidth or simply the noise bandwidth

of the receiver filter function H(f) Given that s0(t) is present at the receiver input, the probability density function (PDF) of v is



   S0  2  2

1 p    s0  e

2



2

(9.5)

Similarly, when s1(t) is present at the receiver input, the conditional probability function of v is



 1 p    s0  e

2

  S1  2

2

2



(9.6)

252

9

Modulation Schemes

From Figure 9.2, given that s0(t) is present, the probability of error is P  e  s0 

 p    s  d

(9.7)

P  e  s1 



(9.8)



0

A

If s1(t) is present, it is A



p    s1  d

If the a priori probability that s0(t) was sent is p, and the a priori probability that s1(t) was sent is q 1  p, then the average probability of error is Pe pP e  s0   qP e  s1 



pe p





   S0  2  2

1

e A

2

2

(9.9)

  d  q A 1 e

 2

  S1  2

2



2

d

(9.10)

If we simplify Equation 9.10, differentiate the result with respect to A, and then set the derivative equal to zero, we can determine the optimum choice for the threshold value, A, to minimize the probability of error Pe [13]: S S

p

0 1

A Aopt ln q 2

S1  S0

(9.11)

2

In most systems, the values of 0 and 1 are equally likely; if they are not, then the designer usually redesigns the encoding method to ensure that they are equally likely. Thus, p q, and S S

0 1 A Aopt

(9.12)

2

For the optimum value of A, with p q, the probability of error is

2

S S

2



S S

0 1 0 1 erfc 1 Pe Q

2



2



(9.13)

where erfc is the complementary error function 1  erf(u) 2Q 2u . The error function, erf(u) is defined as

9.2

Introduction to Modulation

253

p(v s0)

p(v s1)

P(e s0)

P(e s0)

v S0 Figure 9.2

A

S1

Conditional probability density function of the filter output at time t ⴝ T.

e 

2 erf(u)



u

0

t2dt

(9.14)

and

e 2

1 Q(u)





u

x2 2 dx



(9.15)

or u2

e 2 Q(u) , u  1 Gaussian integral

2 u

(9.16)

Next we want to find the filter that provides the minimum probability of error, as expressed by Equation 9.13. At time t0, the sample value consists of a signal-related component g0(t0) and noise component n0(t0). This filter is known as a matched filter and has a transfer function H0(f) optimized to provide the maximum signal-to-noise ratio (SNR) at its output at time t0. Schwartz shows that this filter must be the conjugate match of the signal, s(t). Since we have two signals, s0(t) and s1(t), we will need two filters in our receiver design. If we transmit a signal, s(t), then it has the Fourier transform, S(), which is a complex function. The optimum or matched filter must have a frequency response, H(), where

254

9

Modulation Schemes

H() S*()ejt0

(9.17)

where: S* is the complex conjugate of the Fourier transform of the signal In general this filter is not realizable, since analysis will show that it must have output before there is input. However, we can design filters that approximate the ideal filter. The SNR is defined as 2

g 0(t0) 2

(9.18)

2

It can also be shown that the maximum value for SNR 2 is twice the energy of the input signal (Eg) divided by the single-sided input noise spectral density, regardless of the input signal shape Eg (N0)/2

2max

(9.19)

For a binary system, Equation 9.19 becomes 1 2max



T

N0 0

[s1(t)  s0(t)]2 dt

(9.20)

Since the signals are zero outside the range (0, T), the probability of error corresponding to the optimum receiver filter becomes

1 erfc(

Pe z ) Q( 2z ) 2

(9.21)

where: 1 z



T

4N0 0

[s1(t)  s0(t)]2dt

If the transmitted pulses are allowed to take on any of M transmitted levels with equal probability, then the information rate per transmitted pulse is log2M bits. For a constant information rate, the bandwidth of the transmitted signal can be reduced by the same factor. With M-ary transmission, we will show that the error rates are higher, but if we have sufficient SNR then higher error rates will not matter. Thus, we are using excess SNR to code the signal and reduce its bandwidth.

9.2

Introduction to Modulation

255

When we add additional levels to a baseband system, we are reducing the distance between detection levels in the receiver output. Thus, the error rate of a multilevel baseband system can be determined by calculating the appropriate reduction in error distance. If the maximum amplitude is A, the error distance de between equally spaced levels at the detector is A de

(9.22)

M1

where: M number of levels Setting the error distance A of a binary system to that defined in Equation 9.22 provides the error probability of a multilevel system. 1 M  1 erfc A Pe



log2M

 (M  1)(

M

2 )



(9.23)

where: The factor (M  1)/M reflects that the interior signal levels are vulnerable to both positive and negative noise. The factor 1/log2M arises because the multilevel system is assumed to be coded, so symbol errors produce single bit errors (log2M is the number of bits per symbol), and the probability of multiple bit error is assumed to be small and can be neglected. Equation 9.23 relates error probability to peak signal power A2. To determine Pe with respect to average power, the average power of an M-level system is determined by averaging the power associated with various pulse amplitude levels.



A 2 [A2]avg

M M1



2

M  1

3A 

2

2A [A2]avg 2

2

 …  A2



(9.24)

M/2

(2j  1)2

M(M  1) j 1

(9.25)

where: A the levels  1, 3, 5, …,  M  1  are assumed to be equally M1 likely. If T is the signaling interval for a two-level system, the signaling interval TM for an M-level system providing the same data rate is determined as TM T log2M

(9.26)

256

9

Modulation Schemes

1 . For a raised cosine filter, the noise bandwidth is BN 2TM

From Equation 9.4, we get N

0

2

(9.27)

2TM

2

1/2

T N

0 1





(9.28)

M

Substituting Equation 9.28 into Equation 9.23, we get 1 M1 A Pe  erfc

1/2

 log M 2

M



M

T  N

 1  0 M



(9.29)

The energy per symbol Es Eb log2M A2TM

where: Eb is the bit energy E log M

2 b  A2

(9.30)

TM

Substituting for A from Equation 9.30 into Equation 9.29, we get M  1 erfc   N   log M  M Eb

1 Pe

1/2

 log2 M  1/2 M1

0

2

Eb log2 M   1/TM  Signal power SNR



Noise power

2 T



(9.31)

(9.32)

M N0

N  Eb

 SNR 2 log2 M

(9.33)

0

Another variation of baseband signaling is antipodal baseband signaling (APBBS) in which two signals of opposite polarities are sent. If s0(t) A and s1(t) A for 0  t  T, then s1 t   s0 t  2A.

9.3

Phase Shift Keying

257

We then calculate the value of z from Equation 9.21 as 1 z

 (2A) T

2

4N0 0

2

E

N0

N0

A T b dt

(9.34)

where Eb is the energy in either s0(t) or s1(t), or the bit energy 1 erfc Pe 2

b

b

E 2E Q 

N N 0

(9.35)

0

APBBS is used to modulate some signal; we will compare its SNR with other modulation schemes.

9.3

Phase Shift Keying

For binary phase shift keying (BPSK) [5,7], we transmit two different signals. If the baseband signal is a binary 0, we transmit: A cos(t  ) A cos (t)

(9.36)

and for binary 1, we transmit: (9.37)

A cos (t)

BPSK can be considered a form of ASK where each non-return to zero (NRZ) data bit of value 0 is mapped into a 1 and each NRZ 1 is mapped into a  1. The resulting signal is passed through a filter to limit its bandwidth and then multiplied by the carrier signal cost (see Figure 9.3). The signal constellation for BPSK is shown in Figure 9.4.

cos t

NRZ Data

Figure 9.3

Mapping 1 to 1 0 to 1

BPSK modulator.

Pulse Shaping/ Filter



BPSK Signal

258

9

Modulation Schemes

cos t v

sin t

v

Figure 9.4

Signal constellation for BPSK.

We can also define PSK where there are M phases rather than two phases. In M-ary PSK, every b (where M 2b) bits of the binary bit stream are coded as a signal that is transmitted as A sin (t  j), j 1, . . . , M. The error distance of a PSK system with M phases is A sin(/M) where A is the signal amplitude at the detector. A detector error occurs if noise of the proper polarity is present at the output of either of the two phase detectors. The probability of error is





  E 1/2

 1 Pe erfc sin  log2M 1/2 b M  N log M



2

(9.38)

0

The SNR is given as:

N  E

SNR log2 M b

(for M  2)

0

(9.39)

Each symbol has a length of T. Therefore 2

A T Eb log2M

(9.40)

9.3

Phase Shift Keying

259

The rms noise is: N0 2  2T 

(9.41)

for noise in Nyquist bandwidth. R log2M where R The bandwidth efficiency of the M-ary PSK is given as

Bw

2

is the data rate and Bw is the bandwidth. Next we examine several variations of PSK. Example 9.1 Find Eb/N0 in dB to provide Pe 106 for BPSK and coherent FSK. Solution For BPSK Eb/N0

e Pe

N  E

2  b 0



e 106

2 

where:  Eb/N0   11.32 10.54 dB

FSK requires 3 dB more in terms of Eb/N0 to give the same Pe as BPSK, i.e., 13.54 dB. Example 9.2 The BPSK modulation is used in a channel that adds white noise with single-sided PSD N0 1010 W/HZ. Calculate the amplitude A of the carrier signal to give Pe 106 for a data rate of 100 kbps. Solution From Example 9.1, Eb/N0 10.54 dB 11.32 for Pe 106 Eb N0

A2T 2N0

A2 2N0R



260

9

Modulation Schemes

Quadrature

  

1/2

E

 A 2N0 b R N0

A  2  1010  11.32  105 1/2 1.505 mv

9.3.1

Quadrature Phase Shift Keying (QPSK), Offset-Quadrature Phase Shift Keying (OQPSK) and M-PSK Modulation [5,7,11] For QPSK, M 4, we substitute M 4 in Equation 9.38 to get



N E 1/2

 log 4 1/2 b 1 Pe erfc sin   2 

 log 4  2

1 erfc Pe 2

4

0

2Eb

Eb

Q

N N 0

(9.42)

0

The input binary bit stream  bk , bk 1; k 0, 1, 2, . . . , arrives at the modulator at a rate 1/Tb bps and is separated into two data streams, aI(t) and aQ(t), containing odd and even bits respectively (see Figure 9.5). The modulated QPSK signal s(t) is given as: 1 a  t  cos 2f t    1 a (t) sin 2f t   s t  I

Q  c   c 

(9.43)

  (t) s t  A cos  2fct  

(9.44)

4

2

4

2

4

cos t

odd bits, with durations 2Tb

I



NRZ Data



Serial to Parallel even bits, with durations 2Tb

Q

 sin t

Figure 9.5

QPSK modulator.

QPSK Signal

9.3

Phase Shift Keying

261

where:

A (1/2)(a2I  a2Q) 1 aQ(t)

(t) a tan aI(t)

In the QPSK, the odd and even pulse streams are transmitted at the rate of (1/2)Tb bps and are synchronously aligned. In QPSK, due to the coincident alignment of aI(t) and aQ(t), the carrier phase can change only once every 2Tb. The carrier phase during any 2Tb interval can be any one of the four phases as shown in Figure 9.6, depending on the values of aI(t) and aQ(t) during that interval. During the next interval, if neither pulse stream changes sign, the carrier phase remains the same. If only one of the pulse streams changes sign, a phase shift of /2 occurs. A change in both streams results in a carrier phase shift of . The 180º phase shifts cause the signal envelope to go through zero momentarily (see Figure 9.7). In offset QPSK (OQPSK), the timing of the pulse stream aI(t) and aQ(t) is shifted such that the alignment of the two pulse streams is offset by Tb. In OQPSK, the pulse stream aI(t) and aQ(t) are staggered and thus do not change states simultaneously. The possibility of the carrier changing phase by  degree is

co

s

2

fc t

cos(2fct  4)

(1, 1)

(1, 1) (aQ, aI)

 4 sin(2fct  4)

(1, 1)

(1, 1) n

si f ct

2

Figure 9.6 ␲Ⲑ4-QPSK constellation.

262

9

Modulation Schemes

180 degree phase shift causes a drop in signal level and therefore causes data errors.

Figure 9.7

Signal amplitude of QPSK during 180° phase change. b2 1 b3 1

b0 1 b1 1

b4 1 b5 1

b6 1 b7 1 t

QPSK

O

4T

2T

b2 1

b0 1 b1 1

6T

b4 1 b3 1

8T

b6 1 b5 1

b7 1 t

OQPSK

O Figure 9.8

T

2T

3T

4T

5T

6T

7T

A comparison of QPSK and OQPSK modulation schemes.

eliminated, since only one component can make a transition at one time. Changes are limited to 0º and /2 every Tb seconds (see Figure 9.8). Since the phase transitions of 180º are avoided in OQPSK, the signal envelope does not pass through zero as it does in QPSK. The high-frequency components associated with the collapse of the signal envelope are not reinforced. Thus, out-of-band interference is avoided. In theory, QPSK and OQPSK systems can improve the spectral efficiency of mobile communications. They do, however, require a coherent detector. Coherent detection requires an accurate carrier phase reference. In general, known pilot symbols need to be inserted periodically in the data stream to obtain acceptable performance with a coherent detector on a fading channel. These symbols increase both the bandwidth and the required SNR. In a multipath fading environment,

9.3

Phase Shift Keying

263

the use of a coherent detector is difficult and often results in poor performance over noncoherently based systems. The coherent detection problem can be overcome by using a differential detector, but then QPSK is subject to intersymbol interference which results in poor system performance. The spectrum of OQPSK is given as:



sinf T fcTb

c b psd(f ) Tb



(9.45)

Further improvement is possible if the OQPSK format is modified to avoid discontinuous phase transitions. This was the motivation in designing continuous phase modulation (CPM) schemes. Minimum shift keying (MSK) is one such scheme. A comparison of PSD for QPSK and MSK is shown in Figure 9.9. Since we often encounter Rayleigh fading (see Chapter 3) in the outdoor environment, the BER performance of PSK under flat Rayleigh fading conditions is given by



1 1 1 Pe

2

1  1/(Eb/N0)avg



(9.46)

1.0

Fraction of out-of-band power (GSM) X

MSK 0.20

X Fraction of out-of-band power (IS-54)

X

QPSK

X

0.05

Spectrum

X Fraction

X

QPSK/OQPSK MSK

0.5/ Tb

1/ Tb Frequency

Figure 9.9

Comparison of PSD and QPSK and MSK.

264

9

Modulation Schemes

The symbol error probability for M-PSK under AWGN conditions is given as



N 2bEb

Ps 2Q sin(/M)

(9.47)

0

where: b number of bits per symbol Using Gray code, a symbol error is likely to result in only one out of b bit errors.



2 Q sin(/M)  Pb b

N 2bEb

(9.48)

0

For 8-PSK, M 8 and b 3:



2 Q sin  Pb

3 8



6Eb







Eb 2

Q 0.937 N0 N0 3

(9.49)

For 16-PSK, M 16 and b 4:



Q  · Pb sin 2 16



N 8Eb

0







Eb Q 0.552 N0 2

(9.50)

In M-PSK modulation, the amplitude of the transmitted signal remains constant, thereby yielding a circular constellation.

9.3.2 ␲/4-DQPSK Modulation We can design a PSK system to be inherently differential and thus solve detection problems. /4-DQPSK [11] is a compromise modulation method because the phase is restricted to fluctuate between /4 and (3)/4 rather than the 2 phase changes for OQPSK. It has spectral efficiency about 20% higher than the GMSK modulation used for GSM and DECT. /4-DQPSK is essentially a /4-QPSK with differential encoding of symbol phases. The differential encoding mitigates loss of data due to phase slips. However, differential encoding results in the loss of a pair of symbols when channel errors occur. This can be translated to an approximate 3-dB loss in Eb/N0 relative to coherent /4-QPSK.

9.3

Phase Shift Keying

265

90 /2

135

45

3/4

/4

180 

0

7/4 225

315

5/4

270 3/2 Figure 9.10 ␲/4-DQPSK modulation.

A /4-DQPSK signal constellation (Figure 9.10) consists of symbols corresponding to eight phases. Consider that these eight phase points are formed by superimposing two QPSK signal constellations, offset by 45° relative to each other. During each symbol period, a phase angle from only one of the two QPSK constellations is transmitted. The two constellations are used alternately to transmit every pair of bits (di-bits). Thus, successive symbols have a relative phase difference that is one of the four phases as shown in Table 9.1. Figure 9.10 shows the /4-DQPSK signal constellation. When the phase angle of /4-QPSK symbols are differentially encoded, the resulting modulation is /4-DQPSK. This can be done either by differentially encoding the source bits

266

9

Modulation Schemes

Table 9.1 Phase transitions of ␲/4-DQPSK.

Symbol

␲/4-DQPSK phase transition 45°

00 01

135°

10

ⴚ45°

11

ⴚ135°

Even Bits

Pulse Shaper/ Filter

I

Delay of Tb

Pulse Shaper/ Filter

Q

NRZ Data Serial to Parallel

Odd Bits

Note: Odd bits are delay by Tb

Figure 9.11

OQPSK encoding.

and mapping them onto absolute phase angles or, alternately, by directly mapping the pairs of input bits onto relative phase angles [4, (3)/4] as shown in Figure 9.10. The binary stream bM(t) entering the modulator is converted by a serial to parallel converter into two binary streams b0(t) and be(t). The bits are differentially encoded (see Figure 9.11). The in-phase and quadrature components corresponding to the kth symbol are given as: Ik Ik1 cos(k)  Qk1 sin(k)

(9.51)

Qk Ik1 sin(k)  Qk1 cos(k)

(9.52)

where: Ik and Qk are the in-phase and quadrature components of the /4-DQPSK signal corresponding to the kth symbol, and the amplitudes of Ik and Qk are

1, 0, 1 2

9.3

Phase Shift Keying

267

Since the absolute phase of (k1)th symbol is k1, the in-phase and quadrature components can be expressed as: Ik cosk1 cos(k)  sink1 sin(k) cos(k1  k)

(9.53)

Qk cosk1 sin(k)  sink1 cos(k) sin(k1  k)

(9.54)

These component signals (Ik, Qk) are then passed through a baseband filter having a raised cosine frequency response as:



1



 H f  



Ts 1 1

1  sin  f  2 2Ts

 

 

0

1– 0  f  2Ts

1– 1

 f  2Ts 2Ts

(9.55)

1 f  2Ts

where:  the roll-off factor Ts the symbol duration If g(t) is the response to pulse Ik and Qk at filter input, then the resultant transmitted signal is given as: s t 

g t  kT   cos  s

k

cos t  

k

s t 

g t  kT   sin  s

k

sin t 

(9.56)

k

g t  kT  cos  t    s

k

(9.57)

k

where: 2 the carrier frequency of the transmission The component k results from differential encoding (i.e., k k – 1  k ). Depending on detection method (coherent or differential detection), the error performance of /4-DQPSK can either be the same as or 3 dB worse than QPSK. Example 9.3 The bit stream is 001001110100111010010100 and the transmitted signal is Acos  t  k . Calculate k for the /4-DQPSK modulation method.

268

9

Modulation Schemes

Table 9.2 Phase angle for ␲⁄4-DQPSK. ⌬␾k(degree)

b0be

␾k (degree) 0

00

45

10

45

45 0

01

135

135

11

135

0

01

135

135

00

45

180

11

135

45

10

45

0

10

45

45

01

135

90

01

135

225

00

45

270

Solution See Table 9.2.

9.3.3 MSK and GMSK Modulation The MSK is the constant envelope modulation [5]. It is derived from OQPSK by replacing the rectangular pulse in amplitude with a half-cycle sinusoidal pulse. Let us consider a data stream {ak}, k 0,1,2, where ak 1 at a rate of R 1/Tb, and Tb is the bit duration. The in-phase and quadrature bit streams are: aI t  a0

a2 a4

...

aQ t  a1

a3 a5

...

and

The rate of aI(t) and aQ(t) is (1/2Tb) bit per second. The in-phase, aI(t) and quadrature aQ(t) signals are delayed by interval Tb from each other. The MSK signal is defined as:













(9.58)



(9.59)

 t  2nTb   t  2nTb  s t  aI t  cos cos 2fct  aQ t  sin sin  2fct  2Tb



2Tb

 t  2nTb  s t  cos 2fct  bk t   k 2Tb

9.3

Phase Shift Keying

269

where: n 0, 1, 3, … bk 1 for aI . aQ 1 bk 1 for aI . aQ 1 k 0 for aI 1 k  for aI 1 Note that, since the I and Q signals are delayed by a 1-bit interval, the cosine and sine pulse shapes in Equation 9.58 are actually both in the shape of a sine pulse. MSK has the following properties: • For a modulation bit rate of R, the high frequency, fH f  0.25R when bk 1,

and the low frequency, fL f – 0.25R when bk 1. • The difference between the high frequency and the low frequency is f 1 , where T is the bit interval of the NRZ signal. fH  fL 0.5R b 2Tb

• The signal has a constant envelope.

The error probability for an ideal MSK system is



1 erfc Pe 2

N

Eb

Q 0

 2Eb

N0

(9.60)

This is the same as for QPSK/OQPSK. MSK modulation makes the phase change linear and limited to /2 over a bit interval Tb. This enables MSK to provide a significant improvement over QPSK. Because of the effect of the linear phase change, the PSD has low side lobes that help to control adjacent-channel interference. However, the main lobe becomes wider than QPSK. Thus, it becomes difficult to satisfy the CCIR-recommended value of 60 dB side lobe power levels. The PSD for MSK is given as: 16T

 1  16f T cos2f T

b c b PSD  fc 

2 2 2



c

(9.61)

b

MSK has a narrower bandwidth than BPSK, and is significantly better in terms of side lobe level than any unfiltered linear scheme, but it still retains side lobes at a level that would give unacceptable adjacent channel interference in most practical systems. Any attempt to filter these side lobes out, post-modulation, would regenerate envelope variations. Hence, we require a mean to improve the spectrum of MSK further. This can be done by filtering the data signal before modulation. A wide range of filter responses is possible. The Gaussian filter, which leads to GMSK, is the most popular. The bandwidth (B) of the Gaussian filter is quantified in the time-bandwidth product, BTb. The values of 0.3 to 0.5 are used for BTb.

270

9

Modulation Schemes

The Gaussian response is interesting, in that the frequency response has the same shape as the impulse response. The standard deviation of the impulse response is related to the filter bandwidth B by log 2

0.2206 e B 

(9.62)

where the impulse response is 1 e t h t 

2 / 2 2 

2

The price of the resulting improvement in bandwidth efficiency is a degradation in power efficiency. However, if the time-bandwidth product is not too small (i.e., 0.3 or more), the effect is not excessive (less than 1 dB). In the case of MSK, BER performance for coherent detection is the same as that for deferentially encoded BPSK with coherent detection. BER for GMSK is degraded due to inter-symbol interference (ISI) by the promulgation Gaussian filter. The BER performance of GMSK with coherent detection under AWGN conditions is given by

Pe erfc

 2  Eb

N0

(9.63)

where  is a degradation factor due to premodulation filter,  1 corresponds to the performance for MSK. The BER under flat Rayleigh fading conditions is given as 1 Pe 1 





1 1

(9.64)

 Eb/N0 avg

where (Eb/N0)avg represents an average value Table 9.3 shows relationship between the bandwidth of the premodulation Gaussian filter normalized by bit rate (BTb), the 99.99% bandwidth normalized by a bit duration, and . Example 9.4 The GMSK modulation is used in the GSM system with a channel bandwidth of 200 kHz and a data rate of 270.8 kbps. Calculate (a) the frequency shift between binary 1 and binary 0, (b) the transmitted frequencies if the carrier frequency is 900 MHz, and (c) the bandwidth efficiency in bps/Hz.

9.3

Phase Shift Keying

271

Table 9.3 BTb versus ␤ for GMSK modulation. BTb

99.99% bandwidth



0.20

1.22

0.76

0.25

1.37

0.84

0.30

1.41

0.89

0.40

1.80

0.94

0.50

2.08

0.97

Solution (a) The frequency shift is f fH  fL 0.5R 0.5  270.83 kHz 135.415 kHz

(b) The maximum and minimum frequency are: fH fc  0.25R 900 MHz  67.707 kHz 900.0677 MHz,

and fL fc  0.25R 900 MHz  67.707 kHz 899.9323 MHz

(c) R 270.83 1.35 bps/Hz Bandwidth efficiency

Bw

200

Example 9.5 Determine the 3-dB bandwidth for a Gaussian low-pass filter that is used to generate 0.30 GMSK with a channel data rate of 270 kbps. What is the 99.99% power bandwidth in the RF channel? What is the bit error probability for GMSK if Eb/N0 6 dB? Solution 1 1 Tb

3 3.7 s Rb

270  10

BTb 0.3 0.3 0.3 B 81.08 kHz 6 Tb

3.7  10

272

9

Modulation Schemes

The 3-dB bandwidth is 81.08 kHz. To determine the 99.99% power bandwidth, we use Table 9.3 to find that 1.41 Rb is the required value. The occupied RF spectrum for a 99.99% power bandwidth will be RF bandwidth 1.41 Rb 1.41  270  103 380.7 kHz,  0.89. For Eb/N0 6 dB (3.9811), we get

Pe erfc

9.4

E erfc  2 N b

 2.662 

0

11.5  105

Quadrature Amplitude Modulation

By allowing the amplitude to vary with phase, a modulation scheme referred to quadrature amplitude modulation (QAM) is obtained [1]. Figure 9.12 shows a constellation diagram of 16-QAM. The constellation consists of a square lattice of signal points. The distance d between constellation points can be expressed as d a 2i  K  1   a 2j  K  1  i, j 0, 1, 2, . . . K1

K M

M 4, 16, 64, 256, . . .

Average symbol error probability under AWGN conditions is given as: a Ps nnQ



(9.65)

j

a i a

Figure 9.12

16-QAM constellation.

9.4

Quadrature Amplitude Modulation

where:

273

nn average number of immediate neighbors

2 N0/Ts A the average power

nn and A2 are given as:

2

4 K  2 2  12 K  2   8

n

n

M

and

K/2  1



4Ka2

2

p 0

 2p  1 2

A

M

For 16-QAM, M 16 and K 4: 4(4  2)  12(4  2)  8

n

3 2

n

16

1

2

 2p  1 2 p 0

16a2

16a2 19  16

A

10a2 16



10

A2 a

2 A

2 3Q

 10 

Ps 3Q



2





A Ts

3Q 10N0

2Es 10N0



Since Es 4Eb: 3 Pe Q 4





2Es 3

Q 10N0 4

 

Es 3

Q 5N0 4

  4Eb

5N0

(9.66)

The disadvantage of 16-QAM compared to 16-PSK is that the signal amplitude has an inherent variation, regardless of filtering. This may cause problems in nonlinear amplifiers and on fading channels. Example 9.6 A modulator transmits symbols at a rate of 19,200 symbols per second. Each symbol has 64 different possible states. What is the bit rate?

274

9

Modulation Schemes

Solution Number of bits per symbol: b log2 M log2 64 log2 26 6 Bit rate of the modulator 6  19,200 115.2 kbps Example 9.7 A data stream with data rate Rb 144 Mbps is transmitted on an RF channel with a bandwidth of 36 MHz. Assuming Nyquist filtering and Gaussian channel, determine the modulation scheme that should be used. If the probability of bit error is 3  105, find the required Eb/N0. Solution 144 4 bps/Hz Required spectral efficiency 36

Since the channel is bandwidth limited Rb

4 log2 M Bw

 M 24 16

16-QAM (see Equation 9.66) should be used as it is more efficient than 16-PSK (see Equation 9.50). For a rectangular constellation (see Figure 9.12), with a Gaussian channel and matched filter reception, the bit error probability is given as:

 3  10

3 Pb Q 4

4 Eb 5 N0

5



E

 b 19.5 12.9 dB N0

Example 9.8 Compare the performance of 16-PSK with 16-QAM for a BER probability of 108. Solution 16-PSK (Equation 9.50): 108

E



1 Q 0.552 2

 b 20 dB N0



 Eb

N0

9.5

M-ary Frequency Shift Keying

275

16-QAM (Equation 9.66):

  E

3 4 b 108 Q

5 N0

4

E

 b  16 dB N0

Thus, 16-QAM has an advantage of about 4 dB compared to 16-PSK.

9.5

M-ary Frequency Shift Keying

The PSK and various forms of QPSK do not have a continuous phase, but they provide good bandwidth efficiency to allow more users for a given channel bandwidth [11]. If there is an interest in using low-cost amplifiers that are essentially nonlinear, we need constant-envelope modulation schemes. The frequency shift keying (FSK) modulation is one such schemes. In binary FSK, the carrier frequency is increased or decreased by a fixed value corresponding to a binary 1 or 0. Thus, two separate signals are transmitted. The FSK signals can be detected coherently or noncoherently. The bit error probability Pe for a coherent binary FSK in the AWGN channel is given as:

1 erfc Pe 2

 Eb

(9.67)

2N0

This is 3 dB worse than the coherent BPSK. For noncoherent systems the Pe is: 1 eEb/ 2N0  Pe

(9.68)

2

As the frequency hopping spread spectrum (FHSS) uses orthogonal frequency division multiplexing (OFDM) with MFSK, we provide a brief description of MFSK. In MFSK, signal si(t) at frequency fi is given as

T

2E

si t  s cos  2fit  s

1iM

where: fi fc   2i  1  M  f in which fc carrier frequency f frequency step M number of different signal elements 2b b number of bits per signal element

(9.69)

276

9

Modulation Schemes

Es energy per symbol Eb log2 M Eb energy per bit Ts symbol duration bTb Tb bit duration Rb data rate Bw bandwidth N0 noise density • Total bandwidth required, Bw 2Mf • Minimum frequency separation required 2f 1/Ts

Average probability of symbol error Ps for the coherent orthogonal MFSK signal is given as:

Ps 

M



1 Q



Eb log2M

N0



M2

(9.70)

Average probability of symbol error Ps for the noncoherent orthogonal MFSK signal is given as:



M

 1 i  M  eE /  iN  i  i 1 M

1  eEs / N0  Ps

s

0

M2

(9.71)

where: M!

 Mi  i!  M  i ! The upper-bound of Ps for the orthogonal and noncoherent orthogonal MFSK signal is M  1 eEs /  2N0  Ps 

M2

2

(9.72)

The relationship between bit error probability Pe and Ps for the orthogonal MFSK signal set is given as Pe Ps

2k  1 2k  1

M/2  1



M2

M

P  P

1 lim e

k→

s

2

(9.73)

(9.74)

9.5

M-ary Frequency Shift Keying

277

The channel bandwidth efficiency of a coherent MFSK is Rb

2 log2 M (M  3)



Bw

M2

(9.75)

The channel bandwidth efficiency of a noncoherent MFSK is Rb

2 log2 M M



Bw

M2

(9.76)

The bandwidth efficiency of an MFSK decreases with increasing M, MFSK signals are bandwidth inefficient. However, since all the M signals are orthogonal there is no crowding in the signal space and hence the power efficiency increases with M. MFSK can use nonlinear amplifiers with no performance degradation. Example 9.9 If M 8, fc 250 kHz, f 25 kHz, what is the total bandwidth required? What is the bandwidth efficiency? What is the Eb/N0 required for symbol bit error probability 106 of a coherent MFSK? How many bits per symbol are carried? Solution Total bandwidth required 2Mf 2  8  25 kHz 400 kHz R

2 23 b b 0.5455

2 log M (M  3)

Bw

11

106  M  1  Q z  7Q z 

z 5.08

5.08

N Eb log2 M 0

E

 5.08 2

N0

log 2 8 

 b 9.346 dB

278

9

Low

Modulation Schemes

High

1 bit/symbol binary. Orthogonal Modulation

1 bit/symbol: BPSK

Spectral Efficiency (bit/sec/Hz)

Transmission Quality

2 bits/symbol: QPSK, /4-QPSK, OQPSK 2 bits/symbol: M-PSK, M-QAM High

Figure 9.13

9.6

Low

Modulation method selection.

Modulation Selection

Proper selection of a modulation scheme in mobile communications is needed. Several factors dictate the selection. In general a modulation method should have the following characteristics: • spectrally efficient (i.e., higher bps/Hz); • applicable to cellular/PCS in all environments (urban, suburban, rural); • easy to implement; • low adjacent-channel interference. • good BER performance

As shown in Figure 9.13, transmission quality and spectral efficiency of a modulation method are interrelated. Table 9.4 lists the modulation schemes and their spectral efficiencies used in different mobile systems.

9.7

Synchronization

The demodulation of a signal requires that the receiver be synchronized with the transmitter signal as it is received at the input of the receiver [11]. The synchronization must be for: • Carrier synchronization. The receiver is on the same frequency as the trans-

mitted signal, adjusted for the effects of Doppler shifts. • Bit synchronization. The receiver is aligned with the beginning and end of

each bit interval. • Word synchronization (also known as frame synchronization). The receiver

is aligned with the beginning and end of each word in the transmitted signal.

9.7

Synchronization

279

Table 9.4 Modulation methods in mobile communication.

System

Modulation scheme

Information rate (kbps)

Channel spacing (kHz)

Spectral efficiency(bps/Hz)

IS-136

4-DQPSK

48.6

30

1.62

PDC

4-DQPSK

42.0

25

1.68

GSM

GMSK

270.8

200

1.35

CT-2

GMSK

72

100

0.72

DECT

GMSK (BTb 0.5)

1572

1728

0.67

CDMA

QPSK

1228.8

1230

0.99

16-QAM

384

1600

4.17

UWC-136

Received Signal r(t)

Figure 9.14

Multiply By N

Divide by N

cos t (I )

Divide by N

sin t (Q)

Narrowband PhaseLocked Loop

Carrier recovery for PSK.

If the synchronization in the receiver is not precise for any of these operations, then the BER of the receiver will not be the same as described by equations in previous sections of this chapter. The design of a receiver is an area for which standards are traditionally not specified. It is an art that enables one manufacturer to offer better performance in its equipment compared to a competitor. The methods of achieving the synchronization discussed are the traditional methods. A particular receiver may or may not use any of these methods, and proprietary methods are often used by many manufacturers. For PSK, the carrier signal changes phase every bit interval (see Figure 9.14). If we multiply the received signal by an integer, N, we can convert all of the phase changes in the multiplied signal to multiples of 360º. The new signal then has no phase changes and we can recover it using a narrowband phase-locked loop (PLL). After the PLL recovers the multiplied carrier signal, it is divided by N to recover the carrier at the proper-frequency. By a suitable choice of digital dividing circuits, it is possible to get a precise 90º difference in the output of two dividers and thus generate both the cos  t and sin t signals needed by the receiver. There are also some down-converter integrated circuits that have a precise phase shift network contained within them. The carrier recovery is typically performed at

280

9

Modulation Schemes

some lower intermediate frequency rather than directly at the received frequency. For BPSK we would need an N of 2, but an N of 4 would be used to enable the sine and cosine terms to be generated. For QPSK and its derivatives, an N of 4 is necessary; for /4-DQPSK an N of 8 would be needed. After we recover the carrier, we must reestablish the carrier phase to determine the values of the received bits. Somewhere in the transmitted signal must be a known bit pattern that we can use to determine the carrier phase. The bit pattern can be alternating 0s and 1s that we use to determine bit timing or it could be some other known pattern. The advantage of differential keying (e.g., /4-DQPSK) is that it is not important to know the absolute carrier phase. Only the change in carrier phase from one symbol to the next is important. MSK is a form of frequency modulation; therefore, a different method of carrier recovery is required. In Figure 9.15, the MSK signal has frequency f and deviation f 1/2Tb. We first multiply the signal by 2, thus doubling the deviation and generating strong frequency components at 2f  2f and 2f  2f. We use two PLLs to recover these two signals. s1 t  cos  2ft  ft 

(9.77)

s2 t  cos  2ft  ft 

(9.78)

We then take the sum and difference of s1(t) and s2(t) to generate the desired I(t) and Q(t) signals. I t  s1 t   s2 t  2 cos2ft cosft

(9.79)

Q t  s1 t   s2 t  2 sin2ft sinft

(9.80)

The identical circuit can also be used for carrier recovery for a GMSK system.

r(t) f  f/2, f  f/2

s1(t )

PLL 2f  f

Divide by 2

PLL 2f  f

Divide by 2

I(t )

Multiply by 2

Figure 9.15

Carrier recovery for MSK.

Q(t ) s2(t )

9.7

Synchronization

281

The next step is to recover data timing or bit synchronization (Figure 9.16). Most communications systems transmit a sequence of 1s and 0s in an alternating pattern to enable the receiver to maintain bit synchronization. A PLL operating at bit timing is used to maintain timing. Once the PLL is synchronized on the received 101010 . . . pattern, it will remain synchronized on any other patterns except for long sequences of all 0s or all 1s. MSK uses an additional circuit to achieve bit timing. The s1(t) and s2(t) signals are multiplied together and low-pass filtered (Figure 9.17). s1 t s2 t  cos  2ft  ft   cos  2ft  ft  0.5 cos4ft  0.5 cos2ft

(9.81)

t low  pass filtered  s1 t s2 t  0.5 cos2ft 0.5 cos Tb

(9.82)

The output of a low-pass filter is a clock signal at one-half of the transmitted bit rate. The one-half bit rate clock is the correct rate for demodulation of the signal since the I and Q signals are at one-half of the bit rate. Frame synchronization is determined by the receiver correlating the received signal with a known bit pattern. The receiver performs an autocorrelation function

Received signal r(t)

I(t) or Q(t)

Matched Filter Detector

PLL

Figure 9.16

Differentiator

Generalized data timing recover circuit.

s1(t)

Low-Pass Filter

s2(t)

Figure 9.17

MSK data timing recovery circuit.

Clock Signal 1/2Tb

282

9

Demodulated Signal

Modulation Schemes

Shift Register

Auto-Correlator

Figure 9.18

Frame Timing

Generalized framing recovery circuit.

to determine when the bit pattern is received and then outputs a framing pulse (Figure 9.18).

9.8

Equalization

The received signal in a mobile radio environment travels from the transmitter to the receiver over many paths [7]. The signal fades in and out and undergoes distortion because of the multipath nature of the channel. For a transmitted signal s(t) a(t)cos[t  (t)], we can represent the received signal, r(t), as: n

r t 



xi t  i a t  i cos   t  i    t  i 

i 0

 yi t  i a t  i  sin   t  i    t  i 

(9.83)

We know that the received signal has Rayleigh fading statistics. But what are the characteristics of the xi(t) and yi(t) term in Equation 9.83? If the transmitter signal is narrow enough compared to the multipath structure of the channel, then the individual fading components, xi(t) and yi(t) will also have Rayleigh statistics. If a particular path is dominated by a reflection off a mountain, hill, building, or similar structure, then the statistics of that path may be Rician rather than Rayleigh. If the i (delay spread) is small compared to the bit interval, then a little distortion of the received signal occurs. If the i is greater than the bit interval, then the transmissions from one bit will interfere with transmissions of another bit, resulting in inter-symbol interference (ISI). Spread spectrum systems use wide bandwidth signals and attempt to recover the signals in each of the paths and add them together in a diversity (Rake) receiver. For our discussion here, however, we are transmitting narrowband signals, and the multipath signals are an interference to the desired signal. We need a receiver that removes the effects of the multipath signal or cancels the undesired multipath.

9.8

Equalization

283

Another way to describe the multipath channel is to describe the channel as having an impulse response h(t). The received signal is then written as: r t 







s t h t   d

(9.84)

We can recover s(t) if we can determine a transfer function h1(t) (the inverse of h(t)). One reason it is difficult to perform the inverse function is that it is time varying. The circuit that performs the inverse function is called an equalizer (see Figure 9.19). Generally, we are interested in minimizing the ISI at the same time we do our detection (the sample time in a sample-and-hold circuit). Thus, we can model the equalizer as a series of equal time delays (rather than random as is the general case) with the shortest delay interval being a bit interval. We then construct a receiver that determines req(t), the equalized signal: n

req t 

i t  n r t  n 

(9.85)

i 0

We use the equalized signal req(t) as the input to our detector to determine the value of the ith transmitted bit. We must adjust the values of i to achieve some measure of receiver performance. A typical measure is to minimize the mean square error between the value of the detected bit at the output of the adder and the output of the detector. Other measures for equalizers are possible.

1-bit delay

1-bit delay

2

3

1-bit delay

1-bit delay

detected bit 1

n



i to ith tap

Figure 9.19

Tap Gain Adjustment

Block diagram of an equalizer.

Detector

284

9

9.9

Modulation Schemes

Summary

In this chapter we studied the modulation and demodulation process applicable to cellular systems. Since baseband signals can be transmitted only over short distances with wires and require very long antennas to transmit them without wires, the baseband signals are modulated onto radio frequency carriers. We studied phase modulation and its derivatives /4-QPSK, /4-DQPSK, and OQPSK. We discussed MSK, a form of FM, as a first step toward GMSK which is used for DECT and GSM. Based on the literature, the error performance for GMSK also has the same-shaped curve as ASK with the proper definition of a correcting factor . For GSM, where BTb 0.3,  ~ 0.9; for DECT, where BTb 0.5,  ~ 0.97. We also studied a hybrid of amplitude and phase modulation, called QAM. We discussed the signal constellation and the bit error rate. Finally, we concluded the chapter by providing brief descriptions of the synchronization and equalization processes.

Problems 9.1 A digital signaling system is required to operate at 9.6 kbps. If a signal element encodes a 4-bit word, what is the minimum bandwidth of the channel?

9.2 Given a channel with intended capacity of 20 Mbps, the bandwidth of the channel is 3 MHz. What is the signal-to-noise ratio required to achieve this capacity? 9.3 The receiver in a communications system has a received signal power of 134 dBm, a received noise power spectral density of 174 dBm/Hz, and a bandwidth of 2000 Hz. What is the maximum rate of error-free information for the system? 9.4 Find the minimum required bandwidth for the baseband transmission of an 8-level PAM pulse sequence having a data rate, Rb, of 9600 bps. The system characteristics consist of a raised-cosine spectrum with 50% excess bandwidth. 9.5 Assuming that 0 00, a bit stream 101100 is sent using a /4-DPQSK modulation scheme. The leftmost bits are first applied to the transmitter. Determine phase k, and the values of Ik and Qk during transmission. 9.6 Find the BER probability, Pe, of an M-PSK system for a 14.4 kbps signal. Amplitude is 15 mW, and noise density, N0, is 109 W/Hz. 9.7 A noncoherent orthogonal M-FSK system carries 3 bits per symbol. The system is designed for Eb/N0 6 dB. What is the maximum bit error rate probability? Find the bandwidth efficiency of the system.

References

285

9.8 A noncoherent orthogonal M-FSK system carries 4 bits per symbol. The system is designed to have a maximum probability of symbol-error of 106. What is the required Eb/N0? What is the bandwidth efficiency?

References 1. Burr, Alister. Modulation and Coding for Wireless Communications. Prentice Hall, 2001. 2. Garg, V. K., and Wilkes, J. E. Principles & Applications of GSM. Upper Saddle River, NJ: Prentice Hall, 1999. 3. Glover, I. A., and Grant, P. M. Digital Communications. Prentice Hall, 1998. 4. Hamming, R.W. Error detecting and correcting codes. Bell System Technical Journal, 29:147160, 1950. 5. Lathi, B. P. Modern Digital and Analog Communications Systems, 2nd Edition. Holt, Reinhart and Winston, 1989. 6. Nyquist, H. Certain topics in telegraph transmission theory. AIEE Transactions, 47: 617–644, 1928. 7. Proakis, J. G. Digital Communications, 3rd Edition. McGraw-Hill, 1995. 8. Rappaport, T. S. Wireless Communications: Principles and Practice, 2nd Edition. Prentice Hall, 2000. 9. Shannon, C. E. A mathematical theory of communication. Part 1. Bell System Technical Journal, 27:379, 1948. 10. Shannon, C. E. A mathematical theory of communication, Part 2. Bell System Technical Journal, 27:623, 1948. 11. Skalar, B. Digital Communication — Fundamental and Applications. Englewood Cliffs, NJ: Prentice Hall, 1988. 12. Sweeney, P. Error Control Coding: An Introduction. Prentice Hall, 1991. 13. Ziemer, R. E., and Peterson, R. L. Introduction to Digital Communication. New York: Macmillan Publishing Co., 1992.

This page intentionally left blank

CHAPTER 10 Antennas, Diversity, and Link Analysis 10.1

Introduction

The antenna system used for any radio communication platform is one of the critical and least understood parts of the system. The antenna system is the interface between the radio system and the external environment. Wireless communication systems require antennas at the transmitter and receiver to operate properly. The design and deployment of antennas can make or break a wireless system, and many poorly performing systems can be traced to improperly installed or placed antennas. The antenna system can consist of a single antenna at the base station and one at the mobile station. Primarily the antenna is employed by the base station and the mobile for establishing and maintaining the communication link. There are many types of antennas available, all of which perform specific functions depending on the application at hand. The type of antenna used by a system operator can be colinear, folded dipole, or Yagi, to mention a few [2,8–9]. In this chapter we first define and discuss the roles of antennas. We then examine the interrelationships between antenna beam width, directivity, and gain. After completing our discussion of antennas we explore the concepts of diversity reception where multiple signals are combined to improve the signal-to-noise ratio (SNR) of the system [3,10–13,16]. We describe the noise that a receiver sees and use the concept to perform an analysis of the link between the base station and the mobile station and calculate the SNR at the receiver. Throughout the chapter, we present numerical examples to improve the understanding of each topic.

10.2

Antenna System

The radio communication system requires a reliable communication between a fixed base station and a mobile station. The goal of the system designer is to have the same performance in both the transmitting and receiving directions. This is not always possible since the base station typically has a higher output power than the mobile station. Also, the mobile station antenna is typically on the street at a height of 1 to 2 m compared to the base station antenna height of 50 to 60 m.

287

288

10

Antennas, Diversity, and Link Analysis

This results in the mobile receiver having a high noise level due to interference caused by the ignitions of nearby vehicles or other objects, whereas the base station, with its higher antenna, usually sees a quieter radio environment. These factors can combine to favor one particular direction over the other. The wireless system designer must carefully consider all these factors or the range of the system may be limited by poor performance in one direction. The types of antennas, their gain and coverage patterns, the available power to drive them, the use of simple or multiple antenna configurations, and the polarization (the polarization of a radio wave is the orientation of its electric field vector) are the major factors that can be controlled by the system designer. The system designer has no control over the topography between the base station and mobile station antennas, the speed and direction of the mobile station, and the location of antenna(s) on the mobile station. Each of these factors significantly affects system performance. Sometimes the placement of the mobile antenna can severely limit system performance. While the mobile antenna is usually installed by a knowledgeable technician, the owner of the vehicle may force a nonoptimum placement of the antenna. Furthermore, cellular antennas are vertically polarized physically, but many vehicle antennas are no longer vertical after the vehicle is sent through a car wash. Along with the type of antenna there is the relative pattern of the antenna, indicating in which direction the energy emitted or received will be directed. There are two primary classifications of antennas: omnidirectional and sectorized (directional). The omnidirectional antennas are used when the desire is to have a 360° radiation pattern. The sectorized antennas are used when a more refined pattern is needed. The directional pattern is usually required to facilitate system growth through frequency reuse. The choice of antenna directly impacts the performance of both the base station and the overall network. The antenna selection must consider a number of design issues. Some involve the antenna gain, the antenna pattern, the interface or matching to the transmitter, the receiver used for the site, the bandwidth and frequency range over which the signals will travel, and the power handling capabilities.

10.3

Antenna Gain

The task of a transmitting antenna is to convert the electrical energy travelling along a transmission path into electromagnetic waves in space. The antennas are passive devices, the power radiated by the transmitting antenna cannot be greater than the power entering from the transmitter. It is always less because of losses. Antenna gain in one direction results from a concentration of power in that direction and is accompanied by a loss in other directions. Antenna gain [7] is the most important parameter in the design of an antenna system. A high gain is achieved by increasing the aperture area, A, of the antenna. Antennas obey reciprocity; the

10.3

Antenna Gain

289

transmit gain and receive gain are the same, and the antenna can be analyzed by examining it as either a receive or transmit antenna. The amount of power captured by an antenna is given as: P  pA

(10.1)

where: p  power density (power per unit area) A  aperture area Antenna gain can be defined either with respect to an isotropic antenna or with respect to a half-wave dipole and is usually analyzed as a transmit antenna. An isotropic antenna is an idealized system that radiates equally in all directions. The half-wave dipole antenna is a simple, practical antenna which is in common use. The gain of an antenna in a given direction is the ratio of power density produced by it in that direction divided by the power density that would be produced by an isotropic antenna. The term dBi is used to refer to the antenna gain with respect to the isotropic antenna. The term dBd is used to refer to the antenna gain with respect to a half-wave dipole (0 dBd  2.1 dBi). While most analyses of system performance use a half-wave dipole as the reference, many times antenna gain figures are quoted in dBi to give a falsely inflated gain figure. The system designer must carefully read data sheets on antennas to use the correct gain figure. As a rule of thumb, if the gain is not quoted in either dBd or dBi, the gain is in dBi, with the dBi left out to inflate the gain figures. For an isotropic antenna in free space, the received power density is given as P

T pR   2

4d

(10.2a)

where: PT  transmitter power pR  receiver power density d  distance between transmitter and receiver When a directional transmitting antenna with power gain factor, GT, is used, the power density at the receiver antenna is GT times Equation 10.2a, i.e., P

T pR  GT   2

4d

(10.2b)

290

10

Antennas, Diversity, and Link Analysis

The amount of power captured by the receiver is pR times the aperture area, AR, of the receiving antenna. The aperture area is related to the gain of the receiving antenna by 4A

R GR   2



(10.3)

where:   c f

f  the transmission frequency in Hz c  3  108 m/s is the free-space speed of propagation for electromagnetic waves AR  the effective area of aperture, which is less than the physical area by an efficiency factor R; typical value for R ranges from 60 to 80% The total received power, PR is given as: PR  ARpR

(10.4)

Substituting the value of PR and AR from Equations 10.2b and 10.3 into Equation 10.4, together with the transmitting antenna gain GT, we get  P G G PR   T T R

 4d 

2

(10.5)

Equation 10.5 includes the power loss only from the spreading of the transmitted wave. If other losses are also present, such as atmospheric absorption or ohm losses of the waveguides leading to antennas, Equation 10.5 can be modified as [4]: PR PT

  4d 

 

2

G G L0

T R  

PR GTGR   PT L0Lp

(10.6)

(10.7)

where: 4d 2 denotes the loss associated with propagation of electromagnetic Lp  







waves from the transmitter to the receiver as discussed in Chapter 3. Lp depends on carrier frequency and separation distance d. This loss is always present. L0 is the loss factor for additional losses.

10.3

Antenna Gain

291

When we express Equation 10.7 in terms of decibels, we get  P G G L dB PR  20 log  T T R 0

 4d 

(10.8)

The product PTGT is called the effective isotropic radiated power (EIRP) and  refers to free space path loss (L ) in dB. Another term, effective term 20 log  p

 4d 

radiated power (ERP), is also used. It is the power input multiplied by the antenna gain measured with respect to a half-wave dipole antenna. The EIRP is related to ERP as EIRP  ERP 2.14 dB

(10.9)

In the free space, the path between two antennas has no obstruction (see Figure 10.1) and there is no object where reflection can occur. Thus, the received signal is composed of only one component. When the two antennas are located on the earth, then there are multiple paths from the transmitter to the receiver. The effect of multiple paths is to change the path loss between two points. The simplest case occurs when the antenna heights hT and hR are small compared to their separation distance d and the reflecting earth surface is assumed to be flat (see Chapter 3). The received signal can then be represented by a field that is approximated by a combination of a direct wave and a reflected wave as shown in Figure 10.2. In this case the received power, PR, and transmitted power, PT, are related as (see Chapter 3 and Appendix B for derivation): PR PT



d

Transmitting System PT , Gain: GT

Free-space path-loss model.

G G L0

T R  

(10.10)

Receiving System PR , Gain: GR d

Figure 10.1



hT hR 2

  2

292

10

Dire

ct W

Re

ave

flec

hT

Antennas, Diversity, and Link Analysis

ted

hT hR

Wa

ve

hR hR d

Figure 10.2

Path-loss model with reflection.

Expressing Equation 10.9 in decibels, we get



h h



T R PR  20 log  PT GT GR L0 dB 2

d

(10.11)

comparing it with Equation 10.8 we note that Equation 10.10 is independent of transmitting frequency. Example 10.1 We consider a communication system in which the distance between the transmitter and receiver is 10,000 m. The transmitter EIRP is 30 dBW (GT  20 dBi; PT  10 dBW). The transmitting frequency is 1.5 GHz (  0.2 m). The receiver antenna gain is 3 dBi; and total system losses are 6 dB. Assuming the receiver noise figure, Nf  5 dB and bandwidth, Bw  1.25 MHz, calculate the received signal power at the receiver antenna and the SNR of the received signal. Neglect any feed line losses between the antenna and receiver. Solution

 4  10,000 

0.2 LP : free-space loss  20 log   116 dB

PR  116 30 3 6  89 dBW  59 dBm Noise power, PN  Nf BW N0

10.4

Performance Criteria of Antenna Systems

293

Noise Density, N0  kT  log  1.38  10 23  290   204 dBW  174 dBm PN  5 174 10 log  1.25  106   108 dBm P PN

R SNR    59  108   49 dB

Example 10.2 Using the data in Example 10.1 and antenna heights at the receiver and transmitter units to be 40 m and 2 m, respectively, calculate the received signal power at the receiver antenna and the SNR of the received signal. Solution 40  2 Path loss  20 log   122 dB 2

 10,000 

PR  122 30 3 6  95 dBW  65 dBm PN  108 dBm (from Example 10.1) P PN

R  SNR    65  108   43 dB

10.4

Performance Criteria of Antenna Systems

The performance of an antenna system [7,19] is not restricted to its gain and physical attributes. There are many other parameters that must be considered in evaluating antenna performance. The parameters that define the performance of an antenna system are: • Antenna pattern • Main and side lobe • Radiation efficiency ( ) • Antenna bandwidth • Horizontal beam width • Vertical beam width • Gain (G) • Directivity (D) • Antenna polarization • Input impedance

294

10

Antennas, Diversity, and Link Analysis

• Front-to-back ratio (RFB) • Front-to-side ratio (RFS) • Power dissipation • Intermodulation • Construction • Cost

The antenna pattern chosen should match the coverage requirements for the base station. An antenna has one major lobe and a number of side lobes (see Figure 10.3). The side lobes are important because they create potential problems by generating interference. Ideally there should be no side lobes for the antenna pattern. For down tilting, the side lobes are important because they can create secondary sources of interference. The radiation efficiency is a ratio of total power radiated by an antenna to net power accepted by an antenna from the transmitter. The bandwidth defines the operating range of the frequencies for the antenna. It is the angular separation between two directions in which radiation interest is identical. The half power point for the beam width is the angular separation where there is 3 dB reduction off the main lobe. Normally, the wider the beam width, the lower the gain of the antenna. The gain is the ratio of the radiation intensity in a given direction to that of an isotropically radiated signal. The directivity is the gain calculated assuming a lossless antenna. Real antennas have losses, gain is the directivity multiplied by the efficiency of the antenna. Polarization is important for an antenna because wireless mobile systems use vertical polarization. A vertical antenna is easiest to mount on a vehicle; therefore vertical polarization has been standardized. In general, horizontal or vertical polarization will work equally well. Most cables used as feed line from the transmitter/receiver to the antenna are either 50 or 72/75 ohms. If the input impedance of the antenna is far removed from either of these values, it will be difficult to get an antenna to accept the power delivered to it and its radian efficiency will be low. The front to back ratio is a ratio in respect to how much energy is directed in the exact opposite direction of the main lobe of the antenna (see Figure 10.3). The front to side ratio is a ratio in respect to how much energy is directed in the side lobes of the main lobe of the antenna (see Figure 10.3). The power dissipation is a measure of the total power the antenna can accept at its input terminals. The antenna chosen should be able to handle the maximum envisioned power load without any problem. The amount of intermodulation which the antenna introduces to the network in the presence of strong signals as referenced from manufacturer should be considered in antenna selection. The construction attributes are associated with physical dimensions, mounting requirements, material used, wind loading, and connectors.

10.5

Relationship between Directivity, Gain, and Beam Width of an Antenna

295

Side Lobe 0

45

45 Major Lobe

90

90

Back Lobe 135

135 180

Figure 10.3

10.5

Radiation pattern of an antenna.

Relationship between Directivity, Gain, and Beam Width of an Antenna

Real antennas are not isotropic radiators but have a pattern of more and less power in different directions. The antenna engineer considers the pattern of the power radiated in the horizontal and vertical directions. The shape of the pattern describes the directionality of the antenna. The direction for maximum power is referred to as the primary beam or main lobe, whereas secondary beams are called the minor lobes (back and side lobes), see Figure 10.3 [7]. The pattern of the antenna has two desired effects: concentration of the power in a desired direction to improve the signal strength at the receiver and weakening the power in an undesired direction to reduce interference from or to other receivers. Thus, the minor lobes give undesired radiation or reception. Since the major lobe propagates and receives the most energy, this lobe is called the front lobe. Lobes adjacent to the front lobe are the side lobes and the lobe exactly opposite to the front lobe is called the back lobe. The front-to-back ratio of an antenna is defined as: pmF pmB

RFB  10 log 

where: pmF  maximum power density of the front lobe pmB  maximum power density of the back lobe

(10.12)

296

10

Antennas, Diversity, and Link Analysis

One-half Wavelength

Balanced feed line Figure 10.4

Half-wave dipole antenna.

The front-to-side ratio of an antenna is given as: pmF pmS

RFS  10 log 

(10.13)

where: pmS  maximum power density of the side lobe The half-wave dipole antenna has two parts (see Figure 10.4). The half-wave length is handy for impedance matching. Since the dipole antenna is the simplest one to build, it is often used as the reference to describe the gain of other antennas. The average power density for the dipole vertical antenna is given as: 3P

T  sin 2 pavg   2

8d

(10.14)

In Equation 10.14 it should be noticed that power density at any point depends on the direction and distance d from the dipole. Since there is no dependence on , the antenna pattern is directional in x-z and y-z planes, but omnidirectional in the x-y plane. A dipole mounted in a vertical direction provides an omnidirectional pattern that is useful for the base station antenna of cellular systems. While dipoles are often used as simple antennas, in real systems other types of antennas are used to provide higher gain than a dipole antenna.

10.5.1 The Relationship between Directivity and Gain The directivity of an antenna is the ratio of the maximum power density from the antenna and power density from an isotropic antenna  p max

D

 4d  PT

2

(10.15)

10.5

Relationship between Directivity, Gain, and Beam Width of an Antenna

297

where:  p max  maximum power density from the antenna PT  transmitted power of an isotropic antenna The gain of an antenna is the ratio of maximum power density from the antenna and input power density if the antenna is isotropic.  p max

 p max

 4d 

 4d  

G  PI

2

PT

(10.16a)

 2

Using Equation 10.15 in Equation 10.16a, we get: G D

(10.16b)

where:  efficiency of the antenna PI  input power of an isotropic antenna

10.5.2 Relation between Gain and Beam Width The receiver gain GR can be related to its half-power beam width as [7]: 4 GR  

HPHP

(10.17)

where:

HP and HP are the half-power beam widths in the and  planes The factor 4 is the solid angle subtended by a sphere in steradians (square radians) 180 2  41,250 degree2  solid angle in a sphere 4 steradians  4     

41,250 GR  

HPHP

(10.18)

For an ideal gain antenna, where power density is uniform inside the 3-dB beam width (for both and ) and zero outside the 3-dB beam width, the gain GR can be expressed as: 41,250



GR  

(10.19)

298

10

Antennas, Diversity, and Link Analysis

where:

and  are in degrees For a real antenna with side lobes, the gain should be calculated using Equation 10.20, which includes the effect of side lobes. 32,400



GR  

(10.20)

If an antenna is designed with a circular pattern in one direction, i.e., a linear element is used, the approximate gain can be obtained for the vertical 3-dB beam width from Equation 10.19 as: 41,250 360 

114.6 GR    

(10.21)

The corresponding gain equation for a linear element including side lobes is given as [7]: 101.5 GR  

(10.22)

10.5.3 Helical Antennas Helical antennas provide circular polarized waves in the same direction as that of the helix. A helical antenna can be used to receive circular polarized waves and also receive plane polarized waves with the polarization in any direction. Helical antennas are often used with VHF and UHF satellite transmissions. The gain of a helical antenna is proportional to the number of turns (N). An approximate expression for the gain with respect to an isotropic antenna is given as [7]: 15NS DH 2

G   3 

where: G  antenna gain in dB N  number of turns in the helix, N  3  S  turn spacing in meters, S   4

 DH  diameter of helix in meters, DH      wave length in meters.

(10.23)

10.5

Relationship between Directivity, Gain, and Beam Width of an Antenna

299

The radian pattern for this antenna type has one major lobe and several minor lobes. For the major lobe, the 3 dB beam width (in degrees) is approximately: 

52 

   DH

N  S

(10.24)

Example 10.3 We consider an antenna in which 12 W of power results in 3 W of power being dissipated as resistive losses in the antenna and the rest radiated by the antenna. If the directivity of the antenna is 7 dB (i.e., 5), what is its gain? Solution 12 3   0.75 12

G  D  0.75  5  3.75 (5.74 dB)

Example 10.4 What is the 3-dB beam width of a linear element antenna with a gain of 12 dBi? Solution 101.5

  6.4º 12/10 (10)

Example 10.5 Consider a helical antenna, which has 12 turns and is designed for a frequency of 1.8 GHz. a. calculate the optimum diameter (DH), spacing (S) for the antenna and total length of the antenna, b. calculate the antenna gain, and c. calculate the beam width of the antenna. Solution a. 300  10   c    0.1667 m 6 6

f

1800  10

300

10

Antennas, Diversity, and Link Analysis

0.1667  DH    0.053 m  53 mm  

  0.1667  0.0417 m  41.7 mm S  4

4

L  NS  12  41.7  500 mm

b. 15NS(D  )2

15  12  0.0417  (  0.053)2



(0.1667)

H G      44.92  16.5 dBi 3 3

c. 



52 0.1667   52  0.1667 

      30 degrees DH

10.6

 NS

  0.053

 12  0.0417

Diversity

In Chapter 3, we pointed out that a radio channel is subjected to fading, timedispersion, and other degradations. Diversity techniques are employed to overcome these impairments and improve signal quality [6,13,15,20]. The basic concept of diversity is that the receiver has more than one version of the transmitted signal available, and each version of transmitted signal is received through a distinct channel. When several versions of the signal, carrying the same information, are received over multiple channels that exhibit independent fading with comparable strengths, the chances that all the independently faded signal components experience the same fading simultaneously are greatly reduced. Suppose the probability of having a loss of communications due to fading on one channel is p and this probability is independent on all M channels. The probability of losing communications on all channels simultaneously is then pM. Thus, a 10% chance of losing the signal for one channel is reduced to 0.13  0.001  0.1% with three independently fading channels [5,17]. Typically, the diversity receiver is used in the base station instead of the mobile station, because the cost of the diversity combiner can be high, especially if multiple receivers are necessary. Also, the power output of the mobile station is limited by the battery. Handset transmitters usually lower power than mobilemounted transmitters to preserve battery life and reduce radiation into the human body. The base station, however, can increase its power output or antenna height to improve the coverage to a mobile station.

10.6

Diversity

301

Channel #1

Channel #2 Modulator

M Branch Diversity Combiner

Demodulator

Transmitter Channel #M

Figure 10.5

Receiver

Diversity channel model.

Each of the channels, plus the corresponding receiver circuit, is called a branch and the outputs of the channels are processed and routed to the demodulator by a diversity combiner (see Figure 10.5). Two criteria are required to achieve a high degree of improvement from a diversity system. First, the fading in individual branches should have low crosscorrelation. Second, the mean power available from each branch should be almost equal. If the cross-correlation is too high, then fades in each branch will occur simultaneously. On the other hand, if the branches have low correlation but have very different mean powers, then the signal in a weaker branch may not be useful even though it has less fades than the other branches.

10.6.1 Types of Diversity The following methods are used to obtain uncorrelated signals for combining: 1. Space diversity: Two antennas separated physically by a short distance d can provide two signals with low correlation between their fades. The separation d in general varies with antenna height h and with frequency. The higher the frequency, the closer the two antennas can be to each other. Typically, a separation of a few wavelengths is enough to obtain uncorrelated signals. Taking into account the shadowing effect (see Chapter 3), usually a separation of at least 10 carrier wavelengths is required between two adjacent antennas. This diversity does not require extra system capacity; however, the cost is the extra antennas needed. 2. Frequency diversity: Signals received on two frequencies, separated by coherence bandwidth (see Chapter 3) are uncorrelated. To use frequency diversity in an urban or suburban environment for cellular and personal communications services (PCS) frequencies, the frequency separation must be 300 kHz or more. This diversity improves link transmission quality at the cost of extra frequency bandwidths.

302

10

Antennas, Diversity, and Link Analysis

3. Time diversity: If the identical signals are transmitted in different time slots, the received signals will be uncorrelated, provided the time difference between time slots is more than the channel coherence time (see Chapter 3). This system will work for an environment where the fading occurs independent of the movement of the receiver. In a mobile radio environment, the mobile unit may be at a standstill at any location that has a weak local mean or is caught in a fade. Although fading still occurs even when the mobile is still, the time-delayed signals are correlated and time diversity will not reduce the fades. In addition to extra system capacity (in terms of transmission time) due to the redundant transmission, this diversity introduces a significant signal processing delay, especially when the channel coherence time is large. In practice, time diversity is more frequently used through bit interleaving, forward-error-correction, and automatic retransmission request (ARQ). 4. Polarization diversity: The horizontal and vertical polarization components transmitted by two polarized antennas at the base station and received by two polarized antennas at the mobile station can provide two uncorrelated fading signals. Polarization diversity results in 3 dB power reduction at the transmitting site since the power must be split into two different polarized antennas. 5. Angle diversity: When the operating frequency is 10 GHz, the scattering of signals from transmitter to receiver generates received signals from different directions that are uncorrelated with each other. Thus, two or more directional antennas can be pointed in different directions at the receiving site and provide signals for a combiner. This scheme is more effective at the mobile station than at the base station since the scattering is from local buildings and vegetation and is more pronounced at street level than at the height of base station antennas. Angle diversity can be viewed as a special case of space diversity since it also requires multiple antennas. 6. Path diversity: In code division multiple access (CDMA) systems, the use of direct sequence spread spectrum modulation allows the desired signal to be transmitted over a frequency bandwidth much larger than the channel coherence bandwidth. The spread spectrum signal can resolve in multipath signal components provided the path delays are separated by at least one chip period. A Rake receiver can separate the received signal components from different propagation paths by using code correlation and can then combine them constructively. In CDMA, exploiting the path diversity reduces the transmitted power needed and increases the system capacity by reducing interference.

10.7

Combining Methods

The goal of a combiner is to improve the noise performance of the system [1,11,14]. After obtaining the uncorrelated signals, we need to consider the method of

10.7

Combining Methods

303

processing these signals to obtain the best results. The analysis of combiners is generally performed in terms of SNR. We will examine several different types (selection, maximal-ratio, and equal-gain) of combiners and compare their SNR improvements over no diversity.

10.7.1 Selection Combiner In this case, the diversity combiner selects the branch that instantaneously has the highest SNR (see Figure 10.6). We assume that the signal received by each diversity branch is statistically independent of the signals in other branches and is Rayleigh distributed with equal mean signal power P0. The probability density function of the signal envelope, on branch i, is given as (see Chapter 3) r P0

p(ri)  i e ri /(2P0) 2

(10.25)

where: 2P0  mean-square signal power per branch  r2i  instantaneous power in the ith branch Let i  r2i /(2Ni) and 0  (2P0)/(2Ni), where Ni is the noise power in the i th branch r2

i  i   0

(10.26)

2P0

Monitor SNR

Select Highest SNR

Channel #1

Channel #2 Modulator

Demodulator

Transmitter

Receiver

Channel #M

Figure 10.6

Diversity selection combiner.

304

10

Antennas, Diversity, and Link Analysis

The probability density function for i is given by 1 e (i /0) p(i)  

(10.27)

0

Assuming that the signal in each branch has the same mean, the probability that the SNR on any branch is less than or equal to any given value g is P i  g  



g

0

p(i)di  1 e (g/0)

(10.28)

Therefore, the probability that the SNRs in all branches are simultaneously less than or equal to g is given by PM(g)  P  1, 2, … , M  g    1 e (g/0) M

(10.29)

The probability that at least one branch will exceed the given SNR value of g is given by: P (at least one branch  g)  1 PM(g)

(10.30)

The percentage of time the instantaneous output SNR M is below or equal to the given value, g, is equal to P(M  g). The results of the selection combiner for M  1, 2, and 4 are plotted in Figure 10.7. Note that the largest gain occurs for the 2-branch diversity combiner. By differentiating Equation 10.29 we get the probability density function M 1  g /  0

pM(g)  M g /0   1 e  g /0  

e

(10.31)

The mean value of the SNR can be given as  M 

M     1 e 

0

g

 g/0  M 1  g/0 

e

dg

(10.32)

0





0

M

1 

k k1

M   

(10.33)

10.7

Combining Methods

305

1E 00 M-Selection Combining

Probability of Error

1E 01

M1

1E 02

M2 M=4

1E 03 Gain of 15 dB for M  1 to M  2

Gain of 8 dB For M  2 to M  4

1E 04

1E 05

40

35

30

25

20

15

10

5

0

5

10

Gain  10 log (  g /  0 ) (dB)

Figure 10.7

Probability for different values of M-selection combiners.

Example 10.6 We consider the 3-branch selection combiner in which each branch receives an independent Rayleigh fading signal. If the average SNR is 30 dB (1000), determine the probability that the SNR will drop below 10 dB (10). Compare the result with a case without any diversity. Solution g  10 dB and 0  30 dB 

10  0.01  g   0

1000 M

p3 10 dB    1 e  g/0     1 e 0.01 3  1  10 6

When diversity is not used, M  1 M p1 10 dB    1 e  g/0    1 e 0.01 1  1  10 2

306

10

Antennas, Diversity, and Link Analysis

10.7.2 Switched Combiner The disadvantage with selection combining is that the combiner must be able to monitor all M branches simultaneously. This requires M independent receivers which is expensive and complicated. An alternative is to use switched combining. In this case only one receiver is needed, and it is only switched between branches when the SNR on the current branch is lower than some predefined threshold value g. This is called a switch and stay combiner. The performance of a switch combiner is less than that in selection combining, since unused branches may have SNRs higher than the current branch if the current SNR exceeds the threshold. The threshold therefore has to be carefully selected in relation to the mean power on each branch, which must also be estimated with sufficient accuracy. 10.7.3 Maximal Ratio Combiner In maximal ratio combining, M signals are weighted proportionally to their signalto-noise ratios and then summed (see Figure 10.8). M

rM 

airi(t)

(10.34)

i1

where: ai  weight of i th branch M  number of branches Since noise in each branch is weighted according to noise power,  n2i (t) 

M

M

 ai aj ni(t)nj(t)



j  1i  1

a1

(10.35)

Weight

Channel 1 a2 Channel 2 M Branch Combiner

Modulator Transmitter

Receiver aM Channel M

Figure 10.8

Demodulator

Maximal ratio combining.

10.7

Combining Methods

307

M

NT 

 a2i n2i (t) 

i1

M

 ai 2Ni

i1

(10.36)

where: NT  average noise power

 2

ni (t)  2Ni The SNR at the output is given as:



M

airi(t) i1



2

1 M    M 2

 ai 

(10.37)

2N

i

i1

We want to maximize M. This can be done using the Schwartz inequality.

  M

   M

2

airi 

i1

i1

M

 r 2i 

 ai 2

(10.38)

i1



If ai  ri /(Ni ), then using Equation 10.38 to define Equation 10.39 we get: M

M

M

i

max

1   M 2



i1

M

 M

r 2i

r2i  N i1 i1

max

1 



r 2i



2 i  1 Ni

(10.39)

r 2i

M

i

(10.40)

i1

Thus, the SNR at the combiner output equals the sum of the SNR of the branches.  M

M

max





M

0  M0

i  i

1

(10.41)

i1





M M 

0

(10.42)

308

10

Antennas, Diversity, and Link Analysis

The probability that M  g is given by:

g



p M  g   1 e

g



p M  g   e

0

0



M

g k 1

   0



(k 1)! k1



g k 1

  

M

0



k1

(k 1)!



(10.43)



(10.44)

The plot of p for M  1, 2, and 4 with maximal ratio combining is shown in Figure 10.9.

1E 00 Maximal Ratio Combining 1E 01

Probability of Error

M1 M2

1E 02

M4

1E 03 Gain for M  1 to M  2

Gain for M  2 to M  4

1E 04

1E 05

40

35

30

25

20

15

10

5

0

Gain  10 log (  g /  0 ) (dB)

Figure 10.9

Probability for different values of maximal ratio combiner.

5

10

10.7

Combining Methods

309

10.7.4 Equal Gain Combiner Equal gain combining is similar to maximal ratio combining, but there is no attempt to weight the signal before addition; thus ai  1. The envelope of the output signal is given as: M

r

r t

(10.45)

  i

i1

and mean output SNR is given as:  2 M

  ri



1 i1 M    2

(10.46)

M

Ni

i1

Assuming that mean noise in each branch is the same (i.e., N); Equation 10.46 becomes  2 M

 



1 M   ri 2NM i  1

M



1   rr 2NM i, j  1 i j

(10.47)

  but r2i  2P0; and  ri  (P0)/2 . The various branch signals are uncorrelated,  rjri   rirj   ri  rj, for i not equal to j. Therefore Equation 10.47 will become:

P   1 M   2MP0 M M 1   0  0  1  M 1    2NM



2



4

 M    1 M 1 

0

(10.48)

(10.49)

4

For M  2, the probability p can be written in the closed form as: 2g

 



p M  g   1 e

0

 



g



  e 0

g

    erf

 0

  

g



(10.50)

0

For M  2, the bit error probability should be obtained by a numerical integration technique.

310

10

Antennas, Diversity, and Link Analysis

Table 10.1 SNR ratio in dB. Maximal-ratio combiner

Equal-gain combiner

M

Selection combiner

1

20

20

20

2

10

8.5

9.2

4

4.0

1.0

2.0

6

2.0

2.0

1.5

Table 10.2 SNR improvement in dB at 1% bit-error probability.

M

Selection combiner

Maximal-ratio combiner

Equal-gain combiner

2

10.0

11.5

10.8

4

16.0

19.0

18.0

6

18.0

22.0

21.5

Table 10.1 shows M versus SNR at 1% bit error probability for the selection, maximal ratio, and equal gain combiner. Table 10.2 shows SNR improvement for M  2, 4, and 6 at 1% bit error probability for the selection, maximal ratio, and equal gain combiner. It can be observed that the selection diversity combiner has the poorest performance and the maximal ratio the best. The performance of the equal gain diversity combiner is sightly lower than that of the maximal ratio combiner. The implementation complexity for equal gain combining is significantly less than the maximal ratio combining because of the requirement of correct weighing factors. Performance of the three combining schemes is compared in Figure 10.10.

10.8

Rake Receiver

In 1958, Price and Green proposed a method of resolving multipath using wideband pseudo-random sequences modulated onto a transmitter using other modulation methods (AM or FM) [4]. The pseudo-random sequence (see Chapter 11) has the property that its time-shifted versions are almost uncorrelated. Thus, a signal that propagates from transmitter to receiver over multipath (hence different time delays) can be resolved into separately fading signals by cross-correlating the received signal with multiple time-shifted versions of the pseudo-random sequence. Figure 10.11 shows a block diagram of a typical system. In the receiver, the outputs are time shifted and therefore must be sent through a delay line before entering the diversity combiner. The receiver is called a Rake receiver since the block diagram looks like a garden rake.

10.8

Rake Receiver

311

1E 00 Comparison of Combiners

Probability of Error

1E 01 M4

M2

M1 1E 02 All

1E 03

Selection Maximal Ratio Selection

1E 04

Maximal Ratio

Equal Gain

1E 05

50

40

30

20

10

0

10

20

Gain  10 log (  g /  0 ) (dB)

Figure 10.10

Performance comparison of various combining schemes.

Tapped Delay Line T

a(t)

T

a(t)

T

T

T

T

T

a(t)

T

T

a(t)

Diversity Combiner Combiner Output Figure 10.11

Rake receiver.

When CDMA systems were designed for cellular systems, the inherent wide bandwidth signals with their orthogonal Walsh functions (see Chapter 11) were a natural for implementing a Rake receiver. The Rake receiver reduces the effects of fading and provides spectral efficiency improvement of CDMA over other cellular systems.

312

10

Antennas, Diversity, and Link Analysis

In CDMA systems, the bandwidth (1.25 to 15 MHz) is wider than the coherence bandwidth of the cellular or PCS channel. Thus, when the multipaths are resolved in the receiver, the signals from each tap on the delay line are uncorrelated with each other. The receiver can combine them using any of the combining techniques described in the previous section. Thus, the CDMA system uses the multipath characteristics of the communication channel to its advantage to improve the operation of the system. The performance of the Rake receiver will be governed by the combining scheme used. In the IS-95 and WCDMA, the maximal ratio combining with 3 to 4 branches is used. An important factor in the receiver design is to obtain sychronization of the signals in the receiver. Since in CDMA, adjacent cells are also on the same frequency with different time delays on the Walsh codes, the entire CDMA system is tightly synchronized. Example 10.7 A Rake receiver is used in the wideband CDMA (WCDMA) system (spreading rate 3.84 Mcps) to reduce the multipath effects in the channel. What is the minimum delay difference to successfully resolve the multipath components and operate the Rake receiver? Solution In order to resolve multipath components, the chip duration should be equal to or greater than , where  is defined as: delay distance speed of electromagnetic wave

d

d      s 6

300  10

1 Chip duration of WCDMA system    0.26 s 6 3.84  10

dd

6  0.26  10 6

300  10

 dd  78 m

10.9

Link Budgets

We discussed link margin in Chapter 3. In this section, we present a systematic procedure for developing a link budget of a wireless system [18]. A link budget is the calculation of the amount of power received at a given receiver based on the output power from the transmitter. The link budget considers all of the gains and

10.9

Link Budgets

313

losses that a radio wave experiences along the path from transmitter to receiver. We need to perform the calculation in both directions: from the mobile station to the base station (uplink) and from the base station to the mobile station (downlink). We determine the maximum allowable path loss in each direction and use the lesser of the two to calculate the coverage for the cell and service in question. The link budget should include a margin to allow fading of the signal. In other words, we design the system such that service will be supported even if the signal fades significantly. The greater the fade margin, the greater the reliability

Table 10.3 Uplink budget for speech (outdoor pedestrian) service at 12.2 kbps. Transmitter (Mobile Station)

Note

Mobile transmit power (dBm)

21

Antenna gain (dBi)

0

Body losses (dB)

3.0

Equivalent isotropic radiated power (EIRP)

18

Receiver (Base Station) Thermal noise density (dBm/Hz) Receiver noise figure (dB) Receiver noise power (dBm), calculated for 3.84 Mcps (see Chapter 11) Interference margin (dB) Total noise interference (dBm) Processing gain (dB) (see Chapter 11) Required Eb/N0 (dB) Effective receiver sensitivity (dBm)

174 5.0

103.2 4

Depends on cell loading

25

10 log(3.84  106/12.2  103)

4

Depend on service type

120.2 18

Base station feeder and connector losses (dB)

2

Fast fading margin (dB)

4 7.5

Building penetration loss (dB)

0

Soft handoff gain (dB)

2

Max. allowable path loss (dB)

174 5 10 log(3.84  106) 

103.2 dBm

99.2

Base station antenna gain (dBi)

Log normal fading margin (dB)

10 log(1.38  10 23  290) 

204 dBW  174 dBm

144.7

Total noise interference

Processing gain Required Eb/N0

Enables room for closed loop power control Enable for greater cell-area reliability

18  (120.2)  (18  2  4 7.5  2)  144.7

314

10

Antennas, Diversity, and Link Analysis

Table 10.4 Downlink budget for speech (outdoor pedestrian) service at 12.2 kbps. Transmitter (Base Station)

Note

BS transmit power (dBm)

40

BS antenna gain (dBi)

18

BS feeder and connector losses (dB)

2.0

Equivalent isotropic radiated power (EIRP) (dBm)

56

10 W

Receiver (Mobile Station) Thermal noise density (dBm/Hz) Receiver noise figure (dB) Receiver noise power (dBm), calculated for 3.84 Mcps Interference margin (dB) Total noise interference (dBm) Processing gain (dB) Required Eb/N0 (dB) Effective receiver sensitivity (dBm) Mobile station antenna gain (dBi) Body losses (dB) Fast fading margin (dB)

174 5.0

103.2 0

Depends on cell loading

25

10 log(3.84  106/12.2  103)

4

Service-dependent

124.2

Total noise interference

Processing gain Required Eb/N0

0 3.0 4 7.5

Building penetration loss (dB)

0

Max. allowable path loss (dB)

174 5 10 log (3.84  106) 

103.2 dBm

103.2

Lognormal fading margin (dB)

Soft handoff gain (dB)

10 log(1.38  10 23  290) 

204 dBW  174 dBm

Enables room for closed-loop power control Enable for greater cell-area reliability

2 167.7

(124.2)  56  (3  4  7.5  2)  167.7

of the service. Tables 10.3 and 10.4 provide examples of the uplink and downlink link budget for the WCDMA speech service at 12.2 kbps (outdoor pedestrian). The maximum allowable path loss is 144.7 dB based on uplink budget. The cell coverage must be based on the uplink path loss.

10.10

Summary

In this chapter we discussed the basics of antenna technology. We presented performance criteria of an antenna system and developed relations between the antenna gain, directivity, and beam width. We then moved to discuss diversity, an

References

315

extremely powerful method for improving the quality of communication systems. It is possible to achieve gains equivalent to power savings in excess of 10 dB. These gains are obtained at the expense of extra hardware, particularly in terms of extra antennas and receivers, which must be balanced against the benefits. The key requirements for achieving the maximum benefit are that the multiple branches of the system should have substantially equal mean power and near-zero cross-correlation of the fading signal. We concluded the chapter by presenting link budget calculations for a WCDMA system.

Problems 10.1 Define gain, beam width, and radiation efficiency of an antenna. 10.2 What is diversity? How is it provided in a communication system? 10.3 Among the selection, equal-gain, and maximal ratio combining, which scheme is the best and why? 10.4 In Problem 9.1, calculate the SNR ratio when a frequency of 6 GHz is used for the system. At this frequency, use 10 dB for the receiver noise figure. 10.5 Repeat Problem 9.2 for 3 and 6 GHz and use 10 dB for the receiver noise figure. What can you say about system design needs as frequency is raised? 10.6 If the 3-dB beam width of a linear element antenna is 8°, find its gain. 10.7 A sector antenna has vertical and horizontal beam widths of 50° and 60°, respectively. Calculate its gain (1) ignoring the side lobes and (2) accounting the side lobes. 10.8 Calculate the time separation required for two signals to achieve a high degree of time diversity in a classical Rayleigh channel at 900 MHz with a mobile speed of 20 km/hour. 10.9 Given a 2-branch selection combining system operated with independent Rayleigh fading, calculate the diversity gain for a probability of 10 6. 10.10 Repeat Problem 10.9 for the 2-branch maximal ratio combining system.

References 1. Brennan, D. J. Linear Diversity Combining Techniques. Proceedings of IRE, June 1959, 1075–1101. 2. Bryson, W. B. Antenna System for 800 MHz. IEEE Vehicular Technology Conference, San Diego, May 1982, p. 287. 3. Halpern, S. W. The Theory of Operation of an Equal-Gain Predetection Regenerative Diversity Combiner With Rayleigh Fading Channels. IEEE Transaction on Communications Technology, COM-22(8), August 1974, 1099–1106.

316

10

Antennas, Diversity, and Link Analysis

4. Jakes, W. C. Microwave Mobile Communications. New York: John Wiley & Sons, 1974. 5. Kahn, L. R. Radio Squarer. Proceedings of the IRE, 42, November 1954, 1704. 6. Kozono, S., et al. Base Station Polarization Diversity Reception for Mobile Radio. IEEE Transactions, Vehicular Technology, 33(4), 301–306, 1984. 7. Kraus, J. D. Antennas. New York: McGraw-Hill, 1988. 8. Lee, W. C .Y. Antenna Spacing Requirements for a Mobile Radio Base-Station Diversity. Bell System Technical Journal, 50, no. 6 (July-August 1971). 9. Lee, W. C. Y. Mobile Communication Engineering. New York: McGraw-Hill, 1982. 10. Lee, W. C. Y. Mobile Radio Performance for a Two-Branch Equal-Gain Combining Receiver with Correlated Signals at the Land Site. IEEE Transaction EA, VT-27, November 1978, 239–243. 11. Mahrotra, A. Cellular Radio Performance Engineering. Boston: Artech House, 1994. 12. Mark, J. W., and Zhuang, Weihua. Wireless Communications and Networking. Upper Saddle River, NJ: Prentice Hall, 2003. 13. Parsons, J. D., et al. Diversity Techniques for Mobile Radio Reception. Radio and Electronic Engineer, 45, no. 7 (July 1975): 357–367. 14. Parsons, J. D., and Gardiner, J. G. Mobile Communications Systems. New York: Halsted Press, 1989. 15. Price, R., and Green, P. E., Jr. A Communications Technique for Multipath Channels. Proceeding of the IRE. March 1958, 555–570. 16. Saunders, S. R. Antennas and Propagation for Wireless Communication Systems. New York: John Wiley & Sons, 1999. 17. Schwartz, M., et al. Communication System Techniques. New York: McGraw-Hill, 1966. 18. Smith, C., and Collins, D. 3G Wireless Networks. New York: McGraw-Hill, 2002. 19. Tilston, W. V. On Evaluating the Performance of Communication Antennas. IEEE Communications Society Magazine (September 1981). 20. Vaughan, R. G. Polarization Diversity in Mobile Communications. IEEE Transactions Vehicular Technology, 39(3), 177–186, 1990.

CHAPTER 11 Spread Spectrum (SS) and CDMA Systems 11.1

Introduction

We introduced spread spectrum techniques in Chapter 6. In this chapter, we present details of direct-sequence spread spectrum (DSSS) and frequency-hop spread spectrum (FHSS) systems [1,2,4]. We show how signal spreading and despreading is achieved with binary phase shift keying (BPSK) and quadrature phase shift keying (QPSK) modulation in the DSSS [11,12]. We then address multipath issues in wireless communications and show how code division multiple access (CDMA) takes advantage of multipath in improving system performance with a Rake receiver [6,13]. We conclude the chapter by presenting a summary of the challenges in implementing a CDMA system and providing some highlights of the Telecommunication Industries Association (TIA) IS-95 CDMA system. Those who are not familiar with spreading codes should refer to Appendix D.

11.2

Concept of Spread Spectrum

In a wideband spread-spectrum (SS) system, the transmitted signal is spread over a frequency band that is much larger, in fact, than the maximum bandwidth required to transmit the information bearing (baseband) signal [3]. An SS system takes a baseband signal with a bandwidth of only a few kilohertz (kHz), and spreads it over a band that may be many megahertz (MHz) wide. In SS systems, an advantage in signal-to-noise ratio (SNR) is achieved by the modulation and demodulation process. The SS signal is generated from a data-modulated carrier. The data-modulated carrier is modulated a second time by using a wideband spreading signal. An SS signal has advantages in the areas of security, resistance to narrowband jamming, resistance to multipath fading, and supporting multiple-access techniques. The spreading modulation may be phase modulation or a rapid change of the carrier frequency, or it may be a combination of these two schemes. When spectrum spreading is performed by phase modulation, we call the resultant signal a direct-sequence spread spectrum (DSSS) signal (see Figure 11.1) [15]. When spectrum spreading is achieved by a rapid change of the carrier frequency, we refer to the resultant signal as a frequency-hop spread spectrum (FHSS) signal [14]. When both direct-sequence and frequency-hop techniques are employed, the 317

11

Magnitude of chip

318

TO  Period of Output Wave form

Spread Spectrum (SS) and CDMA Systems

TO  (2n 1) TC TC  chip duration

1

time (t ) 1

Frequency

fn fn1 fn2 f3 f2 f1 time (t ) 0

Tc 2Tc

Transmitted time slots (k bits) One Frame

t 0

t

Tf

M  time slots in each frame; t  Tf /M Note: M  8 in this example Figure 11.1

Spread spectrum techniques.

2Tf

3Tf

11.2

Concept of Spread Spectrum

319

resultant signal is called a hybrid DS-FH SS signal. Another way to also generate an SS signal is the time-hop spread spectrum (THSS) signal. In this case, the transmission time is divided into intervals called frames. Each frame is further divided into time slots. During each frame, one and only one time slot is modulated with a message (details of THSS are not given in this chapter). The DSSS is the averaging technique to reduce interference whereas FHSS and THSS are the avoiding techniques to minimize interference. The spreading signal is selected to have properties to facilitate demodulation of the transmitted signal by the intended receiver, and to make demodulation by an unintended receiver as difficult as possible. These same properties also make it possible for the intended receiver to differentiate between the communication signal and jamming. If the bandwidth of the spreading signal is large relative to the data bandwidth, the spread-spectrum transmission bandwidth is dominated by the spreading signal and is independent of the data signal bandwidth. Consider the channel capacity as given by the Shannon equation [10]: C  Bw log2  1  S/N 

(11.1)

where: Bw  channel bandwidth in Hertz (Hz) C  channel capacity in bits per second (bps) S  signal power N  noise power Equation 11.1 provides the relationship between the theoretical ability of a communication channel to transmit information without errors for a given signalto-noise ratio and a given bandwidth of the channel. The channel capacity can be increased by increasing the channel bandwidth, the transmitted power, or a combination of both. Shannon modeled the channel at baseband. However, Equation 11.1 is applicable to a radio frequency (RF) channel by assuming that the intermediate frequency (IF) filter has an ideal (flat) bandpass response with a bandwidth that is at least 2Bw. This bound assumes that channel noise is additive white Gaussian noise (AWGN). AWGN is often used in the modeling of an RF channel. This assumption is justified since the total noise is generated by random electron fluctuations. The central limit theorem provides us with the assumption that the output of an IF filter has a Gaussian distribution and is frequency independent. For most communications systems that are limited by thermal noise, this assumption is valid. For interference-limited systems, this assumption is not valid and the results may be different. The Shannon equation does not provide a method to achieve the bound. Approaching the bound requires complex channel coding and modulation schemes. In many cases, achieving an implementation that provides performance near this bound is impractical due to the resulting complexity. Spread spectrum

320

11

Spread Spectrum (SS) and CDMA Systems

systems can be engineered to operate at much lower SNRs since the channel bandwidth can be traded for the SNR to achieve good performance at a very low SNR. We rewrite Equation 11.1 as: C Bw

  1.44 loge  1  S/N 

(11.2)

Since 2 N

S 1 S loge  1  S/N      N

2

3 N

1 S  

3

4 N

1 S  

4

 ...

(11.3)

Assuming that the S/N ratio is small (e.g., S/N  0.1), we can neglect the higherorder terms and rewrite Equation 11.2 as: C 1 Bw    1.44

(11.4)

 S/N 

For any given S/N ratio we can have a low information error rate by increasing the bandwidth used to transmit the information. As an example, if we wish a system to operate on an RF link in which the data information rate is 20 kilobits per second (kbps) and the S/N ratio is 0.01, we should use a bandwidth of 20  10 Bw    1.38  106 Hz or 1.38 MHz 3

(11.5)

1.44  0.01

Information can be modulated into the SS signal by several methods. The most common method is to multiply the information with the spread-spectrum code (refer to Appendix D) before it is used to modulate the carrier frequency (see Figure 11.2). This technique applies to any SS system that uses a spreading code to determine RF bandwidth. If the signal that is being sent is analog (voice, for example), the signal must be digitized before being modulated by the spreading code. Sampling

m i (t ) x i (t )

Data Modulator

Demodulator Detection Device Integrator

c i (t ) Direct Sequence (DS) Generator

Figure 11.2

Direct sequence spread spectrum system.

c i (t )

11.3

System Processing Gain

11.3

321

System Processing Gain

In the DSSS system the baseband signal is spread over a large bandwidth by a spreading code. One of the major advantages of a DSSS system is the robustness to interference [8,9]. The system processing gain, Gp, quantifies the degree of interference rejection. The system processing gain is the ratio of spreading rate, Rc, to the information rate, Rb, and is given as: Gp  Rc /Rb

(11.6)

where: Rb  information bit rate Rc  spreading rate Typical processing gains of a DSSS system lie between 10 and 60 dB. With a DSSS system, the total noise level is determined both by thermal noise and by interference. For a given user, the interference is processed as noise. The input and output S/N ratios are related as: (S/N)o  Gp(S/N)i

(11.7)

We express input (S/N)i ratio as: E R

b b 1   b     NS i    N0  Bw i N0 i Gp

E

(11.8)

where: Eb  bit energy N0  noise power spectral density Using Equation 11.8 we rewrite Equation 11.7 as: b  NS o  Gp   NS i    N0 i

E

(11.9)

but Eb  Rb  NS o    N0  Bw 

N  E

 b o

0 o

1 

Gp

(11.10)

Therefore using Equation 11.9 and Equation 11.10 we get,

 N  Eb

0 o

N  E

 Gp  b

0 i

(11.11)

322

11

Spread Spectrum (SS) and CDMA Systems

Example 11.1 A DSSS system has a 1.2288 megachips per second (Mcps) code clock rate and a 9.6 kbps information rate. Calculate the processing gain. How much improvement in information rate is achieved if the code generation rate is changed to 5 Mcps and the processing gain to 256? Solution 1.2288  106 9.6  10

 128  10 log 128  21 dB  3 R Gp

Rb  c  106 Rb  5  19.53 kbps 256

Improvement in information rate  19.53  9.6  9.93 kbps In a DSSS system the information signal, xi(t), is multiplied by a wideband code signal, ci(t), which is the output signal of the direct sequence (DS) generator (see Figure 11.2) [7,9]. The signal xi(t)  ci(t)  mi(t) is modulated and transmitted. This signal occupies a bandwidth far in excess of the minimum bandwidth required to transmit the information signal xi(t). We can observe from Figure 11.3 that the combined signal waveform has more high frequency changes in the data information since 1/Tc 1/Tb, where Tb is the bit interval of the information stream and Tc is the bit interval of the DSSS stream. Tc is called a chip interval. In a DSSS system, the pulse shape is a sequence of short rectangular pulses called chips. The chips are a pseudo-random sequence of 1’s and 1’s known at the receiver and often having a repetition period equal to the symbol period. T

Gp  b Tc

Since Rb  1/Tb and Bw  1/Tc , T Tc

B Rb

w Gp  b  

When xi(t) and ci(t) have the same rate, the product mi(t) contains all the information of xi(t) and has the same rate as ci(t). The spectrum of the signal remains unchanged, and the incoming bit stream is said to be encrypted or scrambled. However, when ci(t) has a higher rate than xi(t), mi(t) contains all the information of xi(t), has a higher bit rate compared to xi(t), and is said to have had its spectrum spread (refer to Figure 11.3).

11.3

System Processing Gain

323

PN Sequence, ci (t ) Chip Rate  Rc  1/Tc Modulo 2 Sum

Base band xi (t ) Information Rate, Rb  1/Tb

cos(2 fct )  cos( ct ) Base band Filter

xi (t )

ci (t )

1

1

mi (t )

Power Amplifier

t

t 1

1

Tb

Tc

Base Band Signal 1

PN Sequence

Tc t xi (t ) ci (t ) [spreading]

1

Figure 11.3

Direct sequence (DS) spreading.

Let us consider the downlink CDMA operation, in which the base station generates a data stream for mobile 1, 2, and 3 . . . xi(t), and multiplies the data stream by an appropriate DS code, ci(t). Next, we add the coded data streams (see Figures 11.4, 11.5, and 11.6): 3

z(t) 

 mj

(11.12)

j1

We modulate the resultant baseband spread spectrum signal, Equation 11.12, by a carrier (frequency c  2 fc) to obtain

 3

zi t   Ai

j1



cj(t)xj(t) cos ct

(11.13)

The spread signal at the receiver for mobile, i, is zi(t)  noise, where zi(t) is transmitted over the allocated bandwidth.

324

11



Spread Spectrum (SS) and CDMA Systems



3

zi t   Ai

cj(t  j)xj(t  j) cos( ct  i)

j1

(11.14)

Equation 11.14 contains the desired user signal and other users’ signals. After multiplication by a coherent carrier phase i estimated by the receiver, a locally generated DS sequence that is the exact replica of the desired code transmitted multiplies the incoming signal — despreading. We assume that this DS sequence is in perfect synchronization (receiver estimates delay j) with the transmitted signal so that: cj(t  j)cj(t  j)  1

(11.15)

The multiplier output yields the desired data signal xi(t  j) plus interfering terms due to other users. Ideally, the integrator, an integrate-and-dump over Tb, should produce a cross-correlation between the desired signal and interferers that is 0. Hence, the output for mobile i is the transmitted data stream, xi(t  j).

Magnitude 1 x1(t)

t

1

Tb

Tc

1 c1(t)

t

1 Tc

1 m1(t)

t

1

Spreading

Figure 11.4

m1(t )  x1(t ) c1(t )

Downlink DSSS operation — signals for mobile 1, 2, and 3.

Magnitude 1 x2(t )

t

1

Tb

Tc

1 t

c2(t ) 1

Tc

1 m2(t )

t

1

m2(t )  x2(t ) c2(t )

Spreading Magnitude 1

t

x3(t ) 1 Tb

1 c3(t )

t

1 Tc

Tc

1 m3(t )

t

1

Spreading Figure 11.4

m3(t )  x3(t ) c3(t )

(Continued).

325

326

11

Spread Spectrum (SS) and CDMA Systems

Magnitude

3 z(t)  m1(t)  m2(t)  m3(t)

2 1 t 1 2 3

Figure 11.5

Tc

Resultant demodulated signal at a mobile.

Example 11.2 We consider signals transmitted from a base station to mobile 1, 2, and 3. The data streams x1(t), x2(t), and x3(t) are multiplied with codes c1(t), c2(t), and c3(t), respectively (see Figure 11.4). The resultant demodulated signal z(t), sum of m1(t), m2(t), and m3(t) is given in Figure 11.5. Show that the transmitted signals to mobiles 1, 2, and 3 are recovered at the mobile receivers by despreading the resultant signal z(t). Neglect propagation delay. Solution From Tables 11.1, 11.2, and 11.3, we observe that the transmitted signals for mobile 1, 2, and 3 are recovered by despreading at the receivers.

Table 11.1 z(t)c1(t) [see Figure 11.6(a)].

Value of integration at end of bit period Bit value (if value of integration is 0, then bit value equals 1)

Tb

2Tb

3Tb

4Tb

5Tb

6

4

6

2

12

1

0

0

1

1

(a) 3 z(t ) 2  m1(t ) 1  m2(t )  m3(t ) 1

t

2 3

Tc 1 t

c1(t ) 1

3

1 Despreading z(t ).c (t )

t

1

1

3

Figure 11.6(a)

Tb

Tb

Tb

Tb

Tb

Despreading of resultant demodulated signal at mobile receivers.

Table 11.2 z(t)c2(t) [see Figure 11.6(b)].

Value of integration at end of bit period Bit value

Tb

2Tb

3Tb

4Tb

5Tb

2

6

8

4

8

1

1

0

1

1

Tb

2Tb

3Tb

4Tb

5Tb

6

4

8

8

8

1

0

1

1

0

Table 11.3 z(t)c3(t) [see Figure 11.6(c)].

Value of integration at end of bit period Bit value

327

328

11

Spread Spectrum (SS) and CDMA Systems

(b) 3 z(t) 2  m1(t) 1  m2 (t)  m3 (t) 1

t

2 3

1 c2 (t)

t

1

Tb

Tb

Tb

Despreading z(t).c2 (t)

t

Tb

Figure 11.6(b)

11.4

Tb

(Continued).

Requirements of Direct-Sequence Spread Spectrum

In the DSSS system, the entire bandwidth of the RF carrier is made available to each user. The DSSS system satisfies the following requirements [16]: • The spreading signal has a bandwidth much larger than the minimum band-

width required to transmit the desired information which, for a digital system, is baseband data. • The spreading of the information is performed by using a spreading signal, called the code signal (see Appendix D). The code signal is independent of the data and is of a much higher chip rate than the data signal.

11.5

Frequency-Hopping Spread Spectrum Systems

329

(c) 3 z (t ) 2  m1(t ) 1  m2 (t )  m3 (t ) 1

t

2 3 1 c3 (t )

t

1 3 2 1

Despreading z (t ). c3 (t )

t

1 2 3

Figure 11.6(c)

Tb

Tb

Tb

Tb

Tb

(Continued).

• At the intended receiver, despreading is accomplished by cross-correlation

of the received spread signal with a synchronized replica of the same code signal used to spread the data.

11.5

Frequency-Hopping Spread Spectrum Systems

In FHSS systems, the binary pseudo-random noise (PN) code generator drives the frequency synthesizer to hop to one of the many available frequencies chosen by the PN sequence generator. When the hopping rate is higher than the symbol rate,

330

11

Spread Spectrum (SS) and CDMA Systems

we have a fast frequency-hop system. If the hopping rate is lower than the symbol rate, i.e., there are several symbols transmitted per frequency hop, we have a slow frequency-hop system. In an FHSS system b bits are used as an integer index for selecting the hop frequency. The term chip is also used in FHSS, but its meaning is different from the DSSS meaning of the word. It depends on whether the system is a slow frequency hop (FH) or a fast FH. In FH/M-ary frequency shift keying (MFSK) a system chip is the tone of shortest duration. The chip rate, Rc, is the maximum of Rs and Rh, where Rs is the symbol rate and Rh is the hop rate. Thus, for slow FHSS systems, the chip rate is equal to the symbol rate (i.e., Rc  Rs), whereas for fast FHSS systems, the chip rate is equal to the hop rate (i.e., Rc  Rh). In a slow FHSS system with MFSK, the selected M frequencies must be an integer number of symbol rates apart. The spacing is required to maintain orthogonality between the frequencies and to allow reliable noncoherent detection. This implies that the minimum bandwidth of an MFSK signal should be about MRs. The modulation scheme commonly used with a fast FHSS system is MFSK, where b  log2 M bits are used to determine which one of M frequencies should be used. The position of the M-ary signal set is shifted pseudo-randomly by the frequency synthesizer over a hopping bandwidth Bw. The frequency synthesizer produces a transmission tone based on the simultaneous instructions of the PN code and the data. At each frequency hop-time a PN generator feeds the frequency synthesizer a frequency word (a sequence of b chips) which decides one Binary Message 1/Tb

Digital modulator

Up-convert

Frequency synthesizer

fc

PN code generator Clock, 1/ Tc Hopping pattern:

fx

M = 2b f1 1 Figure 11.7

FHSS system.

2

3

...

t / Tc

Transmitted signal

11.5

Frequency-Hopping Spread Spectrum Systems

Received signal

331

Digital demodulator

Estimated message

fc

Up-convert

Frequency synthesizer

PN code generator Clock, 1/Tc Figure 11.7

(Continued).

of 2b symbol-set positions (see Figure 11.7). The frequency hopping bandwidth, Bw, and the minimum frequency spacing between consecutive hop positions, f, determine the minimum number of chips necessary in the frequency word. The processing gain, Gp, of an FHSS system is given as: Hopping Bandwidth Minimum Frequency Spacing

M  f f

Gp      M

(11.16)

In the fast frequency-hop systems, there are L frequency hops during a symbol interval (Ts) (i.e., Ts  LTc or Rc  LRs.)  Hopping Bandwidth  (KMf )L

(11.17)

where: K  factor for frequency multiplication M 2b is the number of frequencies produced by frequency synthesizer b  bit in a symbol L  frequency hops per symbol KMfL f

 Gp    MKL

(11.18)

The processing gain of a fast frequency hop system is dependent upon the number of frequencies used (M), the number of hops per symbol (L), and the frequency multiplication factor (K).

332

11

Spread Spectrum (SS) and CDMA Systems

Example 11.3 In an FHSS system, a hopping bandwidth of 100 MHz and a frequency spacing of 10 kHz is used. What is the minimum number of PN chips that are required for each frequency symbol? Solution  106  104 Number of frequency tones in hopping bandwidth  100  3 10

Minimum number of chips  \log2 (104)]  13 chips

Example 11.4 A communication system transmits at 120 kbps and uses 32-FSK (Frequency Shift Keying). A hop rate of 2000 hops per second is used over an available spectrum of 10 MHz. Assuming a negligible synthesizer switching time between hops, calculate (a) data symbol transmitted per hop, and (b) the number of nonoverlapping hop frequencies. Solution For 32-FSK, we have 32  2b i.e., b  5 bits per symbol 120 kbps 5

Symbol rate    24 kbps Since the symbol rate is higher than the hop rate, the system is a slow FHSS system. 24,000 Number of symbols per hop    12 2,000

Minimum bandwidth Bw of an M-ary FSK MRs: Bw  12  24  288 kHz

Number of nonoverlapping hop frequencies: 10 MHz nFH   35 288 kHz

Example 11.5 Consider an FHSS system in which the input data rate is 200 bits per second. The modulation scheme of 32-ary FSK is used to generate the modulation symbol. The frequency hopping rate is 200 hops per second. Calculate: (a) minimum separation between frequency tones; (b) number of frequency tones produced by a frequency synthesizer; (c) processing gain; and (d) hopping bandwidth. Assume a frequency multiplication factor K  1.

11.6

Operational Advantages of SS Modulation

333

Solution With the 32-FSK modulation scheme there are 5 chips per symbol: 200 Symbol rate: Rs    40 symbols/sec 5

The hop rate is higher than symbol rate, the system is a fast FHSS system. 1s Symbol duration    25 ms 40

200 L  5 hops/symbol 40

25 Chip duration    5 ms 5

1 Minimum separation between tones    200 Hz 3 5  10

M  25  32 frequency tones Frequency hopping bandwidth  KMfL  1  32  200  5  64 kHz Gp  MKL  32  1  5  160

11.6

Operational Advantages of SS Modulation

The following are the operational advantages of SS modulation: • Low probability of intercept: Low probability of intercept implies that

a third party cannot easily eavesdrop on the conversation or has to utilize expensive means to accomplish this. A standard communications receiver selects a demodulation circuitry, such as an amplitude demodulator, a frequency demodulator, or a phase demodulator, depending on the modulation scheme used at the transmitter. In an SS system, the receiver demodulates the transmitted energy through some correlation process and effectively combines various components within a wider bandwidth. A single-channel receiver will thus detect a small portion of the transmitted signal which will be too weak for normal detection, and even if it could be magnified to a detectable level would be incomplete and thus not possible to understand. • Low probability of position fix: Conventional radio transmitters are easily pinpointed by simple and inexpensive direction finders. The spectral-spreading concepts make this task much more demanding as greater processing power and integration time will be required within each resolution.

334

11

Spread Spectrum (SS) and CDMA Systems

• Low probability of signal exploitation: This refers to the possibilities which exist

within a communication scenario to exploit the communication link by some manipulation of the waveforms used to carry the message. These could be: • destruction of synchronization messages; • destruction or alteration of the message contents; • invisible or concealed addition of data bits. • High resistance to jamming and interference: SS systems are inherently more robust against jamming and interference than systems not using spreading techniques. This property of SS modulation is used in military systems where the main design objective is to develop a system which is able to deliver a message through a very hostile and impenetrable medium. SS modulation techniques provide an additional factor which is not seen in conventional systems. This is due to the fact that a code is used in the spreading process and unless the jammer/interferer manages to get hold of this code the impact of the jamming/interference is reduced by a significant amount. There are two very important aspects of SS with respect to jamming and interference. The first relates to the actual protection provided by an SS code. The SS receiver provides a post-detection signal-to-noise and signal-to-interference improvement. This means that if the SS signal can be received with sufficient clarity and strength and without interference signals, intelligent jammers might redirect the same waveform with more power and cause problems for the detection process in the data link. This is a practical limitation since a limited number of spreading codes are used in an SS modem and thus the transmission tends to repeat the same codes a large number of times. The vulnerability lies in the fact that the code might be revealed if it sticks out clearly only once—for instance if it is received at a very short distance from the transmitter. This points at the very important requirement of SS systems, that the spectral-spreading scheme should be sufficiently agile and use frequent change of code structures. The next important aspect of SS modulation with respect to jamming and interference is that SS modulation provides a practical way of coping with frequency-band congestion. • High time resolution/reduction of multipath effects: Multipath effects are one of the unavoidable effects in radio communication. Multipath implies that the signal reaching the receiver antenna has travelled by two or more paths. Because these routes inevitably are of different lengths, the time delays of signals that have come along the respective paths are different and the signal will fade in or out with small displacements of the transmitter and receiver (displacements of the order of half a wavelength). Methods exist whereby the signal is coded such that the signals reaching the receiver via different paths add up in phase at the receiver. Adaptive methods capitalizing on the wider bandwidth of SS waveforms make it possible to use radio communication under extremely severe multipath conditions. Example of

11.7

Coherent Binary Phase-Shift Keying DSSS

335

such are wireless local area networks used inside rooms or buildings where the propagation conditions are very poor. • Cryptographic capabilities: The coding aspects of SS modulation have implications for possible security function in a communication link. The spreading codes can be chosen such that they serve the dual purpose of spreading the frequency spectrum of the transmitted signal and making it difficult to decipher the message. This requires that the code provide the required spectral signature (usually reasonably flat), has good anti-cryptographic property, and is sufficiently redundant. As these criteria place rather tough requirements on the coding strategy, a more common approach is to implement SS coding and scramble the code for cryptography independently and usually sequentially.

11.7

Coherent Binary Phase-Shift Keying DSSS

The simplest form of a DSSS communication system uses coherent binary phaseshift keying (BPSK) for both data modulation and spreading modulation. But the most common form uses BPSK for data modulation and quadrature phase-shift keying (QPSK) for spreading modulation. We first consider the simplest case. The ith mobile station is assigned a spreading code signal ci(t) (a periodic spreading code sequence with chips of width Tc). Each mobile station has its own such code signal. Information bits are transmitted by superimposing the data bits onto the code signal. If the ith mobile station transmits the binary data waveform xi(t)[xj(t)  1], it forms the binary sequence. mi(t)  xi(t)ci(t)

(11.19)

Equation 11.19 represents the modulo-2 addition of ci(t) and xi(t) as a multiplication because the binary 0 and 1 represent values of 1 and 1 into the modulator. The transmitted signal from the ith mobile is 

si(t)  xi(t)ci(t) 2P cos( ct  )

(11.20)

where: xi(t)  baseband signal for the ith mobile ci(t)  spreading code for the ith mobile

c  carrier frequency P  signal power

 data phase modulation If Tb is the bit period of si(t), then Tb may correspond to either a full period for ci(t), or to a fraction of a period. If Tb is less than one code period, then the data bits are modulating the polarity of a portion of a code period. The code ci(t) serves as a subcarrier for the source data. Since each mobile station uses the entire

336

11

Spread Spectrum (SS) and CDMA Systems

channel bandwidth and since Equation 11.19 has a code chip rate of I/Tc chips per second, each BPSK carrier uses an RF bandwidth of 1 Bw  

(11.21)

Tc

The available channel RF bandwidth determines the minimum chip width, and the code period determines its relation to the bit time. The number of code chips per bit is given by: T Tc

B Rb

w Gp  b  

(11.22)

The ratio Bw/Rb is the CDMA processing gain, Gp, or simply, the spreading ratio of code modulation. This shows how much the RF bandwidth must be spread relative to the bit rate, Rb, to accommodate a given spreading code length. Each mobile station (MS) uses the same RF carrier frequency and RF bandwidth, but with its own spreading code ci(t). The signal in Equation 11.20 is transmitted using a distortionless path with transmission. The signal is received together with some type of interference and/ or Gaussian noise. Demodulation is performed in part by remodulating with the spreading code appropriately delayed as shown in Figure 11.8. This correlation of the received signal with the delayed spreading waveform is the despreading. This is a critical function in all spread spectrum systems. The signal component of the output of the despreading is: 

xi(t  d) 2P ci(t  d)  ci(t  ˆ d)cos( c(t  d)  )

(11.23)

where: ˆ d  receiver’s best estimate of the transmission delay Since ci(t)  1, the product ci(t  d)  ci(t  ˆ d) will be unity if d  ˆ d, that is, if the code at the receiver is synchronized with the spreading code at the transmitter. When correctly synchronized, the signal component of the output of  the receiver despreading is equal to 2P xi(t  d) cos( c(t  d)  ), which can be demodulated using a conventional coherent phase modulator. The bit error probability, Pe, associated with the coherent BPSK spread spectrum signal is the same as with the BPSK signal and is given as: 

1 erfc Pe   2





     Q     Q 2 Gp      Eb  N0

o

E 2 b N0

Eb N0



o

i

(11.24)

11.8

Quadrature Phase-Shift Keying DSSS

xi 2P cos( ct  ) xi (t )

337

2P cos( ct  ) xi (t )ci (t )

Modulator

ci (t )

 2P cos( ct)

Transmitter

xi (t  d) 2Pci (t  d)ci (t  ˆd)cos [ c (t  d)  ] Band–pass Filter

ci (t  ˆd)

Figure 11.8

Estimated Data Demodulator

Receiver

DSSS system with BPSK.

where: 2 eu /2 Q(u)   , u 1 is known as Q function (see Appendix C)   2 u 

Eb  P/Rb  energy per bit

11.8

Quadrature Phase-Shift Keying DSSS

Sometimes it is advantageous to transmit simultaneously on two carriers which are in phase quadrature. The main reason for this is to save spectrum because, for the same total transmitted power, we can achieve the same bit error probability, Pe, using one-half the transmission bandwidth. The quadrature modulations are more difficult to detect in low probability of detection applications. Also, the quadrature modulations are less sensitive to some types of jamming. We refer to Figure 11.9 and write: 



s(t)  P cI(t) cos[ t  ]  P cQ(t)sin[ t  ]

(11.25)

s(t)  aI(t)  aQ(t)

(11.26)

338

11

Spread Spectrum (SS) and CDMA Systems

cI (t ) P cos( ct  )  Data

aI

Quadrature Hybrid

Modulator

2P cos( ct  ) 



P sin( ct  )  Transmitter

s (t )

aQ cQ(t )

cI (t  ˆd) x (t ) s (t  d )

Power Divider



z (t ) Band–pass Filter ( f)

Demodulator Estimated Data

y (t ) cQ (t  ˆd)

Figure 11.9

Receiver

DSSS system with QPSK.

where: cI(t) and cQ(t) are the in-phase and quadrature spreading codes and aI(t) and aQ(t) are orthogonal. This condition is satisfied in the present case since cI(t) and cQ(t) are independent code waveforms. The receiver for the transmitted signal is shown in Figure 11.9. The bandpass filter is centered at frequency f and has a bandwidth sufficiently wide to pass the data-modulated carrier without distortion. 



x(t)  P/2 cI(t  d)cI(t  ˆ d)cos [ ct  ]  ( P/2 )cQ  (t  d)cI(t  ˆ d)sin [ ct  ] 

(11.27) 

y(t)  P/2 cI(t  d)cQ(t  ˆ d)sin ( ct  )  ( P/2 )cQ  (t  d)cQ(t  ˆ d)cos [ ct  ]

(11.28)

If the receiver-generated replicas of spreading codes are correctly phased then cI(t  d)cI(t  ˆ d)  cQ(t  d)cQ(t  ˆ d)  1

(11.29)

11.9

Bit Scrambling

339 

z(t)  x(t)  y(t)  2P cos[ ct  ]

(11.30)

The signal z(t) is the input to a conventional phase demodulator where data is recovered. When the spreading codes are staggered one-half chip interval with respect to each other, the QPSK is called offset-QPSK (OQPSK). In OQPSK, the phase changes every one-half chip interval, but it does not change more than 90°. This limited phase change improves the uniformity of the signal envelope compared to BPSK and QPSK, since zero-crossings of the carrier envelope are avoided. Neither QPSK nor OQPSK modulation can be removed with a single stage of square-law detection. Two such detectors and the associated loss of signal-to-noise ratio are required. QPSK and OQPSK offer some low probability of detection advantages over the BPSK method.

11.9

Bit Scrambling

Referring to Table 11.4, we consider the following activities at a given transmitter location (see Figure 11.10): 1. 2. 3. 4.

An arbitrary data sequence si(t) is generated by a digital source. An arbitrary code sequence ci(t) is produced by a DS generator. Two sequences are modulo-2 added and transmitted to a distant receiver. At the distant location, the resulting sequence (assuming no propagation delay) is picked up by the receiver (see Figure 11.10).

Table 11.4 Operations with modulo-2 addition. Transmitter

Receiver

1

si (t)

1

1

0

1

0

0

1

1

1

1

2

ci (t)

1

0

0

1

1

1

0

1

0

0

3

si (t)  ci (t)

0

1

0

0

1

1

1

0

1

1

4

si (t)  ci (t)

0

1

0

0

1

1

1

0

1

1

5

ci (t)

1

0

0

1

1

1

0

1

0

0

6

si (t)  ci (t)  ci (t)  si (t)

1

1

0

1

0

0

1

1

1

1

si (t ).ci (t )

si (t )

ci (t ) Figure 11.10

Mobile transmitter.

340

11

Spread Spectrum (SS) and CDMA Systems

5. The code ci(t) used at the transmitter is also available at the receiver. 6. The original data sequence is recovered by modulo-2 adding the received sequence with the locally available code ci (t). Next, referring to Table 11.5, we consider the following set of activities at the given transmitter location. 1. An arbitrary data sequence si(t) is generated by a digital source. In this case, we use 1s and 1s to represent 0s and 1s. 2. An arbitrary code sequence ci(t) is produced by a DS generator. 3. We multiply si(t) and ci(t). The output of the multiplier is transmitted to a distant receiver. 4. At the distant location, the resulting sequence (again assuming no propagation delay) is picked up by the receiver (Figure 11.11). 5. The code ci(t) used at the transmitting location is also available at the receiver. 6. The original data sequence is recovered by multiplying the received sequence by the locally available code, ci(t). From Tables 11.4 and 11.5 we conclude that modulo-2 addition using 1s and 0s binary data is equivalent to multiplication using 1 and 1 binary data as long as we remain consistent in mapping 0s to 1s and 1s to 1s as shown in Table 11.5. (For circuit implementation, modulo-2 addition is preferred since exclusive OR Table 11.5 Operations without modulo-2 addition. Transmitter

Receiver

1

si (t)

1 1

1 1

2

ci (t)

1

1 1 1 1

3

si (t)  ci (t)

4

si (t)  ci (t)

5

ci (t)

1

6

si (t)  ci (t)  ci (t)  si (t)

1 1

si (t )ci (t )

Mobile transmitter.

1

1 1 1 1 1 1

1

1

1

1 1

1

1 1 1 1

1 1

1

1 1

1

1 1 1 1

1 1

1

1

si (t )ci (t )ci (t )

ci (t ) Figure 11.11

1

1 1 1 1 1 1

1

1 1

1

1

1 1 1 1

1

si (t )

11.10

Requirements of Spreading Codes

341

gates are cheaper than multiplication circuits. However, for modeling purposes, the multiplication method is usually easier to formulate and understand than the modulo-2 approach.) We notice that for the output of the receiver to be identical to the original data, the following relationship must be satisfied: si(t)ci(t)ci(t)  si(t)

(11.31)

In other words, ci(t)ci(t) must be equal to 1. Note that ci(t) is a binary sequence made up of 1s and 1s, therefore If ci(t)  1, ci(t)ci(t)  1

(11.32)

If ci(t)  1, ci(t)ci(t)  1

(11.33)

In these discussions we assume that there is no propagation delay and no other processing delay incurred between the transmitter and receiver input. Thus the code copy used at the receiver is perfectly lined up with the initial code used at the transmitter. The two codes are said to be in phase or in synchronization. In practice, however, propagation delay and other processing delays ( i) occur between the transmitter and the receiver. Therefore, the receiver may be timeshifted relative to the initial code at the transmitter. The two codes are no longer in synchronization. As a result, the output of the receiver will no longer be identical to the original data, si(t). In order to recover the original data, si(t), we must “tune” the receiver code sequence to that of the incoming code from the transmitter. In other words, we must time-shift the receiver code in order to line it up with the incoming code. It should be noted that by synchronizing or “tuning” the receiver code to the phase of the incoming code ci(t  i), the original data (shifted by propagation delay) can now be recovered at the output of the receiver. In these discussions, the data sequence and code sequence are assumed to have the same length (one code bit for each data bit) and are used for encrypting the data bits. The process is referred to as bit scrambling.

11.10

Requirements of Spreading Codes

To spread the data sequence, the code sequence must be much faster than the data sequence, and exhibit some random properties. By multiplying the data sequence with the faster code sequence, the resulting product yields a sequence with more transitions than the original data. Using

342

11

Spread Spectrum (SS) and CDMA Systems

suitable random-like codes, the resulting sequence will have the same rate as the code sequence. It is desirable to use a set of orthogonal codes (see Appendix D) to provide good isolation between users. However, in practice, the codes used are not perfectly orthogonal, but they exhibit good isolation characteristics, i.e., they have low cross-correlation.

11.11

Multipath Path Signal Propagation and Rake Receiver

In the absence of a direct line-of-sight signal from the base station (BS) to the mobile station (or the received signal at the mobile station from the base station), the signal is made up of the sum of many signals, each travelling over a separate path. Since these path lengths are not the same, the information carried on the radio link experiences a spread in delay as it travels between the base station and the mobile. In addition, to delay spread, the same multipath environment causes severe local variations in signal strength as these multipath signals are added constructively and destructively at the receiving antenna. This effect is called Rayleigh fading (see Chapter 3). The movement of a mobile causes each received signal to be shifted in frequency as a function of the relative direction and speed of the mobile. This effect is called Doppler shift (see Chapter 3). Multipath is treated as causing delayed versions of the signal to add to the system noise when the differential delay exceeds the chip time, Tc. Substantial performance improvement can occur by detecting each additional path separately, thereby enabling the signals to be combined coherently. A receiver can be implemented to resolve each individual path such that the paths can be combined to produce an overall gain. This type of receiver is known as a Rake receiver. In the Rake receiver for user # 1 (refer to Example 11.2) baseband demodulated signal Z(t) is the sum of N signals which arrives on N different paths (see Figure 11.12). We consider path 2, the multiplication of Z(t) by c1(t  2). The integration starts at time 2 and ends at Tb, to yield the peak response for the path 2. (The output of the integrator is the value of the correlation function of c1(t) for a particular delay. For path 2, this delay is zero, whereas for the other paths the delay exceeds the time duration of a chip.) The contributions from other paths average out to 0, since the differential delays exceed the chip duration, Tc. The response from each path is summed to produce the stronger signal. We illustrate this concept with Example 11.6. Example 11.6 We consider the downlink in Example 11.2 where the demodulated signal at the mobile is z(t)  m1(t)  m2(t)  m3(t) (see Figure 11.13(a)) for mobile 1, 2, and 3. We assume two equal-strength paths and write the demodulated signal as Z(t)  z(t)  z(t2Tc) (see Figure 11.13(c)). The differential delay between these two paths is taken as 2Tc for simplicity. We will show how mobile 1 will

11.11

Multipath Path Signal Propagation and Rake Receiver

Path 1

Tx

343

Rx

Path 2 Path 3 Path N

c1(t)

c1(t2) z(t )

c1(t3)

cos ct

c1(tN)

Figure 11.12

Tb

Integrate and Dump Tb second

Hold Until TbN Tb2

Integrate and Dump Tb second

Hold Until TbN Tb3

Integrate and Dump Tb second

Hold Until TbN



Decide b1(t )

TbN Integrate and Dump Tb second

Hold Until TbN

Simplified Rake receiver for user 1 (equal gain combining).

detect its information using a two-path Rake receiver. The results are shown in Figures 11.13(a), 11.13(b), 11.13(c), and 11.14. Solution Individual path outputs (Tables 11.6 and 11.7) yields an error in a particular bit position. The Rake combining strengthens the signal and removes the error, as shown in Table 11.8. Table 11.6 Z(t)  c1(t). 5Tb

2Tb

4

4

8

8

12

Detected bit value

1

1

0

1

1

Actual bit value

1

0

0

1

1

Value of integration at end of bit period

3Tb

4Tb

Tb

344

11

Spread Spectrum (SS) and CDMA Systems

(a)

3 z(t) 2  m1(t) 1  m2(t)  m3(t)

t

1 2 3

Tb

Tb

Tb

Tb

Tb

(b)

3 2 1 z(t  2Tc)

t

1 2 3

Tb

Tb

Tb

Tb

Path 2: z(t  2Tc) Figure 11.13(a, b)

Mobile #1 sequence locked on path 1.

Tb

11.11

Multipath Path Signal Propagation and Rake Receiver

345

(c)

Z(t )  Path 1  Path 2, i.e., [z(t )  z(t 2Tc)]

6 4 2 Z (t )  z (t )  2 z(t 2Tc)

t

4 6

Tb

Tb

Tb

Tb

Tb

Tc

(d)

1 t

c1(t) 1

4 2 Z(t ).c1(t)

t

2 4 6

Tb

Figure 11.13(c, d)

Tb

(Continued).

Tb

Tb

Tb

346

11

Spread Spectrum (SS) and CDMA Systems

Tc

1 c1(t 2Tc )

t 1

4 2 Z(t).c1(t 2Tc )

t 2 4 6 2Tc

Figure 11.14

Tb

Tb

Tb

Tb

Tb

Mobile #1 sequence locked on path 2.

Table 11.7 Z(t)  c1(t  2Tc). Tb

2Tb

3Tb

4Tb

5Tb

8

8

12

4

12

Detected bit value

1

0

0

0

1

Actual bit value

1

0

0

1

1

Tb

2Tb

3Tb

4Tb

5Tb

Path 1: Integrator output

4

4

8

8

12

Path 2: Integrator output

8

8

12

4

12

12

4

20

4

24

Value of integration at end of bit period

Table 11.8 Sum of path 1 and path 2 integrator.

Sum of integrator outputs (Rake receiver output) Detected bit value

1

0

0

1

1

Mobile 1 bits

1

0

0

1

1

11.13

TIA IS-95 CDMA System

11.12

347

Critical Challenges of CDMA

Code division multiple access (CDMA) is based on DSSS. CDMA is more complex than other multiple access technologies and as such poses several critical challenges. • All users in a given cell transmit at the same time in the same frequency

band. Can they be made not to interfere with each other? • Will a user who is near the base station saturate the base station altogether so that it cannot receive users who are farther away (known as near-far problem)? • CDMA uses a reuse factor of one. This means that the same frequency is used in adjacent cells. Can the codes provide sufficient separation for this to work well in most situations? • CDMA uses soft handoffs where a moving user can receive and combine signals from two or more base stations at the same time. What is the impact on base station traffic handling ability?

11.13

TIA IS-95 CDMA System

Qualcomm proposed the CDMA radio system for digital cellular phone applications. It was optimized under existing U.S. mobile cellular system constraints of the advanced mobile phone system (AMPS). The CDMA system uses the same frequency in all cells and all sectors. The system design has been standardized by the TIA as IS-95 and many equipment vendors sell CDMA equipment that meet the standard. The IS-95 CDMA system operates in the same frequency band as the AMPS using frequency division duplex (FDD) with 25 MHz in each direction.* The uplink (mobile to base station) and downlink (base station to mobile) bands use frequencies from 869 to 894 MHz and from 824 to 849 MHz, respectively. The mobile station supports CDMA operations on the AMPS channel numbers 1013 through 1023, 1 through 311, 356 through 644, 689 through 694, and 739 through 777, inclusive. The CDMA channels are defined in terms of an RF frequency and a code sequence. Sixty-four Walsh codes (see Appendix D) are used to identify the forward channels, whereas unique long PN code offsets are used for the identification of the reverse channels. The modulation and coding features of the IS-95 CDMA system are listed in Table 11.9. Modulation and coding details for the forward and reverse channels differ. Pilot signals are transmitted by each cell to assist the mobile radio to acquire and track the cell site downlink signals. The strong coding helps these radios to operate effectively at an Eb /N0 ratio of a 5 to 7 dB range. The CDMA system (IS-95) uses power control and voice activation to minimize mutual interference. Voice activation is provided by using a variable rate vocoder (see Chapter 8) which for Rate set 1 codec operates at a maximum rate of 8 kbps to a * The frequency spectrum for the A-System cellular service provider is split such that the spectrum is not divisible by 1.25 MHz. Thus the A-System cellular provider cannot partition the spectrum into ten 1.25 CDMA channels. This restriction is not imposed for the B-System, however.

348

11

Spread Spectrum (SS) and CDMA Systems

Table 11.9 Modulation and coding feature of IS-95 CDMA system. Modulation

Quadrature phase-shift keying (QPSK)

Chip rate

1.2288 Mcps

Nominal data rate

9600 bps, full rate with Rate Set 1

Filtered bandwidth

1.25 MHz

Coding

Convolution with Viterbi decoding

Interleaving

With 20-msec span

minimum rate of 1 kbps, depending on the level of voice activity. With the decreased data rate, the power control circuit reduces the transmitter power to achieve the same bit error rate. A precise power control, along with voice activation circuit, is critical to avoid excessive transmitter signal power that is responsible for contributing the overall interference in the system. The Rate set 2 coding algorithms at 13 kbps are also supported. A bit-interleaver with 20 msec span is used with error-control coding to overcome multipath fading and shadowing (see Chapter 3). The time span used is the same as the time frame of voice compression algorithm. A Rake receiver used in the CDMA radio takes advantage of a multipath delay greater than 1 s, which is common in cellular/personal communication service networks in urban and suburban environments.

11.13.1 Downlink (Forward) (BS to MS) The downlink channels include one pilot channel, one synchronization (synch) channel, and 62 other channels including up to 7 paging channels. (If multiple carriers are implemented, paging channels and synch channels do not need to be duplicated). The information on each channel is modulated by the appropriate Walsh code and then modulated by a quadrature pair of PN sequences at a fixed

PN

X

Coherent Carrier cos (2 fct )

Figure 11.15

X

X

Walsh i

Application of PN sequence and Walsh code in CDMA.

11.13

TIA IS-95 CDMA System

349

Walsh i at 1.2288 Mcps

Modulation Symbol 19.2 ksps Traffic Channel Bits

Convolutional Encoder

Symbol, Repetition

Block Interleaver

Power Control Bits

800 bps MUX

Modulo 2 Sum

Modulo 2 Sum 19.2 ksps

Long Code Mask for User m

Long Code Generator

Decimator

I-Channel Pilot PN Sequence 1.2288 Mcps

Signal from one channel

Modulo 2 Sum

I

Base–band Filter

Adder with gain control

cos (2 fct)

I(t )

Adder Modulo 2 Sum

Q

Q -Channel Pilot PN Sequence 1.2288 Mcps

Figure 11.16

800Hz

Decimator

1.2288 Mcps

Base–band Filter Filtered Q signal from other channels

Adder with gain control

cdma(t ) to linear amplifier

Q(t )

sin (2 fct )

Forward traffic channel.

chip rate of 1.2288 Mcps (see Figure 11.15). The pilot channel is always assigned to code channel number zero. If the synch channel is present, it is given the code channel number 32. Whenever paging channels are present, they are assigned the code channel numbers 1 through 7 (inclusive) in sequence. The remaining code channels are used by forward traffic channels (see Figure 11.16). The synch channel operates at a fixed data rate of 1200 bps and is convolutionally encoded to 2400 bps, repeated to 4800 bps, and interleaved. The forward traffic channels are grouped into sets. Rate set 1 has four rates: 9600, 4800, 2400, and 1200 bps. Rate set 2 contains four rates: 14,400, 7200, 3600, and 1800 bps. All radio systems support Rate set 1 on the forward traffic channels. Rate set 2 is optionally supported on the forward traffic channels. When a radio system supports a rate set, all four rates of the set are supported. Speech is encoded using a variable rate vocoder (see Chapter 8) to generate forward traffic channel data depending on voice activity. Since frame duration is

350

11

Spread Spectrum (SS) and CDMA Systems

fixed at 20 ms, the number of bits per frame varies according to the traffic rate. Half rate convolutional encoding is used, which doubles the traffic rate to give rates from 2400 to 19,200 bits per second. Interleaving is performed over 20 ms. A long PN code of 242  1 ( 4.4  1012) is generated using the user’s electronic serial number (ESN) embedded in the mobile station long code mask (with voice privacy, the mobile station long code mask does not use the ESN). The scrambled data is multiplexed with power control information which steals bits from the scrambled data. The multiplexed signal on the traffic channel remains at 19,200 bps and is modulated at 1.2288 Mcps by the Walsh code, Wi, assigned to the ith user traffic channel. The signal is spread at 1.2288 Mcps by quadrature pseudo-random binary sequence signals, and the resulting quadrature signals are then weighted. The power level of the traffic channel depends on its data transmission rate. The paging channel data is processed in a similar manner to the traffic channel data. However, there is no variation in the power level on a per frame basis. The paging channels provide the mobile stations with system information and instructions, in addition to acknowledging messages following access requests on the mobile stations’ access channels. The 42-bit mask is used to generate the long code. The paging channels operate at a data rate of 9600 or 4800 bps. All 64 channels are combined to give single I and Q channels. The signals are applied to quadrature modulators and resulting signals are summed to form a QPSK signal, which is linearly amplified. The pilot CDMA signal transmitted by a base station provides a reference for all mobile stations. It is used in the demodulation process. The pilot signal level for all base stations is much higher (about 4 to 6 dB) than the traffic channel. The pilot signals are quadrature pseudo-random binary sequence signals with a period of 32,768 chips. Since the chip rate is 1.2288 Mcps, the pilot pseudo-random binary sequence corresponds to a period of 26.66 ms, which is equivalent to 75 pilot channel code repetitions every 2 seconds. The pilot signals from all base stations use the same pseudo-random binary sequence, but each base station is identified by a unique time offset of its pseudo-random binary sequence (short code). These offsets are in increments of 64 chips providing 512 unique offset codes. These large numbers of offsets ensure that unique base station identification can be obtained, even in a dense microcellular environment. A mobile station processes the pilot channel to find the strongest multipath signal components. The processed pilot signal provides an accurate estimation of time delay, phase, and magnitude of the multipath components. These components are tracked in the presence of fast fading, and coherent reception with combining is used. The chip rate on the pilot channel and on all frequency carriers is locked to precise system time by using the global positioning system (GPS). Once the mobile station identifies the strongest pilot offset by processing the multipath components from the pilot channel correlator, it examines the signal on its synch channel which is locked to the pseudo-random binary sequence

11.13

TIA IS-95 CDMA System

351

signal on the pilot channel. Since the synch channel is time aligned with its base station’s pilot channel, the mobile station finds the information pertinent to this particular base station. The synch channel message contains time-of-day and long code synchronization to ensure that long code generators at the base station and mobile station are aligned and identical. The mobile station now attempts to access the paging channel and listens for system information. The mobile station enters the idle state when it has completed acquisition and synchronization. It listens to the assigned paging channel and is able to receive and initiate calls.

11.13.2 Uplink (Reverse) (MS to BS) The uplink channel is separated from the downlink channel by 45 MHz at cellular frequencies and 80 MHz at PCS frequencies (1.8 to 1.9 GHz). The uplink uses the same 32,768 chip code as is used on the downlink. The two types of uplink channels are the access channel and reverse traffic channels (see Figure 11.17). The access channel enables the mobile station to communicate nontraffic information, such as originating calls and responding to paging. The access rate is fixed at 4800 bps. All mobile stations accessing a radio system share the same frequency assignment. Each access channel is identified by a distinct access channel long code sequence having an access number, a paging channel number associated with the access channel, and other system data. Each mobile station uses a different PN code; therefore, the radio system can correctly decode the information from an individual mobile station. Data transmitted on the reverse traffic channel is grouped into 20 ms frames. All data on the reverse traffic channel is convolutionally encoded, symbol repeated, block interleaved, and modulated by Walsh symbols transmitted for each six-bit symbol block. The symbols are from the set of the 64 mutually orthogonal waveforms. The reverse traffic channel for Rate set 1 may use either 9600, 4800, 2400, or 1200 bps data rates for transmission. The transmission varies proportionally with the data rate, being 100% at 9600 bps to 12.5% at 1200 bps. An optional second rate set is also supported in the PCS version of CDMA and new versions of cellular CDMA. The actual burst transmission rate is fixed at 28.8 ksps. Since six code symbols are modulated as one of 64 modulation symbols for transmission, the modulation symbol transmission rate is fixed at 4800 modulation symbols per second. This results in a fixed Walsh chip rate of 307.2 kcps. The rate of spreading PN sequence is fixed at 1.2288 Mcps, so that each Walsh chip is spread by 4 PN chips. Table 11.10 provides the signal rates and their relationship for the various transmission rates on the reverse traffic channel. Following orthogonal spreading, the reverse traffic channel and access channel are spread in quadrature. Zero-offset I and Q PN sequences are used for spreading. These sequences are periodic (short code) with 32,768 PN chips in length and are based on characteristic polynomials gI(x) and gQ(x).

352

Primary, Secondary and Signalling Reverse Traffic Channel Information Bits Convolutional Encoder; rate 1/3 9.6 kbps

Frame Data Rate To Modulator Symbol Repetition

28.8 ksps

Block Interleaver

28.8 ksps

64-ary Orthogonal Modulator

28.8 ksps

Data Burst Randomizer

Modulo 2 Sum

4.8 ksps (307.2 kcps)

PN chip 1.2288 Mcps Long Code Generator

Long Code Mask I -Channel Pilot PN Sequence 1.2288 Mcps

cos (2 fct) I

From Either Access Channel or Traffic Channel

Adder

Q Modulo 2 Sum

D

Base–band Filter

Delay of 1/2 PN Chip  406.9 ns Q -Channel Pilot PN Sequence 1.2288 Mcps

Figure 11.17

I(t )

Base–band Filter

Modulo 2 Sum

Reverse traffic channel.

Q(t ) sin (2 fct)

cdma(t )

11.13

TIA IS-95 CDMA System

353

Table 11.10 CDMA reverse traffic channel modulation parameters (rate set 1). Parameter

9600 bps

4800 bps

2400 bps

1200 bps

Units

PN chip rate

1.2288

1.2288

1.2288

1.2288

Code rate

1/3

1/3

1/3

1/3

Mcps bits per code symbol

Transmitting duty cycle

100

50

25

12.5

%

Code symbol rate

3  9600  28,800

28,800

28,800

28,800

Modulation

6

6

6

6

Modulation symbol rate

28,800/6  4800

4800

4800

4800

symbol per second

Walsh chip rate

64  4800  307.2

307.2

307.2

307.2

kcps

Mod. symbol duration

1/4800  208.33

208.33

208.33

208.33

s

PN chips/ code symbol

12,288/288  42.67

42.67

42.67

42.67

PN chip per code symbol

PN chips/ mod. symbol

1,228,800/ 4800  256

256

256

256

PN chip per mod. symbol

PN chips/ Walsh chip

4

4

4

4

PN chips per Walsh chip

symbol per second code symbol per mod. symbol

The maximum-length linear feedback register sequences I(n) and Q(n), based on these polynomials, have a period of 215 1 and are generated by using the following recursions: I(n)  I(n  15)  I(n  8)  I(n  7)  I(n  6)  I(n  2)

(11.34)

based on gI(x) as the characteristic polynomial, and Q(n)  q(n  15)  q(n  12)  q(n  11)  q(n  10)  q(n  9)  q(n  5)  q(n  4)  q(n  3)

(11.35)

based on qQ(x) as the characteristic polynomial, where I(n) and Q(n) are binary numbers (0 and 1) and the additions are modulo-2.

354

11

Spread Spectrum (SS) and CDMA Systems

To obtain the I and Q sequences, a 0 is inserted in I(n) and Q(n) after 14 consecutive 0 outputs (this occurs only once in each period). Therefore, the short PN sequences have one run of 15 consecutive 0 outputs instead of 14. The chip rate for the short PN sequence is 1.2288 Mcps and its period is 26.666 ms. There are exactly 75 repetitions in every 2 seconds. The spreading modulation is OQPSK (see Figure 11.18). The data spread by Q PN sequence is cos(2 fct  /4)

(1, 1) (1, 1)

(aQ, aI)

sin(2 fct  /4)

(1, 1)

(1, 1)

cos(2 fct  /4)

(1, 1) (1, 1)

(aQ, aI)

  /4 sin(2 fct  /4)

(1, 1)

Figure 11.18

(1, 1)

Signal constellation and phase transition of OQPSK and QPSK.

11.13

TIA IS-95 CDMA System

355

delayed by half a chip time (406.901 ns) with respect to the data spread by I PN sequence. Table 11.11 describes the characteristics of OQPSK. Table 11.12 defines the signal rates and their relationship on the access channel. Each base station transmits a pilot signal of constant power on the same frequency. The received power level of the pilot signal enables the mobile station to adjust its transmitted power such that the base station will receive the signal at the requisite power level. The base station measures the mobile station’s received power and informs the mobile station to make the necessary adjustments to its transmitter power. One command every 1.25 ms allows to adjust the transmitted power from the mobile station in steps of 1 dB. The base station uses frame errors reported by the mobile station to increase or decrease its transmitted power. In summary, an IS-95 CDMA system operates with a low Eb /N0 ratio, exploits voice activity and uses the same frequency in all sectors of all cells. Each sector has Table 11.11 Reverse CDMA channel I and Q mapping. I

Q

Phase

0

0

/4

1

0

(3 )/4

1

1

(3 )/4

0

1

/4

Table 11.12 CDMA access channel modulation parameters. Parameter PN chip rate Code rate Code symbol repetition Transmit duty cycle Code symbol rate Modulation

4800 bps 1.2288 1/3 2 100 28,800 6

Units Mcps bits/code symbol symbols/code symbol % sps code symbol/mod. symbol

Modulation symbol rate

4800

sps

Walsh chip rate

307.2

kcps

Mod. symbol duration

208.33

s

PN chips/code symbol

42.67

PN chip/code symbol

PN chips/mod. symbol

256

PN chip/mod. symbol

4

PN chips/Walsh chip

PN chips/Walsh chip

356

11

Spread Spectrum (SS) and CDMA Systems

up to 64 CDMA channels on each carrier frequency. It is a synchronized system with Rake receivers to provide path diversity at the mobile station and at the cell site.

11.14 Power Control in CDMA A proper power control on both the uplink and downlink has several advantages: • System capacity is improved or optimized. • Mobile battery life is extended. • Radio path impairments are properly compensated for. • Quality of service (QoS) at various bit rates can be maintained.

The reverse link (uplink) uses a combination of open loop and closed loop power control to command the mobile station to make power adjustments (see Figure 11.19). The mobile station and the base station receiver measure the received power and use the measurements to maintain a power level for adequate performance. The mobile unit measurement is part of the open loop power control while the base station measurement is part of the closed loop power control. In the closed loop mode, the mobile station transmitter power is controlled by a signal from the base station site. Each base station demodulator measures the received SNR for that mobile station and sends a power command either to increase or decrease mobile station power. The measure-command-react cycle is performed at a rate of 800 times per second for each mobile station in IS-95. The power adjustment command is combined with the mobile’s open loop estimate and the result is used to adjust the transmitter gain. This solves the nearfar interference problem, reduces interference to other mobiles using the same

Weak Signal

Power Control Feedback

Up to 80 dB difference Handset Base Station Power Control Feedback

Signal From Handset Close to Base Station Arrives at Base Station with Little Attenuation

Figure 11.19

Power control in CDMA.

Strong Signal

11.14

Power Control in CDMA

357

CDMA radio channel, helps to overcome fading, and conserves battery power in portable and mobile units. On the uplink, the objective of the mobile station is to produce a nominal received power signal at the base station receiver. Regardless of the mobile’s position or propagation loss, each mobile should be received at the base station with almost the same power level. If the mobile’s signal arrives at the base station with a lower power level than the required power level, its error rate performance will be high. On the other hand, if the mobile’s signal is too high, it will interfere with other users with the same CDMA radio channel causing performance degradation unless the traffic load is decreased. Similarly, a combination of open loop and closed loop power control is used on the forward link (downlink) to keep SNR at the mobile almost constant. Forward link power control mitigates the corner problem. Mobiles at the edges of cells normally require more power than those close to the center of the base station for two reasons: more transmission loss and more interference from adjacent base stations. This is known as the corner problem. Forward link power control minimizes interference to mobiles in the same base station (in multipath environments) as well as mobiles in other base stations. Using the downlink power control, the base station transmits the minimum required power, hence, minimizes the interference to mobiles in the surrounding base stations. The outer loop power control is the finer power control over the closed loop power control. It adjusts the target signal-to-interference ratio (SIR) in the base station according to the needs of the individual radio links and aims at a constant quality, which is usually defined as a certain target bit error rate (BER) or frame error ratio (FER). The required SIR depends on the mobile speed and multipath profile. The outer loop power control is typically implemented by having the base station to each uplink user data frame with frame quality indicator, such as a cyclic redundancy check (CRC) result, obtained during decoding of the particular user data frame.

11.14.1 Open Loop Power Control In the open loop power control, the mobile uses the received signal to estimate the transmission loss from the mobile unit and the base station. p (r) pt

r T(r)  10 log 

(11.36)

L(r)  T(r)  Gt  Gr

(11.37)

where: pt  transmitter’s power output pr(r)  received power at distance r from the transmitter Gt  antenna gain of transmitter

358

11

Spread Spectrum (SS) and CDMA Systems

Gr  antenna gain of receiver L(r)  path loss at distance r T(r)  reverse link transmission power at distance r This estimate is used to allow a rapid response to a sudden improvement in the channel while disallowing a rapid response to a sudden degradation in the channel. This is important because if the channel for one mobile unit suddenly improves, then the signal received at the base station from this mobile unit suddenly increases in power and causes additional interference to all other mobiles sharing the same CDMA channel. The approach to solve this problem is to tolerate a temporary degradation in one mobile unit performance in order to prevent degradation of all mobile units. The signal-to-noise ratio for the mobile at distance r is given as: pm(r)T(r)

t SNR(r)   m

(N0Bw)cell  (M/fr  1)  vf  pt (r)T(r)

(11.38)

where: pm t (r)  mobile power amplifier output T(r)  reverse link transmission power at distance r M  number of users per cell vf  channel activity factor fr  frequency reuse efficiency (N0Bw)cell  cell thermal noise The mobile estimates the T(r) from its measurements of the total received power and the prior knowledge of the base station power amplifier output. The total received power can be expressed as: c c pm r (r)  pt  T(r)  (N0Bw)cell  IocBw  (1  1  2)[pt  L(r)] dB (11.39)

where: Pct  base station power amplifier output (dB) Pm r (r)  mobile received total power (dB) Ioc  spectral density of interference from other cells (dBm/Hz) 1  ratio of power from other cells to power from the mobile’s home cell 2  ratio of thermal noise to power from mobile’s home cell Bw  bandwidth (Hz) To minimize the power transmitted by each mobile, (SNR)min for all reverse channels should be maintained. pm(r)  L(r)

t SNRmin   m

(N0Bw)  (M/fr  1)[(vf  pt )  L(r)]

(11.40)

11.14

Power Control in CDMA

359

m

pr (r) pm t (r) 

SNRmin 

(1  1  2)  pct  m [vf  pm t (r)]  pr (r)

(11.41)

(N0Bw)  ((M/fr)  1) c (1  1  2)  pt

There are two extreme cases: • The mobile is close to its base station in this case, the total received power

is mainly due to its base station and c L(r) (pm r (r))/pt

• The mobile is close to the edge of its base station; in this case the total

received power is the contribution of more than one base station. Worst case conditions are when the mobile is equidistant from three adjacent base stations. Assuming that all base stations transmit the same power, then (1  1  2)  3 The quantity pm(r) (1  1  2!)  pt

r (N0Bw)  ((M/fr)  1)vf  pm t (r)   c

depends only on base station loading ((M/fr)  1)  vf  pm(r)

pm(r) (1  1  2)  pt

t r 1 1     c 

(N0Bw)

1

where:   cell loading  M/Mmax M  number of users in the cell Mmax  maximum capacity of the cell m c SNRmin  pm t (r)  pr (r)  pt  10 log (1  1  2)  (N0Bw)cell

 10 log (1  ) dB

(11.42)

In order to maintain a constant SNR at the base station, the mobile should transmit

360

11

Spread Spectrum (SS) and CDMA Systems

c   pm t (r)  SNR min  pt   N0Bw cell  10 log  1     10 log  1  1  2 

 pm r (r) dB

(11.43)

N  E

b c pm t (r)    10 log Gp  pt   N0Bw cell  10 log  1     10 log  1  1  2  t

 pm r (r)

(11.44) 

pmt (r)  K  prm(r) dB

(11.45)

where: 

N  E

K  b  10 log Gp  pct   N0Bw cell 10 log  1     10 log  1  1  2  t

As an example, let us consider IS-95, in which: Gp  128, Eb/Nt  7 dB,

pct  25 watts, N0  174 dBm/Hz, Bw  1.23 MHz,   0.5, cell noise 

figure Nf  4 dB, 1  1  2  2, then (K)  73 dB (Note: this value is used in Example 11.7). m pm t (r)  73  pr (r) dB

(11.46)

 The constant (K) is the part of the broadcast message that is sent to the mobile by the base station on the paging channel. The speed at which the open loop power control tracks the changes in the channel depends on the time constant of the automatic gain control (AGC) filter.

Example 11.7 An IS-95 CDMA mobile measures the signal strength from the base stations as 97 dBm, what should the mobile transmitter power be set to as a first approximation? After the connection is made with this power level, the base station requires the mobile station to change its power to 18 dBm. How long would it take for the mobile station to make this change? Solution 

m m   pm t (r)  K  pr (r)  73  pr (r)  73  97  24 dBm

Power reduction  24  18  6 dBm

11.15

Softer and Soft Handoff

361

Mobile requires 6 decrements each at 1.25 ms (1/800 sec)  Time required  6  1.25  7.5 ms

11.15

Softer and Soft Handoff

During soft handoff, a mobile station is in the overlapping cell coverage area of two sectors belonging to two different base stations. The communications between mobile station and base station occur concurrently via two air interface channels from each base station separately. Both channels (signals) are received at the mobile station by maximal combining Rake processing (see Figure 11.20). Soft handoff occurs in about 20–40% of calls. Soft handoffs are an integral part of CDMA design. The determination of which pilots will be used in the soft handoff process has a direct impact on the quality of the call and the capacity of the system. Therefore, setting soft handoff parameters is a key element in the system design for CDMA. In the uplink direction, soft handoff differs significantly from softer handoff: the code channel of the mobile station is received from both base stations, but the received data is routed to the base station controller (BSC) for combining. This is done so that the same frame reliability indicator as provided for outer loop power To Other Switch

Soft Handoff Circuit

Switch

R

Base Station Controller

Base Station Controller

R

R Transceiver

Transceiver

Transceiver New Link

Old Link Handset

M : Measurements Are Made by the Handset Using Additional Correlators R : Transfer Requests Are Sent to the Old Cell with a Degraded Link

Figure 11.20

Soft handoff in CDMA.

M Transfer Decision

Transceiver

362

11

Spread Spectrum (SS) and CDMA Systems

control is used to select the better frame between two possible candidates within the BSC. A brief description of each type of pilot set is given below: • The active set is the set of pilots associated with downlink traffic channels

assigned to the mobile units. The active set can contain more than one pilot because a total of three carriers, each with its own pilot, could be involved in a soft handoff process. • The candidate set consists of the pilots that the mobile unit has reported are of a sufficient signal strength to be used. The mobile unit also promotes the neighbor set and remaining set pilots that meet the criteria to the candidate set. • The neighbor set is the list of the pilots that are not currently on the active or candidate pilot lists. The neighbor set is identified by the base station via the neighbor list and neighbor list update messages. • The remaining set contains all possible pilots in the system that can possibly be used by the mobile unit. However, the remaining set pilots that the subscriber unit looks for must be a multiple of Pilot_Inc. The parameters used to control the movement of a pilot from a neighbor to a candidate, to active, and then back to neighbor set are given below: 1. Pilot strength exceeds T_ADD and the mobile unit sends a pilot strength measurement message (PSMM) and transfers the pilot to the candidate set. 2. The pilot strength drops below T_DROP and the mobile unit begins the handoff drop time (T_TDROP). 3. T_COMP is used into decision matrix for adding and removing pilots from the neighbor, candidate, and active set. For more details about power control and soft handoff in CDMA, refer to [5]. Example 11.8 Given the current active set contains pilots P1, P2, P3, and P4. At a particular time, the mobile station measures the signal strength of P1, P2, P3, and P4 as 95 dBm, 100 dBm, 101 dBm, and 105 dBm, respectively. Pilot P5 is in the candidate set. The mobile station measures P5 signal strength as 102 dBm at the same time. Determine a possible pair of the SOFT-SLOPE and ADD-INTERCEPT values that will trigger the mobile station to send a PSMM to the base station. The mobile station is IS-95B compliant and uses the following criterion to add the pilot from the candidate set to the active set.

11.15

Softer and Soft Handoff

363

10 log (Pcj)  Max



  NA

(SOFT-SLOPE)  10 log

 

Pai  (ADD-INTERCEPT) , TADD

i1

where: Pcj  received power of the jth pilot in the candidate set Pai  received power of the ith pilot in the active set Na  number of pilots in the active set TADD  threshold for adding a pilot from the candidate set into the active set Also, if the mobile station is IS-95 compliant, find the value of T-COMP that could trigger the mobile station to generate a PSMM. Assume mobile receiver sensitivity, i.e., noise power  107 dBm and TADD  13 dB. Solution (a) Pa1  P1  Receiver Sensitivity  95  (107)  12 dBm Pa2  P2  Receiver Sensitivity  100  (107)  7 dBm Pa3  P3  Receiver Sensitivity  101  (107)  6 dBm Pa4  P4  Receiver Sensitivity  105  (107)  2 dBm Pc5  102  (107)  5 dBm 10 log  101.2  100.7  100.6  100.2   14.22  (Max {14.22  (SOFT-SLOPE)  (ADD-INTERCEPT), 13}  5)

Thus, we have the following two relations: 14.22  SOFT-SLOPE  ADD-INTERCEPT 13

and 14.22  SOFT-SLOPE  ADD-INTERCEPT  5

Solving these equations, we get SOFT-SLOPE  0.5 and ADD-INTERCEPT  4

364

11

Spread Spectrum (SS) and CDMA Systems

(b) For an IS-95-compliant mobile station (PcjPai)  0.5  T-COMP

Since P1 P2 P3 P4, we replace P4 [102  (105)]  0.5  T-COMP T-COMP  6 dB  4

11.16

Summary

In this chapter we first discussed the concept of spread spectrum systems and provided the main features of the direct sequence spread spectrum and frequency hop spread spectrum systems. A key component of spread spectrum performance is the calculation of the processing gain of the system, which is the relationship between the input and output signal-to-noise ratio of a spread spectrum receiver. Spread spectrum systems trade bandwidth for processing gain, and code division systems use a variety of orthogonal or almost orthogonal codes to allow multiple users in the same bandwidth. Thus, CDMA systems can have a higher capacity than either analog or time division multiple access (TDMA) digital systems. However, because of practical constraints on CDMA systems, it is not possible to achieve the Shannon bound in system design. The upper bound of the capacity of a CDMA system is limited by the processing gain of the system, receiver modulation performance, power control accuracy, interference from other cells, voice activity, cell sectorization, and the ability to maintain accurate synchronization of the system. We concluded the chapter by discussing the high level features of the IS-95 system and listing the challenges in implementing a CDMA system. Two important aspects of the CDMA system, power control and softer and soft handoff, were also discussed.

Problems 11.1 Define open loop, closed loop, and outer loop power control. 11.2 What are the softer and soft handoffs? 11.3 Differentiate between scrambling and spreading. 11.4 What are the requirements for a DSSS? 11.5 What are the basic differences between the DSSS and FHSS systems?

Problems

365

11.6 Which parameters are used to control the movements of the pilot in a CDMA system? 11.7 How many times are power adjustments performed in an IS-95 CDMA system? 11.8 Determine (S/N)o for a spread spectrum system having bandwidth, Bw  1.2288 MHz, information rate, R  9.6 kb/s and (S/N)i  15 dB. Relate your result to Eb/N0. 11.9 Use the data in Example 11.6 and show that the bit pattern at mobile 2 and 3 can be determined from the demodulated signal using Path 1 and Path 2. 11.10 Consider a case where eight chips per bit are used to generate the Walsh codes (see Appendix D). Mobile stations A, B, C, and D are assigned W1, W2, W3, and W4, respectively. The stations use the Walsh code to send a “1” binary bit and use negative Walsh to send a “0” binary bit. Assume all stations are synchronized in time; therefore, chip sequences begin at the same instant. When two or more stations transmit simultaneously, their bipolar signals add linearly. Consider the following four different cases when one or more stations transmit and show that the receiver recovers the bit stream of stations B and C. (Note: dash () means no transmission by the station). Station (A, B, C, D)

Transmitting Stations

10

CD

111

ABC

11

AB

1100

ABCD

11.11 An IS-95 CDMA mobile measures the signal strength from the base station as 100 dBm, what should the mobile transmitter power be set to as a first approximation? 11.12 Given the current active set contains pilots P1, P2, P3, and P4. At a particular time, the mobile station measures the signal strength of P1, P2, P3, and P4 as 98 dBm, 101 dBm, 103 dBm, and 104 dBm, respectively. Pilot P5 is in the candidate set. The mobile station measures P5 signal strength as 103 dBm at the same time. Determine a possible value of the SOFT-SLOPE, that will cause the mobile station to send a PSMM report to the base station, if the ADD-INTERCEPT value is 4. The mobile station is IS-95B compliant and uses the following criterion to add the pilot from the candidate set to the active set.

366

11

Spread Spectrum (SS) and CDMA Systems

10 log (Pcj)  Max



  NA

(SOFT-SLOPE)  10 log

 

Pai  (ADD-INTERCEPT) , TADD

i1

where: Pcj  received power of the jth pilot in the candidate set Pai  received power of the ith pilot in the active set Na  number of pilots in active set TADD  threshold for adding a pilot from the candidate set into the active set Assume mobile receiver sensitivity i.e., noise power  106 dBm and TADD  14 dB.

11.13 Also, if the mobile station is IS-95 compliant in Problem 11.12, find the value of T-COMP that could cause the mobile station to generate a PSMM report.

References 1. Bhargava, V., Haccoum, D., Matyas, R., and Nuspl, P. Digital Communications by Satellite. New York: John Wiley & Sons, 1981. 2. Dixon, R. C. Spread Spectrum Systems, Second Edition. New York: John Wiley & Sons, 1984. 3. Feher, K. Wireless Digital Communications Modulation and Spread Spectrum Applications. Upper Saddle River, NJ: Prentice-Hall, 1995. 4. Garg, V. K., and Wilkes, J. E. Wireless and Personal Communications Systems. Prentice Hall, 1996. 5. Garg, V. K. CDMA IS-95 and cdma2000. Prentice Hall, 2000. 6. Holma, H., and Toskala, A. (editors). WCDMA for UMTS. New York: John Wiley & Sons, 2000. 7. Lee, W. C. Y. Mobile Cellular Telecommunication Systems. New York: McGraw-Hill, 1989. 8. Pahlwan, K., and Levesque, A. H. Wireless Information Networks. New York: John Wiley & Sons, 1995. 9. Smith, C., and Collins, D. 3G Wireless Networks. New York: McGraw-Hill, 2002. 10. Shannon, C. E. Communications in the Presence of Noise. Proceedings of the IRE. nO. 37, pp. 10–21, 1949. 11. Steele, R. Mobile Radio Communications. New York: IEEE Press, 1992.

References

367

12. Skalar, B. Digital Communications — Fundamentals & Applications. Englewood Cliffs, NJ: Prentice Hall 1988. 13. TIA/EIA IS-95. “Mobile Station — Base Station Compatibility Standard for Dual-mode Wideband Spread Spectrum Cellular System,” PN-3422, 1994. 14. Torrien, D. Principle of Secure Communication Systems. Boston: Artech House, 1992. 15. Virterbi, A. J. CDMA. Reading, MA: Addison-Wesley Publishing Company, 1995. 16. Viterbi, A. J. and Padovani, Roberto. Implications of Mobile Cellular CDMA. IEEE Communication Magazine, vol. 30, no. 12, pp. 38–41, 1992.

This page intentionally left blank

CHAPTER 12 Mobility Management in Wireless Networks 12.1

Introduction

Mobility is important in mobile communication [1–3]. It can be categorized as radio mobility and network mobility. Radio mobility is mainly concerned with the handoff process, whereas network mobility mainly deals with mobile location management (i.e., location updating and paging). As mobiles travels across system boundaries, whether they are cell, location, or a mobile switching center (MSC) areas, the network must be able to locate the mobile subscriber and automatically route the call to him or her [4–6]. Security mechanisms, such as subscriber identity authentication and equipment validation (discussed in Chapter 13) provide a level of protection against fraud to subscribers and service providers [8,13]. The public land mobile network (PLMN) is an integrated service digital cellular network providing wireless access for mobile subscribers to other networks and network services, including other PLMNs. A PLMN area is divided into regions called location areas (LAs). Each LA is made up of one or more cell areas. A mobile station (MS) registers with the visitor location register (VLR) each time it enters a new LA. The mobile station is free to move inside a given LA without a registration (see Figure 12.1). Each cell has a unique identification number called a cell global identification (CGI), which contains location area identification (LAI) and cell identification (CI) (see Figure 12.2) [10–12]. The LAI consists of mobile country code (MCC), mobile network code (MNC), and location area code (LAC). The CI is a specific cell identity within a given location area [7]. Mobility management in wireless networks is the primary set of functions that are supported by the network to enable subscriber mobility. Mobility management enables the network to keep track of the subscriber’s status and location to deliver calls to that subscriber. It also enables the network to authorize a subscriber for service in a given service area. The key component to mobility management is the subscribers’ service profile. The service profile is a database record in the network that contains information about each subscriber. This information includes dynamic data such as current location and status of a subscriber as well as permanent data such as the service profile, international mobile subscriber identity (IMSI), and so on of a subscriber. 369

370

12

Mobility Management in Wireless Networks

GPA

Location Area 1

Location Area 3 Location Area 2 Cell Area

Figure 12.1

Wireless PLMN.

MCC

MNC

LAC

CI

LAI Figure 12.2

12.2

Cell global identification.

Mobility Management Functions

Mobility management generally deals with automatic roaming, authentication, and intersystem handoff. Automatic roaming includes a set of network functions that allow a subscriber to obtain service outside the home service provider area. These functions are automatic and do not require special subscriber actions. The automatic roaming functions are divided into: • Mobile station (MS) service qualification • MS location management • MS state management • Home location register (HLR), and VLR fault recovery

The authentication process requires that end users of the system are authenticated, i.e., the identity of each subscriber is verified (see Chapter 13).

12.3

Mobile Location Management

371

Handoff is one of the essential features that guarantees the subscriber mobility in a mobile network, where the subscriber can move around. Maintaining connection with a moving subscriber is possible with the help of the handoff function. The basic concept is simple: when the subscriber moves from the coverage area of one cell to another, a new connection with the target cell has to be set up and connection with the old cell may be released. Controlling the handoff mechanism is, however, quite a complicated issue in cellular systems.

12.3

Mobile Location Management

Location management schemes are based on subscribers’ mobility and incoming call rate characteristics. Using the location update (LU) procedure the network keeps track of the mobile subscriber’s location to direct the incoming call. The paging process transmits paging messages to all those cells where mobile terminals can be located. There is a trade-off between the LU cost and paging cost. If the LU cost is high, the paging cost will be low, as paging is performed over a small area. On the other hand, if the LU cost is low, the paging cost will be high, as paging is performed over a wider area. Location management uses either the periodic LU or the LU-on-LA-crossing. The VLR stores the LAI, and the HLR keeps the VLR identifier. In the periodic LU method, an MS periodically sends its identity to the network. The drawback of this method is its resource consumption. As an example, if the mobile subscriber does not move from an LA for several hours, resources are consumed unnecessarily. In the LU-on-LA-crossing method, each base station (BS) periodically broadcasts the identity of its LA. The MS is required to regularly listen to network broadcast information and store the current LA identity if the received identity differs from the one stored in the MS. An LU procedure is automatically triggered by the MS. The advantage of this method is that it requires LUs only when the MS actually moves. A highly movable mobile generates a large number of LUs, a low mobility mobile triggers only a few. A hybrid location management scheme combines the two methods. The MS generates its LU each time it detects an LA crossing. However, if no communication has occurred between the MS and the network for a specified duration of time, the MS generates an LU. This is a periodic LU that typically allows the network to recover subscriber location data in case of a database failure. Generally LU procedures involve intra-VLR LU, inter-VLR LU using temporary mobile subscriber identity (TMSI), inter-VLR LU using IMSI, and the IMSI-attached procedure that is triggered when the mobile is powered on in the LA where it was powered off. To minimize the location management cost (i.e., LU plus paging traffic processing), the LA must be carefully designed by using a proper subscriber mobility model discussed next.

372

12

Mobility Management in Wireless Networks

12.3.1 Mobility Model A mobility model describes the occurrence of procedures such as LU and handoff. Several models have been proposed. • Fluid model: This model assumes traffic flow to be like the flow of a fluid.

The model suggests that the amount of traffic flowing out of an area is proportional to the population density of the area, the average velocity of movement, and the length of the area boundary. For an area, the average number of crossings per unit time is given as: L

 

(12.1)

where:   number of crossings per unit of time   average population density   average movement velocity in the area L  perimeter of the area One of the limitations of the fluid model is that it describes aggregate traffic and is difficult to apply to situations where individual movement patterns are desired. The fluid model is more applicable to areas with a large population because it uses the average population density and the average movement velocity of the area. • Markovian model: This is also known as the random walk model. The

Markovian model describes individual movement. In this model, a mobile subscriber either remains within the region or moves to an adjacent region according to a transition probability distribution. One of the limitations of this model is that there is no concept of trips. • Gravity model: Variations of the gravity model have been employed in transportation research to model human movement behavior. Gravity models have been applied to regions of varying sizes, from city models to national and international models. The main difficulty in using the gravity model is that many parameters are required in the calculations; it is therefore difficult to model geography with many regions. Next, we show an application of the simple mobility model based on the fluid flow concept for a wireless network in an urban area with small cells and high user density. We evaluate the impact of LUs on RF resource occupancy at the network level and determine the number of transactions processed by MSC/VLR [9]. The transaction is defined here as the message received or transmitted by the MSC/VLR. Assumptions used in our calculations are: (1) cells are hexagonal, (2) maximum blocking probability is 1%, (3) mobiles are uniformly distributed

12.3

Mobile Location Management

373

in the cell area, and (4) movements of the mobiles are uncorrelated; the directions of their movements are uniformly distributed over 0 to 2. The optimal number of cells (Nopt) per LA is given as: 

C

Cpage

 Nopt     LU

(12.2)

R

where: Cpage  cost of paging (in terms of the number of paging messages required to find an MS) CLU  cost of LUs (in terms of the number of LU messages required to update the location of an MS) R  cell radius   mean mobile velocity in LA Using Equation 12.1 for number of LUs, jLU in an LA perimeter of the jth cell per hour will be: L

jLU   

(12.3)

where:

3



1 1 , length (km) of the cell exposed perimeter in an LA L  6R    2Nc 3

R  cell radius (km) Nc  number of cells per location area   mobile density in the cell (mobiles per km2) The resource occupancy RjLU in the jth cell due to MS LUs is given as:

 3

RjLU  jLU

k1

k p j,k  t LU LU



(12.4)

where: pj,k  percent of LUs in the kth case for the jth cell LU k tLU  average duration of an LU in the kth case (k  1: intra-VLR; k  2: inter-VLR with TMSI; and k  3: inter-VLR with IMSI) jLU  number of transactions processed by MSC/VLR in an LA perimeter of the jth cell per hour

374

12

Mobility Management in Wireless Networks

Table 12.1 Number and duration of transaction for different LU types. Transaction type

Number of transaction/LU

Intra-VLR LU

Duration of a transaction

2

600 ms

Inter-VLR LU with TMSI

14

3500 ms

Inter-VLR LU with IMSI

16

4000 ms

The total number of transactions due to LUs generated in the NLA Np location area perimeter cells (which we number from 1 to NLANp) and processed per hour by the MSC/VLR is given as: NLA Np

TNLU 

jLU

 j1

 3

k1

p j,k  t kLU LU



(12.5)

where: NLA  number of LAs managed by MSC/VLR 

3 N

Np  6 c 3  number of cells located on the perimeter of an LA, and Example 12.1 Using the following data for Groupe Special Mobile/Global System for Mobile Communications 1800, evaluate the impact of LUs on the radio resource and calculate the MSC/VLR transaction load using the fluid flow model. • Density of mobiles in the cell  10,000 mobiles/km2 • Cell radius  500 m • Average moving velocity of a mobile  10 km/hour • Number of cells per LA  10 • Number of LAs per MSC/VLR  5 • Number of transactions and duration of each transaction to MSC/VLR per

LU for different LU types are given in Table 12.1. We consider two cases: (1) an optimistic situation in which generated LUs in a cell are only intra-VLR LUs, and (2) a pessimistic situation where generated LUs in a cell are inter-VLR LUs. Solution Case 1: In the first case, the jth cell located at the border of two LAs is related to the same MSC/VLR, only intra-VLR LUs are processed in the cell.

3



3



500 1 1 1 1 L  6R  6   1.902 km      2Nc 3

1000

210 3

12.3

Mobile Location Management

L

375

10 10,000 1.902

jLU    60,543 LUs per hour    





3

RjLU  jLU

60,543(1 600/1000) 3600

pj,k  tkLU    10.1 Erlangs LU

k1

This requires 18 channels at 1% blocking (refer to the Erlang-B table, Appendix A) or 18/8  2.25 traffic channel (about 1/4 of an RF channel, assuming there are 8 traffic channels per RF channel). Case 2: In this case the jth cell is located at the border of two LAs related to two different VLRs. In this case, only inter-VLR LUs will be processed in the cell. We assume 80% of LUs are with TMSI and 20% of LUs are with IMSI.

 3

RjLU  jLU



60,543 3600





3500 4000 p j,k  tkLU   0.8  0.2   60.45 Erlangs LU

k1

1000

1000

This requires 75 channels at 1% blocking (refer to the Erlang-B table, Appendix A) or 75/8  9.38 traffic channels (about 1.25 RF channels).

MSC/VLR Transaction Load We assume that one LA is in the center of the region and the remaining four LAs are on the border of the region. We also assume that, in the perimeter cells at the border LAs, only intra-VLR LUs are generated. For half of the perimeter cells at the border LAs, only inter-VLR LUs are generated. 

Np  6

3

of an LA.

Nc

3  6



10  3

3  7.9545  8 cells located on the perimeter

Number of cells where inter-VLR LUs occur will be: 0.5 4 Np  2 8  16

Number of cells where intra-VLR LUs occur will be: 4 Nc 16  4 10 16  24

376

12

Mobility Management in Wireless Networks

NLA Np

TNLU 

jLU



j1

 3

p j,k  t kLU LU

k1



 60,543 [2 24 16(0.8 14 0.2 16)]  16.85 106 transactions at peak hour

Transactions at Peak Hour The results show that, under heavy traffic conditions, the impact of LUs can be significant. In terms of radio channels used, we find that between 0.25 to 1.25 GSM RF carriers can be used for LA boundary crossings. Although this load cannot directly cause call blocking on the radio interface (in a GSM1800 network with 4 3 reuse cluster, the average number of RF carriers per cell with 3 operators is about 10), it is nevertheless not a negligible impact on traffic channel (TCH), consumption. In terms of processing at the MSC/VLR side, with a processing load of about 16.85 106 transactions per hour, it is obvious that blocking can rapidly occur with the given scenario. The MSC/VLR resources dedicated to LU processing cannot be used to provide other services. A major concern of operators, where objectives are to provide users with rich and sophisticated services, is to have more processing resources at the MSC/VLR.

12.4

Mobile Registration

The detection of a subscriber in a new serving system is an example of a registration event. The registration is a subset of automatic roaming functions. It primarily consists of MS service qualification and MS location management. These functions enable a subscriber to register with the network; that is, to indicate to the network location register functional entities the location and status of the MS. The registration-related functions generally must be performed before the network can provide service to the subscriber; however, there are exceptions, such as access to emergency services by dialing 911. The process of registering a mobile subscriber is unique to wireless networks. The registration is not used for wireline systems because the subscriber location is fixed and does not change. The types of registration supported in a network are dependent on the protocol used for the air interface between the mobile station and radio system and on the algorithms implemented in the serving network. The air interface standards for advanced mobile phone systems (AMPS), time division multiple access (TDMA), and code division multiple access (CDMA) support different types of registration than does GSM.

12.4

Mobile Registration

377

Registration may be initiated by the MS or the network, or may be implied during MS access. Upon receiving the registration request from the MS, the radio system (base station) constructs the Registration Update Request message and sends it to the network (MSC). The Registration Update Request message contains the MS’s identification and location information and may contain authorization parameters. The network (MSC) may respond with a request for authorization (optional) and finally with a Registration Update Response message. The network (MSC) sends a Registration Update Response message to the BS when a registration procedure has been successfully completed. This message indicates whether the MS’s registration has been accepted or rejected. The message may contain additional parameters to be sent to the MS. Upon receipt of this message, the BS sends an appropriate response to the MS. The three possible results in registration requests are: successful registration, unsuccessful registration, and cancellation of registration. A mobile station registering on an access channel may perform any one of the following registration types: • Distance-based registration: when the distance between the current cell and •

• •

• •



the cell where the mobile last registered exceeds a threshold. Geographic-based registration: whenever a mobile enters a new area of the same system. A service area may be segmented into smaller regions, known as location areas, which are groups of one or more cells. The MS identifies the current location area via parameters transmitted by the MS on the forward control channel. Location-based registration reduces the paging load of the system by allowing the network to page only in the location area(s) where a mobile station is registered. Parameter change registration: when specific operating parameters in the mobile are changed. Periodic registration: when the system sets parameters on the forward control channel to indicate that all or some of the mobile stations must register. The registration can be directed to a specific mobile or a class of mobile stations. Power-down registration: when the mobile is switched off. This allows the network to deregister a mobile immediately upon its power-down. Power-up registration: when power is applied to the mobile, and used to notify the network that the mobile is now active and ready to place or receive calls. Timer-based registration: when a timer expires in the mobile. This procedure allows the database in the network to be cleared if a registered MS does not register after a fixed time interval. The time interval can be varied by setting parameters on the control channel.

378

12

BS

MS

Mobility Management in Wireless Networks

New MSC/VLR

HLR

Old VLR

Registration Determination

1

2

Global Challenge Registration Request

3 Validate RAND

4

REGISTER

5

IS-41 REGNOT

6

Database Update

7

IS-41 REGCANC

8

Confirm

9 REGNOT Response

10 REGISTER

11 12

REGISTER Confirm

Figure 12.3

Call flows for MS registration of all mobiles listening to a control channel.

It should be noted that all registration types are not supported in a network. The following are the call flows for the registration (ANSI-41 standard) of all MSs listening to a control channel (see Figure 12.3): 1. The MS determines that it must register with the system. 2. The MS listens on the control channel for the global challenge, random number (RAND). 3. The MS sends a message to the BS with IMSI, RAND, and other parameters, as needed, in the MS Registration Request. 4. The BS validates RAND. 5. The BS sends a REGISTER message to the new MSC/VLR. 6. If the MS is not currently registered to the serving VLR, the VLR sends a REGistration NOTification (REGNOT) message to the user’s HLR containing the IMSI and other data as needed.

12.4

Mobile Registration

379

7. The MS’s HLR receives the REGNOT message and updates its database accordingly (stores the location of the VLR that sent the REGNOT message). 8. The MS’s HLR sends an IS-41 REGistration CANCel (REGCANC) message to the old VLR where the MS was previously registered so that the old VLR can cancel the MS’s previous registration. 9. The old VLR returns a Confirmation message that includes the current value of the call count (CHCNT). 10. The user’s HLR then returns a REGNOT Response message to the new VLR and passes along information that the VLR needs (e.g., user’s profile, interexchange carrier ID, shared secret key for authentication, and current value of CHCNT). If the registration is a failure (due to invalid IMSI, service not permitted, nonpayment of bill, or other reason), then the REGNOT message will include a failure indication. 11. Upon receiving the successful REGNOT message from the user’s HLR, the VLR assigns a TMSI. The MSC receives the message, retrieves the data and sends a REGISTER message to the BS. 12. The BS receives the REGISTER message and forwards it to the MS to Confirm Registration. Note: For details of various messages refer to the ANSI-41 standard.

12.4.1 GSM Token-Based Registration When an MS registers with the new network, it sends its TMSI and LAI. The LAI informs the system where to find the old VLR. The network then queries the old VLR for data and uses the data to authenticate the MS. The new VLR then communicates with the HLR to update the location of the MS. The HLR sends a registration cancellation message to the old VLR. The operation of the token-based system is slightly different from the ANSI-41 system in that the segmentation of call processing between BS, MSC, and VLR is defined differently. Therefore the call flow does not, in most cases, distinguish where call processing is done. The call flows for token-based registration are shown in Figure 12.4. The following are the steps in the call flows: 1. The MS sends a Registration message to the visiting system with old TMSI and old LAI. 2. The visiting system queries the old VLR for data. 3. The old VLR returns security-related information (e.g., unused triplets and location of HLR). 4. The visiting system issues a challenge to the MS.

380

12

Mobility Management in Wireless Networks

BS, MSC, New VLR MS 1

HLR

Old VLR

Register Query

2

Response 4

3

Unique Challenge Challenge Response

5 6

Assign TMSI Location Update

7

Data Base Update

8 Acknowledgment

9

Registration Cancellation 11 12

Figure 12.4

10

New TMSI ACK

Token-based registration.

5. The MS responds to the challenge. 6. The visiting system assigns a new TMSI. 7. The visiting system sends a message to the HLR with location update information. 8. The HLR updates its location database with the new location of the MS. 9. The HLR acknowledges the message and may send additional securityrelated data (additional security triplets). 10. The HLR sends a Registration Cancellation message to the old system. 11. The visiting system sends an encrypted message to the MS with new TMSI. 12. The MS acknowledges the message. Note: Steps 7–10 and 11–12 can occur in any order. If the old VLR is not reachable for any reason, then the network will request the MS to send its IMSI, and communications with the HLR will then occur.

12.4

Mobile Registration

MS

381

BSS

MSC

VLR

HLR

Mobile turns on RIL3-RR channel request on RACH RIL3-RR IMM assign on AGCH SABM ≤ Identify of the message ≥ (SDCCH)

Causes the mobile to seize a dedicated SDCCH

Establishes the signaling link

UA (SDDCH) RIL3-MM IMSI attach (SDCCH)

MAP/B Attach IMSI MAP/B IMSI Attach ACK

RIL3-RR IMSI attach ACK (SDCCH) RIL3-RR channel release (SDCCH)

IMSI attach ACK BSSMAP clear command

Clear command MSC asks the BSS to release the allocated dedicated resources

The VLR is requested to mark IMSI as active System can now page MS for PSTN/ISDN calls

RIL3-RR DISC (SDCCH) BSSMAP clear command UA (SDCCH) RACH: Random Access Channel AGCH: Access Grant Channel SDCCH: Stand-alone Dedicated Control Channel

Figure 12.5

MAP: Mobile Application Part (See Ch.7) TMSI: Temporary Mobile Subscriber Identity RIL3: Radio Interface Layer 3 BSSMAP: Base Station Subsystem Management Application part

IMSI attach process in GSM.

12.4.2

IMSI Attach and IMSI Detach (Registration and Deregistration) in GSM In GSM, mobile power-up implies IMSI attach which causes a mobile registration. On the other hand, mobile power-down implies IMSI detach causing mobile deregistration. Figures 12.5 and 12.6 show call flows for IMSI attach and detach (for details see GSM standards). 12.4.3 Paging in GSM In this scenario we assume that the mobile is already registered with the system and has acquired a TMSI. It is also assumed that the mobile is located in its home

382

12

MS

Mobility Management in Wireless Networks

BSS

MSC

VLR

HLR

Mobile turns off RIL3-RR channel request on RACH RIL3-RR IMM assign on AGCH

SABM ≤ Identify of the message ≥ (SDDCH)

On AGCH; assigns SDCCH

Establishes the signaling link

UA (SDDCH) RIL3-MM IMSI detach (SDCCH)

BSSMAP complete

MAP/B detach IMSI

The VLR sets IMSI detached flag MAP/D deregister MS MAP/D deregistration accepted

MAP/B ACK IMSI detachment BSSMAP clear command

BSSMAP clear complete

Figure 12.6 IMSI detach process in GSM.

network. A land subscriber dials the directory number of the mobile subscriber. Figure 12.7 shows the call flow for this scenario. 1. The public switched telephone network (PSTN) routes the call to the MSC assigned this directory number. The directory number in the Initial Address Message (IAM) is the MS ISDN Number (MSISDN). 2. The MSC sends the Send Routing Information Message to the HLR to provide the routing information for the MSISDN. 3. The HLR acknowledges the Send Routing Information Message to the MSC. This message contain the MS routing number (MSRN). If the MS is roaming within the serving area of this MSC, the MSRN returned by the HLR will most likely be same as MSISDN. In this scenario we assume that the mobile is not roaming.

12.4

Mobile Registration

MS

383

BSS

MSC

VLR

HLR

PSTN

IAM

1

Send routing information 2 Send routing information (ACK)

3

Send info for incoming call

4 5 6

Page Page Page request

7 Channel request 8 Immediate assignment 9 10

Page response Page response

11 Process access request 12 Complete call

13 14 15 16 17 18 19 20 21

Setup Call confirmed Alert Answer complete (ACM) Connect Connect acknowledge Answer Send info for incoming Call (Acknowledge)

Figure 12.7 Mobile-terminated call in GSM.

4. The MSC informs its VLR about the incoming call using a Send Info for Incoming Call Message that includes MSRN. 5. The VLR responds to the MSC through a Page Message that specifies the LAI and TMSI of the MS. If the MS is barred from receiving the calls, the VLR informs the MSC that a call cannot be directed to the MS. The MSC

384

12

6.

7. 8.

9.

10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.

12.5

Mobility Management in Wireless Networks

would connect the incoming call to an appropriate announcement (e.g., “The mobile phone that you have called is not permitted to receive calls”). The MSC uses the LAI provided by the VLR to determine which base station subsystems (BSSs) will page the MS. The MSC sends the Page Message to each of the BSSs to perform the paging of the MS. Each BSS broadcasts the TMSI of the MS in the Page Request Message on the paging channel. When the MS hears its TMSI broadcast on the paging channel, it responds to the BSS with a Channel Request Message over the common access channel, random access channel (RACH). On receiving the Channel Request Message from the MS, the BSS allocates an SDCCH (Stand-alone Dedicated Control Channel) and sends the Immediate Assignment Message to the MS over the access grant channel (AGCH). It is over the SDCCH that the MS communicates with the BSS and the MSC until a traffic channel is assigned. The MS sends a Page Response Message to the BSS over the SDCCH. The message contains the mobile TMSI and LAI. The BSS forwards the Page Response Message to the MSC. The MSC sends a Process Access Request Message to the VLR. The VLR responds with a Complete Call Message. The MSC then sends a Setup Message to the MS. The MS responds with a Call Confirmed Message. The MSC then sends an Alert Message to the MS. The MSC sends an Address Complete Message (ACM) to the PSTN. When the user answers, the MS sends a Connect Message to the MSC. The MSC sends a Connect Acknowledge Message to the MS. The MSC sends an Answer Message to the PSTN. The two parties can now talk. The VLR closes the dialog with the MSC by sending a Send Info for Incoming Call (acknowledge) Message to the MSC.

Handoff

As the mobile moves from one cell area to another, an active call must undergo a switch from one channel to another. This process is called a handover or handoff. In the TDMA system, a handoff usually involves both a change of channel carrier frequency and time slot. It also may require a reconfiguration of the wireline facilities by dropping the connection to the serving (old) BSS and switching to a connection to the new (target) BSS as shown in Figure 12.8.

12.5

Handoff

385

MSC B

234

MSC A

0

PSTN

89 2

01 0

234

BSC1

1

89 2

10 34 0

0

00 5

01 1

234

BSC2

234

2

234

Figure 12.8

2

89 2

01 0

0 52

1 10

0 234

1

00 5

01 1

3

00 5

01 1

BSC3 234

4

00 5

01 1

0 55

1 10

0 234

Inter-MSC, inter-BSS handoff.

In this example the mobile tunes from carrier frequency 6 on time slot 3 served by the old BSS to carrier frequency 9 on time slot 7 served by the new BSS. At the same time, the MSC switches the call from the old BSS to the new BSS. All this process may take less than 150 ms and go unnoticed by the subscriber. Handoff is initiated because of a variety of reasons. Signal strength deterioration is the most common cause for handoff at the edge of a cell. Other reasons may include load balancing where the handoff is network initiated to relieve traffic congestion by shifting calls in a highly congested cell to a lightly loaded cell. The handoff could be synchronous or it may be asynchronous. Because the mobile does not have to resynchronize itself in the synchronous situation, handoff delay is much smaller (100 ms in synchronous versus 200 ms in asynchronous). Handoffs are either initiated by the base station controller (BSC), based upon radio subsystem criteria such as RF level (RXLEV), signal quality (RXQUAL) or distance, or they are a result of network traffic loading. Appropriate decisions are made by a handoff algorithm. The measurements performed by the MS and base station transceiver (BTS) which are collated by the BSC include: (1) by MS — RXLEV, RXQUAL (for serving downlink and adjacent BSs), and (2) by BTS — RXLEV, RXQUAL, Distance (uplink for serving BS).

386

12

Mobility Management in Wireless Networks

Within the BSC these measurements are compared with a set of threshold values. If any one of the thresholds is exceeded, the BSC attempts to rectify the situation by means of power control. If the adjustment of power control cannot bring the parameters within the thresholds, a handoff is required. There are two types of handoff: internal and external. If the serving and target base stations are located within the same BSS, the BSC for the BSS performs the handoff without the involvement of MSC. This type of handoff is referred to as intra-BSS handoff. However, if the serving and target base stations do not reside within the same BSS, an external handoff is performed. In this case, the MSC coordinates the handoff and performs the switching tasks between the serving and target base stations. The external handoffs can be either within the same MSC (intra-MSC) or between different MSCs (inter-MSC).

12.5.1 Handoff Techniques Handoff techniques can be classified as mobile-controlled handoff (MCHO), networkcontrolled handoff (NCHO), and mobile-assisted handoff (MAHO). MCHO is the most popular technique for low-tier radio systems (indoor). It has been used by European digital enhancements of cordless telephony (DECT) and North American personal access communications system (PACS). In this case, the mobile continuously monitors the signal strength and signal quality from the serving base station and several handoff candidate base stations. When the handoff criteria are met, the mobile checks the best candidate base station for an available traffic channel and launches a handoff request. The combined control of automatic link transfer (ALT) and time slot transfer (TST) by the mobile is considered desirable because of the following reasons. • Offload the handoff task from the network. • Ensure robustness of the radio link by allowing reconnecting of calls even

when radio channels suddenly become poor. • Prevent simultaneous triggering of the two processes by control of both ALT and TST. Network-controlled handoff is employed by a low-tier CT-2 and a high-tier AMPS system. In this case, the base station monitors signal strength and signal quality from the mobile. When these fall below some thresholds, the network arranges for a handoff to another base station. The network asks all surrounding base stations to monitor the signal from the mobile and report the measurements back to the network. The network selects a new base station for handoff and informs both the mobile and the new base station. Mobile-assisted handoff strategy has been employed by high-tier 2G (GSM, IS-95 CDMA, IS-136 TDMA) and 3G (WCDMA, cdma2000) systems. In this approach, the network provides the mobile with a list of base station frequencies (those of nearby base stations). The network asks the mobile to measure the signal strengths and signal quality (usually determined from bit error rate) from the surrounding base stations (as well as the serving base station) and report the measurements back to the

12.5

Handoff

387

serving base station so that the network can decide whether a handoff is required and to which base station. Since MAHO has been widely used for the high-tier cellular systems, we focus only on this technique in subsequent sections.

12.5.2 Handoff Types Handoff can be categorized as hard handoff, soft handoff, and softer handoff. The hard handoff can be further divided into intrafrequency and interfrequency hard handoffs. During the handoff process, if the old connection is terminated before making the new connection, it is called a hard handoff. In the case of an interfrequency hard handoff, the carrier frequency of the new radio access is different from the old carrier frequency to which the MS was connected. On the other hand, if the new carrier frequency, to which the MS is accessed after the handoff procedure, is the same as the original carrier then it is referred to as an intrafrequency handoff. In the 2G TDMA systems, the majority of handoffs are intrafrequency hard handoffs. Interfrequency handoffs may occur between two different radio access networks, for example, between GSM and Universal Mobile Telecommunications Services. In this case, it can also be called intersystem handoff. An intersystem handoff is always a type of interfrequency, since different frequencies are used in different systems. The handoff is referred to as soft handoff if the new connection is established before the old connection is released. In the 3G systems, the majority of handoffs are intrafrequency soft handoffs. A soft handoff performed between two sectors belonging to different base stations but not necessarily to the same BSC is called a 2-way soft handoff. A soft handoff may be more than 2-way if the number of sectors involved in the handoff process is more than two. In softer handoff, the BS transmits through one sector but receives from more than one sector. In this case the MS has active uplink radio connections with the network through more than one sector belonging to the same BS. When soft and softer handoffs occur simultaneously, the term soft-softer handoff is usually used. 12.5.3 Handoff Process and Algorithms A basic handoff process consists of three main phases including measurements, decision, and execution phase (see Figure 12.9). The overall handoff process discussed here is related to MAHO strategy. The MS continuously measures the signal strength of the serving and the neighboring cells, and reports the results to the network. From a system performance standpoint, handoff measurement phase is an important task. The signal strength of the radio channel may vary significantly due to fading and signal path loss, resulting from the cell environment and user mobility. Also, an excess of measurement reports by MS or handoff execution by the network increases the overall signaling load, which is not desired. The decision phase consists of an assessment of the overall quality of service (QoS), of the connection and comparing it with the requested QoS attributes and estimates from neighboring cells. Depending on the outcome of this comparison, the handoff procedure may or may not be triggered.

388

12

Mobility Management in Wireless Networks

Measurement Criteria Measurement Measurement Reports

Algorithm Parameters Decision Handoff Criteria

Handoff Signaling Execution Radio Resource Allocation

Figure 12.9 Handoff process.

The execution phase involves handoff signaling and radio resource allocation. Radio signal strength (RSS) measurements from the serving cell and neighboring cells are primarily used in most of the networks. Alternatively or in conjunction, the path loss, signal-to-interference ratio (SIR), bit error rate (BER), and block error rate (BLER) have been used in certain voice and data networks. The following parameters are generally used in the handoff algorithm: • Upper threshold is the level at which the signal strength of the connection is

at the maximum acceptable level with respect to the required QoS. • Lower threshold is the level at which the signal strength of the connection

is at the minimum acceptable level to satisfy the required QoS. Thus, the signal strength of the connection must not fall below this level. • Handoff margin is a predefined parameter that is set at the point where the signal strength of the neighboring cell has exceeded the signal strength of the serving cell by a certain amount and/or a certain time. Some of the traditional handoff algorithms are as follows: • RSS type: The BS with the largest signal strength is selected. • RSS plus threshold type: A handoff is performed if the RSS of a new BS

exceeds that of the serving BS and the signal strength of the serving BS is below the lower threshold value. • RSS plus handoff margin type: A handoff is performed if the RSS of a new BS is more than that of the serving BS by a handoff margin.

12.5

Handoff

389

12.5.4 Handoff Call Flows In the following sections we discuss intra-MSC and inter-MSC handoff scenarios for GSM. Intra-MSC Handoff The MS constantly monitors the signal quality of the BSS-MS link. The BSS may also optionally forward its own measurements to the MS. When the link quality is poor, the MS will attempt to maintain the desired signal quality of the radio link by requesting a handoff. The following steps occur in the handoff process (see Figure 12.10). 1. The MS determines that a handoff is required; it sends the Measurement Report Message to the serving BSS. The message contains the RSS measurements. 2. The serving BSS sends a Handoff Request Message to the MSC. The message contains a rank-ordered list of the target BSSs that are qualified to receive the call. Um MS

1

A BSS

Signal strength measurement SIG, MEAS

2

A´ MSC

Rank-ordered list of target BSSs HAND, REQ

BSS

Kc, Identity of trunk, LAI HAND, REQ

3 New radio channel ID HAND, REQ, ACK

4 5 6 7 8

New radio channel ID HAND, REQ, ACK New radio channel ID HAND, REQ, ACK HAND, ACC On new radio channel PHYS, INFO HAND, DET

9 10

HAND, COMP HAND, COMP

11 12 13

REL, RCH REL, RCH, COMP

Figure 12.10 Call flow for intra-MSC handoff.

390

12

Mobility Management in Wireless Networks

3. The MSC reviews the global cell identity (GCI) associated with the best candidate to determine if one of the BSSs that it controls is responsible for the cell area. In this scenario the MSC determines that the cell area associated with the target BSS is under its control. To perform an intra-MSC handoff, two resources are required: a trunk between the MSC and the target BSS and a radio channel in the new cell. The MSC reserves a trunk and sends a Handoff Request Message to the target BSS. The message includes the desired cell area for handoff, the identity of the MSC-BSS trunk that was reserved, and the mobile encryption key (Ki). 4. The target BSS selects and reserves the appropriate resources to support the handoff pending the connection execution. The target BSS sends a Handoff Request Acknowledgment to the MSC. The message contains the new radio channel identification. 5. The MSC sends the Handoff Command Message to the serving BSS. In this message the new radio channel identification supplied by the target BSS is included. 6. The serving BSS forwards the Handoff Command Message to the MS. 7. The MS retunes to the new radio channel and sends the Handoff Access Message to the target BSS on the new radio channel. 8. The target BSS sends the Physical Information Message to the MS. 9. The target BSS informs the MSC when it begins detecting the MS handing over with the Handoff Detected Message. 10. The target BSS and the MS exchange messages to synchronize/align the MS’s transmission in the proper time slot. On completion, the MS sends the Handoff Completed Message to the target BSS. 11. At this point the MSC switches the voice path to the target BSS. Once the MS and target BSS synchronize their transmission and establish a new signaling connection, the target BSS sends the MSC the Handoff Completed Message to indicate that the handoff is successfully completed. 12. The MSC sends a Release Message to the serving BSS to release the old radio channel. 13. At this point the serving BSS releases all resources with the MS and sends the Release Complete Message to the MSC. In hard handoff, the open interval gap starts when the MS retunes to the new radio channel and ends after synchronization without any loss in voice/data transmission in the BSS or MSC. It should be noted that GSM recommendations require that the open interval gap during a handoff should not exceed 150 ms for 90% handoffs. Inter-MSC Handoff In this scenario we assume that a call has already been established. The serving BSS is connected to the serving MSC and the target BSS to the target MSC (see Figure 12.11).

12.5

Handoff

391

Serving BSS

MS

1

Target MSC

Serving MSC

Target BSS

Signal strength measurement SIG, MEAS Rank-ordered list of target BSSs

2

HAND_PER 3 HAND_ NUM

4

HAND_NUM_COMP

5

HAND_REQ

6

New radio channel ID HAND_REQ_ACK

7 HAND_PER_ ACK

8

NET_SETUP

9

10

11

12

New radio channel ID HAND_COMP

SETUP_COMP

New radio channel ID HAND_COMP

13

14

On new radio channel HAND_ACC

PHYS_INFO

HAND_DET

15

16

Figure 12.11

HAND_COMP

Call flow for inter-MSC handoff.

Target VLR

392

12

MS

Serving BSS

Mobility Management in Wireless Networks

Serving MSC

Target MSC

Target BSS

Target VLR

HAND_COMP

17

18

SEND_ENDSIG

19

ANSWER

REL_RCH

20

REL_RCH_COMP

21

22

END_SIGNAL

23

NET_REL

REL_HAND_NUM

24

Figure 12.11

(Continued).

1. Same as step 1 in the intra-MSC handoff: handoff required 2. Same as step 2 in the intra-MSC handoff: rank-ordered list of target BSSs 3. When the call is handed over from the serving MSC to the target MSC via PSTN, the serving MSC sets up an inter-MSC voice connection by placing the call to the directory number that belongs to the target MSC. When the serving MSC places this call, the PSTN is unaware that the call is a handoff and follows the normal call routing procedures, delivering the call to the target MSC. The serving MSC sends a Prepare Handoff Message to the target MSC. 4. The target MSC sends a handoff directory number (HAND_NUM) message to its VLR to assign TMSI. 5. The target VLR sends the TMSI in the directory number completion (HAND_NUM_COMP) message. 6. Same as step 3 in the intra-MSC handoff: handoff request to target BSS 7. Same as step 4 in the intra-MSC handoff: target BSS acknowledges with new radio channel

12.6

Summary

393

8. The target MSC sends the handoff perform request acknowledgment (HAND_PER_ACK) message to the serving MSC indicating that it is ready for the handoff. 9. The serving MSC sends the network setup (NET_SETUP) message to the target MSC to set up the call. 10. The target MSC acknowledges this message with the network setup completion (SETUP_COMP) message to the serving MSC. 11. Same as step 5 in the intra-MSC handoff: serving BSS handoff 12. Same as step 6 in the intra-MSC handoff: mobile station handoff 13. Same as step 7 in the intra-MSC handoff: mobile retunes and handoff accepted 14. Same as step 8 in the intra-MSC handoff: channel information to mobile 15. Same as step 9 in the intra-MSC handoff: target BSS detects handoff 16. Same as step 10 in the intra-MSC handoff: synchronization of new radio channel 17. Same as step 11 in the intra-MSC handoff: voice path to target BSS established 18. The target MSC sends the send end signal (SEND_ENDSIG) message to the serving MSC as a reminder to inform the target MSC of the call release. 19. At this point the handoff has been completed, a new voice path is set up between the MS and target BSS. The target MSC sends an ANSWER message to the serving MSC. 20. Same as step 12 in the intra-MSC handoff: old radio channel release request 21. Same as step 13 in the intra-MSC handoff: old radio channel release complete 22. The serving MSC sends the message to the target MSC: call released 23. The serving MSC releases the network resources and sends the NET_REL message to the target MSC. 24. The target MSC sends the release handoff directory number (REL_HAND_ NUM) message to the VLR to release the connection.

12.6

Summary

In this chapter we focused on mobility management in wireless wide area networks (WWANs). Mobility is concerned with radio mobility and network mobility. Radio mobility involves a handoff process, whereas network mobility deals with mobile paging and location update. We presented an application of the simple

394

12

Mobility Management in Wireless Networks

mobility model based on a fluid flow concept for a wireless network in an urban area with small cells and high user density. Next, we discussed mobile registration procedures and message flows of ANSI-41 standards for North American cellular systems and highlighted differences with the token-based registration scheme used in GSM. We concluded the chapter by explaining different types of handoff (hard, soft, and softer), handoff procedure, and message flows for the intra- and inter-MSC handoff.

Problems 12.1 Describe mobility models that are used in cellular systems to determine location updates and number of handoffs.

12.2 Repeat Example 12.1 using the following data and other data in the example: • Density of mobile in the cell  6000 mobiles/km2 • Cell radius  1 km • Average moving velocity of a mobile  20 km/hour • Number of cells per LA  8 • Number of LAs per MSC/VLR  4

12.3 What is mobile registration? Why is it used? 12.4 Describe different types of registration schemes that can be used in a cellular system. 12.5 Describe the main features of the GSM token-based registration scheme. 12.6 What is handoff in a cellular system? Why is handoff used? 12.7 Describe various handoff techniques used in mobile networks. 12.8 What are the differences between hard and soft handoff methods? 12.9 Describe the steps used in the handoff process of a cellular system. 12.10 Develop a flow diagram for an intra-MSC handoff in the CDMA IS-95 network.

References 1. Akyildiz, I. F., et al. Mobility Management in Current and Future Communication Networks. IEEE Networks, July/August, 1998. 2. Bhagwat, P., Perkins, C., and Tripathi, S. Network Layer Mobility: An Architecture and Survey. IEEE Personal Communications, vol. 3, no. 3, June 1996, pp. 54–64.

References

395

3. Brown, T. X., and Mohan, S. Mobility Management for Personal Communications Systems. IEEE Transactions Vehicular Technology, vol. 46, no. 2, May 1997, pp. 269–278. 4. 3GPP, http://www.3gpp.org. 5. 3GPP2, http://www.3gpp2.org. 6. GSM Specification Series 8.01–8.60, “BSS-MSC Interface, BSC-BTS Interface.” 7. GSM Specification Series 9.01–9.11, “Network Interworking, MAP.” 8. Kaaranen, H., et al. UMTS Networks: Architecture, Mobility and Services. John Wiley & Sons, 2001. 9. Lam, D., Cox, D. C., and Widom, J. Teletraffic Modeling for Personal Communications Services. IEEE Communications Magazine, February 1997, pp. 79–87. 10. Lin, Y-B., and Chlamtac, I. Wireless and Mobile Network Architecture. John Wiley & Sons, 2001. 11. Markoulidakis, G., et al. Mobility Modeling in Third-Generation Mobile Telecommunications Systems. IEEE Personal Communications, vol. 4, August 1997, pp. 41–56. 12. McNair, J., Akyildiz, I. F., and Bender, M. An Intersystem Handoff Technique for IMT2000 System. Tel Aviv, Israel: Infocom00. March 2000. 13. Mouly, M., and Pautet, M. The GSM System for Mobile Communications. Palaiseau, France, 1992.

This page intentionally left blank

CHAPTER 13 Security in Wireless Systems 13.1

Introduction

Although radio has existed for almost 100 years, most of the population uses wireline phones. Only over the last 30 years have large numbers of people used wireless or cordless phones. With this exposure, users of wireless phones and the news media have challenged two bedrocks of the telecommunications industry: privacy of conversation and billing accuracy. The current concepts of privacy of communications and accuracy of billing are based on the telephone company’s ability to route an individual pair of wires to each residence and office. Thus, when a call is placed on a pair of wires, the telephone company can correctly associate the call on a wire with the correct billing account [1–4]. Similarly, since there is a pair of wires from a home to the telephone company central office, no one can easily listen to the call. For most people, a wiretap is an abstract concept that only concerns someone who is involved in illegal activities. Communications on shared media can be intercepted by any user of the media. When the media are shared, anyone with access to the media can listen to or transmit on the media. Thus, communications are no longer private. In shared media, the presence of a communication request does not uniquely identify the originator, as it does in a single pair of wires per subscriber. In addition, all users of the network can overhear any information that an originator sends to the network and can resend the information to place a fraudulent call. The participants of the phone call may not know that their privacy is compromised (see Figure 13.1). When the media are shared, privacy and authentication are lost unless some method is established to regain it. Cryptography provides the means to regain control over privacy and authentication [5]. In the past, there have been attempts to control privacy and authentication through noncryptographic means. These have failed thus far. The designers of the original cellular service in the United States implemented authentication of the mobile telephone using a number assignment module (NAM) and an electronic serial number (ESN). The NAM would be implemented in a programmable read only memory (PROM) for easy replacement when the phone number changed. The ESN would be implemented in a tamper-resistant module that could not be changed without damaging the cellular telephone. In practice, many manufacturers 397

398

13

Mobile User

Security in Wireless Systems

Mobile User Mobile Network

Which of these calls are private?

Wireline Network Wireline User Figure 13.1

Mobile system privacy.

implement the NAM and the ESN in either battery-backed random access memory (RAM) or electrically erasable PROM (EEPROM). The manufacturer and the installer place the data in the phone via external programming. Similarly, the designers assumed that privacy of cellular communications would occur because 900-MHz scanners would be difficult and too expensive to build. When those scanners became easily available, the U.S. Congress passed the Electronic Communications Privacy Act in 1986, and in 1992 the FCC banned the importation and manufacture of scanners covering cellular phone bands. In practice, the laws do not help since there are millions of scanners in existence today. Furthermore, cellular test equipment is easy to build or buy, and most cellular phones can be placed in a maintenance mode that allows them to monitor any channel. Any cellular phone can be easily converted to a cellular scanner. To provide the proper privacy and authentication for a mobile station, a cryptographic system is essential. Some of the cryptographic requirements are in the air interface between the mobile station and base station. Other requirements are on databases stored in the network and on information shared between systems in the process of handoff to provide service for roaming units. In this chapter we examine the requirements needed for privacy and authentication of wireless systems, and then we discuss how each of the cellular and personal communications services (PCS) systems supports these requirements. The chapter discusses four levels of voice privacy. We then identify requirements in the areas of privacy, theft resistance, radio system requirements, system lifetime, physical requirements as implemented in mobile stations, and law enforcement needs. We will examine different methods that are in use to meet these needs.

13.2

Security and Privacy Needs of a Wireless System

399

13.2 Security and Privacy Needs of a Wireless System 13.2.1 Purpose of Security Most frauds result in a loss to the service provider. It is important to recognize that this loss may be in terms of: • No direct financial loss, but results in lost customers and an increase in use

of the system with no revenue. • Direct financial loss, where money is paid out to others, such as other network carriers and operators of value-added networks such as a premium rate service line. • Potential loss of business, where customers may move to another service provider because of the lack of security. • Failure to meet legal and regulatory requirements, such as license conditions, or data protection legislation. The objective of security for most wireless systems is to make the system as secure as the public switched telephone network. The use of radio as the transmission medium allows a number of potential threats from eavesdropping on the transmissions. It was soon apparent in the threat analysis that the weakest part of the system was the radio path, as this can be easily intercepted. The technical features for security are only a small part of the security requirements; the greatest threat is from simpler attacks such as disclosure of the encryption keys, an insecure billing system, or corruption. A balance is required to ensure that these security processes meet these requirements. At some point in time judgment must be made of the cost and effectiveness of the security measure limitation.

13.2.2 Privacy Definitions When most people think of privacy, they think of either of two levels [6,13]: none, and privacy that is used by military users. However, as we describe here, there are four levels of privacy that need to be considered. • Level 0: None. With no privacy enabled, anyone with a digital scanner could

monitor a call. • Level 1: Equivalent to wireline. As discussed earlier, most people think wireline communications are secure. Anyone in the industry knows that they are not, but the actions to tap a line often show the existence of the tap. With wireless communications, the tap can occur without anyone’s knowledge. Therefore, the actions to tap a wireline call must be translated into a different requirement for a wireless system. With this level of security, the types of conversations that would be protected are the routine everyday conversations of most people. These types of communications would be personal discussions that most people would not want exposed to the

400

13

Security in Wireless Systems

general public — for example, details of a recent operation or other medical procedure, family financial matters, mail order using a credit card, family discussions, request for emergency services (911), discussions of vacation plans (thus revealing when a home will be vacant). • Level 2: Commercially secure. This level would be useful for conversations in which the participants discuss proprietary information — for example, stock transactions, lawyer-client discussions, mergers and acquisitions, or contract negotiations. A cryptography system that allows industrial activities to be secure for about 10–25 years would be adequate. If one particular conversation was broken, the same effort would be needed to break other conversations. • Level 3: Military and government secure. This is the level that an average person thinks of when cryptography is discussed. This would be used for the military activities of a country and nonmilitary government communications. The appropriate government agency would define requirements for this level.

13.2.3 Privacy Requirements In this section we discuss the privacy needs of a wireless telephone user. Figure 13.2 is a high-level diagram of a wireless system that shows areas where intruders can compromise privacy. A user of a mobile system needs privacy in the following areas: • Privacy of call setup information. During a call setup, the mobile station

will communicate information to the network. Some of the information that a user or mobile station could send includes calling number, calling card number, or type of service requested. The system must send all this information in a secure fashion. • Privacy of speech. The system must encrypt all spoken communications so that intruders cannot intercept the signals by listening on the airwaves. • Privacy of data. The system must encrypt all user communications so that intruders cannot intercept the data by listening on the airwaves. • Privacy of user location. A user should not transmit information that enables an eavesdropper to determine the user’s location. The usual method to meet this requirement is to encrypt the user ID. Three levels of protection are often needed: 1. Eavesdropping of radio link 2. Unauthorized access by outsiders to the user location information stored in the network visitor location register (VLR) and home location register (HLR) 3. Unauthorized access by insiders to the user location information stored in the network. This level is difficult to achieve, but not impossible

13.2

Security and Privacy Needs of a Wireless System

401

Downlink Billing Database Uplink Location Database

Signaling Voice Data Location Identification Calling Pattern Financial Transaction - Keypad - Spoken

Switch

Network Interconnect

Roaming Service Roaming Service Provider

Figure 13.2

Privacy requirements.

• Privacy of user identification. When a user interacts with the network, the

user ID is sent in a way that does not show user identification. This prevents analysis of user calling patterns based on user ID. • Privacy of calling patterns. No information must be sent from a mobile that enables a listener of the radio interface to do traffic analysis on the mobile user. Typical traffic analysis information is: • Calling number • Frequency of use of the mobiles • Caller identity • Privacy of financial transactions

If the user transmits credit card information over any channel, the system must protect the data. Users may order items from mail order houses via a telephone that is wireless. Users may choose to voice their credit card numbers rather than dialing them via touch-tone phone. Users may access bank voice response systems, where they send account data via tone signaling. Users may access calling card services of carriers and may speak or use tone signaling to send the card number.

402

13

Security in Wireless Systems

All these communications need to be private. Since the user can send the information on any channel — voice, data, or call control — the system must encrypt all channels.

13.2.4 Theft Resistance Requirements The system operator may or may not care if a call is placed from a stolen mobile station as long as the call is billed to the correct party. The owner of a mobile station will care if the unit is stolen. The mobile terminal design should reduce theft of the mobile station by making reuse of a stolen mobile station difficult. Even if the mobile station is registered to a new legitimate account, the use of the stolen mobile station should be stopped. The mobile station design should also reduce theft of services by making reuse of a stolen mobile station unique information difficult. Requirements needed to accomplish the reduction in theft are: • Clone-resistant design. In the current wireless systems, cloning of mobile

stations is a serious problem; methods must be put in place to reduce or eliminate fraud from cloning. To achieve fraud reduction, mobile station unique information must not be compromised by any of the following means: 1. Over the air: someone listening to a radio channel should not be able to determine information about the mobile station and then program it into a different mobile station; 2. From the network: The databases in the network must be secure. No unauthorized person should be able to obtain information from those databases. 3. From network interconnect: Systems will need to communicate with each other to verify the identity of roaming mobile stations. A system operator could perpetrate fraud by using the security information about roaming mobile stations to make clone mobile stations. 4. The communication scheme used between systems to validate roaming mobile stations should be designed so that theft of information by a fraudulent system does not compromise the security of the mobile station. 5. Thus, any information passed between systems for security checking of roaming mobile stations must have enough information to authenticate the roaming mobile station. It must also have insufficient information to clone the roaming mobile station. 6. From users cloning their own mobile station: Users can perpetrate fraud on the system. Multiple users could use one account by cloning mobile stations. The requirements for reducing or eliminating this fraud are the same as those to reduce repair and installation fraud described below.

13.2

Security and Privacy Needs of a Wireless System

403

• Installation and repair fraud. Theft of service can occur when the service

is installed or when a terminal is repaired. Multiple mobile stations can be programmed with the same information (cloning). The cryptographic system must be designed so that installation and repair cloning is reduced or eliminated. • Unique user ID. More than one person may use a handset. It is necessary to identify the correct person for billing and other accounting information. Therefore, the user of the system must be uniquely identified in the system. • Unique mobile station ID. When all security information is contained in a separate module (smart card), the identity of the user is separate from the identity of the mobile station. Stolen mobile stations can then be valuable for obtaining service without purchasing a new (full price) mobile station. Therefore, the mobile station should have unique information contained within it that reduces or eliminates the potential for stolen mobile stations to be registered with a new user.

13.2.5 Radio System Requirements When a cryptographic system is designed, it must function in a hostile radio environment characterized by bit errors caused by: • Multipath fading and thermal noise. The characteristics of the radio

channel affect the choice of cryptographic algorithms. The radio signals will take multiple diverse routes from the mobile station to the base station. The effect of multiple diverse routes that can be severe and cause burst errors is fading. Although the system may be interference limited, there may be conditions when the limiting factor on performance is thermal noise. The choice of cryptographic modes must include both of these channel characteristics. • Interference. The mobile systems may initially share a radio spectrum with other users. The modulation scheme and cryptographic system must be designed so that interference with shared users of the spectrum does not compromise the security of the system. • Jamming. Although usually thought about only in the context of military communications, civilian systems can also be jammed. As wireless communication becomes ubiquitous, jamming of the service can also be a method of breaking the security of the system. Therefore, cryptographic systems must work in the face of jamming. • Support of handoff. When the call handoff occurs to another radio port in the same or adjacent mobile system, the cryptographic system must maintain synchronization.

404

13

Security in Wireless Systems

13.2.6 System Lifetime Requirements It has been estimated that computing power doubles every 18 months. An algorithm that is secure today may be breakable in 5 to 10 years. Since any system being designed today must work for many years after design, a reasonable requirement is that the procedures must last at least 20 years. The algorithm must have provisions to be upgraded in the field. 13.2.7 Physical Requirements Any cryptographic system used in a mobile station must work in the practical environment of a mass-produced consumer product. Therefore, the cryptographic system must meet the following requirements: • Mass production. It can be produced in mass quantities (million of units

per year). • Exported and/or imported. The security algorithm must be capable of being

exported and imported. Two problems are solved with export and import restrictions lifted: 1. It can be manufactured anywhere in the world. 2. It can be carried on trips outside the United States. As an alternative, if an import/export license for the algorithm cannot be obtained, the following restrictions must apply: • Either only U.S. manufacturing or two-stage manufacturing • All mobile stations must be made in the United States or all mobile stations made outside the United States will have final assembly in the United States • All mobile stations must be impounded on leaving the United States. • Basic handset requirements. Any cryptographic system must have mini-

mum impact on the following mobile station requirements: • • • • • • •

Size Weight Power drain Heat dissipation Microprocessor speed Reliability Cost

• Low-cost level 1 implementation. Level 1 implementation would be expected

as a baseline for most mobile systems. Therefore, level 1 implementation must

13.2

Security and Privacy Needs of a Wireless System

405

be low cost. Designers obtain low-cost solutions by implementations that can be done either in software or in low-cost hardware. Software solutions are attractive. Often mobile stations have spare read only memory (ROM), RAM, and central processing unit (CPU) cycles in microprocessors.

13.2.8 Law Enforcement Requirements When a valid court order is obtained in the United States, current telephones (either wired or wireless) are relatively easy to tap by the law enforcement community. The same requirements described in this chapter to ensure privacy and authentication of wireless mobile communications make it more difficult to execute legitimate court wiretap orders. The law enforcement community can wiretap mobile stations after properly obtaining court orders. When an order is obtained, there are several ways a mobile system operator can meet the needs of the order. Any method used must not compromise the security of the system. Figure 13.3 shows possible approaches to tapping the call. The tap can be done over the air or at a central switch. This discussion assumes that only the radio portion of the link is encrypted and the call appears in the clear in the wired portion of the network. If end-to-end

Hacker

Wiretap at Switch

No Encryption Radio System

Switch

Weak Encryption Court order

Strong Encryption Wiretap Vehicle

Figure 13.3

Law enforcement requirements.

Key Escrow Agency

406

13

Security in Wireless Systems

encryption is used, other means must be considered to obtain the information since the call never appears clear except at the end points. Over-the-Air Tap When the tap is done over the air, a wiretap van is required. The van is driven to inside the cell where the call is placed. A centrally located base station (BS) receives interference from mobile stations in many cells or may not be able to receive a low-power mobile station at all. In a large-cell mobile system, wiretap stations could be deployed in each cell, but in a small cell system, the number of tap points would be too high. Therefore, a wiretap van is needed and is driven to the correct cell where the call is placed. After the van is driven to the correct cell, it needs to be close to the mobile station. A van might have an antenna that is a maximum of 6 to 10 feet high versus a BS antenna has a height of 25 to 100 feet or more. Thus, the van must be closer to the mobile station than a cell radius. A quick rule of thumb for the wiretap van is that if the mobile station is in line of sight, then the wiretap van can receive the mobile station transmission. If a wiretap van is used, then the transmissions of the mobile station must be decrypted. The following are possibilities: • No encryption: This approach makes tapping the easiest; if no encryption

is used, anyone can listen to a call over the airwaves. Thus, law enforcement personnel can listen to and record a call, and so can anyone else. • Breakable algorithms: If the algorithm is weak enough, law enforcement agencies can break the algorithm when permitted to do so by an appropriate court order. Unfortunately, given the proliferation of desktop/laptop personal computers, any algorithm that can be easily broken by the law enforcement community will also be quickly broken by anyone else. • Strong encryption: Strong encryption makes it difficult, if not impossible, for the wiretap van to decrypt the transmission. One method to resolve this dilemma is to use a key escrow system where all cryptographic keys would be available from an appropriate key escrow agency. With a court order, the information could be obtained by law enforcement agencies so that they could listen to and record a call. Wiretap at Switch Since all mobile calls must be routed through a central switch, those calls that use radio-link-only encryption can be tapped at the central switch under a court order. This is the preferred method for low-power wireless calls. This method leaves it to the user and system provider to have appropriate levels of security in the wireless portion of the call.

13.4

Methods of Providing Privacy and Security in Wireless Systems

13.3

407

Required Features for a Secured Wireless Communications System

For wireless communications to be secure the following features must be available [8–12]: • User authentication proves that the users are who they claim to be. • Data authentication consists of data integrity and data origin authentication.

With data integrity the recipient can be sure that the data has not changed. Data origin authentication proves to the recipient that the stated sender has originated the data. • Data confidentiality means the data is encrypted so that it is not disclosed

while in transit. • Non-repudiation corresponds to a security service against denial by either

party of creating or acknowledging a message. • Authorization is the ability to determine whether an authenticated entity

has the permission to execute an action. • Audit is a history of events that can be used to determine whether anything

has gone wrong and, if so, what it was, when it went wrong, and what caused it. • Access control enables only authorized entities to access resources. • Availability ensures that resources or communications are not prevented

from access or transmission by malicious entities. • Defense against denial of service is the attack corresponding to the security

service of availability.

13.4

Methods of Providing Privacy and Security in Wireless Systems

North American and European cellular and PCS systems support a variety of air interface protocols. They include: • The Advanced Mobile Phone System (AMPS) • The IS-54/IS-136 TDMA protocol • The IS-95 CDMA • The cdma2000 • The Global System for Mobile communications (GSM) • The Wideband CDMA (WCDMA) system

Across these protocols, there are four security models that have been used for cellular/ PCS phones in the United States and Europe.

408

13

Security in Wireless Systems

1. MIN/ESN: The original AMPS used a 10-digit mobile identification number (MIN) and a 32-bit ESN. All data is sent in clear text. Data is shared between systems with bad (incorrect) MINs, ESNs, and MIN/ESN pairs. When a mobile station (MS) roams into a system, first the bad list is checked, and then a message is sent to the home system to validate the MIN/ESN pair. The intersystem communications are sent via Signal System 7 (SS7) using an ANSI IS-41 protocol. As an improvement to this approach, some systems require that a user enter a PIN before placing the calls. The main advantage of the Personal identification number (PIN) is that it can be changed in the network when it is compromised, and the user can continue to have the same phone number. Cellular phones that are cloned must have their phone number (MIN) changed to stop the fraudulent use. 2. Shared secret data (SSD): The TDMA and CDMA systems in the United States use SSD stored in the network and the mobile phone. At service initiation time, a secret key is stored in the phone and the network. AMPS, IS-95 CDMA, IS-54/IS-136 TDMA, and cdma2000 all support SSD. The intersystem communications are sent via SS7 using an ANSI IS-41 protocol. All mobile stations are assigned an ESN at the time of manufacturing. They are also assigned a 15-digit international mobile subscriber identity (IMSI), that is unique worldwide, an A-key, and other data at the time of service installation. When the MS is turned on, it must register with the system. When it registers, it sends its IMSI and other data to the network. The VLR in the visiting system then queries the HLR for the security data and service profile information. The VLR then assigns a temporary mobile subscriber identity (TMSI) to the MS. The MS uses the TMSI for all further access to that system. The TMSI provides anonymity of communications since only the MS and the network know the identity of the MS with a given TMSI. When the MS roams into a new system, some air interfaces use the TMSI to query the old VLR and then assigns a new TMSI; other air interfaces request that the MS send its IMSI and then assign a new TMSI. Each time an MS places or receives a call, a call counter (CHCNT) is incremented. The counter is also used for clone detection since clones will not have a call history identical to the legitimate phone. 3. Security triplets (token based): GSM uses its own unique algorithm and does not share secrets between cellular or PCS systems. It uses a token-based authentication scheme. When an MS roams into a system, a message is sent to the home system asking for sets (3 to 5 typically) of triplets (unique challenge, response to the challenge, and a voice privacy key derived from the challenge). Each call that is placed or received uses one triplet. After all triplets are used up, the visited system must send a new message to the home system to get another set of triplets. The intersystem communications use the CCITT SS7 and GSM mobile application part (MAP) protocol. Each system operator can choose its own authentication method. The MS and the HLR each support the same method and have common data. Each MS sends a

13.6

IEEE 802.11 Security

409

registration request; then the network sends a unique challenge. The MS calculates the response to its challenge and sends a message back to the network. The VLR contains a list of triplets; the network compares a triplet with responses it receives from the MS. If the response matches, the MS is registered with the network. The just-used triplet is discarded. 4. Public key: The public key system is analogous to the lock and its combinations. Public key algorithm relies on two cryptographic keys, intimately related to each other but each not derivable from the other. Public key systems do not need communications to the home system to validate the MS. The intersystem communications are still needed to validate the account and get user profile information.

13.5

Wireless Security and Standards

The National Institute of Standards and Technology (NIST) expects that future IEEE 802.11 (and possibly other wireless technologies) products will offer advanced encryption standard (AES)-based data link level cryptographic services that are validated under the U.S. Federal Information Processing Standard (FIPS) 140-2 [7]. As these will mitigate most concerns about wireless eavesdropping or active wireless attacks, their use is strongly recommended when they become available. • IEEE 802.11 — WLAN. Data security using encryption is an optional

functionality of medium access control (MAC). The functionality is called wired equivalent privacy (WEP). Encryption is only supplied between stations and not on an end-to-end basis. No key management is specified. Authentication is performed by assigning an Extended Service Set ID (ESSID) to each access point (AP) in the network and by using the ESSID in a challenge-response authentication scheme. WEP was shown to have severe security weaknesses. Wi-Fi protected access (WPA) was introduced by the Wi-Fi Alliance as an intermediate solution to WEP insecurities. WPA implemented a subset of IEEE 802.11i specifications which will be discussed in the following section. • European and North American Systems. Almost all information being sent between an MS and the network is encrypted, and sensitive information is not transmitted over a radio channel.

13.6

IEEE 802.11 Security

The IEEE 802.11 Wi-Fi wireless local area network (WLAN) (see Chapter 18) standard addressed security with the WEP protocol, which proved relatively easy to crack and was shown to have major security weaknesses. IEEE 802.11i, also known as Wi-Fi protected access 2 (WPA2), is an improved security protocol for IEEE 802.11. IEEE 802.11i includes stronger encryption, authentication, and key

410

13

Security in Wireless Systems

management strategies that go a long way toward guaranteeing data and system security. The new data-confidentiality protocols in 802.11i are the temporal key integrity protocol (TKIP) and counter-mode/block chaining message authentication code protocol (CCMP). 802.11i also uses an 802.1X key distribution system to control access to the network. Because 802.11 handles unicast and broadcast traffic differently, each traffic type has different security concerns. 802.11i uses a negotiation process to select the correct confidentiality protocol and key system for each traffic type. Other features introduced in 802.11i include key caching and preauthentication. The TKIP is a data confidentiality protocol, which improves the security of products using WEP. Among WEP’s numerous flaws are its lack of a message integrity code and its insecure data-confidentiality protocol. The message integrity code enables devices to authenticate that the packets are coming from the claimed source. This authentication is important in a wireless system where traffic can be easily injected. The TKIP uses a mixing function to defeat weak-key attacks. The mixing function creates a per frame key to avoid the WEP weaknesses. The CCMP is a data-confidentiality protocol to handle packet authentication as well as encryption. For confidentiality CCMP uses AES in counter mode. For authentication and integrity, CCMP uses a cipher block chaining message authentication code (CBC-MAC). In 802.11i, CCMP uses a 128-bit key. The block size is 128 bits. The CBC-MAC size is 8 octets, and nonce size is 48 bits. There are two bytes of 802.11 overhead. The CBC-MAC, the nonce, and the 802.11 overhead make the CCMP packet 16 octets larger than an unencrypted 802.11 packet. Although slightly slower, the larger packet is not a bad exchange for increased security. The CCMP protects some fields that are not encrypted. The additional parts of the 802.11 frame that are protected are known as additional authentication data (AAD). AAD includes the packet source and destination and protects against attackers replaying packets to different destinations. The 802.1X provides a framework to authenticate and authorize devices connecting to the network. It prevents access to the network until such devices pass authentication. The 802.1X also provides a framework to transmit key information between authenticator and supplicant. For 802.11i, the access point takes the role of the authenticator and the client card the role of supplicant. The supplicant authenticates with the authentication server through the authenticator. In 802.1X, the authenticator enforces authentication. The remote authentication dial-in user service (RADIUS) protocol (see Section 13.9) is typically used between authenticator and authentication server. Once the authentication server concludes authentication with the supplicant, the authentication server informs the authenticator of the successful authentication and passes established keying material to the authenticator. At that point, the supplicant and authenticator share

13.7

Security in North American Cellular/PCS Systems

411

established key material through extensive authentication protocol over LANs (EAPOL)-key exchange. If all exchanges have been successful, the authenticator allows traffic to flow through the controlled port giving the client access to the network. The 802.11i EAPOL-key exchange uses a number of keys and has a key hierarchy to divide initial key material into useful keys. The two key hierarchies are: pair-wise key hierarchy and group key hierarchy. In the 802.11i specification, these exchanges are referred to as the 4-way handshake and the group key handshake. The 4-way handshake does several things: • Confirms the pairwise master key (PMK) between the suppliant and • • • •

authenticator Establishes the temporal keys to be used by the data-confidentiality protocol Authenticates the security parameters that were negotiated Performs the first group key handshake Provides keying material to implement the group key handshake

Wireless clients often roam back and forth between access points. This has a negative effect on the system performance. Key caching reduces the load on the authentication server and reduces the time required to get connected to the network. The basic concept behind the key caching is for a client and access point to retain a security association when client roams away from the access point. When the client roams back to the access point, the security association can be restarted. Preauthorization enables a client to establish a PMK security association to an access point with which the client has yet not been associated. Preauthorization provides a way to establish a PMK security association before a client associates. The advantage is that the client reduces the time that it is disconnected from the network. Preauthorization has limitations. Clients performing preauthorization will add load to the authorization server. Also, since preauthorization is done at the IEEE 802 layer, it does not work across IP subnets.

13.7

Security in North American Cellular/PCS Systems

The ANSI-41 authentication features are independent of the air-interface protocol used to access the network, and subscribers are never involved in the process. A successful outcome of authentication occurs when it can be shown that the MS and the network possess identical results of a calculation performed in both the MS and the network. The authentication center (AC) is the primary functional entity in the network responsible for performing this calculation (see Chapter 7), although the serving system (i.e., the VLR) may also be allocated certain responsibilities. The authentication calculations are based on a set of algorithms, collectively known as the cellular authentication and voice encryption (CAVE) algorithm.

412

13

Security in Wireless Systems

The authentication process and algorithm are based on the following two secret numbers: 1. Authentication key (A-key) (64-bit) 2. Shared secret data (SSD) (128-bit) The A-key is a 64-bit secret number that is the permanent key used by the authentication calculations in both the MS and the AC. The A-key is permanently installed into the MS and is securely stored at the AC in the network when a new subscription is obtained. Once the A-key is installed in the MS, it should not be displayed or retrievable. The MS and the AC are the only functional entities ever aware of the A-key; it is never transmitted over the air or passed between systems. The primary function of the A-key is as a parameter used in calculation to generate the SSD. The COUNT is a 6-bit parameter that is intended to provide additional security in case the A-key or SSD is compromised. The current value of the COUNT is maintained by both the MS and the authentication controller. The respective counts should generally be the same — they may not always match exactly due to radio transmission problems or system failures in the network. If the respective counts differ by a large enough range, or frequently do not match, the AC may assume that a fraudulent condition exists and take corrective action. Note that a COUNT mismatch detection does not conclusively indicate that the particular MS accessing the system is fraudulent — only that a clone may exist.

13.7.1 Shared Secret Data Update The SSD is a 128-bit secret number that is essentially a temporary key used by authentication calculations in both the MS and the AC. The SSD may also be shared with the serving system via a number of ANSI-41 messages. The SSD is a semipermanent value. It can be modified by the network at any time, and the network can command the MS to generate a new value. The SSD is obtained from calculations using the A-key, the ESN, and a random number shared between the MS and the network. SSD calculation results in two separate 64-bit values, SSD_A and SSD_B. SSD_A is the value used for the authentication process, whereas SSD_B is used for encryption algorithms for privacy and to encrypt and decrypt selected messages on the radio traffic channel. Figure 13.4 shows the SSD generation process. At any time, the network can order the MS to update the SSD by generating the new SSD with a new SSD random number for security purposes. 13.7.2 Global Challenge For a global and unique challenge authentication process, the ANSI-41 standard is used [8,9]. In a global challenge the serving system presents a numeric authentication challenge to all mobile stations that are using a particular radio control

13.7

Security in North American Cellular/PCS Systems

Serving System

413

HLR/AC

MS

SSD Random #

MIN, ESN SSD Random #

MIN, ESN SSD Random #

SSD Random # ESN A-key SSD Random #

A-key ESN

Generate SSD

Generate SSD

New SSD

New SSD Figure 13.4

SSD generation.

ESN

Serving System

MS

SSD_A MIN

RAND

HLR/AC

1 MIN

CAVE SSD_A 2 MS Auth. Results

MS Auth. Results

MS Auth. Results, RAND

ESN

RAND

3 4

CAVE AC Auth. Result Result  Pass or fail

Figure 13.5 Global challenge authentication process (no SSD sharing with serving system).

channel. The ANSI-41 AC verifies that the numeric authentication response from an MS attempting to access the system is correct. This is called a global challenge because the challenge indicator and random number used for the challenge are broadcast on the radio control channel and are used by all mobile stations accessing that control channel. The authentication process flow diagram (when SSD is not shared with the serving system) is given in Figure 13.5. 1. The serving system generates a random number (RAND) and sends it to the MS in the overhead message on the control channel.

414

13

Security in Wireless Systems

2. MS calculates an authentication result using CAVE and transmits that result back to the serving system when it accesses the system for registration, call origination, or paging response purposes. 3. The serving system forwards the authentication result and the random number to AC. 4. The AC independently calculates an authentication result and compares it to the result received from the MS. If the results match, the MS is considered successfully authenticated. If the results do not match, the MS may be considered fraudulent and service may be denied. If the SSD is shared, then the serving system performs the calculations.

13.7.3 Unique Challenge In the ANSI-41 unique challenge, the authentication controller directs the serving system to present a numeric authentication challenge to a single MS that either is requesting service from the network or is already engaged in a call. The serving system presents the numeric authentication challenge to the MS and verifies that the numeric authentication response provided by the MS is correct. The unique challenge is so named because the challenge indicator and the random number used for the challenge are directed to a particular MS, whereas a global challenge is required by each MS. Figure 13.6 shows the basic unique challenge procedure for authentication when SSD is not shared.

MS

Serving System

HLR/AC

SSD_A

SSD_A AC authentication result, random #

2 ESN MIN

random #

random # 1

ESN MIN

random #

CAVE

CAVE

MS authentication result

MS authentication result 3

AC authentication result result  pass or fail 4

Figure 13.6

result

Basic unique-challenge authentication process when SSD is not shared.

13.8

Security in GSM, GPRS, and UMTS

415

1. The AC generates a random number and uses it to calculate an authentication result. The AC sends both the random number and authentication result to the serving system. 2. The serving system forwards the random number to the MS. 3. The MS calculates an authentication result and sends it to the serving system. 4. The serving system compares the result from the AC with the result from the MS. If the results match, the MS is considered to have successfully responded to the challenge. If they do not match, the MS may be considered fraudulent and service may be denied. Either way, the serving system reports the results to the AC. If SSD is shared, the serving system may initiate the unique challenge process and would report a failure to the AC.

13.8 Security in GSM, GPRS, and UMTS 13.8.1 Security in GSM GSM allows three-band phones to be used seamlessly in more than 160 countries. In GSM, security is implemented in three entities: • Subscriber identity module (SIM) contains IMSI, TMSI, PIN, MSISDN,

authentication key Ki (64-bit), ciphering key (Kc) generating algorithm A8, and authentication algorithm A3. SIM is a single chip computer containing the operating system (OS), the file system, and applications. SIM is protected by a PIN and owned by an operator. SIM applications can be written with a SIM tool kit. • GSM handset contains ciphering algorithm A5. • Network uses algorithms A3, A5, A8; Ki and IDs are stored in the authentication center. Both A3 and A8 algorithms are implemented on the SIM. The operator can decide which algorithm to use. Implementation of an algorithm is independent of hardware manufacturers and network operators. A5 is a stream cipher. It can be implemented very efficiently on hardware. Its design was never made public. A5 has several versions: A5/1 (most widely used today), A5/2 (weaker than A5/1; used in some countries), and A5/3 (newest version based on the Kasumi block cipher). The authentication center contains a database of identification and authentication information for subscribers including IMSI, TMSI, location area identity (LAI), and authentication key (Ki). It is responsible for generating (RAND), response (RES), and ciphering key (Kc) which are stored in HLR / VLR for authentication and encryption processes. The distribution of security credentials and encryption algorithms provides additional security.

416

13

Security in Wireless Systems

GSM uses information stored on the SIM card within the phone to provide encrypted communications and authentication. GSM encryption is only applied to communications between a mobile phone and the base station. The rest of the transmission over the normal fixed network or radio relay is unprotected, where it could easily be eavesdropped or modified. In some countries, the base station encryption facility is not activated at all, leaving the user completely unaware of the fact that the transmission is not secure. GSM encryption is achieved by the use of a shared secret key. If this key is compromised it will be possible for the transmission to be eavesdropped and for the phone to be cloned (i.e., the identity of the phone can be copied). The shared secret key could easily be obtained by having physical access to the SIM, but this would require the attacker to get very close to the victim. However, it has been shown by research that the shared secret key can be obtained over the air from the SIM by transmitting particular authentication challenges and observing the responses. If the base station can be compromised then the attacker will be able to eavesdrop on all the transmission being received. The attacker will also have access to the shared secret keys of all the mobile phones that use the base station, thus allowing the attacker to clone all of the phones. Authentication in the GSM system is achieved by the base station sending out a challenge to the mobile station. The MS uses a key stored on its SIM to send back a response that is then verified. This only authenticates the MS, not the user. A 64-bit key is divided to provide data confidentiality. It is not possible to encrypt all the data; for example, some of the routing information has to be sent in clear text. GSM Token-based challenge The security-related information consisting of triplets of RAND, signature response (SRES), and Kc are stored in the VLR. When a VLR has used a token to authenticate an MS, it either discards the token or marks it used. When a VLR needs to use a token, it uses a set of tokens that is not marked as used in preference to a set that is marked used. When a VLR successfully requests a token from the HLR or an old VLR, it discards any tokens that are marked as used. When an HLR receives a request for tokens, it sends any sets that are not marked as used. Those sets shall then be deleted or marked as used. The system operator defines how many times a set may be reused before being discarded. When HLR has no tokens, it will query the authentication center for additional tokens. The token-based challenge can be integrated into various call flows (e.g., registration, handoff). It is described separately here for clarity. Figures 13.7 and 13.8 show the call flows of token-based challenges.

13.8

Security in GSM, GPRS, and UMTS

MS

New VLR

HLR

Old VLR

Unique Challenge

1 2

Radio System, MSC

417

CALC SRES Challenge Response

3 4

Auth. Request CALC SRES and Compare

5

Auth. Response

6 7

Success or Failure Call Proceeds or Terminates

Figure 13.7

GSM token-based unique challenge.

1. The serving system sends a RAND to the MS. 2. The MS computes the SRES using RAND and the authentication key (Ki) in the encryption algorithm. 3. The MS transmits the SRES to the serving system. 4. The MSC sends a message to the VLR requesting authentication. 5. The VLR checks the SRES for validity. 6. The VLR returns the status to the MSC. 7. The MSC sends a message to the MS with a success or failure indication. Both GSM and North American systems use the international mobile equipment identity (IMEI) stored in the equipment identity register (EIR) (see Chapter 7) to check malfunctions and fraudulent equipment. The EIR contains a valid list (list of valid mobiles), a suspect list (list of mobiles under observation), and a fraudulent list (list of mobiles for which service is barred) (see Figure 13.9 for call flow).

13.8.2 Security in GPRS The general packet radio service(GPRS) allows packet data to be sent and received across a mobile network (GSM). GPRS can be considered an extension to the GSM network to provide 3G services. GPRS has been designed to allow users to connect to the Internet, and as such is an essential first step toward 3G networks

418

13

B MS

MSC

3

D

Request IMSI IMSI Acknowledge IMSI Acknowledge

4 5

Get Authentication Parameters IMSI Get Authentication Parameters IMSI

6

Authentication Parameters

7 Authentication Parameters

8 9 10 11

AUC

HLR

Request IMSI

1 2

New VLR

Security in Wireless Systems

Authenticate Mobile Station RAND

Authenticate Mobile Station RAND

RAND, SRES, Kc

RAND, SRES, Kc

Authenticate Response SRES

12

Figure 13.8

Authenticate Response SRES

GSM token-based unique challenge with ciphering.

for all mobile operations. In GPRS, TMSI is replaced by P-TMSI and P-TMSI signature as alternative identities. The HLR GPRS register maps between internet protocol (IP) addresses and IMSI. GPRS security functionality is equivalent to the existing GSM security. Authentication and encryption setting procedures are based on the same algorithms, keys, and criteria as in GSM systems. GPRS provides identity confidentiality to make it difficult to identify the user. This is achieved by using a temporary identity where possible. When possible, confidentiality also protects dialed digits and addresses. As in GSM, the device is authenticated by a challenge-response mechanism. This only verifies that the smart card within the device contains the correct key. GPRS does not provide end-to-end security so there is a point where the data is vulnerable to eavesdropping or attack. If this point can be protected, e.g., in a physically secure location, this is not a problem. However, if end-to-end security

13.8

Security in GSM, GPRS, and UMTS

MS

419

MSC

EIR

IMEI Request

1

IMEI Response

2

3

Check IMEI (IMEI)

IMEI: International Mobile Equipment Identity

4

Figure 13.9

IMEI Check Results

Equipment identity check.

is required, there are other standards that can be used over GPRS; such as the wireless application protocol (WAP) and Internet protocol security (IPSec). In GPRS authentication is performed by serving GPRS support node (SGSN) instead of VLR. The encryption is not limited to radio part, but it is up to SGSN. An IP address is assigned after authentication and ciphering algorithm negotiation.

13.8.3 Security in UMTS The security in universal mobile telecommunications services (UMTS) is built upon the security of GSM and GPRS. UMTS uses the security features from GSM that have proved to be needed and robust. UMTS security tries to ensure compatibility with GSM in order to ease interworking and handoff between GSM and UMTS. The security features in UMTS correct the problems with GSM by addressing its real and perceived security weaknesses. New security features are added as necessary for new services offered by UMTS and to take into account the changes in network architecture. In UMTS the SIM is called UMTS SIM (USIM). UMTS uses public keys. In UMTS mutual authentication between the mobile and BS occurs; thus there is no fake BS attack. UMTS has increased key lengths

420

13

Security in Wireless Systems

and provides end-to-end security. The other security features of UMTS are listed below: • Subscriber individual key K; • Authentication center and USIM share • User-specific secret key K, • Message authentication functions f1, f2 • Key generating functions f3, f4, f5 • The authentication center has a random number generator. • The authentication center has a scheme to generate fresh sequence numbers. • USIM has a scheme to verify freshness of received sequence numbers. • Authentication functions f1, f2 are: • MAC (XMAC); • RES (XRES). • Key generating functions f3, f4, f5 are: • f3: ciphering key CK (128 bit); • f4: integrity key IK (128 bit) and • f5: anonymity key AK (128 bit). • Key management is independent of equipment. Subscribers can change

handsets without compromising security. • Assure user and network that CK / IK have not been used before. • For operator specific functions, UMTS provides an example called Milenage based on the Rijndael block cipher. • Integrity function f9 and ciphering function f8 are based on the Kasumi block cipher.

13.9

Data Security

The primary goals in providing data security are confidentiality, integrity, and availability. Confidentiality deals with the protection of data from unauthorized disclosures of customers and proprietary information. Integrity is the assurance that data has not been altered or destroyed. Availability is to provide continuous operations of hardware and software so that parties involved can be assured of uninterrupted service. In this section, we focus upon commonly used data security methods including firewalls, encryption, and authentication protocols.

13.9.1 Firewalls Firewalls have been used to prevent intruders from securing Internet connection and making unauthorized access and denial of service attacks to the organization

13.9

Data Security

421

network. This could be for a router, gateway, or special purpose computer. The firewalls examine packet flowing into and out of the organization network and restrict access to the network. There are two types of firewalls: (1) packet filtering firewall, and (2) application-level gateway. The packet filter examines the source and destination address of packets passing through the network and allows only the packets that have acceptable addresses. The packet filter also examines IP addresses and TCP (transmission control protocol) ports. The packet filter is unaware of applications and what an intruder is trying to do. It considers only the source of data packets and does not examine the actual data. As a result, malicious viruses can be installed on an authorized user computer, giving the intruder access to the network without authorized user knowledge. The application-level gateway acts as an intermediate host computer between the outside client and the internal server. It forces everyone to login to the gateway and allows access only to authorized applications. The application-level gateway separates a private network from the rest of the Internet and hides individual computers on the network. This type of firewall screens the actual data. If the message is deemed safe, then it is sent to the intended receiver. These firewalls require more processing power than packet filters and can impact network performance.

13.9.2 Encryption Encryption is one of the best methods to prevent unauthorized access of an intruder. Encryption is a process of distinguishing information by mathematical rules. The main components of an encryption system are: (1) plaintext (not encrypted message), (2) encryption algorithm (works like a locking mechanism to a safe), (3) key (works like the safe’s combination), and (4) ciphertext (produced from plaintext message by encryption key). Decryption is the process that is the reverse of encryption. It does not always use the same key or algorithm. Plaintext results in decryption. The following types of keys are used in encrypting data. Secret Key (symmetric encryption) Both sender and recipient share a knowledge of the same secret key. The scrambling technique is called encryption. The message is referred to as plaintext or clear text, and the encrypted version of it is called ciphertext. The encryption of a plaintext x into a ciphertext y using a secret key ek is given as (see Figure 13.10): y  ek(x)

Ciphertext

The corresponding decryption yields x  dk(y)

where: dk is the decryption key

Plaintext

422

13

Security in Wireless Systems

Plaintext x Encryption Key ek

Plaintext x

dk(y)

ek(x)

Decryption Key dk

Intruder Listening/ Eavesdropping

Masquerading Network

Figure 13.10

Encryption using secret key.

Ideally, the encryption scheme should be such that it cannot be broken at all. Because there are no practical methods of achieving such an unconditional security, encryption schemes are designed to be computationally secure. The encryption and decryption algorithms use the same key, and, hence, such algorithms are called symmetric key algorithms. The symmetric key algorithm is vulnerable to interception and key management is a challenge. The strength of this algorithm depends upon length of key. Longer keys are more difficult to break. If the length of a secret key is n bits, at least 2n1 steps would be required to break the encryption. The data encryption standard (DES) defined by US NIST performs encryption in hardware thereby speeding up the encryption and decryption operation. Additional features of DES are: 1. DES is a block cipher and works on a fixed-size block of data. The message is segmented into blocks of plaintext, each comprising 64 bits. A unique 56-bit key is used to encrypt each block of plaintext into a 64-bit block of ciphertext. The receiver uses the same key to perform the decryption operation on each 64-bit data block it receives, thereby reassembling the blocks into a complete message. 2. The larger the key, the more difficult it is for someone to decipher it. DES uses a 56-bit key and provides sufficient security for most commercial applications. Triple-DES is the extended version of DES which applies DES three times with two 56-bit keys. International data encryption algorithm (IDEA) is a block cipher method similar to DES. It operates on 64-bit blocks of plaintext and uses a 128-bit key. The algorithm can be implemented either in hardware or software. It is three times faster than DES and is considered superior to DES.

13.9

Data Security

423

The key sizes used in current wireless systems are not sufficiently large enough for good security. IS-136 uses a 64-bit A-key that is secure, but is still considered to be weak. Public Key (or asymmetric encryption) Public key encryption uses longer keys than does symmetric encryption. The key management problem is greatly reduced because the public key is publicized and the private key is never distributed. There is no need to exchange keys. In a public key system, two keys are used, one for encrypting and one for decrypting. The two keys are mathematically related to each other but knowing one key does not divulge the other key. The two keys are called the “public key” and the “private key” of the user. The network also has a public key and a private key. The sender uses a public key to encrypt the message. The recipient uses its private key to decrypt the message. Public key infrastructure (PKI) is a set of hardware, software, organizations, and policies to public key encryption work on the Internet. There are security firms that provide PKI and deploy encrypted channels to identify users and companies through the use of certificates — VeriSign Inc. Xcert offers products based on PKI. Public Key Algorithms Rivet-Shamir-Adleman (RSA) Algorithm The RSA algorithm [7] is based on public key cryptography. The pretty good privacy (PGP) version of RSA is a public domain implementation available for noncommercial use on the Internet in North America. It is often used to encrypt e-mail. Users make their public keys available by posting them on web pages. Anyone wishing to send an encrypted message to that person copies the public key from the web page into the PGP software and sends the encrypted message using the person’s public key. Two interrelated components of the RSA are (see Figure 13.11): 1. Public key and the private key 2. The encryption and decrypting algorithm

Source: S Kp : public key Plaintext m

Ciphertext: Kp(m)

RSA encryption

Receiver: R Ks: secret key RSA decryption Plaintext: m  Ks [Kp(m)]

Figure 13.11

RSA algorithm operation.

424

13

Security in Wireless Systems

Steps in the RSA algorithm are: • Choose two large prime numbers, p and q (RSA labs recommend that the

• • • •

product of p and q be on the order of 768 bits for personal use and 1024 bits for corporate use). Compute n  pq and z  (p  1)  (q  1). Choose a number, e, less than n, which has no common factors (other than 1) with z (in this case e and z are the prime numbers). Find a number d such that ed  1 is exactly divisible by z. The public key available to the world is the pair of numbers (n, e); and the private key is the pair of numbers (n, d). Encrypted value

me mod(n)  C

(13.1)

Plaintext

m  Cd mod(n)

(13.2)

Example 13.1 Using the prime numbers p  5 and q  7, generate public and private keys for the RSA algorithm. Solution • n  pq  5  7  35, z  (p  1)·(q  1)  4  6  24 • choose e  5, because 5 and 24 have no common factors except 1 • choose d  29 since ed  1  5  29  1  144. This is exactly divisible

by z (24). • Public key (35, 5) • Private key (35, 29) If the sender sends a letter e that has a numeric representation of 5, show that the receiver gets the letter e. The calculations are shown below. Sender: Plaintext letter e Receiver: Ciphertext 10

m: numeric representation 5 cd 1029

me

ciphertext: Cme mod n

55  3125 m  cd mod n 5

10 plaintext letter e

Diffie-Hellman (DH) Algorithm The Diffie-Hellman key exchange algorithm was proposed in 1976. It is a widely used method for key exchange and is based on cyclic groups. In practice, multiplicative groups of prime field Zp or the group

13.9

Data Security

425

of an elliptic curve are most often used. If the parameters are chosen carefully, the DH protocol is secure against passive (i.e., attacker can only eavesdrop) attacks. The DH key exchange is a cryptographic protocol that allows two parties that have no prior knowledge of each other to jointly establish a shared secret key over an insecure communications channel. This key can then be used to encrypt subsequent communications using a symmetric key cipher. The implementation of protocol uses the multiplicative groups of integers modulo p, where p is prime and g is primitive mod p. The algorithm works as follows (see Figure 13.12): 1. Ron and Mike agree to use a prime number p and base g. 2. Mike chooses a secret integer a  {2, 3, 4, …, p  1} and sends Ron g a mod p. 3. Ron chooses a secret integer b  {2, 3, 4, …, p  1} and sends Mike g b mod p. 4. Ron computes (g a mod p)b mod p  K. 5. Mike computes (g b mod p)a mod p  K. 6. Mike and Ron use K as the secret key for encryption. It should be noted that only a, b and gab  gba are kept secret. All other values are sent in clear. Once Mike and Ron compute the shared secret key they can use it as an encryption key, known to them only, for sending messages across the same open communications channel. Example 13.2 Determine the secret encrypting key, K, using the Diffie-Hellman key exchange algorithm, if two parties agree to use a prime number p  23 and base g  5. Party A selects its secret number a  6 and party B chooses its secret number b  15.

Mike a, g, p

Ron

g, p, A

B  g b mod p

A  g a mod p K  B a mod p

b

B

K  Ab mod p

K  Ab mod p  (g a mod p)b mod p  (g b mod p)a mod p  B a mod p Figure 13.12

Diffie-Hellman key exchange algorithm.

426

13

Security in Wireless Systems

Solution 1. 2. 3. 4. 5.

Party A sends to party B ga mod p  56 mod 23  8. Party B sends to party A gb mod p  515 mod 23  19. Party A computes (gb mod p)a mod p  196 mod 23  2. Party B computes (g a mod p)b mod p  815 mod 23  2. Party A and B use K  2 as the secret key for encryption.

One-time Key Method The one-time key method is based on the generation of a new key every time data is transmitted. A single-use key is transmitted in a secure (encoded) mode and, once used, becomes invalid. In some implementations, the central system does not issue a key for a new connection until the user supplies the previously used key. Elliptic Curve Cryptography (ECC) The features of the ECC are discussed below: • ECC is a public key encryption technique that is based on elliptic curve theory. • ECC can be used in conjunction with most public key encryption methods,

such as RSA and Diffie-Hellman. • ECC can yield a level of security with a 164-bit key, while other systems require a 1024-bit key. • Because ECC helps to establish equivalent security with lower computing power and battery resources, it is widely used for mobile applications. • Many manufacturers (3COM, Cylink, Motorola, Pitney Bowes, Siemens, TRW, and VeriFone) have included support for ECC in their products. Digital Signature A digital signature provides a secure and authenticated message transmission (enabled by public key enabling (PKE)). It provides proof identifying the sender. The digital signature includes the name of the sender and other key contents (e.g., date, time, etc). The features of the digital signature method are discussed below: • A digital signature can be used to ensure that users are who they claim to be. • The signing agency signs a document, m, using a private decryption key, dB,

and computes a digital signature dB(m). • The receiver uses the agency’s public key, eB, and applies it to the digital signature, dB(m), associated with the document, m, and computes eb[db(m)] to produce m. • This algorithm is very fast, especially with hash functions.

13.9

Data Security

427

• It is only used in message authentication codes when a secure channel is used

to transmit unencrypted messages, but needs to verify their authenticity. • It is also used in the secure channels of a secure socket layer (SSL).

13.9.3 Secure Socket Layer SSL is a protocol that uses a session-level layer in the Internet to provide a secure channel. SSL is widely used on the web. In SSL, the server sends its public key and encryption technique to be used to the browser. The browser generates a key for the encryption technique and sends it to the server. Communications between server and browser are encrypted using the key generated by the browser. The features of SSL are: • Negotiate cipher suite which is a collection of encryption and authentication

algorithms. • Bootstrapped secure communication, which eliminates the need for third

parties, and uses unencrypted communications for initial exchanges. • Public-key crypto for secret keys and secret-key crypto for data.

13.9.4 IP Security Protocol (IPSec) IPSec is a widely used protocol that can be employed with other application layer protocols (not just for web applications such as SSL). The operations of IPSec between A and B involve: • A and B generate and exchange two random keys using Internet key

exchange (IKE). • A and B combine the two numbers to create an encryption key to be used between them. • A and B negotiate the encryption technique to be used such as DES or 3DES. • A and B then begin transmitting data using either the transport mode in which only the IP payload is encrypted or tunnel mode in which the entire IP packet is encrypted.

13.9.5 Authentication Protocols Authentication of a user is used to ensure that only the authorized user is permitted into the network and into the specific resource inside the network. Several methods used for authentication are user profile, user account, user password, biometrics, and network authentication. The user profile is assigned to each user account by the manager. The user profile determines the limits of a user in accessing the network (i.e., allowable login day and time of day, allowable physical locations, allowable number of incorrect login attempts). The user profile specifies access details such as data and network resources that a user can access and type of access (e.g., read, write, create, delete). The form of access to the network may be based on the password, card, or one-time password.

428

13

Security in Wireless Systems

With a biometric-based form of access, the user can gain access based on finger, hand, or retina scanning by a biometric system. It is convenient and does not require remembering a password. Biometric-based methods are used in high security applications. Network authentication requires a user to login to an authentication server, which checks user ID and password against a database, and issues a certificate. The certificate is used by the user for all transactions requiring authentications. Kerberos is one of many commonly used authentication protocols. Two other authentication protocols that have been used are remote authentication dial-in user service (RADIUS) and terminal access controller access control system(TACACS). Kerberos is a secret-key network authentication protocol that uses a DES cryptographic algorithm for encryption and authentication. It was designed to authenticate requests for network resources. Kerberos is based on the concept of a trusted third party that performs secure verification of users and services. The primary use of Kerberos is to verify that users and the network services they use really are who and what they claim to be. To accomplish this, a trusted Kerberos server issues tickets to users. These tickets, which have a limited life span, are stored in a user’s credential cache. The tickets are used in place of standard user name and password authentication mechanisms. RADIUS is a distributed client/server system that secures the network against unauthorized access. In the Cisco implementation, RADIUS clients run on Cisco routers and send authentication requests to a server. The central server contains all user authentication and network service access information. RADIUS is the only security protocol supported by wireless authentication protocol. TACACSⴙ(improved TACAS) is a security application that provides centralized validation of users attempting to gain access to a router or network access server. TACACS services are maintained in a database on a TACACS daemon running on a UNIX, Windows NT, Window 2000 workstation. TACACS provides for separate and modular authentication, authorization, and accounting facilities. A network administrator may allow remote users to have network access through public services based on remote-access solutions. The network must be designed to control who is allowed to connect to it, and what they are allowed to do once they get connected. The network administrator may find it necessary to configure an accounting system that tracks who logs in, when they log in, and what they do once they have logged in. Authentication, authorization, and accounting (AAA) security services provide a framework for these kinds of access control and accounting functions. The user dials into an access server that is configured with challenge handshake authentication protocol (CHAP). The access server prompts the user for a name and password. The access server authenticates the user’s identity by requiring the user name and password. This process of verification to gain access is called authentication. The user may now be able to execute commands on that server once it has been successfully authenticated. The server uses a process for authorization to determine which commands and resources should be made available to that particular user. Authorization asks

13.10

Air Interface Support for Authentication Methods

429

the question, what privileges does this user have? Finally, the number of login attempts, the specific commands entered, and other system events can be logged and time-stamped by the accounting process. Accounting can be used to trace a problem, such as a security breach, or it may be used to compile usage statistics or billing data. Accounting asks questions such as: what did this user do and when was it done? The following are some of the advantages in using AAA: • AAA provides scalability. Typical AAA configurations rely on a server or

group of servers to store user names and passwords. This means that local databases don’t have to be built and updated on every router and access server in the network. • AAA supports standardized security protocols — TACACS, RADIUS, and Kerberos. • AAA lets the administrator configure multiple backup systems. For example, an access server can be configured to consult a security server first and a local database second. • AAA provides an architectural framework for configuring three different security features: authentication, authorization, and accounting.

13.10

Air Interface Support for Authentication Methods

The various air interfaces used for PCS and cellular systems in Europe and North America support one or more of the different authentication methods. Only the older AMPS supports MIN/ESN as the authentication method. All of the digital systems in North America, except for GSM1900, support SSD. GSM supports only token-based authentication. UMTS supports token-based authentication along with some advanced security features. cdma2000 supports SSD. Table 13.1 summarizes this information. Table 13.1 Summary of authentication methods for PCS and cellular systems in Europe and North America. Type of Authentication

Air interface AMPS

MIN/ESN

SSD

X

TDMA IS-136

X

GSM

UMTS

Public key

None

X

CDMA IS-95

cdma2000

Token-based

Type of voice privacy supported

Strong Strong X

Strong Strong

X X

Strong

430

13.11

13

Security in Wireless Systems

Summary of Security in Current Wireless Systems

Each of the security methods satisfies the security needs for a wireless system in different ways. The older AMPS has poor security. The digital systems using either SSD or tokens meet most of the security needs of the wireless systems except full anonymity. The public-key-based security system meets all the requirements, including anonymity, but is not yet fully implemented. Privacy of communications is maintained via encryption of signaling, voice, and data for the digital systems. The AMPS sends all data in the clear and has no privacy unless the user adds it to the system. The following is a summary of the support for security requirements for the PCS and cellular systems in North America and Europe (see Table 13.2). Table 13.2 Summary of support for security requirements for PCS and cellular systems in Europe and North America.

Feature

MIN/ESN (AMPS)

SSD

Token-based

Public key

Privacy of Communication • Signaling

None

High: messages are encrypted

High: messages are encrypted

High: messages are encrypted

• Voice

None

High: voice is encrypted

High: voice is encrypted

High: voice is encrypted

• Data

None

High: data is encrypted

High: data is encrypted

High: data is encrypted

High: if authentication is done

High: if authentication is done

High: if authentication is done

Billing Accuracy • Accuracy

None: phones can be cloned

Privacy of User Information • Location

None

Moderate: using IMSI/ TMSI

Moderate: using IMSI/ TMSI

High: public key provides full anonymity

• User ID

None

Moderate: using IMSI/ TMSI

Moderate: using IMSI/ TMSI

High: public key provides full anonymity

• Calling Pattern

None

High: using TMSI and encryption

High: using TMSI and encryption

High: public key provides full anonymity

Theft Resistance of MS • Over the air • From Network

None Depends on system design

High Depends on system design

High Depends on system design

High Depends on system design

(Continued)

13.11

Summary of Security in Current Wireless Systems

Feature • From interconnection • Cloning

MIN/ESN (AMPS) Depends on system design

SSD Depends on system design

Token-based Depends on system design

431

Public key Depends on system design

None

High

Medium

High

Handset Design

Algorithm run in microprocessor of handset

Algorithm run in microprocessor of handset

Algorithm run in microprocessor of handset

Microprocessor speed may be fast enough for some algorithms

Law Enforcement Needs

Easily met on the air interface (if van is nearby to MS or at the switch)

Must wiretap at the switch

Must wiretap at the switch

Must wiretap at the switch

13.11.1 Billing Accuracy Since AMPS phones can be cloned from data intercepted over the radio link, billing accuracy for AMPS is low to none. For other systems, when authentication is done, billing accuracy is high. If a system operator gives service before authentication or even if authentication failure occurs, then billing accuracy will be low. 13.11.2 Privacy of Information Privacy of user information is high for the public key system, moderate for the SSD and token-based systems (since sometimes IMSI is sent in cleartext) and low for the AMPS. 13.11.3 Theft Resistance of MS MS theft resistance is high over-the-air transmission for all systems except the AMPS. Since the token-based system in GSM doesn’t support a call history count, it has a lower resistance to cloning than the SSD or public key systems. Earlier AMPS phones using MIN/ESN have no resistance to cloning, but now they support SSD. The resistance of stealing data from network interconnects or from operations systems (OS) in the network depends on the system design. 13.11.4 Handset Design All of the authentication and privacy algorithms easily run in a standard 8-bit microprocessor used in mobile stations, except the public key systems. 13.11.5 Law Enforcement The AMPS is relatively easy to tap at the air interface. The digital systems will require a network interface since privacy is maintained over the air interface.

432

13

Security in Wireless Systems

The network requirements currently meet most of the needs of the law enforcement community doing legal wiretaps.

13.12

Summary

In this chapter, we discussed the requirements for strong privacy and authentication of wireless systems, and outlined how each of the cellular and PCS systems supports these requirements. Four levels of voice privacy were presented. We then identified requirements in the areas of privacy, theft resistance, radio system requirements, system lifetime, physical requirements as implemented in mobile stations, and law enforcement needs. We also examined different methods of authentication that are in use to satisfy these needs. The chapter described the requirements that any cryptographic system should meet to be suitable for use in a ubiquitous wireless network. We also examined security models and described how they met security requirements.

Problems 13.1 Define the equivalent to wireline and commercially secure systems. 13.2 What are the privacy requirements of a wireless telephone user? 13.3 Define the methods that are used in wireless systems to provide privacy and security. 13.4 What are the main differences between global and unique challenge methods used in North American cellular systems? 13.5 Describe the token-based authentication scheme used in the GSM system. 13.6 Describe the public-key-based authentication scheme. 13.7 Describe how you would design a mobile station so that the security data stored in the terminal is tamper resistant. 13.8 Describe data encryption keys. 13.9 Describe the call flows that would be necessary for a shared secret key registration when the old VLR must first be queried for IMSI before the correct HLR can be queried. 13.10 Describe the call flows that would be necessary for a token-based registration in GSM when the old VLR cannot be queried for IMSI and the network must request that the mobile station send its IMSI before messages can be exchanged with the HLR.

References

433

References 1. D’Angelo, D. M., McNair, B., and Wilkes, J. E. Security in Electronic Messaging Systems. AT&T Technical Journal, 73, no. 3, May/June 1994. 2. JTC(AIR)/94.03.25-257R1. “Minimum Requirements for PCS Air Interface Privacy and Authentication.” 3. Karygiannis, T., and Owens, L. Wireless Network Security. NIST Special Publication 800-48, November 2002. 4. Owens L., and Crowe, David. Wireless Security Perspectives. Calgary, Canada: Cellular Networking Perspectives Ltd., 1999–2001. 5. Paar, C. Lectures Notes — Applied Cryptography and Data Security, version 2.5, January 2005. 6. Report of the Joint Experts Meeting on Privacy and Authentication for PCS. Phoenix, Arizona, November 8–12, 1993. 7. Rivet, R. L., Shamir, A., and Adleman, L. A Method for Obtaining Digital Structures and Public-Key Crypto Systems. Communications ACM 21, no. 2, February 1978, 120–127. 8. Snyder, R. A., and Gallagher, M. D. Wireless Telecommunications Networking with ANSI41, second edition. New York: McGraw-Hill, 2001. 9. TIA Interim Standard, IS-41 C, “Cellular Radio Telecommunication Intersystem Operations.” 10. TR-46 P&A ad hoc/94.04.17.01R5, “TR-46 PCS Privacy and Authentication, Volume 1, Common Requirements,” Version 6, November 1994. 11. TR-46 P&A ad hoc/94.04.17.02R4, “TR-46 PCS Privacy and Authentication, Volume 2, PCS1900 Based Requirements.” 12. TR-46 P&A ad hoc/94.04.17.02R3, “TR-46 PCS Privacy and Authentication, Volume 3, Shared Secret Data Requirements.” 13. Wilkes, J. E. Privacy and Authentication Needs of PCS. IEEE Personal Communications, 2, no. 4, August 1995.

This page intentionally left blank

CHAPTER 14 Mobile Network and Transport Layer 14.1

Introduction

In this chapter, we first introduce concepts of the Internet and briefly discuss IPv4 and IPv6. We then focus on mobile internet protocol (MIP) and session initiation protocol (SIP) designed to provide a variety of IP data services in mobile networks [1–9]. We conclude the chapter by discussing transmission control protocol (TCP) enhancements for mixed networks (wireline and wireless). The chapter is primarily intended for those readers who are not familiar with the Internet and TCP/IP. The chapter may be omitted by those who have sufficient exposure to TCP/IP and the Internet. The Internet is a global network that supports a variety of interpersonal and interactive multimedia applications and operates in a packet-switched mode [15,20]. The Internet consists of a large number of different access networks that are interconnected by a global internetwork. Associated with each access network is the Internet Service Provider (ISP) network, intranet, enterprise network, local area network (LAN), and so on, that is referred to as a gateway. The global internetwork consists of an interconnected set of regional, national, and international networks all of which are joined together using high bit rate leased lines and routers. Each of the routers in the interconnection network has a number of access networks attached to it by means of access gateways. Each access network is an LAN with a single netid and the address resolution protocol (ARP) in each host is used to carry out routing within the access network. Hence, the ARP in each gateway acts as an agent (proxy ARP) on behalf of all hosts at the site/campus to relay packets to and from the interconnection network. ARP and reverse address resolution protocol (RARP) are specialized protocols used only with certain types of network interfaces (such as Ethernet and token-ring) to convert between the addresses used by the IP layer and the addresses used by the network interfaces. The interconnection network comprises four routers (R1–R4) that are interconnected by, say, leased lines (refer to Figure 14.1). Each leased line has a pair of numbers associated with it. The first number is used as a leased line identifier and the second is referred to as the cost of the line. Let us assume that the host

435

436

14

Netid 2

Mobile Network and Transport Layer

Interconnection Network

Netid 3 Host

G2

G3

B

2,1 R3

R2

5,1 3,2

1,2 R1

Netid 1

4,2

G1

Host

Access network (e.g., site/campus LAN)

R4

Netid 4 G4

A

R  Router G  Access Gateway

Figure 14.1

Example of internetwork topology.

A on netid 1 wants to send a datagram to host B on netid 3. On determining that the destination netid in the datagram header is for a different netid from its own, gateway G1 forwards the datagram to router R1 over the connecting line/ link. On receipt of the datagram, R1 proceeds to forward it first to R3 over the interconnection network and then it is forwarded by R3 to G3. At that point, the IP in G3 knows how to route the datagram to host B using the hostid part of the destination IP address and the related media access control (MAC) address of the host B in its ARP cache. Thus, each IP address has two parts: netid that is centrally managed by the Internet information center and hostid that is allocated by the local administrator of the access network to which the host is attached.

14.2 Concept of the Transmission Control Protocol/Internet Protocol Suite in Internet Assume a process associated with port 1 at host A wishes to send a message to another process associated with port 3 at host B (see Figure 14.2). The process at host A hands the message to the TCP with instructions to send it to host B, port 3. TCP hands the message to IP with instructions to send it to host B. Note that IP need not to be told the identity of the destination port. All it needs to know is that the data is intended for host B. Next, IP hands the message down to the network access layer (NAP) (e.g., Ethernet logic) with instructions to send it to router J that is the first hop on the way to B. To control this operation, control information as well as user data must be transmitted.

14.2 Concept of the Transmission Control Protocol/Internet Protocol Suite in Internet 437

Host A

App X App Y

Host B

Port or Service access point (SAP)

App Y App X

Logical Connection (TCP Connection) TCP

3

1

IP

Global Network address Logical connection (e.g., virtual circuit)

Network Access Protocol (NAP) #1 Physical

Router J

TCP IP

Network Access Protocol #2 Physical

IP Sub network attachment point address Network 1

Figure 14.2

NAP1

NAP2

Phy

Phy

Network 2

TCP/IP concept.

We assume the network interface card in all the hosts that are attached to an access network communicate with other hosts using the TCP/IP protocol stack. The various access networks have different operational parameters associated with them in terms of their bit rate, frame format, maximum frame size, and type of address that are used. Therefore, the routing and forwarding operations associated with gateway are performed at the network layer. In the TCP/IP protocol stack the network layer protocol is IP. In order to transfer packets of information from one host to another, the IP in the two hosts together with the IP in each Internet gateway and router are involved in performing the routing and other harmonization functions necessary. The IP in each host has a unique address assigned to it. TCP/user datagram protocol (UDP) passes the block of information to its local IP together with the IP address of the intended recipient. The source IP adds the destination and source IP addresses to the head of the block, together with an indication of the source protocol (TCP or UDP), to form an IP datagram. The IP forwards the datagram to its local gateway. The TCP/IP protocol suite allows computers of all sizes, supplied by many different computer vendors, running totally different operating systems, to communicate with each other. TCP/IP is a 4-layer system (PL, LL, NL, TL)

438

14

Mobile Network and Transport Layer

Source Host

Destination Host Application PDUs

AL

Application

Application TCP/UDP PDUs

TL

TCP/UDP

TCP/UDP Access Network Gateway

Router

Access Network Gateway

IP

IP

IP

NL

IP

LL

LPX

LPX

LPI

LPI

LPI

LPI

LPY

LPY

PL

PLX

PLX

PLI

PLI

PLI

PLI

PLY

PLY

Access Network X

Internet Global Internetwork

Logical communication path of PDUs

IP

Access Network Y

Actual path

IP : Internet Protocol LP : Link Protocol PL : Physical layer LL : Link Layer AL : Application layer NL : Network layer TL : Transport layer

Figure 14.3

Internet networking components and protocols.

(see Figure 14.3). The link layer (LL) normally includes the device driver in the operating system and the corresponding network interface card in the computer. Together they handle all the hardware details of physically interfacing with cable (or whatever type media is being used). The network or internet layer handles the movement of packets around the network. Internet protocol (IP), Internet control message protocol (ICMP), and Internet group management protocol (IGMP) are used at the network layer. IP is the workhorse protocol of the TCP/IP protocol suite. All TCP, UDP, ICMP, and IGMP data get transmitted as IP datagrams [10,11,19,30]. IP provides an unreliable connectionless datagram delivery service. By unreliable we mean there are no guarantees that an IP datagram successfully gets to its destination. IP provides a best effort service when something goes wrong, such as a router temporarily running out of buffers. IP has a simple error handling algorithm: throw away the datagram and try to send an ICMP message back to source. The term connectionless means that the IP does not maintain any state information about successive datagrams. Each datagram is handled independently from all other datagrams. IP datagrams are delivered out of order.

14.3

Network Layer in the Internet

439

ICMP is an adjunct to IP. It is used by an IP layer to exchange error messages and other vital information with the IP layer in another host or router. IGMP is used with multicasting: sending a UDP datagram to multiple hosts. The transport layer (TL) provides a flow of data between two hosts, for the application layer above. In the TCP/IP protocol suite two different transport protocols, TCP and UDP, are used. TCP provides a reliable flow of data between two hosts. It is concerned with data segmentation passed to it from application into appropriately sized segments for the network layer below, acknowledging received packets, setting timeouts to make certain the other end acknowledges packets that are sent, and so on. UDP provides a much simpler service to the application layer. It sends packets of data called datagrams from one host to the other, but there is no guarantee that the datagrams will reach the other end. Any desired reliability must be added by the application layer.

14.3

Network Layer in the Internet

The IP provides the basis for the interconnections of the Internet [16–18]. IP is a datagram protocol. The packets contain an IP header. The basic header, without options, is shown in Figure 14.4. The version field contains the version of IP—IPv4 or IPv6. The internet header length (IHL) field specifies the actual length of the header in multiples of 32-bit words. The minimum length is 5. The maximum permissible length is 15. The type of service (TOS) field allows an application protocol/process to specify the relative priority of the application data and the preferred attributes associated with the path to be 4 Version

8

16 Type of Service IHL (TOS) Identification D M

Time-to-Live

Protocol

32 Total Length Fragment Offset Header Checksum

Source IP Address Destination IP Address Options Payload (65,535 bytes)

3

1 1 1

Precedence D T R Priority (0 – 7) Low Delay Figure 14.4

IP header.

2 Unused

Type of Service

High reliability High Throughput

440

14

Mobile Network and Transport Layer

followed. It is used by each gateway and router during the transmission and routing of the packet to transmit packets of higher priority first and to select a line/route that has the specified attributes should a choice be available. The total length field defines the total length of the initial datagram including the header and payload parts. When the contents of the initial datagram need to be transferred in multiple packets, then the value in this field is used by the destination host to reassemble the payload contained within each smaller packet — known as a fragment — into the original payload. The identification field enables the destination host to relate each received packet fragment to the same original datagram. Don’t fragment or D-bit is set by a source host and is examined by routers. A D-bit indicates that the packet should not be fragmented. More fragment or M-bit is used during the reassembly procedure associated with data transfers involving multiple smaller packets/fragments. It is set to 1 in all but the last packet/fragment in which it is set to 0. The fragment offset field is used to indicate the position of the first byte of the fragment contained within a smaller packet in relation to the original packet payload. All fragments except the last one are in multiples of 8 bytes. The time-to-live field defines the maximum time for which a packet can be in transit across the Internet. The value is in seconds and is set by the IP in the source host. It is decremented by each gateway and router by a defined amount and should the value become zero, the packet is discarded. The protocol field is used to enable the destination IP to pass the payload within each received packet to the same (peer) protocol that sent the data. This can be an internal network layer protocol such as the ICMP or a higher-layer protocol such as TCP or UDP. The header checksum applies just to the header part of the datagram and is a safeguard against corrupted packets being routed to incorrect destinations. The source and destination Internet addresses indicate the sending host and the intended recipient host for this datagram. The options field is used in selected datagrams to carry additional information relating to security, source routing, loose source routing, route recording, stream identification, and time-stamp. The last field is the payload. A symbolic address, or name, of the form [email protected] can be used instead of an Internet address. It is translated into an Internet address by directory tables that are organized along the same hierarchy as the addressing. Typically, the domain is of the form machine.institution.type.country. The type is edu for educational institutions, com for companies, gov for governmental agencies, org for nonprofit organizations, and mil for military. The country field is omitted for the United States and is a two-letter country code for the other countries (e.g., fr for France). For instance, the author’s address is [email protected] With best-effort delivery service (optional quality of service (QoS)), IP packets may be lost, corrupted, delivered out-of-order, or duplicated. The upper layer entities should anticipate and recover on an end-to-end basis.

14.3

Network Layer in the Internet

1

7

441

24 bits

0

Class A

hostid

netid  1 to 127  126 networks; hostid  0.0.0 to 255.255.255  16,777,214 hosts

2

14

16 bits

1 0

hostid

Class B unicast

netid  128 to 191.255  16,382 networks; hostid  0.0.0 to 255.255  65,534 hosts

3

21

8 bits

1 1 0

hostid

Class C

netid  192.0 to 223.255.255  2,097,152 networks; hostid  0 to 255  256 hosts

4

28 bits

1 1 1 0

Multicast Address

Class E  reserved

1 1 1 1

Figure 14.5

Class D  multicast

Internet addresses.

14.3.1 Internet Addresses Three classes of Internet addresses (unicast) are used (see Figure 14.5): • Class A — 7 bits for netid and 24 bits for hostid, they are used with networks

having a large number of hosts (224) • Class B — 14 bits for netid and 16 bits for hostid, they are used with networks having a medium number of hosts (216) • Class C — 21 bits for netid and 8 bits for hostid, they are used with networks having a small number of hosts (28) It should be noted that the netid and hostid with all 0s or all 1s have special meaning. • An address with hostid of all 0s is used to refer to the network in netid part

rather than a host • An address with a netid of all 0s implies the same network as the source

network/netid • An address of all 1s means broadcast the packet over the source network • An address with a hostid of all 1s means broadcast the packet over the

destination network in netid part A class A address with a netid of all 1s is used for test purposes within the protocol stack of the source host. It is known as the loop-back address.

442

14

AL

AP

TL

AP

Mobile Network and Transport Layer

AP

TCP

ICMP NL

AP

UDP

IP OSPF

ARP

LL

IGMP RARP

Link Layer Protocol Physical Layer

PL

Network Point of Attachment

AP: Application Protocol/process ARP: Address Resolution Protocol RARP: Reverse ARP ICMP: Internet Control Message Protocol IGMP: Internet Group Message Protocol OSPF: Open Shortest Path First UDP: User Datagram Protocol TCP: Transmission Control Protocol

Figure 14.6

Adjunct protocols.

14.3.2 IP Adjunct Protocols Figure 14.6 shows the IP adjunct protocols [25–29]. • Address resolution protocol (ARP) and reverse ARP (RARP) are used by IP

in hosts that are attached to a broadcast LAN (such as Ethernet or token ring) in order to determine the physical MAC address of a host or gateway given its IP address (ARP), and, in case of the RARP, the reverse function. • Open shortest path first (OSPF) protocol is a routing protocol used in the global internetwork. Such protocols are present in each internetwork router. They are used to build up the contents of the routing table used to route packets across the global internetwork. • Internet control message protocol (ICMP) is used by the IP in a host or gateway to exchange errors and other control messages with IP in another host or gateway. • Internet group message protocol (IGMP) is used with multicasting to enable a host to send a copy of a datagram to the other hosts that are part of the same multicast group.

14.3

Network Layer in the Internet

443

The ICMP forms an integral part of all IP implementations. It is used by hosts, routers, and gateways for a variety of functions, and especially by network management. The main functions associated with the ICMP are as follows: • Error reporting • Reachability testing • Congestion control • Route-change notification • Performance measuring • Subnet addressing

The standard way to send an IP packet over any point-to-point link is either dial-up modems (e.g., asynch framing), leased lines (e.g., bit synchronous framing), or ISDN, IS-99 CDMA (e.g., octet-synchronous framing). The link control protocol (LCP) runs during initial link establishment and negotiates link-level parameters (e.g., maximum frame size, etc.). The IP control protocol (IPCP) establishes the IP address of the client (the point-to-point (PPP) server, allocates a temporary address, or the client notifies the server of the fixed address) and negotiates for the use of TCP/IP header compression.

14.3.3 QoS Support in the Internet QoS requirements include a defined minimum mean packet throughput rate and a maximum end-to-end packet transfer delay. To meet the varied set of QoS requirements, two schemes have been standardized: • Integrated Services (IntServ) • Differentiated Services (DiffServ)

Packets relating to different types of call/session are each allocated a different value in the precedence bits of the type of service (TOS) field of the IP packet header. This is used by routers within the Internet to differentiate between the packet flows relating to different types of calls. Integrated Services (IntServ) The IntServ solution defines three different classes of service: • Guaranteed: It specifies maximum delay and jitter, and an assured level of

bandwidth is guaranteed. • Controlled load (also known as predictive): No firm guarantee is provided but the flow obtains a constant level of service equivalent to that obtained with the best-effort service at light loads. • Best-effort: This is intended for text-based applications. To cater for three different types of packet flows within each router, three separate output queues are used for each line — one for each class. Appropriate

444

14

Mobile Network and Transport Layer

scheduling mechanisms are used to ensure that the QoS requirements of each class are met. Differentiated Service (DiffServ) Incoming packet flows relating to individual calls are classified by the router/gateway at the edge of the DiffServ compliant net/Internet into one of the defined service/traffic classes by examining selected fields in various headers in the packet. The TOS field in the IP packet header is replaced by a new field called the differentiated service (DS) field. Within the DiffServ network, a defined level of resources in terms of buffer space within each router and the bandwidth of each output line is allocated to each traffic class. Internet Protocol version 6 (IPv6) Today’s Internet operates over the common network layer datagram protocol, Internet Protocol version 4 (IPv4). In the early 1990s, a new design of addressing scheme was initiated within the Internet Engineering Task Force (IETF) due to the recognized weaknesses of IPv4. The result was IPv6 (see Figure 14.7). The single most significant advantage IPv6 offers is increased destination and source addresses. IPv6 quadruples the number of network address bits from 32 bits in 32 bits Version

Traffic Class

Payload Length

Flow Level Next Header

Hop Limit

Source Address (128 bits) Destination Address (128 bits) Payload Version: 4-bit field to identify the IP version number Traffic class: 8-bit field is similar to Type of Service field in IPv4 Flow level: This 20-bit field is used to identify a “flow” of datagrams Payload length: 16-bit is treated as an unsigned integer to give the number of bytes in the datagram Next header: This field identifies the protocol to which the contents of this datagram will be delivered (for example TCP or UDP). The field uses the same values as Protocol field in IPv4 header Hop limit: The contents of this field are decremented by one by each router that forwards the datagrams. If the hop limit count reaches zero, the datagram is discarded Source and destination address: 128-bit field

Figure 14.7

IPv6.

14.3

Network Layer in the Internet

445

IPv4 to 128 bits, which provides more than enough globally unique IP addresses for every network device on the planet. This will lead to network simplification, first, through less need to maintain a routing state within the network and second, through reduced need for address translation; hence, it will improve the scalability of the Internet. IPv6 will allow a return to a global end-to-end environment where the addressing rules of the network are transparent to applications. The current IP address space is unable to satisfy the potentially large increase in number of users or the geographical needs of Internet expansion, let alone the requirements of emerging applications such as Internet-enabled personal digital assistants (PDAs), personal area networks (PANs), Internet-connected transportation, integrated telephony services, and distributed gaming. The use of globally unique IPv6 addresses simplifies the mechanisms used for reachability and end-to-end security for network devices, functionally crucial to the applications and services driving the demand for the addresses. The lifetime of IPv4 has been extended using techniques such as address reuse with translation and temporary use allocations. Although these techniques appear to increase the address space and satisfy the traditional client/server setup, they fail to meet the requirements of new applications. The need for an always-on environment to be connectable precludes these IP address conversion, pooling, and temporary allocation techniques, and the “plug and play” required by consumer Internet applications further increases address requirements. The flexibility of the IPv6 address space provides the support for private addresses but should reduce the use of network address translation (NAT) because global addresses are widely available. IPv6 reintroduces end-to-end security that is not always readily available throughout an NAT-based network. The success of IPv6 will depend ultimately on the innovative applications that run over IPv6. A key part of IPv6 design is its ability to integrate into and coexist with existing IP networks. It is expected that IPv4 and IPv6 hosts will need to coexist for a substantial time during the steady migration from IPv4 to IPv6, and the development of transition strategies, tools, and mechanisms has been part of the basic IPv6 design from the start. Selection of a deployment strategy will depend on current network environment, and factors such as the forecast of traffic for IPv6 and availability of IPv6 applications on end systems. IPv6 does not allow for fragmentation and reassembly at an intermediate router; these operations can be performed only by the source and destination. If an IPv6 datagram received by a router is too large to be forwarded over the outgoing link, the router simply drops the datagram and sends a packet too big ICMP message back to sender. The checksum field in IPv4 was considered redundant and was removed because the transport layer and data link layer protocols perform checksum.

446

14

Telnet

FTP

HTTP

SMTP

Mobile Network and Transport Layer

Voice over IP

DNS

TCP

ICMP

NFS

UDP

Internet Protocol (IP)

ARP

PPP

Other Subnets

Enet FTP: File transfer protocol SMTP: Simple mail transport protocol HTTP: Hyper text transfer protocol PPP: Point-to-point protocol ARP: Address resolution protocol ICMP: Internet control message protocol UDP: User datagram protocol TCP: Transmission control protocol

Figure 14.8

14.4

TCP/IP protocol suite.

TCP/IP Suite

The TCP/IP suite (Figure 14.8) occupies the middle five layers of the 7-layer open system interconnection (OSI) model (see Figure 14.9) [30]. The TCP/IP layering scheme combines several of the OSI layers. From an implementation standpoint, the TCP/IP stack encapsulates the network layer (OSI layer 3) and transport layer (OSI layer 4). The physical layer, the data-link layer (OSI layer 1 and 2, respectively) and application layer (OSI layer 7) at the top can be considered non-TCP/ IP-specific. TCP/IP can be adapted to many different physical media types. IP is the basic protocol. This protocol operates at the network layer (layer 3) in the OSI model, and is responsible for encapsulating all upper layer transport and application protocols. The IP network layer incorporates the necessary elements for addressing and subnetting (dividing the network into subnets), which enables TCP/IP packets to be routed across the network to their destinations. At a parallel level, the ARP serves as a helper protocol, mapping physical layer addresses typically referred to as MAC-layer addresses to network layer (IP) addresses. There are two transport layer protocols above IP: the UDP and TCP. These transport protocols provide delivery services. UDP is a connectionless

14.4

TCP/IP Suite

447

OSI Model

TCP/IP Layers

User application process

Distributed information services File transfer, mail exchange, management functions, remote access

Application Layer

Application Layer— SMTP, HTTP, etc.

TCP/IP Applications

Syntax independent data transfer Syntax Management, Data representation translation

Presentation Layer

Synchronization and dialog control

Session Layer

Network independent data exchange

Socket Layer

End-to-end message transfer connection management, error control, and flow control

Transport Layer

Transport Layer UDP, TCP

Network routing & addressing

Network Layer

Network Layer IP (ARP)

Framing, data transparency, and error control

Data Link Layer

Data Link Layer (10Base-T)

Electrical and mechanical network interface definitions

Physical Layer

TCP/IP Stack Core

Physical Connection to Network Data Communications Network

Figure 14.9

Data Communications Network

A comparison of the OSI model and TCP/IP protocol layers.

delivery transport protocol and used for message-based traffic where sessions are unnecessary. TCP is a connection-oriented protocol that employs sessions for ongoing data exchange. File transfer protocol (FTP) and Telnet are examples of applications that use TCP sessions for their transport. TCP also provides the reliability of having all packets acknowledged and sequenced. If data is dropped

448

14

Mobile Network and Transport Layer

or arrives out-of-sequence, the stack’s TCP layer will retransmit and resequence. UDP is an unreliable service, and has no such provisions. Applications such as the simple mail transport protocol (SMTP) and hyper text transfer protocol (HTTP) use transport protocols to encapsulate their information and/or connections. To enable similar applications to talk to one another, TCP/IP has what are called “well-known port numbers.” These ports are used as sub-addresses within packets to identify exactly which service or protocol a packet is destined for on a particular host. TCP/IP serves as a conduit to and from devices, enabling the sharing, monitoring, or controlling those devices. A TCP/IP stack can have a tremendous effect on a device’s memory resources and CPU utilization. Interactions with other parts of the system may be highly undesirable and unpredictable. Problems in TCP/IP stacks can render a system inoperable.

14.5

Transmission Control Protocol

The TCP [31,32] is the connection-oriented transport layer protocol designed to operate on the top of the datagram network layer IP. The two widely used protocols are known under the collective name TCP/IP. TCP provides a reliable end-to-end byte stream transport. The segmentation and reassembly of the messages are handled by IP, not by TCP. TCP uses the selective repeat protocol (SRP) with positive acknowledgments and time-out. Each byte sent is numbered and must be acknowledged. A number of bytes can be sent in the same packet, and the acknowledgment (ACK) then indicates the sequence number of the next byte expected by the receiver. ACK carrying sequence number m provides acknowledgment for all packets up to, and including, packets with sequence number m  1. If a packet is lost, the receiver sends duplicate ACK for a subsequent correctly received packet. The TCP header is at least 20 bytes, and has 16 error detection bits for the data and the header. The error detection bits are calculated by summing the 1’s complements of the groups of 16 bits that make up the data and the header, and by taking the 1’s complement of that sum. The number of data that can be sent before being acknowledged is the window size (Wmax) which can be adjusted by either the sender or the receiver to control the flow based on the available buffers and the congestion. Initial sequence numbers are negotiated by means of a threeway handshake at the outset of connection. Connections are released by means of a three-way handshake. TCP transmitter (Tx) uses an adaptive window based transmit strategy. Tx does not allow more than Wmax unacknowledged packets outstanding at any given time. With the congestion window lower limit at time t equal to X(t), packets up to X(t)1 have been transmitted and acknowledged. Tx can send starting from

14.5

Transmission Control Protocol

449

X(t). X(t) has a nondecreasing sample path. With congestion window width at time t equal to W(t), this is the amount of packets Tx is allowed to send starting with X(t). W(t) can increase or decrease (because of window adaptation), but never exceed Wmax. Transitions in X(t) and W(t) are triggered by receipt of ACK. Receiver (Rx) of an ACK increases X(t) by an amount equal to the amount of data acknowledged. Changes in W(t), however, depend on the version of TCP and the congestion control process. Tx starts a timer each time a new packet is sent. If the timer reaches a round trip time-out (RTO) value before the packet is acknowledged, a time-out occurs. Retransmission is initiated on time-out. RTO value is derived from a round trip timer estimation procedure. RTO is sent only in multiples of a timer granularity. The window adaptation procedure is as follows: 1. Slow start phase. At the beginning of the TCP connection, the sender enters the slow start phase, in which window size is increased by 1 maximum segment size (MSS) for every ACK received; thus, the TCP sender window grows exponentially in round trip timer. If W  Wth, W ← W  1 for each ACK received; Wth is the slow-start threshold.

2. Congestion avoidance phase. When the window size reaches Wth, the TCP sender enters the congestion avoidance phase. TCP uses a sliding windowbased flow control mechanism allowing the sender to advance the transmission window linearly by one segment upon reception of an ACK, which indicates that the last in-order packet was received successfully by the receiver. If W ← W  1/W for each ACK received

3. Upon time-out. When packet loss occurs at a congested link due to buffer overflow at the intermediate router, either the sender receives duplicate ACKs, or the sender’s RTO timer expires. These events activate TCP’s fast retransmit and recovery, by which the sender reduces the size of the congestion window to half and linearly increases the congestion window as in the congestion avoidance phase, resulting in a lower transmission rate to relieve the link congestion. W  ← 1 and Wth ← W/2

Assuming long running connections and large enough window sizes, the upper bound on throughput, R, of a TCP connection is given by: 0.93MSS R  RTTp

(14.1)

450

14

Mobile Network and Transport Layer

where: MSS  maximum segment size RTT  average end-to-end round trip time of the TCP connection p  packet loss probability for the path. Equation 14.1 neglects retransmissions due to errors. If the error rate is more than 1%, these retransmissions have to be considered. This leads to the following formula: MSS R    

RTT1.33p  RTO · p · (1  32p2) · min(1, 30.75p )

(14.2)

where: RTO  retransmission time-out ~ 5RTT For a given object, the latency is defined as the time from when the client initiates a TCP connection until the time at which the client receives the requested object in its entirety. Using the following assumptions, we provide expressions for latency with a static and dynamic congestion window. • The network is not congested. • The amount of data that can be transmitted is dependent on the sender’s • • • •



congestion window size. Packets are neither lost nor corrupted. All protocol header overheads are negligible and ignored. The object to be transferred consists of an integer number of segments of size MSS. The only packets with non-negligible transmission times are packets that carry maximum-size TCP segments. Request messages, acknowledgments, and TCP connection establishment segments are small and have negligible transmission times. The initial threshold in the TCP congestion-control mechanism is a large value that is never attained by the congestion window. Static Congestion Window O L  2RTT    {(K  1)[S/R  RTT  (WS)/R]}* R

where: L  latency of the connection {x}*  max(x, 0) K  O/(WS) round-up to the nearest integer W  congestion window size

(14.3)

14.5

Transmission Control Protocol

451

O  Size of the object to be transmitted R  Transmission rate of the link from the server to the client Maximum Segment Size (MSS)  S RTT  round-trip time Dynamic Congestion Window





O S S L  2RTT    P RTT    (2P  1) R

R

R

(14.4)

where: P  min {Q, K  1} in which Q  log2[1  RTT/(S/R)]  1 ISO defined five classes (0 to 4) of connection-oriented transport services (ISO 8073). We briefly describe class 4, which transmits packets with error recovery and in the correct order. This protocol is known as Transport Protocol Class 4 (TP4) and is designed for unreliable networks. The basic steps in the TP4 connection are given below: • Connection establishment: This is performed by means of a three-way

handshake to agree on connection parameters, such as a credit value that specifies how many packets can be sent initially until the next credit arrives, connection number, the transport source and destination access points, and a maximum time-out before ACK. • Data transfer: The data packets are numbered sequentially. This allows resequencing. ACKs may be done for blocks of packets. There is a provision for expedited data transport in which the data packets are sent and acknowledged one at a time. Expedited packets jump to the head of the queues. Flow is controlled by windows or by credits. • Clear connection: Connections are released by an expedited packet indicating the connection termination. The buffers are then flushed out of the data packets corresponding to that connection. In practice, TCP has been tuned for a traditional network consisting of wired links and stationary hosts. TCP assumes that congestion in the network is the primary cause of packet losses and unusual delay. TCP performs well over wired networks by adapting to end-to-end delays and congestion losses. TCP reacts to packet losses by dropping its transmission (congestion) window size before retransmitting packets, initiating congestion control or avoidance mechanisms. These measures result in a reduction in the load on the intermediate links, thereby controlling the congestion in the network. While slow start is one of the most useful mechanisms in wireline networks, it significantly reduces the efficiency of TCP when used together with mobile receivers or senders.

452

14

Mobile Network and Transport Layer

Example 14.1 We consider sending an object of size O  800 kilobytes from a server to a client. If maximum segment size (S)  536 bytes and RTT  100 msec, and the transport protocol uses a static window with window size W, determine the minimum possible latency for a transmission rate of 1 Mbps. What is the minimum window size that achieves this latency? Solution O 800  1000 Lmin  2RTT    0.2    1.0 sec. 6 1  10

R

S WS For minimum latency   RTT   0 R

R

RTT  1  100  103  1  23.3  24.3 segments ⬖W1  536  8 10

 RS 

 6

Example 14.2 Find the upper bound of the throughput for a TP connection if RTT  100 msec, maximum segment size (MSS)  536 bytes, and packet loss probability for the path is 1%. What is the throughput with retransmissions due to errors? Solution 0.93  (536  8)

R    0.3988 Mbps  0.100   0.01

With retransmission due to error: 536  8 R    0.3349 Mbps   0.10.0133  0.5  0.01  [1  32(0.01)2] min(1, 30.75  0.01 )

14.5.1 TCP Enhancements for Wireless Networks TCP was primarily designed for wired networks. Its parameters were selected to maximize its performance on wired networks where packet delays and losses are caused mainly by congestion. In wired networks, random bit error rate is negligible. In a wireless network, packet losses occur due to handoff or fading and can be random. When TCP responds to packet losses by invoking congestion control or an avoidance algorithm, a degraded end-to-end performance in wireless network results. A wireless environment violates many of the assumptions made by TCP. Several approaches have been suggested to improve end-to-end TCP performance over wireless links [12–14]. They can be classified into three categories.

14.5

Transmission Control Protocol

453

1. end-to-end TCP protocols, where loss recovery is performed by the sender, such as explicit loss notification (ELN) option 2. link-layer protocols that provide local reliability using techniques such as forward error correction (FEC) and retransmission of lost packets in response to automatic repeat request (ARQ) messages 3. split TCP connection protocol that breaks the end-to-end TCP connection into two parts at the base station, one between the sender and the base station, and the other between the base station and the receiver. All wireless networks face a high bit error rate. In heterogeneous networks, to explicitly differentiate the cause of packet loss is the primary goal of TCP design. Such efforts aim to find an explicit way to inform the sender of the cause of packet loss, be it congestion or random errors. Thus, the sender is able to make appropriate decisions on how to adjust the congestion window. In the end-to-end case, the link-layer ARQ mechanism is used to improve the error rate seen by TCP. The IS-95 CDMA data stack uses this approach. In the link-layer case, network layer software is modified at the base station to monitor every passing packet in either direction. Cache packets at the base station are used and local retransmissions across wireless links are performed. In the split, the TCP mode wireless portion is separated from the fixed portion. With split TCP, TCP may get ACK even before the packet is successfully delivered to the receiver. It also involves software overhead. Several schemes that have been used with the goal to improve TCP’s performance in wireless and mobile environment are: • Indirect TCP (I-TCP) (see Figure 14.10) • Snooping TCP (see Figure 14.11) • Mobile TCP (M-TCP) • Fast retransmit/fast recovery Sender

TCP #1

Base Station TCP #2

Fixed Host

Mobile host

Router Land-based Network Figure 14.10

Antenna Air-based Network

Split connection (indirect) TCP in wireless environment.

454

14

Mobile Network and Transport Layer

• Transmission/time-out freezing • Selective retransmission

Table 14.1 gives comparisons of these schemes. Sender

TCP #1

Base Station (snooper) TCP #2

Fixed Host

Mobile host

Router

Antenna

Note: Base station snoops and acts like TCP (generates ACKs, delete ACKs, resent segments)

Figure 14.11 Table 14.1

Snooping agent TCP in wireless environment.

Comparison of TCP enhancements for mobility.

Approch

Mechanism

Advantages

Disadvantages

I-TCP

Splits TCP connection into two connections

Isolation of wireless link, simple

Loss of TCP semantics, higher latency at handoff security problem

Snooping TCP

Snoops data and ACKs, local retransmission

Transparent for end-to-end connection, MAC integration possible

Insufficient isolation of wireless link, security problem

M-TCP

Splits TCP connection, chokes sender via window size

Maintains end-toend semantics, handles longterm and frequent disconnections

Bad isolation of wireless link, overhead due to bandwidth management, security problem

Fast retransmit/fast recovery

Avoids slow-start after roaming

Simple and efficient

Mixed layers, not transparent

Transmission/ time-out freezing

Freezes TCP state at disconnection, resumes after reconnecting

Independent of content, works for longer interruptions

Changes in TCP required, MAC independent

Selective retransmission

Retransmits only lost data

Very efficient

Slightly more complex receiver software, more buffer space required

14.5

Transmission Control Protocol

455

The current TCP for 2.5G/3G wireless networks describes a profile to optimize TCP for wireless wide-area networks (WWANs) such as GSM/GPRS, UMTS, or cdma2000. The following characteristics have been considered in deploying applications over 2.5/3G wireless links: • Data rates • Latency • Jitter • Packet loss

Based on these characteristics, the following configuration parameters for TCP in a wireless environment have been suggested: • Large window size: TCP should use large enough window sizes based

• •

• •





on the bandwidth delay experienced in wireless systems. A large initial window size (more than the typical 1 segment) of 2 to 4 segments may increase performance particularly for short transmissions (a few segments in total). Limit transmit: This is an extension of fast retransmission/fast recovery and is useful when small amounts of data are to be transmitted. Large maximum segment size: The larger the MSS the faster TCP increases the congestion window. Link-layers fragment packet data units (PDUs) for transmit according to their needs and a large MSS may be used to increase performance. MSS path discovery should be used for larger segment sizes instead of assuming the small default MSS. Selective ACK (SACK): SACK allows the selective retransmission of packets. It is beneficial compared to the standard cumulative scheme. Explicit congestion notification (ECN): ECN allows a receiver to inform a sender of congestion in the network by setting the ECN-echo flag on receiving an IP packet that has experienced congestion. This scheme makes it easier to distinguish packet loss due to retransmission errors from packet loss due to congestion. Time-stamp: With the help of time-stamps, higher delay spikes can be tolerated by TCP without experiencing a spurious time-out. The effect of bandwidth oscillation is also reduced. No header compression: Header compression mechanism does not perform well in the presence of packet losses. It should not be used.

14.5.2 Implementation of Wireless TCP Traditional TCP schemes suffer from severe performance degradation in a mixed wired and wireless environment. Modifications to standard TCP to remove its deficiency in wireless communications have been proposed.

456

14

Data Packet Fixed Host

Data Packet Base Station

ACK

Figure 14.12

Mobile Network and Transport Layer

Mobile Host ACK

End-to-end connection TCP in wireless environment.

The design of wireless TCP should consider the characteristics of a particular type of wireless network and its need; for example, a satellite network has a long propagation delay, and an ad hoc network is infrastructureless. Wireless TCP algorithms can be designed in either split mode or end-to-end mode. The split mode divides the TCP connection into a wireless and wired portion, and ACKs are generated for both portions separately (see Figure 14.10). By doing so, the performance of the wired portion is not affected by the relatively unreliable wireless portion. The end-to-end mode treats the route from the sender to the receiver as an end-to-end path, and the sender is acknowledged directly by the receiver (see Figure 14.12). This approach maintains the end-to-end semantics of the original TCP design. The split mode TCP attempts to shield the wireless portion from the fixed network by separating the flow control at the intermediate router (or a base station), so that the wireless behavior has the least impact on the fixed network. The intermediate router acts as a terminal in both the fixed and wireless portions. Both end hosts communicate with the intermediate router independently, without knowledge of the other end. The intermediate router is provided with functionality to coordinate the transaction between two network portions. In indirect TCP (I-TCP) the mobile support router (MSR) connects the mobile host to the fixed host and establishes two separate TCP connections with the fixed host and mobile host, respectively. The mobile support router communicates with the fixed host on behalf of the mobile host. The congestion window is maintained separately for wireless and fixed connections. When the mobile host switches cells, a new MSR takes over the communication with the fixed host seamlessly. Thus, the fixed host is protected from the unreliable feature of wireless connections. In split mode, an intermediate router has to reveal the information in the TCP packet and process related data before it reaches its destination. This violates the end-to-end semantics of the original TCP. In the end-to-end approach, only the end hosts participate in flow control. The receiver provides feedback reflecting the network condition, and the sender makes decisions for congestion control. In the end-to-end approach, the ability to accurately probe for the available bandwidth is the key to better performance. The available bandwidth of a flow is the minimum unused link capacity of the flow’s fair share along the path.

14.6

Mobile IP and Session Initiation Protocol

457

The end-to-end approach can have its congestion control mechanism realized in two ways: reactive and proactive. By reactive congestion control, the sender rectifies the congestion window when the network situation becomes marginal or has crossed a threshold. By proactive congestion control, feedback from the network guides the sender to reallocate network resources in order to prevent congestion. In cellular networks, where the base station interconnects a fixed network and a mobile network, modifications of TCP algorithms focus on cellular characteristics such as handoff and high bit error rate. The end-to-end solution is used to improve TCP performance. It imposes no restrictions on routers and only requires code modifications at the mobile unit or receiver side. This approach addresses the throughput degradation caused by frequent disconnections (and reconnections) due to mobile handoff or temporary blockage of radio signals by obstacles. The assumption is that the mobile has knowledge of radio signal strength, and therefore can predict the impending disconnections. In this approach, the receiver on the mobile proactive sets the window size to zero in the ACK packets in the presence of impending disconnections. The zero window size ACK packet forces the sender into persist mode, where it ceases to send more packets while keeping its sending window unchanged. To prevent the sender from exponentially backing off when it detects the reconnection, the receiver sends several positive ACK packets to the sender acknowledging the last received packet before disconnection so that the transfer can resume quickly at the rate before the disconnection occurs. To implement the scheme proposed in reference [11], cross-layer information is required to be exchanged, and the TCP layer protocol must be exposed to some details of roaming and handoff algorithms implemented by network interface card vendors on the interface devices.

14.6

Mobile IP (MIP) and Session Initiation Protocol (SIP)

Third-generation (3G) mobile networks are designed to provide a variety of IP data services such as voice over IP (VoIP) and instant messaging (IM). Both IPv4 and IPv6 are supported in order to provide future-proof solutions. Mobility is supported through both mobile-specific and IP mechanisms. Mobile IP is a key technology for managing mobility in wireless networks [21–25]. At the same time, the SIP is the key to realizing and provisioning services in IP-based mobile networks. The need for mobility of future real-time service independent of terminal mobility requires SIP to seamlessly interwork with MIP operations. 3G networks introduced support of IP mobility through MIP. In particular, the cdma2000 network specified by 3GPP2 deploys MIP to support terminal mobility between points of attachment to the network. MIP will also be supported in 3GPP networks. The SIP (defined by the IETF) is key to service provisioning in 3G networks beyond plain IP connectivity. 3GPP has defined and standardized a network infrastructure called the IP multimedia service (IMS) based on SIP for

458

14

Mobile Network and Transport Layer

supporting a multitude of services to 3G users. Examples are VoIP, instant messaging, and streaming. There are substantial differences between 3GPP and 3GPP2 packet core networks that have an impact on how SIP services can be provided through an IP multimedia service. In particular, 3GPP2 networks use MIP to support terminal mobility. The adoption of SIP and MIP in 3G networks introduces the need for SIP and the IMS to interwork with MIP. In this section, we briefly introduce MIP and SIP and discuss the issues related to interworking between SIP and MIP, with a focus on IPv6 and the applicability to 3G networks.

14.6.1 Mobile IP IP packets do not require mechanisms to set up a dedicated bandwidth or channel. The IP address serves a dual purpose — for routing packets through the Internet and as an end-point identifier for applications in end-hosts. The connections in an IP network use sockets to communicate between clients and servers. A socket consists of source IP address, source port, destination IP address, and destination port. A TCP connection cannot survive any address change because it relies on the socket to determine a connection. However, when a terminal moves from one network to another, its address changes. A mobile node (MN) is a terminal than can change its location and thus its point of attachment. The partner for communication is called the correspondent node that can be either a fixed or an MN. Mobile IP is designed to support host mobility on the Internet. In order for an MN to move across different connection points while maintaining connectivity with other nodes on the Internet, the MN needs to maintain the same address. Two versions of MIP are defined depending on IP version used in the network: MIPv4 for IPv4 networks and MIPv6 for IPv6 networks. Mobile IP implies that a user is connected to one or more applications across the Internet, that the user’s point of attachment changes dynamically, and that all connections are automatically maintained despite the change. When MN moves its attachment point to another network, it is considered a foreign network for this host. Once the mobile is reattached, it makes its presence known by registering with a network node, typically a router, on the foreign network known as a foreign agent (FA). The mobile then communicates with a similar agent on the user’s home network, known as a home agent (HA), giving the home agent the care-of address (CoA) of the mobile node; the care-of address identifies the foreign agent’s location. A home agent tracks a mobile host’s location. The mobile host is affiliated with a static IP address on the home network and a foreign agent supports mobility on a foreign network by providing routing to a visiting mobile host. Network supporting mobile IP will have to create foreign agents to deliver packets of information to the mobile host. Mobile IP is fundamental for the paradigm to provide the successful model for wireless data, which takes the

14.6

Mobile IP and Session Initiation Protocol

459

Mobile node A

3 Home Network For A

4 Foreign Network

Home Agent 1

Foreign Agent

2 Internet or other topology of routers and links 5

Server X

Figure 14.13

Mobile IP scenario.

connection one has into the corporate intranet and makes it wireless. Figure 14.13 shows an MIP scenario that includes the following steps: 1. Server X transmits an IP datagram destined for mobile node A, with A’s home address in the IP header. The IP datagram is routed to A’s home network. 2. At the home network, the incoming IP datagram is intercepted by the home agent. The home agent encapsulates the entire datagram inside a new IP datagram which has the A’s care-of address in the header, and retransmits the datagram. The use of an outer IP datagram with a different destination IP address is known as tunneling. This IP datagram is routed to the foreign agent. 3. The foreign agent strips off the outer IP header, encapsulates the original IP datagram in a network-level packet data unit (PDU), and delivers the original datagram to A across the foreign network. 4. When A sends the IP datagram to X, it uses X’s IP address. This is a fixed address; that is, X is not a mobile node. Each IP datagram is sent by A to a router on the foreign network to X. Typically, this router is also the foreign agent. 5. The IP datagram from A to X travels directly across the Internet to X, using X’s IP address. In MIPv4, MN registers with an FA that becomes the point of contact for the MN. Subsequently, the MN updates its HA, which is a router on the home network that forwards packets meant for the MN’s home address (HoA) to the MN’s current point of attachment (i.e., the CoA of the FA). This allows the MN to remain “always on” — always reachable at its HoA.

460

14

Mobile Network and Transport Layer

MIPv6 also supports direct peer-to-peer communication, called route optimization, between the MN and its core networks without having to traverse the HA. In this way, the MN uses the HoA for communication with a core network (CN) and the CoA for routing purposes. Since MIP operates at the network layer, any change of CoA is transparent to the transport protocols and applications. Hence, all applications in the MN and CN can ignore the mobility of the MN and do not have to deal with a change of network attachment. MIPv4 has been a standard for some years; MIPv6 is currently becoming a standard. Mobile IP Capabilities Three basic capabilities of MIP are registration, discovery, and tunneling (see Figure 14.14). These are discussed in more detail in the following paragraphs. • Registration: A mobile node uses an authenticated registration procedure

to inform its home agent of its care-of address. • Discovery: A mobile node uses a discovery procedure to identify a prospective

home agent and foreign agent. • Tunneling: Tunneling is used to forward IP datagrams from a home address

to a care-of address. The discovery process in MIP is similar to the router advertisement process defined in ICMP. The agent discovery makes use of ICMP router advertisement messages, with one or more extensions specific to MIP. A mobile node is responsible for an ongoing discovery process to determine if it is attached to its network (in which case a datagram may be received without forwarding) or to a foreign network. A transition from the home network to a foreign network can occur at any time without notification to the network layer (i.e., the IP layer). The discovery for a mobile node is a continuous process. Figure 14.15 shows the flow diagram for the agent discovery procedure. Location management in MIP is achieved via a registration process and an agent advertisement. FAs and HAs periodically advertise their presence using Registration

Discovery

User Datagram Protocol (UDP)

Internet Control Message Protocol (ICMP)

Internet Protocol (IP)

Figure 14.14

Mobile IP capabilities.

Tunneling

14.6

Mobile IP and Session Initiation Protocol

461

Listen for advertisements

No

Advertisement Detected?

Solicit advertisement

Yes Current CoA  previous CoA? No

Deregister

Figure 14.15

Yes

Current CoA  home net?

No

Register with Current CoA

Agent discovery procedure.

agent advertisement messages. The same agent may act as both an HA and an FA — mobility extensions to ICMP messages which are used for agent advertisements. The messages contain information about the CoA associated with the FA, whether the agent is busy, whether minimal encapsulation is permitted, whether registration is mandatory, and so on. The agent advertisement packet is a broadcast message on the link. If the mobile node gets an advertisement from its HA, it must deregister its CoA and enable a gratuitous ARP. If a mobile node does not hear any advertisement, it must solicit an agent advertisement using ICMP. Once an agent is discovered, the MN performs either registration or deregistration with the HA, depending on whether the discovered agent is an HA or an FA. The MN sends a registration request using UDP to HA through the FA (or directly, if it is a co-located CoA).The HA creates a mobility binding between the MN’s home address and the current CoA that has a fixed lifetime. The mobile node should register before expiration of the binding. A registration reply indicates whether the registration is successful. A rejection is possible by either the HA or FA for such reasons as insufficient resources, the home agent is unreachable, there are too many simultaneous bindings, failed authentication, and so on. Each FA maintains a list of visiting mobiles containing the following information: • Link-layer address of the mobile node • Mobile node home IP address

462

14

Mobile Network and Transport Layer

• UDP registration request source port • Home agent IP address • An identification field • Registration lifetime • Remaining lifetime of pending or current registration

IP Tunneling. Tunneling is used to forward IP datagrams from a home address to a care-of-address. Types of IP tunneling are: • IP-within-IP encapsulation — simplest approach, defined in RFC 2003 • Minimal encapsulation — involves fewer fields, defined in RFC 2004 • Generic routing encapsulation (GRE) — procedure developed prior to mobile

IP, defined in RFC 1701 IP-within-IP encapsulation (see Figure 14.16): The entire IP datagram becomes the payload in a new IP datagram. The inner, original IP header is

New IP Header

Version ⴝ4

Type of Service

IHL Identification

Total Length Flags

Protocol  4

Time to Live

Fragment Offset Header Checksum

Source address (home agent address) Destination address (care-of address)

Old IP Header

Version ⴝ4

Type of Service

IHL Identification

Time to Live

Protocol

Total Length Flags

Fragment Offset

Header Checksum

Source address (original sender) Destination address (home address)

IP payload (e.g., TCP segment)

Note: Shaded fields are copied from the Inner IP header to the Outer IP header.

Figure 14.16

IP-within-IP encapsulation.

14.6

Mobile IP and Session Initiation Protocol

463

unchanged except to decrement time-to-live (TTL) by 1. The outer header is a full IP header in which: • Two fields, version number and type of service, are copied from an inner header • The source address typically is the IP address of the home agent, and the

destination address is the CoA for the intended destination Minimal encapsulation (see Figure 14.17): This results in less overhead and can be used if the mobile node, home agent, and foreign agent all agree to do so. A new header with the following fields is used between the original IP header and the original IP payload: • Protocol • Header checksum • Original destination address • Original source address

The following fields in the original IP header are modified to form the new outer IP header: • Total length • Protocol

Modified IP Header

Version ⴝ4

Type of Service

IHL

Identification Time to Live

Protocol = 55

Total Length Flags

Fragment Offset Header Checksum

Source address (home agent address)

Minimal Forwarding header

Destination address (care-of address) Protocol

reserved

S

Header Checksum

Destination address (home address) Source address (original sender)

IP payload (e.g., TCP segment)

Note: Shaded fields in the inner IP header are copied from the original IP header. Shaded fields in the outer IP header are modifed from the original IP header.

Figure 14.17

Minimal encapsulation.

464

14

Mobile Network and Transport Layer

• Header checksum • Source address • Destination address

The encapsulation (home agent) prepares the encapsulated datagram which is now suitable for tunneling and delivery across the Internet to the care-of address. The fields in the minimal forwarding header are restored to the original IP header and the forwarding header is removed from the datagram. The total length field in the IP header is decremented by the size of the minimal forwarding header and the checksum field is recomputed.

14.6.2 Session Initiation Protocol (SIP) SIP is used for provisioning services in IP-based mobile networks. SIP specifications define an architecture of user agents and servers (proxy server, redirect server, register) that support communications between SIP peers through user tracking, call routing, and so on. In SIP, each user is uniquely identified by an SIP universal resource indicator, which is used as the identifier to address the called user when the sending session initiation requests. However, an IP address is associated with the user in order to route SIP signaling from the SIP register. A SIP user registers with the SIP register to indicate its presence in the network and its willingness to receive incoming session initiation requests from other users. A typical session in SIP begins with a user sending an INVITE message to a peer through SIP proxies. When the recipient accepts the request and the initiator is notified, the actual data flow begins, usually taking a path other than the one taken by the SIP signaling messages. An INVITE message typically carries a description of the session parameters. In particular, each media component of the SIP session is described in terms of QoS parameters. The user can modify the parameters regarding an existing session by adding or removing media components or modifying the current QoS using a re-INVITE message. SIP also supports personal mobility by allowing a user to reregister with an SIP register on changing its point of attachment to the network, in particular on changing its IP address. A user could also change point of attachment during an active session provided the user reinvites the session providing the new parameters.

14.7

Internet Reference Model

Although many useful protocols have been developed in the context of OSI, the overall 7-layer model has not flourished. The TCP/IP architecture has come to dominate. There are a number of reasons for this outcome. The most important is that the key TCP/IP protocols were mature and well tested at a time when similar OSI protocols were in the development stage. When businesses began to recognize the need for interoperability across networks, only TCP/IP was available and ready to go. Another reason is that the OSI model is unnecessarily complex, with

Problems

465

Application Transport (end-to-end)

Figure 14.18

e-mail, HTTP, FTP, SMTP

TCP, UDP

Network

IP, ICMP

Link

Subnets

Internet reference model.

seven layers to accomplish what TCP/IP does with four layers. The Internet reference model based on TCP/IP is shown in Figure 14.18. The model includes: • The application layer that covers OSI application and presentation layers.

Some of the application layer protocols are HTTP, FTP, SMTP, post office protocol (POP), etc. • The transport (end-to-end) layer includes OSI transport and session layers. The end-to-end protocols are TCP and UDP. • The network (Internet) layer corresponds to the upper part of the OSI network layer. The protocol is IP. • The link (subnets) layer includes the lower part of the OSI network layer, link layer, and physical layer.

14.8

Summary

In this chapter, we first presented the workings of the Internet and discussed the addressing scheme used in the Internet. We then discussed IP, TCP, and mobile IP (MIP); presented the IP suite and compared it with the OSI model. We also discussed TCP in a wireless environment.

Problems 14.1 Which two addresses are used in an IP address? 14.2 What are the maximum number of class A, B, and C network IDs? 14.3 Describe four layers of the TCP/IP protocol suite. 14.4 Describe the characteristics of IP in the TCP/IP protocol suite. 14.5 What is the ICMP and what is its role in the TCP/IP protocol suite? 14.6 Describe the IGMP in the TCP/IP protocol suite.

466

14

Mobile Network and Transport Layer

14.7 What are the ARP and RARP in the TCP/IP protocol suite? 14.8 What are the roles of the TCP and UDP in the TCP/IP protocol suite? 14.9 Discuss split-connection and end-to-end implementation in TCP for a wireless environment. 14.10 Discuss IntServ and DiffServ in the Internet. 14.11 Define MIP in the Internet. 14.12 An object of size (O) 1200 kilobytes is transmitted from a server to a client. The maximum segment size (S) is 600 bytes. The round trip time (RTT) is 120 msec. If the transport protocol uses a static window with window size W, what is the maximum possible latency for a transmission rate, R, of 2 Mbps? How many segments are required to achieve this latency?

References 1. Almquist, P. Types of Service in the Internet Protocol Suite. RFC 1349, July 1992. 2. Al-Quliti, K., “Mobile IP Overview with an Implementation over CDMA2000 Packet Core Network.” Southern Methodist University, Dallas, [email protected] 3. Bakre, A., and Badrinath, B. R. “I-TCP: Indirect TCP for Mobile Hosts.” Proc ICDCS’95, May 1995, pp. 136–143. 4. Braden, R. T., ed. Requirements for Internet Hosts — Communication Layers. RFC 1122, October 1989. 5. Braden, R. T., ed. Requirements for Internet Hosts — Application and Support. RFC 1123, October 1989. 6. Braden, R. T., ed. Extending TCP for Transactions — Concepts. RFC 1379, November 1992. 7. Clark, D. D. Window and Acknowledgement Strategy in TCP. RFC 813, July 1982. 8. Deering, S. E. Host Extensions for IP Multicasting. RFC 1112, August 1989. 9. Deering, S. E., and Hinden, R. Internet Protocol, Version 6 (IPv6) Specification. RFC 1883, December 1995. 10. Deering, S. E. ed. ICMP Router Discovery Messages. RFC 1256, September 1991. 11. Goff, T., et al. “Freeze-TCP: A true end-to-end enhancement mechanism for mobile environment.” Proc. IEEE INFOCOM2000, 2000, pp. 1537–1545. 12. Hiller, T., et al. CDMA 2000 Wireless Data Requirements. RFC 3141, June 2001. 13. Jacobson, V. Congestion Avoidance and Control. Computer Communication Review, vol. 18, no. 4, August 1988, pp. 314–329. 14. Jacobson, V., Braden, R. T., and Borman, D. A. TCP Extensions for High Performance. RFC 1323, May 1992.

References

467

15. Kleinrock, L. The Latency/Bandwidth Trade-off in Gigabit Networks. IEEE Communication Magazine, vol. 30, no. 4, April 1992, pp. 36–40. 16. LaQuey, T. The Internet Companion: A Beginner’s Guide to Global Networking. Reading, MA: Addison-Wesley, 1993. 17. Mockapetris, P. V. Domain names: Implementation and Specification. RFC 1035, November 1987. 18. Mogul, J. C. IP Network Performance. In Internet System Handbook, eds., D. C. Lynch and M. T. Rose. Reading, MA: Addison-Wesley, 1993, pp. 575–675. 19. Nagle, J. Congestion Control in IP/TCP Internetworks. RFC 896, January 1984. 20. Partridge, C. Gigabit Networking. Reading, MA: Addison-Wesley, 1994. 21. Perkins, C. E. Mobile IP. IEEE Communication Magazine, May 1997, pp. 84–99. 22. Perkins, C. E., ed. Mobile IP Joins Forces with AAA. IEEE Personal Communications, August 2000. 23. Perkins, C. E. “Mobile Networking Through Mobile IP.” http://www.computer.org/ internet/v2n1/perkins.htm. 24. Perkins, C. E. IP Encapsulation within IP. RFC 2003, May 1996. 25. Perkins, C. E. Minimal Encapsulation within IP. RFC 2004, May 1996. 26. Postel, J. B. User Datagram Protocol. RFC 768, August 1980. 27. Postel, J. B. Internet Control Message Protocol. RFC 792, September 1981. 28. Postel, J. B., ed. Transmission Control Protocol. RFC 793, September 1981. 29. Simpson, W. A. Point-to-Point Protocol (PPP). RFC 1548, December 1993. 30. Stevens, W. R. TCP/IP Illustrated Volume 1 — Protocols, Reading, MA: Addison-Wesley, 1994. 31. Tanenbaum, A. S. Computer Network. Second Edition, Englewood Cliffs, NJ: PrenticeHall, 1989. 32. Tian, Y., Xu, K., and Ansari, N. TCP in Wireless Environments: Problem and Solution. IEEE Radio Communications, March 2005, pp. 527–531.

This page intentionally left blank

CHAPTER 15 Wide-Area Wireless Networks (WANs) — GSM Evolution 15.1

Introduction

Third-generation (3G) wireless systems [2,3,9] offer access to services anywhere from a single terminal; the old boundaries between telephony, information, and entertainment services are disappearing. Mobility is built into many of the services currently considered as fixed, especially in such areas as high speed access to the Internet, entertainment, information, and electronic commerce (e-commerce) services. The distinction between the range of services offered via wireline or wireless is becoming less and less clear and, as the evolution toward 3G mobile services speeds up, these distinctions will disappear in the first decade of the new millennium. Applications for a 3G wireless network range from simple voice-only communications to simultaneous video, data, voice, and other multimedia applications. One of the main benefits of 3G is that it allows a broad range of wireless services to be provided efficiently to many different users. Packet-based Internet Protocol (IP) technology is at the core of the 3G services. Users have continuous access to on-line information. E-mail messages arrive at hand-held terminals nearly instantaneously and business users are able to stay permanently connected to the company intranet. Wireless users are able to make video conference calls to the office and surf the Internet simultaneously, or play computer games interactively with friends in other locations. Figure 15.1 shows the data rate requirement for various services. In 1997, the TIA/EIA IS-136 community through the Universal Wireless Consortium (UWC) and the Telecommunications Industry Association (TIA) TR 45.3 adopted a three-part strategy for evolving its IS-136 TDMA-based networks to 3G wireless networks to satisfy International Mobile Telephony-2000 (IMT-2000) requirements. The strategy consists of: • Enhancing the voice and data capabilities of the existing 30 kHz carrier

(IS-136) • Adding a 200 kHz carrier for high-speed data (384 kbps) in high mobility applications • Introducing a 1.6 MHz carrier for very high-speed data (2 Mbps) in low-mobility applications 469

470

15

Wide-Area Wireless Networks (WANs) — GSM Evolution

Bits per second

1M 100 k 10 k Voice

Images, Audio, Text

Video (High Quality); (Medium Quality); (Slow Scan)

1k Service Figure 15.1

User data requirements.

The highlight of UWC strategy was the global convergence of IS-136 time division multiple access (TDMA) with a global system for mobile communications (GSM) through the evolution of the 200 kHz GSM carrier for supporting high-speed data applications (384 kbps) while also improving a 30 kHz carrier for voice and mid-speed data applications. In this chapter we focus first on GSM evolution to packet data services and present the details of the general packet radio service (GPRS) and the enhanced data for GSM evolution (EDGE) service. We then provide details of 3G systems including wideband code division multiple access (CDMA) (WCDMA) (i.e., universal mobile telecommunications services (UMTS)). We conclude the chapter by outlining the details of high-speed downlink packet access (HSDPA).

15.2

GSM Evolution for Data

From a radio access perspective, adding 3G capabilities to 2G systems mainly means supporting higher data rates. Possible scenarios depend on spectrum availability for the network service provider. Depending on the spectrum situation, two different migration paths can be supported: • Reframing of existing spectrum bands • New or modified spectrum bands

Two 3G radio access schemes are identified to support the different spectrum scenarios: 1. Enhanced data rates for GSM evolution (EDGE) with high-level modulation in a 200 kHz TDMA channel is based on plug-in transceiver equipment, thereby allowing the migration of existing bands in small spectrum segments. 2. Universal mobile telecommunications services (UMTS) is a new radio access network based on 5 MHz WCDMA and optimized for efficient support of 3G services. UMTS can be used in both new and existing spectra.

15.2

GSM Evolution for Data

471

From a network point of view, 3G capabilities implies the addition of packet switched (PS) services, Internet access, and IP connectivity. With this approach, the existing mobile networks reuse the elements of mobility support, user authentication/service handling, and circuit switched (CS) services. With packet switched services, IP connectivity can then be added to provide a mobile multimedia core network by evolving the existing mobile network. GSM is moving to develop enhanced cutting-edge, customer-focused solutions to meet the challenges of the new millennium and 3G mobile services [29]. When GSM was first introduced, no one could have predicted the dramatic growth of the Internet and the rising demand for multimedia services. These developments have brought about new challenges to the world of GSM. For GSM operators, the emphasis is now rapidly changing from that of instigating and driving the development of technology to fundamentally enabling mobile data transmission to that of improving speed, quality, simplicity, coverage, and reliability in terms of tools and services that will boost mass market take-up. Users are increasingly looking to gain access to information and services wherever they are and whenever they want. GSM should provide that connectivity. Internet access, web browsing and the whole range of mobile multimedia capability are the major drivers for development of higher data speed technologies. Current data traffic on most GSM networks is modest, less than 5% of total GSM traffic. But with the new initiatives coming to fruition during the course of the next two to three years, exponential growth in data traffic is forecast. The use of messaging-based applications may reach up to about 90% by the year 2008. GSM data transmission using high-speed circuit switched data (HSCSD) and GPRS may reach a penetration of about 80% by 2008 [1]. GSM operators have two nonexclusive options for evolving their networks to 3G wideband multimedia operation: (1) using GPRS and EDGE in the existing radio spectrum, and in small amounts of the new spectrum; or (2) using WCDMA in the new 2 GHz bands, or in large amounts of the existing spectrum. Both approaches offer a high degree of investment flexibility because roll-out can proceed in line with market demand with the extensive reuse of existing network equipment and radio sites. In the new 2 GHz bands, 3G capabilities are delivered using a new wideband radio interface that offers much higher user data rates than are available today — 384 kbps in the wide area and up to 2 Mbps in the local area. Of equal importance for such services is the high-speed packet switching provided by GPRS and its connection to public and private IP networks. GSM and digital (D)AMPS (IS-136) operators can use existing radio bands to deliver some of the 3G services, even without the new wideband spectrum by evolving current networks and deploying GPRS and EDGE technologies. In the early years of 3G service deployment, a large proportion of wireless traffic will still be voice-only and low-rate data. So whatever the ultimate capabilities

472

15

Wide-Area Wireless Networks (WANs) — GSM Evolution

of 3G networks, efficient and profitable ways of delivering more basic wireless services are still needed. The significance of EDGE for today’s GSM operators is that it increases data rates up to 384 kbps and potentially even higher in a good quality radio environment using current GSM spectrum and carrier structures more efficiently. EDGE is both a complement and an alternative to new WCDMA coverage. EDGE also has the effect of unifying the GSM, D-AMPS and WCDMA services through the use of dual-mode terminals.

15.2.1 High Speed Circuit Switched Data High-speed circuit switched data (HSCSD) [1,4,5] is a feature that enables the co-allocation of multiple full rate traffic channels (TCH/F) of GSM into an HSCSD configuration. The aim of HSCSD is to provide a mixture of services with different air interface user rates by a single physical layer structure. The available capacity of an HSCSD configuration is several times the capacity of a TCH/F, leading to a significant enhancement in air interface data transfer capability. Ushering faster data rates into the mainstream is the new speed of 14.4 kbps per time slot and HSCSD protocols that approach wireline access rates of up to 57.6 kbps by using multiple 14.4 kbps time slots. The increase from the current baseline of 9.6 kbps to 14.4 kbps is due to a nominal reduction in the error-correction overhead of the GSM radio link protocol (RLP), allowing the use of a higher data rate. For operators, migration to HSCSD brings data into the mainstream, enabled in many cases by relatively standard software upgrades to base station (BS) and mobile switching center (MSC) equipment. Flexible air interface resource allocation allows the network to dynamically assign resources related to the air interface usage according to the network operator’s strategy, and the end-user’s request for a change in the air interface resource allocation based on data transfer needs. The provision of the asymmetric air interface connection allows simple mobile equipment to receive data at higher rates than otherwise would be possible with a symmetric connection. For end-users, HSCSD enables the roll-out of mainstream high-end segment services that enable faster web browsing, file downloads, mobile video-conference and navigation, vertical applications, telematics, and bandwidth-secure mobile local area network (LAN) access. Value-added service providers will also be able to offer guaranteed quality of service and cost-efficient mass-market applications, such as direct IP where users make circuit-switched data calls straight into a GSM network router connected to the Internet. To the end-user, the value-added service provider or the operator is equivalent to an Internet service provider that offers a fast, secure dial-up Internet protocol service at cheaper mobile-to-mobile rates. HSCSD is provided within the existing mobility management. Roaming is also possible. The throughput for an HSCSD connection remains constant for the

15.2

GSM Evolution for Data

473

duration of the call, except for interruption of transmission during handoff. The handoff is simultaneous for all time slots making up an HSCSD connection. Endusers wanting to use HSCSD have to subscribe to general bearer services. Supplementary services applicable to general bearer services can be used simultaneously with HSCSD. Firmware on most current GSM PC cards needs to be upgraded. The reduced RLP layer also means that a stronger signal strength is necessary. Multiple time slot usage is probably only efficiently available in off-peak times, increasing overall off-peak idle capacity usage. HSCSD is not a very feasible solution for bursty data applications.

15.2.2 General Packet Radio Service The general packet radio service (GPRS) [6,7] enhances GSM data services significantly by providing end-to-end packet switched data connections. This is particularly efficient in Internet/intranet traffic, where short bursts of intense data communications are actively interspersed with relatively long periods of inactivity. Because there is no real end-to-end connection to be established, setting up a GPRS call is almost instantaneous and users can be continuously on-line. Users have the additional benefits of paying for the actual data transmitted, rather than for connection time. Because GPRS does not require any dedicated end-to-end connection, it only uses network resources and bandwidth when data is actually being transmitted. This means that a given amount of radio bandwidth can be shared efficiently among many users simultaneously. The next phase in the high-speed road map is the evolution of current short message service (SMS), such as smart messaging and unstructured supplementary service data (USSD), toward the new GPRS, a packet data service using TCP/IP and X.25 to offer speeds up to 115 kbps. GPRS has been standardized to optimally support a wide range of applications ranging from very frequent transmissions of medium to large data volume. Services of GPRS have been developed to reduce connection set-up time and allow an optimum usage of radio resources. GPRS provides a packet data service for GSM where time slots on the air interface can be assigned to GPRS over which packet data from several mobile stations is multiplexed. A similar evolution strategy, also adopting GPRS, has been developed for DAMPS (IS-136). For operators planning to offer wideband multimedia services, the move to GPRS packet-based data bearer service is significant; it is a relatively small step compared to building a totally new 3G IMT-2000 network. Use of the GPRS network architecture for IS-136 packet data service enables data subscription roaming with GSM networks around the globe that support GPRS and its evolution. The IS-136 packet data service standard is known as GPRS-136. GPRS-136 provides the same capabilities as GSM GPRS. The user can access either X.25 or an IP-based data network.

474

15

Wide-Area Wireless Networks (WANs) — GSM Evolution

GPRS provides a core network platform for current GSM operators not only to expand the wireless data market in preparation for the introduction of 3G services, but also a platform on which to build IMT-2000 frequencies should they acquire them. The implementation of GPRS has a limited impact on the GSM core network. It simply requires the addition of new packet data switching and gateway nodes, and an upgrade to existing nodes to provide a routing path for packet data between the wireless terminal and a gateway node. The gateway node provides interworking with external packet data networks for access to the Internet, intranet, and databases. A GPRS architecture for GSM is shown in Figure 15.2 and network element interfaces in Figure 15.3. GPRS supports all widely used data communications

Base Station Subsystem

Mobile Station

ME

SIM

BTS

BSC

Network Subsystem

MSC/ VLR

EIR

GMSC

HLR

AUC

Other Networks

PSTN

PLMN

UE BSS SD



GGSN

SGSN

CN

SGSN: Serving GPRS Support Node GGSN: Gateway GPRS Support Node HLR: Home Location Register VLR: Visitor Location Register MSC: Mobile Switching Center BSS: Base Station System GMSC: Gateway MSC EIR: Equipment Identity Register ME: Mobile Equipment SIM: Subscriber Identity Card PLMN: Public Land Mobile Network

Figure 15.2

A GPRS architecture in GSM.

Internet

15.2

GSM Evolution for Data

475

SMS-GMSC SMS-IWMSC

MAP-H

SM-SC MAP-C

Gd

MSC/VLR

HLR MAP-D

A

Gs

Gc

Gi

Gr TE

BSS

MT

SGSN

Um

R

Gb Gn

Signaling + Information Signaling

Gp GGSN Other PLMN

PDN TE

GGSN Gn

Gf EIR MAP-F

SGSN SGSN: Serving GPRS Support Node GGSN: Gateway GPRS Support Node MAP: Mobile Application Part HLR: Home Location Register VLR: Visitor Location Register MSC: Mobile Switching Center BSS: Base Station System GMSC: Gateway MSC IWMSC: Interworking MSC TE: Terminal Equipment MT: Mobile Terminal EIR: Equipment Identity Register SMS: Short Message Service

Figure 15.3

GPRS interfaces for different network elements.

protocols, including IP, so it is possible to connect with any data source from anywhere in the world using a GPRS mobile terminal. GPRS supports applications ranging from low-speed short messages to high-speed corporate LAN communications. However, one of the key benefits of GPRS — that it is connected through the existing GSM air interface modulation scheme — is also a limitation, restricting its potential for delivering higher data rates than 115 kbps. To build even higher rate data capabilities into GSM, a new modulation scheme is needed. GPRS can be implemented in the existing GSM systems. Changes are required in an existing GSM network to introduce GPRS. The base station subsystem (BSS) consists of a base station controller (BSC) and packet control unit (PCU). The PCU supports all GPRS protocols for communication over the air interface. Its function is to set up, supervise, and disconnect packet switched calls. The packet control unit supports cell change, radio resource configuration, and channel assignment. The base station transceiver (BTS) is a relay station without protocol functions. It performs modulation and demodulation.

476

15

Wide-Area Wireless Networks (WANs) — GSM Evolution

The GPRS standard introduces two new nodes, the serving GPRS support node (SGSN) and the gateway GPRS support node (GGSN). The home location register (HLR) is enhanced with GPRS subscriber data and routing information. Two types of services are provided by GPRS: • Point-to-point (PTP) • Point-to-multipoint (PTM)

Independent packet routing and transfer within the public land mobile network (PLMN) is supported by a new logical network node called the GPRS support node (GSN). The GGSN acts as a logical interface to external packet data networks. Within the GPRS networks, protocol data units (PDUs) are encapsulated at the originating GSN and decapsulated at the destination GSN. In between the GSNs, IP is used as the backbone to transfer PDUs. This whole process is referred to as tunnelling in GPRS. The GGSN also maintains routing information used to tunnel the PDUs to the SGSN that is currently serving the mobile station (MS). All GPRS user related data required by the SGSN to perform the routing and data transfer functionality is stored within the HLR. In GPRS, a user may have multiple data sessions in operation at one time. These sessions are called packet data protocol (PDP) contexts. The number of PDP contexts that are open for a user is only limited by the user’s subscription and any operational constraints of the network. The main goal of the GPRS-136 architecture is to integrate IS-136 and GSM GPRS as much as possible with minimum changes to both technologies. In order to provide subscription roaming between GPRS-136 and GSM GPRS networks, a separate functional GSM GPRS HLR is incorporated into the architecture in addition to the IS-41 HLR. The European Telecommunication Standards Institute (ETSI) has specified GPRS as an overlay to the existing GSM network to provide packet data services. In order to operate a GPRS over a GSM network, new functionality has been introduced into existing GSM network elements (NEs) and new NEs are integrated into the existing service provider’s GSM network. The BSS of GSM is upgraded to support GPRS over the air interface. The BSS works with the GPRS backbone system (GBS) to provide GPRS service in a similar manner to its interaction with the switching subsystem for the circuit-switched services. The GBS manages the GPRS sessions set up between the mobile terminal and the network by providing functions such as admission control, mobility management (MM), and service management (SM). Subscriber and equipment information is shared between GPRS and the switched functions of GSM by the use of a common HLR and coordination of data between the visitor location register (VLR) and the GPRS support nodes of the GBS. The GBS is composed of two new NEs — the SGSN and the GGSN. The SGSN serves the mobile and performs security and access control functions. The SGSN is connected to the BSS via frame-relay. The SGSN provides

15.2

GSM Evolution for Data

477

packet routing, mobility management, authentication, and ciphering to and from all GPRS subscribers located in the SGSN service area. A GPRS subscriber may be served by any SGSN in the network, depending on location. The traffic is routed from the SGSN to the BSC and to the mobile terminal via a BTS. At GPRS attach, the SGSN establishes a mobility management context containing information about mobility and security for the mobile. At PDP context activation, the SGSN establishes a PDP context which is used for routing purposes with the GGSN that the GPRS subscriber uses. The SGSN may send in some cases location information to the MSC/VLR and receive paging requests. The GGSN provides the gateway to the external IP network, handling security and accounting functions as well as the dynamic allocation of IP addresses. The GGSN contains routing information for the attached GPRS users. The routing information is used to tunnel PDUs to the mobile’s current point of attachment, SGSN. The GGSN may be connected with the HLR via optional interface Gc. The GGSN is the first point of public data network (PDN) interconnection with a GSM PLMN supporting GPRS. From the external IP network’s point of view, the GGSN is a host that owns all IP addresses of all subscribers served by the GPRS network. The PTM-SC handles PTM traffic between the GPRS backbone and the HLR. The nodes are connected by an IP backbone network. The SGSN and GGSN functions may be combined in the same physical node or separated — even residing in different mobile networks. A special interface (Gs) is provided between MSC/VLR and SGSN to coordinate signaling for mobile terminals that can handle both circuit-switched and packet-switched data. The HLR contains GPRS subscription data and routing information, and can be accessible from the SGSN. For the roaming mobiles, the HLR may reside in a different PLMN than the current SGSN. The HLR also maps each subscriber to one or more GGSNs. The objective of the GPRS design is to maximize the use of existing GSM infrastructure while minimizing the changes required within GSM. The GSN contains most of the necessary capabilities to support packet transmission over GSM. The critical part in the GPRS network is the mobile to GSN (MS-SGSN) link which includes the MS-BTS, BTS-BSC, BSC-SGSN, and the SGSN-GGSN link. In particular, the Um interface including the radio channel is the bottleneck of the GPRS network due to the spectrum and channel speed/quality limitations. Since multiple traffic types of varying priorities are supported by the GPRS network, the quality of service criteria as well as resource management is required for performance evaluation. The BSC will require new capabilities for controlling the packet channels, new hardware in the form of a packet control unit, and new software for GPRS mobility management and paging. The BSC also has a new traffic and signaling interface from the SGSN.

478

15

Wide-Area Wireless Networks (WANs) — GSM Evolution

The BTS has new protocols supporting packet data for the air interface, together with new slot and channel resource allocation functions. The utilization of resources is optimized through dynamic sharing between the two traffic types handled by the BSC. MS-SGSN Link The logical link control (LLC) layer (see Figure 15.4) is responsible for providing a link between the MS and the SGSN. It governs the transport of GPRS signaling and traffic information from the MS to the SGSN. GPRS supports three service access points (SAPs) entities: the layer 3 management, subnet dependent convergence, and short message service (SMS). On the MS-BSS link, the radio link control (RLC), the medium access control (MAC), and GSM RF protocols are supported. The main drawback in implementing GPRS on an existing GSM infrastructure is that the GSM network is optimized for voice transmission (i.e., the GSM channel quality is designed for voice which can tolerate errors at a predefined level). It is therefore expected that GPRS could have varied transmission performance in a different network or coverage area. To overcome this problem, GPRS supports multiple coding rates at the physical layer. A GPRS could share radio resources with GSM circuit switched (CS) service. This is governed by a dynamic resource sharing based on the capacity of demand criteria. A GPRS channel is allocated only if an active GPRS terminal exists in the SGSN

BSS Application

IP/ X.25

IP/ X.25

Relay

SNDCP

SNDCP

PSPDN UDP/TCP Specific Protocols IP

LLC

LLC Relay

LLC

GPRS Tunnel

GPRS Tunnel

UDP/TCP

RLC

RLC

BSSGP

BSSGP

IP

MAC

MAC

L2

Network Service L1 bis

L2

L2

L1

L1

GSM RF

GSM RF L1 bis

Um

Gb

SNDCP: Sub Network Dependent Convergence Protocol BSSGP: Base Station System GPRS Protocol LLC: Logical Link Control RLC: Radio Link Control MAC: Medium Access Control

Figure 15.4

Protocol stack in GPRS.

Gn

15.2

GSM Evolution for Data

479

network. Once resources are allocated to GPRS, at least one channel will serve as the master channel to carry all necessary signaling and control information for the operation of the GPRS. All other channels will serve as slave and are only used to carry user and signaling information. If no master channel exists, all the GPRS users will use the GSM common control channel (CCCH) and inform the network to allocate GPRS resources. A physical channel dedicated to GPRS is called a packet data channel (PDCH). It is mapped into one of the physical channels allocated to GPRS (see Figure 15.5). A PDCH can either be used as a packet common control channel (PCCCH), a packet broadcast control channel (PBCCH), or a packet traffic channel (PTCH). Uplink PRACH PCCCH

PPCH Downlink PAGCH PNCH

Signaling Channels

PBCCH

PBCCH

Uplink Downlink

PACCH

Uplink

Packet Data Logical Channels

PDCCH

PTCCH/U Downlink PTCCH/D Uplink

Traffic Channels

PDTCH/U

PDTCH PDTCH/D Downlink

PCCCH: Packet Common Control Channel PBCCH: Packet Broadcast Control Channel PDCCH: Packet Dedicated Control Channel PDTCH: Packet Data Traffic Channel PRACH: Packet Random Access Channel PPCH: Packet Paging Channel PAGCH: Packet Access Grant Channel PNCH: Packet Notification Channel PACCH: Packet Associated Control Channel PTCCH: Packet Timing Advance Control Channel

Figure 15.5

GPRS logical channels.

480

15

Wide-Area Wireless Networks (WANs) — GSM Evolution

The PCCCH consists of: • Packet random access channel (PRACH) — uplink • Packet access grant channel (PAGCH) — downlink • Packet notification channel (PNCH) — downlink

On the other hand, the PTCH can either be: • Packet data traffic channel (PDTCH) • Packet associated control channel (PACCH)

The arrangement of GPRS logical channels for given traffic characteristics also requires the combination of PCCCHs and PTCHs. Fundamental questions such as how many PDTCHs can be supported by a single PCCCH is needed in dimensioning GPRS. RLC/MAC Layer The multiframe structure of the PDCH in which GPRS RLC messages are transmitted is composed of 52 TDMA frames organized into RLC blocks of four bursts resulting in 12 blocks per multiframe plus four idle frames located in the 13th, 26th, 39th, and 52nd positions (see Figure 15.6). B0 consists of frames 1, 2, 3 and 4, B1 consists of frames 5, 6, 7, and 8 and so on. It is important that the mapping of logical channels onto the radio blocks is done by means of an ordered set of blocks (B0, B6, B9, B1, B7, B4, B10, B2, B8, B5, B11). The advantage of ordering the blocks is mainly to spread the locations of the control channels in each time slot reducing the average waiting time for the users to transmit signaling packets. It also provides an interleaving of the GPRS multiframe. GPRS uses a reservation protocol at the MAC layer. Users that have packets ready to send request a channel via the PRACHs. The random access burst consists of only one TDMA frame with duration enough to transmit an 11-bit signaling message. Only the PDCHs carrying PCCCHs contain PRACHs. The blocks used as PRACHs are indicated by an uplink state flag (USF  free) by the downlink pair channel. 52 TDMA Frames B0

B1

B2

T

B3

B4

B5

I

B6

I  Idle Frame T  Frame used for PTCCH B0……B11  Radio Blocks

Figure 15.6

Idle frame location in GPRS multiframe.

B7

B8

T

B9

B10

B11

I

15.2

GSM Evolution for Data

481

Alternatively, the first K blocks following the ordered set of blocks can be assigned to PRACH permanently. The access burst is transmitted in one of the four bursts assigned as PRACH. Any packet channel request is returned by a packet immediate assignment on the PRACHs whose locations are broadcast by PBCCH. Optionally, a packet resource request for additional channels is initiated and returned by a packet resource assignment. The persistence of random access is maintained by the traffic load and user class with a back-off algorithm for unsuccessful attempts. In the channel assignment, one or more PTCHs (time slot) will be allocated to a particular user. A user reserves a specific number of blocks on the assigned PTCH as indicated by the USF. It is possible to accommodate more than one user per PTCH. User signaling is also transmitted on the same PTCH using the PAGCH whose usage depends on the necessity of the user. The performance of the MAC layer depends on the logical arrangement of the GPRS channels (i.e., allocation of random access channels, access grant channels, broadcast channels, etc.) for given traffic statistics. This is determined by the amount of resources allocated for control and signaling compared to data traffic. A degree of flexibility of logical channels is also achieved as the traffic varies. The arrangement of logical channels is determined through the PBCCH. LLC Layer The LLC layer is responsible for providing a reliable link between the mobile and the SGSN. It is based on the LAPD (link access protocol D) protocol. It is designed to support variable length transmission in a PTP or PTM topology. It includes the layer function such as sequence control, flow control, error detection, ciphering, and recovery as well as the provision of one or more logical link connections between two layer 3 entities. A logical link is identified by a DLCI (data link control identity) which consists of a service access point identity (SAPI) and terminal equipment identity (TEI) mapped on the LLC frame format. Depending on the status of the logical link, it supports an unacknowledged or an acknowledged information transfer. The former does not support error recovery mechanisms. The acknowledged information transfer supports error and flow control. This operation only applies to point-to-point operations. The LLC frame consists of an address field (1 or 5 octets), control field (2 or 6 octets), a length indicator field (2 octets maximum), information fields (1500 octets maximum), and a frame check sequence of 3 octets. Four types of control field formats are allowed including the supervisory functions (S format), the control functions (U), and acknowledged and unacknowledged information transfer (I and UI). In the performance evaluation, the objective is to determine delay during the exchange of commands and responses involved in various operations supported by the LLC in relation to the transfer of an LLC PDU. The LLC commands and responses are exchanged between two layer 3 entities in conjunction with a service primitive invoked by the mobile or the SGSN.

482

15

Wide-Area Wireless Networks (WANs) — GSM Evolution

Data Packet Routing in the GPRS Network The following discusses data packet routing for the mobile originated and mobile terminated data call scenarios. In mobile originated data routing, the mobile gets an IP packet from an application and requests a channel reservation. The mobile transmits data in the reserved time slots. The packet switched public data network (PSPDN) PDU is encapsulated into a sub-network dependent convergence protocol (SNDCP) unit that is sent via LLC protocol over the air interface to the SGSN currently serving the mobile (see Figure 15.7). For mobile terminated data routing (see Figure 15.7), we have two cases: routing to the home GPRS network, and routing to a visited GPRS network. In the first case, a user sends a data packet to a mobile. The packet goes through the local area network (LAN) via a router out on the GPRS context for the mobile. If the mobile is in a GPRS idle state, the packet is rejected. If the mobile is in standby or active mode, the GGSN routes the packet in an encapsulated format to SGSN. In the second case, the home GPRS network sends the data packet over the inter-operator backbone network to the visiting GPRS network. The visiting GPRS network routes the packet to the appropriate SGSN. The PTP and PTM applications of GPRS are listed below: • Point-to-point • Messaging (e.g., e-mail) • Remote access to corporate networks (b) (a)

Visited GPRS Network

Home GPRS Network BTS

BSC

MSC

MSC

BSC

BTS

GPRS Mobile

GPRS Mobile SGSN

Intra-operator Backbone Network

GGSN

Host

GPRS Register Intra-operator Backbone Network

Data Network

LAN

SGSN

Intra-operator Backbone Network

GGSN

Router

(a) Mobile Originated Data Call Routing (b) Mobile Terminated Data Call Routing to Visited GPRS Network

Figure 15.7

GPRS Register

Data call routing in GPRS network.

15.2

GSM Evolution for Data

483

• Access to the Internet • Credit card validation (point-of-sales) • Utility meter readings • Road toll applications • Automatic train control • Point-to-multipoint • PTM-multicast (send to all) • News • Traffic information • Weather forecasts • Financial updates • PTM-group call (send to some) • Taxi fleet management • Conferencing

GPRS provides a service for bursty and bulky data transfer, radio resources on demand, shared use of physical radio resources, existing GSM functionality, mobile applications for a mass application market, volume dependent charging, and integrated services, operation and management.

15.2.3 Enhanced Data Rates for GSM Enhancement The enhanced data rates for GSM enhancements (EDGE) [8] provides an evolutionary path from existing 2G systems (GSM, IS-136, PDC) to deliver some 3G services in existing spectrum bands. The advantages of EDGE include fast availability, reuse of existing GSM, IS-136 and PDC infrastructure, as well as support for the gradual introduction of 3G capabilities. EDGE is primarily a radio interface improvement, but it can also be viewed as a system concept that allows GSM, IS-136, and PDC networks to offer a set of new services. EDGE has been designed to improve S/I by using link quality control. Link quality control adapts the protection of the data to the channel quality so that for all channel qualities an optimal bit rate is achieved. EDGE can be seen as a generic air interface for efficiently providing high bit rates, facilitating an evolution of existing 2G systems toward 3G systems. The EDGE air interface is designed to facilitate higher bit rates than those currently achievable in existing 2G systems. The modulation scheme based on 8-PSK is used to increase the gross bit rate. GMSK modulation as defined in GSM is also part of the EDGE system. The symbol rate is 271 ksps for both GMSK and 8-PSK, leading to gross bit rates per time slot of 22.8 kbps and 69.2 kbps, respectively. The 8-PSK pulse shape is linearized by GMSK to allow 8-PSK to fit into the GSM

484

15

3

58 bits

Wide-Area Wireless Networks (WANs) — GSM Evolution

26 bits

58 bits

3

8.25

0.577 ms Figure 15.8

Burst format for EDGE with 8-PSK.

spectrum mask. The 8-PSK burst format is similar to GSM (see Figure 15.8). EDGE reuses the GSM carrier bandwidth and time slot structure. EDGE (also known as the 2.5G system) has been designed to enhance user bandwidth through GPRS. This is achieved through the use of higher-level modulation schemes. Although EDGE reuses the GSM carrier bandwidth and time slot structure, the technique is by no means restricted to GSM systems; it can be used as a generic air interface for efficient provision of higher bit rates in other TDMA systems. In the Universal Wireless Communications Consortium (UWCC) the 136 high speed (136 HS) radio transmission technology (RTT) radio interface was proposed as a means to satisfy the requirements for an IMT-2000 RTT. EDGE was adopted by UWCC in 1998 as the outdoor component of 136 HS to provide 384 kbps data service. The standardization effort for EDGE has two phases. In the first phase of EDGE the emphasis has been placed on enhanced GPRS (EGPRS) and enhanced CSD (ECSD). The second phase is defined with improvements for multimedia and real-time services. In order to achieve a higher gross rate, new modulation scheme, quaternary offset quadrature amplitude modulation (QOQAM) has been proposed for EDGE, since it can provide higher data rates and good spectral efficiency. An offset modulation scheme is proposed because it gives smaller amplitude variation than 16-QAM, which can be beneficial when using nonlinear amplifiers. EDGE co-exists with GSM in an existing frequency plan and provides link adaptation (modulation and coding are adapted for channel conditions). Radio Protocol Design The radio protocol strategy in EDGE is to reuse the protocols of GSM/GPRS whenever possible, thus minimizing the need for new protocol implementation. EDGE enhances both the GSM circuit-switched (HSCSD) and packet-switched (GPRS) mode operation. EDGE includes one packet-switched and one circuitswitched mode, EGPRS and ECSD, respectively. Enhanced GPRS (EGPRS). The EDGE radio link control (RLC) protocol is somewhat different from the corresponding GPRS protocol. The main changes are related to improvements in link quality control scheme. A link adaptation scheme regularly estimates the link quality and subsequently selects the most appropriate modulation and coding scheme for the transmission to maximize the user bit rate. The link adaptation scheme offers

15.2

GSM Evolution for Data

485

mechanisms for choosing the best modulation and coding scheme for the radio link. In GPRS only the coding schemes can be changed between two consecutive link layer control (LLC) frames. In the EGPRS even the modulation can be changed. Different coding and modulation schemes enable adjustment for the robustness of the transmission according to the environment. Another way to handle link quality variations is incremental redundancy. In this scheme, information is first sent with very little coding, yielding a high bit rate if decoding is immediately successful. If decoding is not successful, additional coded bits (redundancy) are sent until decoding succeeds. The more coding that has to be sent, the lower the resulting bit rate and the higher the delay. EGPRS supports combined link adaptation and incremental redundancy schemes. In this case, the initial code rate of the incremental redundancy scheme is based on measurements of the link quality. Benefits of this approach are the robustness and high throughput of the incremental redundancy operation in combination with lower delays and lower memory requirements enabled by the adaptive initial code rate. In EGPRS the different initial code rates are obtained by puncturing a different number of bits from a common convolutional code R  1/3. The resulting coding schemes are given in Table 15.1. Incremental redundancy operation is enabled by puncturing a different set of bits each time a block is retransmitted, whereby the code rate is gradually decreased toward 1/3 for every new transmission of the block. The selection of the initial modulation and code rate is based on regular measurements of link quality.

Table 15.1 Channel coding scheme in EDGE (PS transmission).

Coding scheme

Gross bit rate (kbps)

Code rate

Modulation

Radio interface rate per time-slot (kbps)

CS-1

22.8

0.49

GMSK

11.2

CS-2

22.8

0.63

GMSK

14.5

CS-3

22.8

0.73

GMSK

16.7

CS-4

22.8

1.0

GMSK

22.8

PCS-1

69.2

0.329

8-PSK

22.8

Radio interface rate on 8 time-slots (kbps)

182.4

PCS-2

69.2

0.496

8-PSK

34.3

274.4

PCS-3

69.2

0.596

8-PSK

41.25

330.0

PCS-4

69.2

0.746

8-PSK

51.60

412.8

PCS-5

69.2

0.829

8-PSK

57.35

458.8

PCS-6

69.2

1.000

8-PSK

69.20

553.6

486

15

Wide-Area Wireless Networks (WANs) — GSM Evolution

Actual performance of modulation and coding scheme together with channel characteristics form the basis for link adaptation. Channel characteristics are needed to estimate the effects of a switch to another modulation and coding combination