Get the full ebook with Bonus Features for a Better Reading Experience on ebookluna.

com

(eBook PDF) Fundamentals of Communication Systems,

Global Edition 2nd Edition

https://ebookluna.com/product/ebook-pdf-fundamentals-of-
communication-systems-global-edition-2nd-edition/

OR CLICK HERE

DOWLOAD NOW

Download more ebook instantly today at https://ebookluna.com

 Instant digital products (PDF, ePub, MOBI) ready for you
Download now and discover formats that fit your needs...

(eBook PDF) Fundamentals of Communication Systems, Global

Edition 2nd Edition

https://ebookluna.com/product/ebook-pdf-fundamentals-of-communication-
systems-global-edition-2nd-edition-2/

ebookluna.com

(eBook PDF) Fundamentals of C# Programming for Information

Systems 2nd Edition

https://ebookluna.com/product/ebook-pdf-fundamentals-of-c-programming-
for-information-systems-2nd-edition/

ebookluna.com

(eBook PDF) Fundamentals of C# Programming for Information

Systems 2nd Edition

https://ebookluna.com/product/ebook-pdf-fundamentals-of-c-programming-
for-information-systems-2nd-edition-2/

ebookluna.com

(eBook PDF) Fundamentals of Information Systems 9th

Edition

https://ebookluna.com/product/ebook-pdf-fundamentals-of-information-
systems-9th-edition/

ebookluna.com
(eBook PDF) Fundamentals of Information Systems 8th
Edition

https://ebookluna.com/product/ebook-pdf-fundamentals-of-information-
systems-8th-edition/

ebookluna.com

(eBook PDF) Comparative Health Systems: A Global

Perspective 2nd Edition

https://ebookluna.com/product/ebook-pdf-comparative-health-systems-a-
global-perspective-2nd-edition/

ebookluna.com

(eBook PDF) Fundamentals of Organizational Communication,

Updated Edition 9th Edition

https://ebookluna.com/product/ebook-pdf-fundamentals-of-
organizational-communication-updated-edition-9th-edition/

ebookluna.com

(eBook PDF) Fundamentals of Investing 13th Global Edition

https://ebookluna.com/product/ebook-pdf-fundamentals-of-
investing-13th-global-edition/

ebookluna.com

(eBook PDF) Fundamentals of Hydraulic Engineering Systems

(5th Edition)

https://ebookluna.com/product/ebook-pdf-fundamentals-of-hydraulic-
engineering-systems-5th-edition/

ebookluna.com
 Global Global
edition edition

edition
Global
Fundamentals of

Communication Systems
For these Global Editions, the editorial team at Pearson has
Communication Systems

Fundamentals of
collaborated with educators across the world to address a
wide range of subjects and requirements, equipping students
SECOND edition
with the best possible learning tools. This Global Edition
preserves the cutting-edge approach and pedagogy of the
original, but also features alterations, customization and
adaptation from the North American version.

second edition
This is a special edition of an established

Salehi
Proakis
title widely used by colleges and universities
throughout the world. Pearson published this
exclusive edition for the benefit of students
outside the United States and Canada. If you
purchased this book within the United States
or Canada you should be aware that it has
been imported without the approval of the
Publisher or Author. John G. Proakis • Masoud Salehi
Pearson Global Edition

PROAKIS_1292015683_mech.indd 1 21/07/14 9:14 PM

A01 PROA5682 02 GE FM July 17, 2014 page 7

Contents
PREFACE 17

1 INTRODUCTION 21

1.1 Historical Review 21

1.2 Elements of an Electrical Communication System 24
1.2.1 Digital Communication System, 27
1.2.2 Early Work in Digital Communications, 30
1.3 Communication Channels and Their Characteristics 32
1.4 Mathematical Models for Communication Channels 38
1.5 Summary and Further Reading 40

2 SIGNALS AND LINEAR SYSTEMS 41

2.1 Basic Concepts 41

2.1.1 Basic Operations on Signals, 41
2.1.2 Classification of Signals, 43
2.1.3 Some Important Signals and Their Properties, 51
2.1.4 Classification of Systems, 58
2.1.5 Analysis of LTI Systems in the Time Domain, 61
2.2 Fourier Series 63
2.2.1 Fourier Series and Its Properties, 64
2.2.2 Response of LTI Systems to Periodic Signals, 74
2.2.3 Parseval’s Relation, 76
2.3 Fourier Transform 78
2.3.1 From Fourier Series to Fourier Transforms, 78
2.3.2 Basic Properties of the Fourier Transform, 84
2.3.3 Fourier Transform for Periodic Signals, 98
2.3.4 Transmission over LTI Systems, 101
2.4 Filter Design 105
2.5 Power and Energy 109
2.5.1 Energy-Type Signals, 109
2.5.2 Power-Type Signals, 112

2.6 Hilbert Transform and Its Properties 115

2.7 Lowpass and Bandpass Signals 118

7
A01 PROA5682 02 GE FM July 17, 2014 page 8

8 Contents

2.8 Summary and Further Reading 120

Problems 121

3 AMPLITUDE MODULATION 137

3.1 Introduction to Modulation 138

3.2 Amplitude Modulation 139
3.2.1 Double-Sideband Suppressed-Carrier AM, 139
3.2.2 Conventional Amplitude Modulation, 146
3.2.3 Single-Sideband AM, 152
3.2.4 Vestigial-Sideband AM, 154
3.3 Implementation of Amplitude Modulators and Demodulators 157
3.4 Signal Multiplexing 164
3.4.1 Frequency-Division Multiplexing, 164
3.4.2 Quadrature-Carrier Multiplexing, 165

3.5 AM Radio Broadcasting 166

3.6 Summary and Further Reading 169
Appendix 3A: Derivation of the Expression for SSB-AM Signals 169
Problems 171

4 ANGLE MODULATION 181

4.1 Representation of FM and PM Signals 181

4.2 Spectral Characteristics of Angle-Modulated Signals 186
4.2.1 Angle Modulation by a Sinusoidal Signal, 186
4.2.2 Angle Modulation by an Arbitrary Message Signal, 190

4.3 Implementation of Angle Modulators and Demodulators 191

4.4 FM Radio Broadcasting 199
4.5 Summary and Further Reading 201
Problems 202

5 PROBABILITY AND RANDOM PROCESSES 210

5.1 Review of Probability and Random Variables 210

5.1.1 Sample Space, Events, and Probability, 210
5.1.2 Conditional Probability, 211
5.1.3 Random Variables, 214
5.1.4 Functions of a Random Variable, 221
5.1.5 Multiple Random Variables, 223
5.1.6 Sums of Random Variables, 228
A01 PROA5682 02 GE FM July 17, 2014 page 9

Contents 9

5.2 Random Processes: Basic Concepts 229

5.2.1 Statistical Averages, 232
5.2.2 Wide-Sense Stationary Processes, 235
5.2.3 Multiple Random Processes, 237
5.2.4 Random Processes and Linear Systems, 238
5.2.5 Power Spectral Density of Stationary Processes, 240
5.2.6 Power Spectral Density of a Sum Process, 245
5.3 Gaussian and White Processes 246
5.3.1 Gaussian Processes, 246
5.3.2 White Processes, 248
5.3.3 Filtered Noise Processes, 250
5.4 Summary and Further Reading 255
Problems 256

6 EFFECT OF NOISE ON ANALOG COMMUNICATION SYSTEMS 275

6.1 Effect of Noise on Amplitude Modulation Systems 275

6.1.1 Effect of Noise on a Baseband System, 276
6.1.2 Effect of Noise on DSB-SC AM, 276
6.1.3 Effect of Noise on SSB AM, 278
6.1.4 Effect of Noise on Conventional AM, 279
6.2 Effect of Noise on Angle Modulation 283
6.2.1 Threshold Effect in Angle Modulation, 291
6.2.2 Preemphasis and Deemphasis Filtering for FM, 294
6.3 Comparison of Analog-Modulation Systems 297
6.4 Effects of Transmission Losses and Noise in Analog Communication
Systems 298
6.4.1 Characterization of Thermal Noise Sources, 299
6.4.2 Effective Noise Temperature and Noise Figure, 300
6.4.3 Transmission Losses, 303
6.4.4 Repeaters for Signal Transmission, 304
6.5 Summary and Further Reading 307
Problems 308

7 ANALOG-TO-DIGITAL CONVERSION 316

7.1 Sampling of Signals and Signal Reconstruction from Samples 317

7.1.1 The Sampling Theorem, 317

7.2 Quantization 321

7.2.1 Scalar Quantization, 322
7.2.2 Vector Quantization, 329
A01 PROA5682 02 GE FM July 17, 2014 page 10

10 Contents

7.3 Encoding 331

7.4 Waveform Coding 332
7.4.1 Pulse Code Modulation, 333
7.4.2 Differential Pulse Code Modulation, 336
7.4.3 Delta Modulation, 338
7.5 Analysis–Synthesis Techniques 341
7.6 Digital Audio Transmission and Digital Audio Recording 345
7.6.1 Digital Audio in Telephone Transmission Systems, 345
7.6.2 Digital Audio Recording, 347

7.7 The JPEG Image-Coding Standard 352

7.8 Summary and Further Reading 355
Problems 356

8 DIGITAL MODULATION METHODS IN AN ADDITIVE WHITE

GAUSSIAN NOISE CHANNEL 367

8.1 Geometric Representation of Signal Waveforms 368

8.2 Binary Modulation Schemes 372
8.2.1 Binary Antipodal Signaling, 372
8.2.2 Binary Orthogonal Signaling, 376
8.3 Optimum Receiver for Binary Modulated Signals in Additive White
Gaussian Noise 381
8.3.1 Correlation-Type Demodulator, 382
8.3.2 Matched-Filter-Type Demodulator, 391
8.3.3 The Performance of the Optimum Detector for Binary Signals, 399
8.4 M-ary Digital Modulation 404
8.4.1 The Optimum Receiver for M-ary Signals in AWGN, 404
8.4.2 A Union Bound on the Probability of Error, 416

8.5 M-ary Pulse Amplitude Modulation 418

8.5.1 Carrier-Modulated PAM for Bandpass Channels (M-ary ASK), 420
8.5.2 Demodulation and Detection of Amplitude-Modulated
PAM Signals, 423
8.5.3 Probability of Error for M-ary PAM, 423
8.6 Phase-Shift Keying 426
8.6.1 Geometric Representation of PSK Signals, 428
8.6.2 Demodulation and Detection of PSK Signals, 430
8.6.3 Probability of Error for Phase-Coherent PSK Modulation, 431
8.6.4 Differential Phase Encoding and Differential Phase Modulation
and Demodulation, 436
8.6.5 Probability of Error for DPSK, 438
A01 PROA5682 02 GE FM July 17, 2014 page 11

Contents 11

8.7 Quadrature Amplitude-Modulated Digital Signals 439

8.7.1 Geometric Representation of QAM Signals, 441
8.7.2 Demodulation and Detection of QAM Signals, 443
8.7.3 Probability of Error for QAM, 444
8.8 Carrier-Phase Estimation 449
8.8.1 The Phase-Locked Loop, 449
8.8.2 The Costas Loop, 457
8.8.3 Carrier-Phase Estimation for PAM, 459
8.8.4 Carrier-Phase Estimation for PSK, 460
8.8.5 Carrier-Phase Estimation for QAM, 464
8.9 Symbol Synchronization 466
8.9.1 Early–Late Gate Synchronizers, 467
8.9.2 Minimum Mean Square Error Method, 470
8.9.3 Maximum-Likelihood Method, 471
8.9.4 Spectral-Line Method, 472
8.9.5 Symbol Synchronization for Carrier-Modulated Signals, 475
8.10 Regenerative Repeaters 476
8.11 Summary and Further Reading 477
Problems 479

9 MULTIDIMENSIONAL DIGITAL MODULATION 505

9.1 M-ary Orthogonal Signals 505

9.1.1 Probability of Error for M-ary Orthogonal Signals, 508
9.1.2 A Union Bound on the Error Probability of M-ary Orthogonal
Signals, 511
9.2 Biorthogonal Signals 512
9.2.1 Probability of Error for M-ary Biorthogonal Signals, 515

9.3 Simplex Signals 517

9.3.1 Probability of Error for M-ary Simplex Signals, 518

9.4 Binary-Coded Signals 519

9.4.1 Probability of Error for Binary-Coded Signals, 521
9.5 Frequency-Shift Keying 521
9.5.1 Demodulation of M-ary FSK, 523
9.5.2 Optimum Detector for Noncoherent Binary FSK, 527
9.5.3 Probability of Error for Noncoherent Detection of M-ary FSK, 530
9.6 Modulation Systems with Memory 533
9.6.1 Continuous-Phase FSK, 533
9.6.2 Spectral Characteristics of CPFSK Signals, 544
A01 PROA5682 02 GE FM July 17, 2014 page 12

12 Contents

9.7 Comparison of Modulation Methods 545

9.8 Summary and Further Reading 552
Problems 553

10 DIGITAL TRANSMISSION THROUGH BANDLIMITED AWGN CHANNELS 563

10.1 Characterization of Bandlimited Channels and Signal Distortion 563

10.1.1 Intersymbol Interference in Signal Transmission, 567
10.1.2 Digital Transmission through Bandlimited Bandpass Channels, 569

10.2 The Power Spectrum of Digitally Modulated Signals 572

10.3 Signal Design for Bandlimited Channels 576
10.3.1 Design of Bandlimited Signals for Zero ISI—The Nyquist
Criterion, 578
10.3.2 Design of Bandlimited Signals with Controlled ISI—Partial-Response
Signals, 584
10.4 Detection of Partial-Response Signals 586
10.4.1 Symbol-by-Symbol Detection, 587
10.4.2 Probability of Error for Symbol-by-Symbol Detection, 590
10.4.3 Maximum-Likelihood Sequence Detection of Partial-Response
Signals, 593
10.4.4 Error Probability of the Maximum-Likelihood Sequence
Detector, 596
10.5 System Design in the Presence of Channel Distortion 597
10.5.1 Design of Transmitting and Receiving Filters for a Known
Channel, 598
10.5.2 Channel Equalization, 602
10.6 Summary and Further Reading 619
Appendix 10A: Power Spectrum of Modulated Signals 621
10A.1 The Power Spectrum of the Baseband Signal, 621
10A.2 The Power Spectrum of the Carrier Modulated Signals, 623

Problems 624

11 MULTICARRIER MODULATION AND OFDM 641

11.1 Orthogonal Frequency-Division Multiplexing 641

11.2 Modulation and Demodulation in an OFDM System 642
11.3 An OFDM System Implemented via the FFT Algorithm 646
11.4 Spectral Characteristics of OFDM Signals 649
A01 PROA5682 02 GE FM July 17, 2014 page 13

Contents 13

11.5 Peak-to-Average Power Ratio in OFDM Systems 651

11.6 Applications of OFDM 653
11.6.1 Digital Subscriber Lines, 653
11.6.2 Wireless LANs, 655
11.6.3 Digital Audio Broadcasting, 656
11.7 Summary and Further Reading 656
Problems 657

12 AN INTRODUCTION TO INFORMATION THEORY 661

12.1 Modeling Information Sources 662

12.1.1 Measure of Information, 664
12.1.2 Joint and Conditional Entropy, 667
12.1.3 Mutual Information, 670
12.1.4 Differential Entropy, 670
12.2 The Source Coding Theorem 672
12.3 Source Coding Algorithms 675
12.3.1 The Huffman Source Coding Algorithm, 675
12.3.2 The Lempel–Ziv Source Coding Algorithm, 679

12.4 Modeling of Communication Channels 681

12.5 Channel Capacity 684
12.5.1 Gaussian Channel Capacity, 689

12.6 Bounds on Communication 691

12.7 Summary and Further Reading 694
Problems 695

13 CODING FOR RELIABLE COMMUNICATIONS 709

13.1 The Promise of Coding 709

13.2 Linear Block Codes 714
13.2.1 Decoding and Performance of Linear Block Codes, 720
13.2.2 Some Important Linear Block Codes, 727
13.2.3 Error Detection versus Error Correction, 728
13.2.4 Burst-Error-Correcting Codes, 729
13.3 Convolutional Codes 731
13.3.1 Basic Properties of Convolutional Codes, 732
13.3.2 Maximum Likelihood Decoding of Convolutional Codes—The Viterbi
Algorithm, 737
13.3.3 Other Decoding Algorithms for Convolutional Codes, 742
13.3.4 Bounds on the Error Probability of Convolutional Codes, 742
A01 PROA5682 02 GE FM July 17, 2014 page 14

14 Contents

13.4 Good Codes Based on Combination of Simple Codes 745

13.4.1 Product Codes, 747
13.4.2 Concatenated Codes, 748
13.5 Turbo Codes and Iterative Decoding 748
13.5.1 MAP Decoding of Convolutional Codes—The BCJR Algorithm, 751
13.5.2 Iterative Decoding for Turbo Codes, 757
13.5.3 Performance of Turbo Codes, 759
13.6 Low-Density Parity-Check Codes 761
13.6.1 Decoding LDPC Codes, 765
13.7 Coding for Bandwidth-Constrained Channels 767
13.7.1 Combined Coding and Modulation, 768
13.7.2 Trellis-Coded Modulation, 769

13.8 Practical Applications of Coding 776

13.8.1 Coding for Deep-Space Communications, 776
13.8.2 Coding for Telephone-Line Modems, 778

13.9 Summary and Further Reading 779

Problems 780

14 DATA TRANSMISSION IN FADING MULTIPATH CHANNELS 789

14.1 Characterization of Physical Wireless Channels 789

14.2 Channel Models for Time-Variant Multipath Channels 791
14.2.1 Frequency Nonselective Fading Channel, 794
14.2.2 Frequency Selective Fading Channel, 797
14.2.3 Models for the Doppler Power Spectrum, 798
14.2.4 Propagation Models for Mobile Radio Channels, 801
14.3 Performance of Binary Modulation in Rayleigh Fading Channels 803
14.3.1 Probability of Error in Frequency Nonselective Channels, 803
14.3.2 Performance Improvement through Signal Diversity, 806
14.3.3 The RAKE Demodulator and Its Performance in Frequency Selective
Channels, 812
14.3.4 OFDM Signals in Frequency Selective Channels, 814
14.4 Multiple Antenna Systems 815
14.4.1 Channel Models for Multiple Antenna Systems, 816
14.4.2 Signal Transmission in a Slow Fading Frequency Nonselective MIMO
Channel, 817
14.4.3 Detection of Data Symbols in a MIMO System, 819
14.4.4 Error Rate Performance of the Detectors, 820
14.4.5 Space–Time Codes for MIMO Systems, 822
14.5 Link Budget Analysis for Radio Channels 830
14.6 Summary and Further Reading 833
Problems 835
A01 PROA5682 02 GE FM July 17, 2014 page 15

Contents 15

15 SPREAD-SPECTRUM COMMUNICATION SYSTEMS 845

15.1 Model of a Spread-Spectrum Digital Communication System 846

15.2 Direct Sequence Spread-Spectrum Systems 847
15.2.1 Effect of Despreading on a Narrowband Interference, 850
15.2.2 Probability of Error at the Detector, 851
15.2.3 Performance of Coded Spread-Spectrum Signals, 856
15.3 Some Applications of DS Spread-Spectrum Signals 856
15.3.1 Low-Detectability Signal Transmission, 856
15.3.2 Code Division Multiple Access, 857
15.3.3 Communication over Channels with Multipath, 858
15.3.4 Wireless LANs, 859
15.4 Generation of PN Sequences 860
15.5 Frequency-Hopped Spread Spectrum 863
15.5.1 Slow Frequency-Hopping Systems and Partial-Band Interference, 864
15.5.2 Fast Frequency Hopping, 867
15.5.3 Applications of FH Spread Spectrum, 868
15.6 Synchronization of Spread-Spectrum Systems 869
15.6.1 Acquisition Phase, 869
15.6.2 Tracking, 872
15.7 Digital Cellular Communication Systems 876
15.7.1 The GSM System, 878
15.7.2 CDMA System Based on IS-95, 882
15.7.3 Third Generation Cellular Communication Systems and Beyond, 886
15.8 Summary and Further Reading 888
Problems 889

REFERENCES 897

INDEX 906
A01 PROA5682 02 GE FM July 17, 2014 page 16
A01 PROA5682 02 GE FM July 17, 2014 page 17

Preface
This book is intended as a senior-level undergraduate textbook on communication systems
for Electrical Engineering majors. Its primary objective is to introduce the basic techniques
used in modern communication systems and to provide fundamental tools and methodolo-
gies used in the analysis and design of these systems. Although the book is mainly written
as an undergraduate-level textbook, it can be equally useful to the practicing engineer, or
as a self-study tool.
The emphasis of the book is on digital communication systems, which are treated
in detail in Chapters 7 through 15. These systems are the backbone of modern commu-
nication systems, including new generations of wireless communication systems, satellite
communications, and data transmission networks. Traditional analog communication sys-
tems are also covered in due detail in Chapters 3, 4, and 6. In addition, the book provides
detailed coverage of the background required for the course in two chapters, one on linear
system analysis with emphasis on the frequency-domain approach and Fourier techniques,
and one on probability, random variables, and random processes. Although these topics are
now covered in separate courses in the majority of Electrical Engineering programs, it is
the experience of the authors that the students frequently need to review these topics in a
course on communications, and therefore it is essential to have quick access to the relevant
material from these courses. It is also assumed that the students taking this course have a
background in calculus, linear algebra, and basic electric circuits.

NEW TO THIS EDITION

The following are the major new features in the Second Edition of Fundamentals of Com-
munication Systems:

– Major reorganization of basic digital modulation methods based on geometric repre-

sentation of signals
– Expanded coverage of carrier phase estimation and symbol synchronization
– New chapter on multicarrier modulation and OFDM
– New and expanded coverage of iterative decoding of turbo codes and LDPC codes
– New section on multiple antenna (MIMO) systems for radio channels
– New chapter on spread spectrum signals and systems

ORGANIZATION OF THE BOOK

The book starts with a brief review of communication systems in Chapter 1, followed by
methods of signal representation and system analysis in both time and frequency domains

17
A01 PROA5682 02 GE FM July 17, 2014 page 18

18 Preface

in Chapter 2. Emphasis is placed on the Fourier series and the Fourier transform represen-
tation of signals and the use of transforms in linear systems analysis.
Chapters 3 and 4 cover the modulation and demodulation of analog signals. In
Chapter 3, amplitude modulation (AM) is covered. In Chapter 4, frequency modulation
(FM) and phase modulation (PM) are covered. AM and FM radio broadcasting are also
treated in these chapters.
In Chapter 5, we present a review of the basic definitions and concepts in proba-
bility and random processes. Special emphasis is placed on Gaussian random processes,
which provide mathematically treatable models for additive noise disturbances. Both time-
domain and frequency-domain representations of random signals are presented.
Chapter 6 covers the effects of additive noise in the demodulation of amplitude-
modulated (AM) and angle-modulated (FM, PM) analog signals and a comparison of these
analog signal modulations in terms of their signal-to-noise ratio performance. We also
present the characterization of thermal noise and the effect of transmission losses in analog
communication systems.
Chapter 7 is devoted to analog-to-digital conversion. The sampling theorem and
quantization techniques are treated first, followed by waveform encoding methods includ-
ing PCM, DPCM, and DM. This chapter concludes with brief discussion of LPC speech
decoding and the JPEG standard for image compression.
Chapter 8 treats basic digital modulation methods and their performance in AWGN
channels. The methods described are binary antipodal and orthogonal signals, and M-ary
pulse amplitude modulation (PAM), phase-shift keying (PSK), and quadrature amplitude
modulation (QAM). These types of digital signals are characterized in terms of their geo-
metric representation. The optimum demodulation of these signals is derived based on
the maximum a posteriori and maximum-likelihood criteria. In addition, we also describe
methods for carrier phase estimation using a phase-locked loop (PLL), and symbol syn-
chronization.
In Chapter 9, we treat multidimensional digital modulation signals based on a geo-
metric representation of such signals and derive their performance when transmitted in
an AWGN channel. Signal types considered include orthogonal signals, biorthogonal sig-
nals, simplex signals, binary-coded signals, and frequency-shift keying (FSK). Continuous-
phase FSK (CPFSK) and its spectral characteristics are also treated.
In Chapter 10, we consider the transmission of digital signals in bandlimited AWGN
channels. The effect of channel distortion on the transmitted signals is shown to result
in intersymbol interference (ISI). Then, the design of signals that eliminate or control the
effect of ISI is described. Finally, we treat the design of adaptive equalizers for suppressing
ISI in the channel distorted received signal.
The focus of Chapter 11 is on digital signal transmission via multicarrier modula-
tion and orthogonal frequency-division multiplexing (OFDM). The implementation of the
OFDM modulator and demodulator, based on the use of the FFT algorithm, is described.
Additional topics treated include the spectral characteristics of OFDM signals and methods
for reducing the peak-to-average power ratio (PAR) in OFDM signals. Finally, we present
several applications of OFDM in current digital communication systems.
A01 PROA5682 02 GE FM July 17, 2014 page 19

Preface 19

In Chapter 12, we present the basic limits on communication of information, includ-

ing the information content of memoryless sources, efficient coding of the source out-
put, and the capacity of the AWGN channel. Two widely used algorithms for encoding
the output of digital sources, namely, the Huffman coding algorithm and the Lempel–Ziv
algorithm, are also described.
In Chapter 13, we treat channel coding and decoding. Linear block codes and con-
volutional codes are described for enhancing the performance of a digital communication
system in the presence of AWGN. Both hard-decision and soft-decision decoding of block
and convolutional codes are treated. Coding for bandwidth-limited channels (trellis-coded
modulation), turbo codes, and low-density parity check codes are also treated.
In Chapter 14, we treat the characterization of physical wireless channels and the
construction of mathematical models for time-varying, fading multipath channels. The per-
formance of binary modulation in Rayleigh fading channels is determined and the benefits
of signal diversity for combating signal fading is demonstrated. The RAKE demodulator is
described and its performance on frequency selective channels is evaluated. Also treated in
this chapter is the use of multiple transmit and receive antennas for increasing the transmis-
sion rate and obtaining signal diversity in wireless communication systems. Methods for
mapping digital signals for transmission on multiple antennas are also presented, including
block coding methods such as the Alamouti code and trellis codes. The final topic treated
in this chapter is link budget analysis for radio channels.
The final chapter of this book introduces the reader to spread-spectrum digital com-
munication techniques and their use in combating interference, both intentional (jamming)
and unintentional, the latter arising from other users of the channel. In particular, we treat
direct sequence (DS) spread spectrum and frequency-hopping (FH) spread spectrum, and
their performance characteristics in the presence of interference. Also treated is the gener-
ation of pseudo-noise (PN) sequences for use in spreading the spectrum of the transmitted
signal. The final topic of this chapter describes the use of spread spectrum signals in digital
cellular communication systems, including 2nd-, 3rd-, and 4th-generation (2G, 3G, 4G)
cellular systems.
Throughout the book many worked examples are provided to emphasize the use of
the techniques developed in theory. Following each chapter are a large number of prob-
lems at different levels of difficulty. The problems are followed by a selection of com-
puter problems, which usually ask for simulation of various algorithms developed in that
chapter using MATLAB. The solutions to the MATLAB problems are made available at
www.pearsonglobaleditions.com.

Acknowledgments

We wish to thank the reviewers of the first and second editions of this book (Selin Aviyente,
Michigan State University; Costas Georghiades, Texas A&M University; Dennis Goeckel,
University of Massachusetts, Amherst; Bijan Mobasseri, Villanova University; Robert
Morelos-Zaragoza, San Jose State University; Ahmad Safaai-Jazi, Virginia Tech; Lei Wei,
A01 PROA5682 02 GE FM July 17, 2014 page 20

20 Preface

University of Central Florida; and Stephen Wilson, University of Virginia) for their com-
ments and recommendations. Their suggestions have resulted in significant improvements
to the various topics covered in this book.

John G. Proakis
Adjunct Professor,
University of California at San Diego
and Professor Emeritus,
Northeastern University,
Masoud Salehi
Northeastern University.

Pearson would like to thank the following people for their work on the Global Edition:

Contributors:
Amlan Chakrabarti
University of Calcutta.

Sanjay H.S.
M.S. Ramaiah Institute of Technology.

Reviewer:
Vijay Kumar Chakka
Shiv Nadar University.
Random documents with unrelated
content Scribd suggests to you:
into holes, or stumbled and were carried away. The majority were
saved, but ten of the 51st and rather more of the Brunswickers were
swept right down-stream and drowned. Yet, despite mishaps, the
hussars got over, and, advancing rapidly, surprised the French
cavalry picquet at the village of Val de Perdrices, a little way up-
stream, taking it whole—an officer and 32 men. This was certainly
about the most extraordinary instance of carelessness on the part of
outposts during the war, and reflects as much discredit on Digeon,
whose dragoons were supposed to be watching the lower Esla, as
on the wretched officer in charge of the picquet. How was it possible
that such a large body as a brigade could approach in daylight the
best-known ford of the neighbourhood, at a spot where the course of
the high road showed the convenience of the passage, without
finding a single vedette on the bank? And why was such an
important point watched (or not watched) by a half-troop, instead of
by a force which could have offered at least a momentary opposition,
and have passed the alarm to its regiment, which, as Digeon’s report
shows, was at Iniesta, only four miles to the rear?
Meanwhile Bock’s Germans and D’Urban’s Portuguese dragoons
had an equally difficult, though equally unopposed, passage at the
ford of Palomilla, four miles up-stream. The river was running
furiously, and seven or eight horses and three or four men were
washed away and drowned. Anson’s light dragoons at the head of
the third column had an even worse experience at San Vincente del
Barco, opposite San Cebrian: the ford was found utterly
impracticable, and the brigade was ordered back toward Tabara: on
its way it was turned off to a second projected crossing-spot farther
south. This was also discovered to be hopeless, and finally the
whole northern column was ordered to cross at Almendra, behind
the southern column. After a day’s profitless countermarching[481], it
came down thither, to find the pontoons laid, the infantry of the
southern column all across the river, and well forward, while that of
the centre column was crossing rapidly. In the end all Graham’s
troops save the leading cavalry brigades had to use the bridge.
Meanwhile Grant’s hussars, advancing on Zamora, found that the
French had evacuated it in haste on the first news of the crossing of
the Esla. Digeon, with his two regiments of dragoons, his half-battery
of horse artillery, and the four companies of voltigeurs, had gone off
to Toro. Wellington therefore was able to occupy Zamora without
opposition on the night following the passage of the fords[482], and
moved his head-quarters thither next day. The moment that he knew
that Daricau and Digeon had absconded, he sent orders to Hill to
march not on Zamora, or the ford of Villa al Campo, as he had at first
intended, but directly on Toro, which would save twenty miles
marching for the right wing. For it was clear that Digeon and Daricau
could not hold Toro with some 6,000 men against Graham’s 40,000,
now across the Esla and in close pursuit of them. And there was no
great body of French available to reinforce them within a couple of
days’ march—at most Reille’s horse and one infantry division could
have come up. On June 1st the signs were that the enemy would not
stand—Digeon’s dragoons were falling back on the Toro road—
Reille’s cavalry, heard of at Belver in the morning, had gone off
eastward towards Medina del Rio Seco in haste. Wherefore
Wellington pushed on fast on that day, sending on Graham’s
columns by the two roads Zamora-Toro and Zamora-Rio Seco, which
do not diverge sufficiently to make it impossible to concentrate the
troops on them in a few hours. But caution proved unnecessary. The
French evacuated Toro in the afternoon, so that the junction-point
with Hill was safely secured, and Graham’s divisions were present in
full force next morning, to cover the passage of the southern wing
across the Douro.
Wellington moved his head-quarters to Toro on the morning of
June 2nd, and sent out his cavalry on all the roads which branch out
from it: Anson’s and Bock’s brigades going north-east occupied
Vezdemarben. Grant’s hussars pushing along the river-road toward
Tordesillas came up with Digeon’s rear at Morales, six miles outside
Toro, and fell upon it vigorously. The French dragoons—two
regiments—were drawn up in a defensive position, with a swamp in
front and bridge over a ravine behind. Grant charging furiously with
the 10th Hussars and the 18th in support on the flank, broke the front
regiment, whereupon the enemy went to the rear in disorder, and
was chased for two miles, prisoners being captured in considerable
numbers. At last pursuers and pursued ran in on the rear brigade of
Daricau’s infantry, drawn up in good order, with a battery across the
road, on the heights of Pedroso del Rey. Grant had therefore to call
off his men, and to wait for the enemy to retire, which they presently
did in good order. Two officers and 208 men, all of the 16th
Dragoons, were captured[483], about a hundred of them wounded[484];
the rest were mainly taken owing to the bad condition of their
mounts, ‘raw-boned horses with evident marks of bad provender,
escort duties, and counter-marches—nearly the whole of them had
horrible sore backs.’ The 10th Hussars had only 16 casualties—one
of them an officer who had pursued incautiously and ridden into the
French infantry, by whom he was wounded and taken prisoner. This
was a good day’s record for the 10th Hussars, who started their first
Peninsular service—they had only landed in February—with a very
handsome success.
During the day of the combat of Morales Hill’s infantry had started
off from their billets in the villages ten miles north of Salamanca to
move on Toro. A forced march of over twenty miles, through
Fuentesauco, across a very bare and desolate country, brought the
head of the column down to the Douro, where they encamped
among well-watered fields and vegetable gardens. ‘Officers and
men, after the long sultry day, devoured with zest and relish the raw
cabbages, onions, and melons.’ Next morning (June 3rd) the whole
column began to cross the river at Toro, the artillery and baggage by
a ford no more than knee-deep, the infantry by the fine but broken
bridge. Only one arch of it had been blown up, and a resourceful
engineer, Lieutenant Pringle, had contrived an easy method of
utilizing it. A row of long ladders had been laid against each side of
the gap in the roadway: their feet inclined together and united in the
shallow water below. Long and stout planks had then been laid
across, resting at each end on the rungs of each pair of
corresponding ladders, and making a sort of platform. The men went
down the upper rungs of one set of ladders, walked a few steps over
the planks, and ascended by the rungs of the ladders on the other
side. This was rather tedious for the passage of four divisions, and
took the whole day and the following morning. But by noon on June
4 the entire army was concentrated in the vicinity of Toro on the
north bank. Its cavalry was many miles in advance, at Pedroso del
Rey on the Tordesillas road, Almaraz and Villavelid on the Rio Seco
road. Parties sent out northward had got in touch with Giron’s
Galician army, which had passed Benavente on June 1 and reached
Villalpando on the 3rd, with Penne Villemur’s squadrons out in its
front.
Thus every man of Wellington’s striking force, 80,000 sabres and
bayonets, was concentrated north of Toro on the night of June 3—all
the British in a single mass, the Galicians some 18 miles off on the
flank, but easily available. Nothing was now south of the Douro save
Carlos de España’s Spanish division, left to garrison Salamanca,
and Julian Sanchez’s horse, who were searching the roads south of
the river, and had just captured a large French cavalry patrol at
Castro Nuño, near Pollos. It seemed to Wellington incredible that the
enemy would reply to his stroke at their communications by a similar
stroke at his on the Salamanca-Rodrigo line. Indeed, all reports
showed them moving north, in order to form opposite him on the
north bank of the Douro. Moreover, it was clear that they would have
the greatest difficulty in concentrating a sufficient force to fight him,
for the possession of Valladolid and the defence of the great
northern chaussée. The first stage of his plan had been completed
with entire success.
 SECTION XXXVI: CHAPTER IV
MOVEMENTS OF THE FRENCH:
MAY 22-JUNE 4

At the moment when Wellington launched his two great columns

into Spain, the French head-quarters staff was in a condition of
nervous expectation. The spring was so far advanced that it had
been expected that the Allies would have been already on the move,
and their long quiescence was supposed, very reasonably, to cover
some new plan which it was impossible to divine. The position of the
French armies was very unfavourable, entirely owing to the
continued absence from the front of the whole infantry of the Army of
Portugal, which, by Napoleon’s desire and by the detailed
instructions of the Minister of War at Paris, had been lent to the Army
of the North, and sent backward into Biscay, Navarre, and Aragon to
hunt the guerrilleros. Five and a half out of the six divisions of
Reille’s command were still occupied on these marches and counter-
marches in the rear, when May was far spent, and when the
offensive of the Allies must be expected at every moment.
The French army, available for immediate operations, was
therefore short of one-third of its strength, and Jourdan and King
Joseph disliked the situation. Jourdan confesses in his Memoirs that
he and his master ought to have ordered Clausel to suspend his
operations, however incomplete they might be, and to send back all
the borrowed infantry to the valley of the Douro. But the minister’s
letters kept repeating so often that the campaign north of the Ebro
must be completed at all costs, that the King considered that he
could do no more than transmit to Paris the warnings that he was
receiving, urging on the minister that it was time to suspend those
operations, and that General Clausel should be ordered to come
back in haste towards Burgos[485]. This self-exculpation of Jourdan
shows clearly enough the miserable consequences of the system of
double-command which Napoleon had always kept up in Spain. The
habit of sending orders direct from Paris to fractions of the Army of
Spain was so deeply ingrained, that the titular Commander-in-Chief
and his Chief of the Staff dared not issue instructions of primary
importance to one of the generals under them without obtaining
leave from the Emperor! And at this moment the Emperor was not
even at Paris—he had long been at the front in Germany, and had
fought the battle of Lützen on May 2. What came from Paris was not
even the orders of Napoleon, but the orders of Clarke, transmitting
his impression of the imperial will from dispatches already many
days old, which would be doubly out of date before they reached
Valladolid. The supreme master must take the responsibility of the
fact that on May 15 or May 18 his representatives in Spain were
asking for leave to modify his arrangements, by petitions which could
receive no reply—for mere reasons of space and time—till the crisis
which they were fearing had burst upon them. The whole system
was ruinous—in 1813 as it had been in 1812. The only rational
method would have been to turn over the whole conduct of affairs in
Spain to some local authority, supreme in everything and responsible
for everything. Yet stronger men than Joseph and Jourdan would
perhaps have taken the risk of offending their master, and have
issued peremptory orders, which Clausel, Foy, and the other outlying
generals would probably have obeyed.
On May 20 the distribution of the Army of Spain was as follows:
King Joseph and the 2,500 men of his guards, horse and foot, lay at
general head-quarters at Valladolid. Of the two infantry divisions of
D’Erlon’s Army of the Centre, one, that of Darmagnac, had been lent
to Reille, when all the infantry of the Army of Portugal had been
borrowed from him, and was lying at Medina de Rio Seco, in the rear
of Reille’s cavalry, who were watching from a discreet distance the
Army of Galicia and the roads in the direction of Astorga and Leon. It
will be remembered that on May 20 Reille, hearing vague rumours of
allied movements beyond the Esla, had executed a great sweep with
Boyer’s dragoons across the bridge of Benavente, and for five
leagues beyond it, had found nothing, and had reported that there
were no British troops in the direction, and no Spaniards nearer than
Astorga[486].
The other division of the Army of the Centre, that of Cassagne,
was at Segovia, far south of the Douro, keeping touch with the large
garrison still left in Madrid, which, as long as it was maintained there,
could not be left in a state of absolute isolation.
Of the Army of the South, Gazan had his head-quarters at
Arevalo, not very far from the King at Valladolid, but retained with
him only Tilly’s cavalry division. The rest of his troops were woefully
dispersed. Daricau’s division and Digeon’s dragoons lay at Zamora
and Toro—nearly 100 miles from head-quarters—maintaining a
loose touch with Reille’s cavalry. Villatte—with one detached
regiment of Digeon’s dragoons to keep watch for him—was at
Salamanca—fifty miles from Daricau. Conroux, with a third division,
was at Avila, with a whole block of mountains between him and
Villatte. He was supposed to be watching Hill, and had detachments
as far out as Monbeltran on the borders of Estremadura. He was
also separated from the force at Madrid by several sierras and the
formidable pass of the Guadarrama. In the capital itself lay Leval
with one more division, while the independent brigade of
Maransin[487], had been intermittently holding Toledo, and was
actually there on the 20th, in company with a brigade of light cavalry
under Pierre Soult[488]. The extreme outer flanks of the Army of the
South were as far apart therefore as Zamora and Toledo—160 miles
as the crow flies—and a Spanish crow has rough country to fly over
—and each of them was some 90 miles by road from the head-
quarters in the centre.
As to the infantry of the Army of Portugal, the only element of it
which lay anywhere near head-quarters was a brigade of Maucune’s
at Palencia on the Carrion, which was guarding a large dépôt of
stores and transport. The other brigade of Maucune’s was at Burgos.
Lamartinière’s division (late that commanded for so long by Bonnet)
was watching the great chaussée from Burgos by Briviesca to
Miranda del Ebro, in order to keep touch with Clausel in Navarre.
Similarly Sarrut’s division, after its long and fruitless chase of Longa,
was keeping safe the roads from Bilbao to Miranda, in order to link
Foy with the main army. Foy himself was on the shores of the Bay of
Biscay, where he had stormed Castro-Urdiales on May 12, and was
trying to make an end of the local guerrilleros. The other two
divisions of Reille’s infantry were lost to sight in the mountains of
Navarre, where they were marching and countermarching with
Clausel, in pursuit of the elusive Mina. These were the divisions of
Taupin and of Barbot: the troops which were working along with them
were the two ‘active’ divisions of the Army of the North, those of
Abbé and Vandermaesen. Clausel’s great flying column, of 15,000
men or more, was so continually on the move through regions where
cross-communication was impossible, owing to the insurrection, that
it could not be located with certainty on any given date, or receive
instruction without a delay that might run to eight or ten days.
Obviously it would take the French army a week to concentrate
on Valladolid or Arevalo, if Wellington should be aiming at
Salamanca and the Central Douro. But if he were about to attack
one of the extreme wings, at Zamora or Madrid, the time required
would be much greater. And Joseph and Jourdan were not at all sure
that the secret plan of Wellington might not be a thrust, with Hill’s
force as the spear-head, at Madrid. One of the many false rumours
sent to head-quarters was that forage had been ordered for Long’s
cavalry at Escalona on the Tietar, many miles in front of Talavera.
While another report truly chronicled the concentration of Hill’s
brigades at Bejar, and the gathering of stores there, but interpreted
their meaning as preparation for a march on Avila by the Puente de
Congosto.
Uncertain as to what was Wellington’s plan, the French Higher
Command finally resolved to await its manifestation before giving the
final orders—a system which was certain to lead to some initial loss
of territory at the opening of the campaign, since the enemy would
have a week in hand, before he could be opposed by a sufficient
mass on the line which he might select. On May 18th Jourdan
ordered Gazan to push exploring reconnaissances towards the
Portuguese frontier, and, if Wellington should be met advancing, to
order Leval to evacuate Madrid, to send forward all his cavalry to
support Villatte at Salamanca, and, when the latter should be driven
in, to have his whole army ready to receive him behind the line of the
Trabancos river, except Daricau’s division, which should remain at
Toro. Reille was to bring forward Darmagnac’s infantry from Rio
Seco to support his cavalry on the Esla, and to unite Maucune’s
scattered force by calling up his rear brigade from Burgos, so as to
provide a reserve for Darmagnac. If Wellington should advance, as
was thought most probable, from Ciudad Rodrigo, marching on
Salamanca in full strength, and driving in Villatte, it was intended to
bring in all the troops of Gazan and D’Erlon to the north bank of the
Douro, and to defend the course of that river from Toro to Tordesillas,
as Marmont had done in June 1812. If the British showed strength
on the Esla, which was thought unlikely, Reille would reinforce
Daricau and Digeon with all his cavalry and the divisions of
Darmagnac and Maucune. But no definite orders were sent either to
Clausel[489] or to Foy, Sarrut, and Lamartinière—the directions to be
given them would depend on the strength which the British
displayed, and the front on which they appeared. Thus a week or
more after Wellington should have shown his hand, the two armies
of the South and Centre, with the cavalry and one infantry division of
the Army of Portugal, ought to be concentrated on the required
points north of the Douro, with the possibility of bringing up later the
five missing divisions of the Army of Portugal, and Clausel’s two
disposable divisions from the Army of the North. But these last might
take a very long time to appear, and meanwhile there would only be
45,000 infantry and 10,000 cavalry in hand opposite Wellington[490].
Considering the latter’s estimated numerical superiority, it seemed
that he might be brought to a stand, but hardly beaten, by the forces
which the King could collect. A defensive campaign on the Douro
was really Jourdan’s forecast of the game; it might perhaps be
turned into an offensive, when the missing divisions of the Army of
Portugal and Clausel should come up, with some 30,000 men of
reinforcements.
It is a nervous business to wait for the move which finally betrays
the enemy’s intention, as every one knows who remembers August
1914. It was probably with some feeling of relief that Gazan at
Arevalo and King Joseph at Palencia heard, on the 24th, that
Wellington had passed the Agueda on the 22nd, and was apparently
marching on Salamanca in great strength. Conroux sent in news that
Hill was at the same time marching up from Bejar northward—not
turning east. The campaign therefore, as it appeared, was to take
the form which had seemed most probable, and the hypothetical
orders for concentration which had been issued to the generals on
the 18th became valid. But Gazan lost a day in evacuating Madrid,
by riding over from Arevalo to Valladolid and formally requesting the
King’s authorization for the retreat of Leval from the capital. So the
aide-de-camp bearing the dispatch of recall only started from
Valladolid on the morning of the 25th, instead of from Arevalo on the
afternoon of the 24th: it was clear therefore that Leval would be late
at the concentration point.
The other parts of the French scheme were carried out according
to plan between the 25th and the 29th. When Villatte found himself
attacked at Salamanca, and got away with some loss eastward on
the 26th, he fell back on Medina del Campo, and found waiting for
him there on the 28th Conroux’s division from Avila and the division
of cavalry which Gazan had been keeping at his head-quarters at
Arevalo—Tilly’s dragoons. D’Erlon was in march from Segovia with
Cassagne’s division and Treillard’s horse, and had reached Olmedo,
only fifteen miles from Medina, so that he would be available next
day. The whole made up a force of 16,000 foot and 4,000 horse. But
Leval was late, and would not be up for three days more, nor had
Maucune’s brigade yet come down from Burgos, while Daricau and
Digeon began to look uncomfortably remote in their position at Toro
and Zamora, where only Reille could reach them.
Joseph and Jourdan had intended to take position behind the
Zapardiel on the 30th, there to demonstrate against Wellington, who
ought by this time to be coming down upon them. It was necessary
to hold him in check till Leval should have got in from Madrid to join
the main body, and then all would retire beyond the Douro, to the
intended defensive position. But this day two disquieting facts
became notable: one was that Wellington was not advancing from
Salamanca: the troops that he had brought thither had not moved
since the 26th. The other was that Digeon sent word that there were
allied forces approaching the lower Esla from the direction of
Braganza, although Reille had reported ten days back that there
were absolutely no signs of allied movement from this direction—
their numbers were as yet incalculable. The suspicion, then for the
first time, arose that the Salamanca advance might be a mere feint,
despite of Wellington’s personal appearance in that direction.
On the morning of June 1st suspicion became certainty. Digeon
reported that the British had crossed the Esla in great force, by
several fords, and that he and Daricau were retiring on Toro at once.
Joseph would have liked to go behind the Douro without delay, but
could not possibly do so until Leval, Maransin, and Pierre Soult
should have come up from the south. And the Madrid column was
not due on the Douro for two more days, owing to its late start—no
fault of its own, but due to Gazan’s dilatory conduct on May 24th.
Leval had only got the order to evacuate the capital on the 26th.
Luckily for him Maransin had come in with his flying column from
Toledo on the 21st, so that there was no need to wait to pick up this
detachment. The garrison was directed to be under arms for
marching at daybreak on the 27th, and with them went a
considerable train of Afrancesado refugees, for many of Joseph’s
ministers and courtiers had refused to start by the earlier convoys,
which had gone to Valladolid in March and April, hoping that
emigration might never again become necessary. The remembrance
of the miserable march to Valencia in July 1812 weighed on their
minds, and made them unwilling to face a second hegira of the same
sort. Now all had to go; and quantities of carriages, carts, wagons,
and mules belonging to civilians were mixed with, or trailed behind,
the columns of Leval’s infantry. The convoy could not travel very fast;
it only reached the foot of the mountains on the night of the 27th,
crossed the Guadarrama pass on the 28th, and reached Espinar on
the northern descent on the 29th. Here Leval turned off the baggage
and refugees on to the road of Segovia, under charge of an escort
commanded by General Hugo, late Governor of Madrid. He himself
with the fighting men went on to Arevalo by the great chaussée,
reached the Douro on the 2nd of June, and joined Gazan at
Tordesillas. The latter had crossed the river on May 31st, the
moment that it became clear that the Madrid column was not going
to be intercepted by any British force from the Salamanca direction.
D’Erlon passed the Douro only on June 2, the same day as Leval,
having waited behind to cover the arrival of the column of refugees
and transport from Madrid, which had been directed on Segovia and
Cuellar.
On June 2nd therefore the south bank of the Douro had been at
last evacuated by all the French forces, but Jourdan’s and Joseph’s
plan for defending the northern bank was obviously out of date and
impossible. For Wellington was already at Toro with Graham’s
40,000 men, and there was no means of preventing Hill’s 30,000 and
the Galicians from joining him north of the river. The Salamanca
column, if only the French had known it, was already marching north,
to fall in alongside of the other and larger mass of the allied army,
and was due at Toro next day (June 3rd).
This last fact was, of course, unknown at French head-quarters
on the afternoon of June 2nd—all that had transpired was that
Wellington was at Toro with a very large force, and the Galicians
close by his flank on the Benavente road. The exact position of the
imperial army was that Gazan, with four and a half infantry divisions
of the Army of the South and all the cavalry of Digeon, Tilly, and
Pierre Soult, was concentrated on a ten-mile front between
Tordesillas and Torrelobaton: Reille, with Darmagnac’s division and
Boyer’s cavalry, was at Medina de Rio Seco, twelve miles farther
north. D’Erlon, with Cassagne’s division and Treillard’s dragoons,
was at Valladolid, fifteen miles behind Tordesillas; the King and his
Guards at Cigales, ten miles north of D’Erlon, twelve miles east of
Reille. Maucune’s two brigades had at last concentrated at Palencia
—some 25 miles north of Valladolid. There was now a very solid
mass of troops, which could be united at one spot—Torrelobaton for
example—by a concentric march, taking up one day, or could on the
other hand string itself out in a well closed-up front behind the
Carrion and Pisuerga rivers, from Palencia to Simancas, to defend
the line of those streams. The force was formidable—over 40,000
bayonets, quite 10,000 sabres. It was physically possible either to
mass and offer Wellington a battle, or to stand fast to defend the line
of the Pisuerga. But would either course be prudent, considering that
Hill could join Wellington with ease at Toro, so that the entire Anglo-
Portuguese army could be concentrated for a general action, not to
speak of the Galicians of Giron on his flank? Exact calculations as to
the allied strength were impossible—but it would certainly exceed
70,000 men: some of the French generals put it at 90,000. To
engage in an open battle west of the Pisuerga against such superior
numbers would be insane. But what of the idea of taking up a
defensive line east of that river, in the hope of rallying Foy, Clausel,
and other outlying forces within a week or ten days? Marmont had
done well with his position behind a river in July 1812, and had only
been ruined at Salamanca because he had left his strong line and
attacked with insufficient numbers.
There was much debate at head-quarters on the afternoon of
June 2. Several policies were discussed, even, as it would appear, a
desperate suggestion of Jourdan’s to bring the whole army back
southward across the Douro, to Medina del Campo and Olmedo, and
defy Wellington to cross its front so as to cut the Burgos road and
the communication with France. ‘It is doubtful whether Wellington
would have dared to continue his march to the Carrion, and to
abandon his line of connexion with Portugal,’ writes Jourdan; ‘more
probably he would have repassed the Douro, to follow the French
army, which could then have retired up-stream to Aranda, and from
thence either on Burgos or on Saragossa. Time would have been
gained—Clausel would have come up, and we could have fought on
ground more suitable for cavalry[491].’ This most hazardous plan
would have commenced by abandoning to the enemy all Old Castile
and Biscay, with the French forces scattered in them; for even if
Wellington had followed the King with his Anglo-Portuguese army,
Giron’s Galicians and the insurgents of the North—Longa, Porlier, El
Pastor, and the rest—would have been left free to clear the whole
country up to the Pyrenees. The French detachments in the North
must have retired on Bayonne. King Joseph and the generals
rejected the scheme at once, on the ground that Napoleon’s orders
had always insisted on the retention of the direct road to France, by
Burgos and Vittoria, as the most important of all considerations. This
was undoubtedly a correct decision. Not only was the road along the
south bank of the Douro to Aranda a very bad one, through an
exhausted country, but it is clear that, if they took it, the French
would never have got back again on to the Burgos line, but would
have been forced to take the Saragossa line. The way from Aranda
by Soria to Saragossa was through a rough country infested by
Duran, the Empecinado, and other active guerrillero leaders. It could
never have served as the main artery of communication for an army
of the size of King Joseph’s. But supposing that the King should
have reached Saragossa, his only touch with France would be
through Barcelona and Roussillon, for there is no decent carriage
road from Saragossa across the Central Pyrenees—the difficult pass
by Jaca having often been tried and found wanting, save for small
and unencumbered detachments. To base the whole army of Spain
on Perpignan instead of Bayonne would have been hopelessly
impracticable.
The question as to whether the line along the Pisuerga, from
Palencia to opposite Simancas, could not be held was the really hard
problem. It was 35 miles long, well marked, but with local faults; the
water was low; there were points where the west bank commanded
the east; from near Palencia to Cabezon the only good road was on
the western side of the river, and so out of control; while the bad path
on the eastern side was cut up by three streams coming in from the
high plateaux of the province of Burgos. In the rear there was the
defile of Torquemada, where the road and the Pisuerga itself came
out of the upland. After pondering over the question, Joseph decided
that the Pisuerga was too dangerous a line to hold, that he was still
too weak in numbers, and that he had better fall back on Burgos,
and wait in position there, on ground much more defensible than the
broad plains of the Pisuerga and the Carrion, for the arrival of the
missing divisions of the Army of Portugal and of Clausel.
On the evening of the 2nd orders were issued that all the
impedimenta of the Army of Spain should move off northward at
once. The Grand Park and other transport, the Spanish refugees
with all their carriages and lumber, the French civil administrators,
the King’s ministers with his private baggage and treasure, and
much miscellaneous stuff from the royal palaces—pictures, books,
and antiquities—were to start off at once on the road to Burgos. One
great convoy was dispatched that night, a second and still larger one
the following morning. Escort was found for them from the King’s
Spanish troops, the so-called ‘division’ of Casa Palacios, and other
detachments, making up 4,000 men in all. On the afternoon of the
3rd the army executed a general movement of retreat, leaving only a
cavalry screen behind, to observe Wellington’s advance. The Army
of the South came back through Valladolid, after blowing up all the
bridges on the lower Pisuerga and the neighbouring streams, and
then marched up its western bank to Cabezon. Reille evacuated
Medina de Rio Seco and fell back on Palencia, where he picked up
Maucune. Head-quarters and the King’s guard moved to Magaz, just
south of Palencia. D’Erlon marched from Valladolid to Dueñas, ten
miles farther south than Magaz, and made ready to blow up the
bridge there and cross to the other side of the Pisuerga. The whole
army was collected in a space of 15 miles, and halted for two days,
in a safe position for retreat on Burgos, when it should be pressed.
For Joseph and Jourdan naturally wished to gain time—every day
that passed made it more likely that the missing divisions from the
North would be heard of, or even be reported as approaching. Yet it
would seem that even yet no direct and peremptory orders had been
sent to Clausel, for Jourdan writes in his Memoirs that ‘the King
suspended the retrograde movement because he thought to gain
time, and hoped that the Minister of War would have given orders to
General Clauzel to come down on Burgos[492].’ Again the wretched
system of double-command and constant reference to Paris was
working! If it be remembered that it was only on the 18th of May that
the King had made his final appeal to the minister to let loose
Clausel from the northern operations, and that the news of
Wellington’s actual advance had only been sent to France on the
24th, when could it be expected that the directions from Paris would
reach the scattered columns lost in the mountains of Navarre? And
was there any certainty that Clarke would visualize the full danger of
the position, and give at once the kind of orders that the King
desired? He might send instead a lecture on the Emperor’s
intentions, if past experience was to be trusted, and some
suggestions which the events of the last ten days would have put
completely out of date.
 SECTION XXXVI: CHAPTER V
THE OPERATIONS AROUND BURGOS:
JUNE 4-14, 1813

On June 3rd Wellington had halted the infantry of Graham’s

column at Toro, in order to allow the whole of Hill’s column to cross
the bridge and fords, and to complete the junction of the army. He
gave as his reasons that he expected to meet Gazan, D’Erlon, and
the King, on the Hormija river, when he should have debouched from
Toro, and that he intended to fight them with his whole force, not with
the northern column only. ‘I do not think,’ he wrote, ‘that we are so
close up or so well concentrated as we ought to be, to meet the
enemy in the strength in which he will appear on the Hormija,
probably to-morrow; and therefore I propose to halt the heads of the
different columns to-morrow, and close up the rear of each, moving
Hill in this direction preparatory to our further movement[493].’ Only
the cavalry continued to press forward on the 3rd; they found no
signs of resistance, but only vedettes, which retired cautiously on
their approach. Wellington was so far right in his apprehension, that
the idea of fighting in front of Valladolid had been one of the three
plans which the French Head-quarters Staff had considered on June
2. But, as we have already seen, that scheme had been rejected as
rash, and during Wellington’s halt upon June 3 the enemy were
evacuating Valladolid and all its neighbourhood, and commencing
their retreat behind the Pisuerga. By eight o’clock on the evening of
that day the cavalry reports, all along the front, showed that the
enemy was retiring on every road, and it appeared certain that they
did not intend to fight; the army therefore advanced in three main
columns, Wellington himself with the 3rd, 4th, 6th, and 7th Divisions,
followed by Hill, by the high road on Valladolid through La Mota:
Graham with the 1st and 5th Divisions, Pack and Bradford, by the
more northern parallel road, by Villavelid and Villar de Frades, on
Medina de Rio Seco. The Galicians were converging on the same
point as Graham, by the high road from Villalpando and Villafrechos,
but were directed not to approach too close to Medina till further
notice. The reason for their delay was a curious one—Giron had just
written on the previous day to say that he had no reserve of infantry
cartridges, and only 60 rounds a man in the soldiers’ pouches. He
wanted to borrow a supply from the British army. Wellington replied
in considerable heat: when he was on the move he could not give
away his own reserve ammunition: no doubt Giron was not
personally to blame—nor his men—but the administration in Galicia.
Yet the unfortunate result would be that the Galicians must never be
put in a position where there would be a heavy consumption of
cartridges, i. e. must be held in reserve, or used for subsidiary
operations only. Giron was directed to make small regimental
reserves, by taking away a proportion of the cartridges from each
man and carrying them on mules. The general got off easily—but
Wellington thundered that night on the War Minister O’Donoju, ‘Here
is an army which is clothed, armed, and disciplined, but cannot be
brought into action with the enemy: I am obliged to keep it in the
rear. How can troops march without provisions, or fight without
ammunition? The cause of the country may be lost unless the
Government establish in the provinces some authority to which the
people will pay obedience, and which will insure their resources for
the purposes of the war[494].’ It will be noted that all through the rest
of the campaign Giron’s corps was used for flanking movements,
and never put in the forefront of the fighting—though the other
Spanish troops with the army, Morillo’s and later Longa’s divisions,
were used freely.
On the 4th June Wellington moved with the ‘Head-Quarters
Column’ as far as La Mota, Hill getting no farther than Morales. At La
Mota the information received showed that the French were in
general retreat—the cavalry advance got into Valladolid and found
some undestroyed stores of ammunition there. Julian Sanchez,
scouting far towards the south, discovered a considerable magazine
of grain at Arevalo, where Gazan’s head-quarters had been. In view
of the new situation it was necessary to recast the movements of the
army—there was to be no fight on the Hormija, or for the possession
of Valladolid. The initial strategical success of the passage of the
Esla had settled the event of the first phase of the campaign, and
cleared the French not only out of New Castile, but out of the
kingdom of Leon and great part of Old Castile. The fact that the
enemy had abandoned Valladolid and the lower crossings of the
Pisuerga, would seem to show that he intended to fall back on the
Burgos country: he could hardly be intending to defend a position
behind the middle Pisuerga and the Carrion, since he had
surrendered the passages at and about Valladolid, by which the
southern flank of such a position could be turned.
Now nine months ago Wellington had been in these same
regions, with the army of Clausel retreating before him from
Valladolid on Burgos. On that occasion he had followed the enemy in
a straightforward pursuit along the great high road, by Cabezon and
Torquemada. On this occasion his strategy was entirely different.
Leaving only a cavalry screen between himself and the enemy, he
proceeded to move his whole army towards the north-west by
secondary roads, marching in four parallel columns not on Burgos
but on the upper Pisuerga north-west of that fortress, so as to turn
entirely any position that the enemy might take up on the Hormaza,
the Arlanzon, or the Urbel. It can hardly be doubted that he had
already in his mind the great manœuvre which he was to accomplish
during the next fortnight—that of outflanking the French right wing by
a wide sweeping movement, which would not only force Joseph and
Jourdan to evacuate Burgos and its neighbourhood, but would cut
them off from the royal road to Bayonne and their main
communication with France. The first order in his dispatch book
which definitely reveals this intention does not appear till June 10[495],
but the facts which it contains prove that the plan must have been
laid long ere Wellington started from Portugal.
This plan was no less a scheme than the transference of the
base of the British army from the port of Lisbon to the Bay of Biscay,
so that when, in its wide turning movement, it should have passed
the Ebro and neared the Cantabrian coast, it should find stores and
munitions awaiting it, and be no longer tied to the long line of
communication with Portugal, which it had always used hitherto. It
was an astonishing example of the strategical use of sea-power, to
which there had hitherto been no parallel in the Napoleonic wars.
The only manœuvre at all resembling it was that by which Sir John
Moore in December 1808 had changed his line of communication
from Lisbon to Corunna. But this was for a sudden hasty retreat,
made by a small army—a very different thing from an advance made
by a large army. For Moore had only wanted transports on which to
abscond from the Peninsula; Wellington had planned the arrival of a
fleet of supply vessels, on whose contents he was to rely for the
future sustenance of his army during a long campaign.
Their existence was his second great secret in 1813—the first
had been his plan for the crossing of the Esla. But the latter might
have been guessed at as a possibility by any intelligent French
general: the former, it is safe to say, had never been dreamed of as a
conceivable move. Santander is so remote from Portugal, and the
French were so firmly rooted in Northern Spain in the spring of 1813,
that it could never enter into the head of any of Napoleon’s
subordinates that the British Commander-in-Chief was planning such
an elaborate surprise. They were bound to believe that all
Wellington’s operations would be founded on, and circumscribed by,
the basic fact that his line of communication was on Lisbon. A
manœuvre which presupposed the complete abandonment of that
line was not conceivable by officers reared in Continental
campaigns, and unused to contemplating the correlation of land
operations and naval strategy. Yet the fact that the supply fleet had
been gathered at Corunna long weeks before, and kept there till the
conquest of Northern Spain was well on its way to completion, is
conclusive evidence as to Wellington’s intentions. The crucial
dispatch of June 10 ran as follows—it was addressed to Colonel
Bourke in charge of the British dépôts in Galicia:
‘There are at Corunna certain ships loaded with biscuit and flour,
and certain others loaded with a train of heavy artillery and its
ammunition, and some with musket ammunition, and I shall be much
obliged if you will request any officer of the navy who may be at
Corunna, when you receive this letter, to take under his convoy all
the vessels loaded as above mentioned, and to proceed with them to
Santander. If he should find Santander occupied by the enemy, I beg
him to remain off the port till the operations of this army have obliged
the enemy to abandon it.
‘If the enemy is not occupying Santander, I beg him to enter the
port, but to be in readiness to quit it again if the enemy should
approach the place, until I shall communicate with him[496]’.
It should be noted that Wellington had carefully avoided calling
attention to the accumulation of ships and stores at Corunna, even in
his letters to the ministers at home[497]. Their existence there was
plausibly explained by the need for supplying the Spanish army of
Galicia—which had indeed received much material during the spring.
The provision of heavy artillery shows that he was contemplating as
a probability the siege of San Sebastian and the other northern
fortresses. He had failed at Burgos in 1812 for lack of the 24-
pounders which Home Popham had vainly offered to send him from
Santander: now he was quite ready to receive them from that port.
As things stood on the Pisuerga, at the moment that he sent this
prescient dispatch to Bourke, it was certainly a bold prophecy to
write that he was about to clear the Cantabrian coast by his next
move. The French were still unbeaten, and he was more than a
hundred miles from the Bay of Biscay, from which he was separated
by a most rough and complicated land of mountains. But he had
already taken his measure of the capacity of the hostile
commanders, and his hopes were high.
The great flank movement’s initial stages can be best traced by
following on the map Wellington’s nightly head-quarters. They were
La Mota on the 4th, Castromonte on the 5th, Ampudia on the 6th—
all these carefully avoid the neighbourhood of the Pisuerga, though
the cavalry pushed on to feel the lines of the French along the river.
But the army, instead of making for the crossings at Valladolid,
Cabezon, or Dueñas, kept steadily on in four parallel columns, on
roads far from the river. Hill, still as always on the right, went by
Torrelobaton, Mucientes, and Dueñas; Graham, on the left, by
Medina de Rio Seco and Grixota; ‘the Head-Quarters Column’ under
Wellington himself, by Castromonte and Ampudia—while, far off, the
Galician army on the north moved by Villaramiel and Villoldo on
Aguilar de Campos. The flanking cavalry kept up a continual