Digial processing of signals: Theory and practice

2 downloads 0 Views 258KB Size Report
signal processing, but as one of the best introductory books on this subject. The major merits of .... As can be seen from the list of topics covered by the various chapters, the book ... distribution transformers; and topics relating to the design of.

Book Reviews The following reviews were selected from those recently published in various /€E€ TRANSACTIONS, Magazines, and Newsletters. They are reprinted here to make them conveniently available to the many readers who otherwise might not have ready access to them. Each review is followed by an identification o f its original source.

Digital Processing of Signals: Theory and Practice-Maurice Bellanger. (New York,NY: Wiley, 1984, 375 pp.) Reviewed by Enrico Del Re, Dipartimento di lngegneria Elettronica, Universiti di Firenze, 50739 Florence, Italy. This is the English translation of the original French book TraitementNumerique Du Signal: Theorie Et Pratique by the same author. Many books, dealing with the basic theory and/or the applications of the modern digital processing of signals,are presently available i n English and some of them are of a good level. Therefore, the publication of a new book should be motivated by i t s specific merits, such as a particularly clear exposition ofthe theory presented and/or the relevance of the applications dealt with. In the reviewer’s opinion this book by M. Bellangermeets these conditions to so large an extent that it can be considered not only a valuable addition to the English-written literature on digital signal processing, but as one of the best introductory books on this subject. The major merits of the book are: the structure; that is, the subdivision of the subject within the 13 chapters; the clearness and rigorousness of the exposition; and the inclusion of practically all the topics of interest in the digital signal processing field (that is not easily available in other books). The book deals entirely with one-dimensional signals and systems. The treatment of the basic theory as well as the applications of both one-dimensional and multi-dimensional digital signals and systems would have constituted an apparently more complete book. However, the author’s choice is fully motivated by the fact that most properties of mutlidimensional signals and systems can be generally derived from those of the one-dimensional case. Chapter 1 introduces the concepts and techniques of signal sampling and coding that are the fundamental operations preceding any digital processing function. After a brief reviewof the fundamentals of Fourier analysis, distributions, some properties of deterministic and random signals, the sampling and quantization operations are considered in detail. The effect of sampling on the signal spectrum and the sampling theorem are clearly discussed. Some useful considerations on the sampling operation of sinusoidal and random signalsare presented and a methodof generating discrete noise is described. Both linear and nonlinear (A-law) and optimal codings of signals are dealt with. This first chapter is very useful and wellpositioned, making the reader familiar with the practical problemswhich are to be faced before entering the digital world. Chapters 2 and 3 deal with transformations. Chapter 2 introduces the discrete Fourier transform and clearly explains various algorithms of the fastFourier transform (decimation-in-time, decimation-in-frequency, radix 4). Then a well-written and useful section on the finite wordlength effects on the fast Fourier transform follows. The analysis of the filtering function ofthe discrete Fourier transform and its application to the power spectrum estimation are described. The chapter ends with the description of the fast con-


volution and the chirp-z transform algorithms. Chapter 3 gives the reader valuable and, i n a certain sense, unique material. By using the Kronecker product of matrices and permutations, the author gives a unified representation of the various fast Fourier algorithms and introduces some new ones (e.g., the odd-time odd-frequency discrete Fourier transform). The chapter alsodescribes other fast algorithms, e.g., the Winograd and the number theoretic transforms. The topic presented in this chapter is at a higher level than is required for a basic course and can be omitted at first reading. However, Chapter 3 is undoubtedly an interesting and useful p 2 of the book which more interested readers and researchers can profitably refer to. Chapter 4 is the first one dealing with the properties of timeinvariant digital linear systems. It introduces the z-transform, defines and evaluates the power of discrete signals, gives the filtering relations for random signals, introduces the most interesting class of linear time-invariant discrete systems defined by difference equations, and ends reviewing thestate variable analysis of discrete systems. Chapter 5 discusses finite impulse response digital filters. The main methods of calculating the filter coefficients are described by Fourier series expansion from frequency specifications (window method), by the leastsquares criterion, and by the Chebyshev criterion. Some structures for implementing finite-impulse-response filters are described together with the effects of the finite wordlength. The last section introduces the property of minimum-phase filters to reduce the computational complexity. Chapters 6 and7 deal with infinite-impulse-response digital filters. Chapter 6 describes the basic elements (first-order and second-order sections)used for their implementation, evaluates the effects of coefficient and internal data wordlength limitation, and discusses the stability and limit-cycles conditions. Chapter 7 describes the properties of more general infinite-impulse-response filters and the methods for calculating their coefficients: direct calculation from analog models by the bilinear transformation and iterative techniques for specified frequency response. Implementation structure are presented for high-order filters. A detailed analysis of the effects of the coefficient and internal data wordlength limitation is explained, giving valuable tools for assessing the filter design. Chapter 8 deals with another interesting and up-to-date topic in the area of digital processing of signals. The properties and implementation of digital ladder filters are described, starting from the review of analog two-port circuits. This chapter includes sections dealing with switched capacitor filters, digital wave filters, and digitallattice filters. The subject of Chapters 4 to 8 allows the reader to get familiar with the most important methods and techniques of analyzing and designing digital filters. These chapters are also a good reference material for researchers. Chapter 9 deals with complex signals, i n particular analytic signals, and minimum-phase filters, that are two important topics for many applications. Starting from the definition and properties of the Fourier transform of real and causal sequences, the concept of analytic signal is introduced and its properties are outlined. As a fundamental part of the generation of an analytic signal, quadrature filters are considered, both of the nonrecursive and recursive types. The chapter ends describing the characteristics of minimum-phase filters. Chapter 10 considers the problem of efficient digital processing of signals by multirate filtering, presently a topic of great interest. The decomposition of finite-impulse-response filters to obtain effi-


VOL. 75, NO. 2, FEBRUARY 1957

cient processing algorithms and structures is described, with special emphasis on half band filters, that offer a valuable solution for some applications. Multirate filtering with infinite-impulseresponse filters is also presented. Multirate filtering implementation structures are illustrated. The chapter ends with the applicationofmultirate signalprocessing to filter banksby using the polyphase networks and the discrete Fourier transform. Chapter 11 is devoted to adaptive filtering. It describes adaptive filtersofthe most common type, where the coefficients vary according to the gradient algorithm and the approximation criterion is the leastmeansquareerrorsignal. After reviewing the propertiesofthe autocorrelation matrix of a random signal and pointing out the fundamental role of its eigenvalues, the gradient algorithm is clearly introduced and the convergence conditions are studied. The two major adaptation parameters, the time constant and the residual error, are then analyzed, together with the arithmetic complexity for the filter assessment from an implementation point of view. Sign algorithms and cascade offinite-impulseresponse adaptive filters and adaptive second-order sections of infinite-impulse-response filters arealso considered. Finally, the important case of the linear prediction is analyzed and its implementation is discussed. Theassessment of the implementation complexity of a desired digital signal processing i s one of the most important and difficult problems a design engineer is faced with and few quantitative parameters are available for an a priori accurate estimate of the system complexity (in addition, there is no definite quantity or set of quantities that, in a certain sense, measure the implementation complexity,which remainsan essentially qualitative evaluation parameter). A resonable evaluation of system complexity can be carried out by experience design engineers only and will be mainly based on the personal sensitivity gained by working in this field. However, some basic starting points for a digital signal processing system complexity assessmentcan undoubtedly be stated.These points are the subject of Chapter 12, which reviews the types of circuits to be used in the digital implementations (including distributedarithmetic and types of digital signalprocessors) and indicateswhatkindof quantities must be considered for the evaluation of the system complexity. It is a good starting point to help the reader who is getting hisorher own approach to the system complexity evaluation. Chapter 13presents the most important applications of digital signal processing in telecommunications. Although the section does not consider the successful applications of digital signal processing to many other areas, however, this proves to be an interesting part, giving details of concrete designs and implementationsofdigital signal processing equipment in one of the most challenging areas. I think the reader will appreciate this section and will.gain more familiarity with the digital signal processing problems and solutions, looking at really operating implementations. As can be seen from the list of topics covered by the various chapters, the book introduces the fundamentals of all the aspects of digital processing of current interest in the research and application fields for one-dimensional signals.Some of these topics are not available in other basic books on digital signal processing. The exposition is concise, rigorous, and clear. Many useful examples are disseminated throughout the text, which considerably help the reader to familiarize and master the theory and subjects dealt with. Moreover, exercises, often based on concrete examples, are added at the end of each chapter, which allows the reader to test his or her understandingof the material and to become more familiar with its use. Lastly, a valuable idea of the author is to provide the book owner of a pocket memorandum that, for a quick design, summarizes the main relations among the digital filter parameters. In conclusion, this book is well-written and successful book on the digital signal processing methods, techniques and applications and an importantcontribution to the literature of the field. It clearly benefits of the wide knowledge and practice acquired by the author in his long-term research activity i n this field. By its nature the book can be strongly recommended as a text bookfor students of courses in digital signalprocessing, as a self-study book for practicing engineers, and as a reference book for researchers. The reviewer is persuaded that users will award the book the success it certainly deserves. Reprinted from /€E€ Trans. Acoustics, Speech, and Signal Processing, vol. A S P - 3 3 . no. 4, pp. 1062-1063, Aug. 1985.


Book Alert

The following descriptionsof recent bookswere prepared b y the staff of the EngineeringSocieties Library, 345 East 47th Street, New York, N Y 70077-2394.These books are available in the Library for loan or reference use. The prospective buyer should contact the listed publishers or his local technical book store.

Computer-Aided Design in Magnetics-D. A. Lowther and P. P. Silvester. (New York: Springer-Verlag, 1986, 324 pp., bound, $45.00, ISBN 0-387-15756-5.) This book addresses physical problem formulation and engineering design techniques necessary for the effective use of CAD systems specifically meant for magnetics.These modern design techniques offer immensely powerful tools that blend electromagnetic field theory, numerical mathematics, and computer graphics. Concentrating on the art of design, the book contains ample case studies to illustrate all aspects of the design process. It is intended to serveusers of computational analysis and design systems in magnetics and electromagnetics.

R o b o t Sensors. (International Trends in Manufacturing Technolo g y ) 2 Volumes-Alan Pugh,Ed. (Kempston, Bedford MK42 7BT, UK: IFS (Publications) Ltd., and New York, NY: Springer-Verlag, 1986, bound, $49.50,ISBN0-948507-01-2 IFS (Publications) Ltd, 0-387-16125-2 Springer-Verlag.)

Volume one-Vision-includes an overview of imaging sensors and papers on robot visionssensors. Also included is information on fiber-optic sensors,sensors exploiting laser technology, and scene illumination/structuredlight in connectionwithrobotvision. Volume two-Tactile and Non-Vision-covers passive and active force sensors, tactile arraysensors, tactile transducers, and ultrasonic sensors.

Electric Power Distribution System Engineering-Turan Gonen. (New York, NY: McCraw-Hill, 1986,739pp., bound, price not given, ISBN 0-07-023707-7.) This textbook concentrates on modern electrical powerdistriution engineering. The author coverssuch topics as distribution system planning; distribution system automation; load characteristics; application of distribution transformers; and topics relating to the design of subtransmission lines, distribution substations, primary, and secondary systems.There are alsodiscussions of voltage drop and power loss calculations, application of capacitors, voltage regulation, reliability, and distribution system harmonics.

Modern Communications and Spread Spectrum-George R. Cooper and Clare D. McCillem. (New York, NY: McCraw-Hill, 1986, 436 pp., bound, $51.95, ISBN 0-07-012951-7.) The first sevenchapterscover topics relating to analog and digitalcommunications. They review modulation theory, noise analysis, coherent and noncoherent detection of analog, binary, and m-ary signals, and they develop methods of comparing the performances of communication systems. The remaining chapters provide adetailed, understandable introduction to spread-spectrum communications, including a coherent and unified development of spread-spectrum theory.