This book provides a good coverage of the spectral estimation field. Chapter 1 is a general presentation of the spectral estimation problem. Chapters 2 and 3 review the needed basic results of matrix algebra and of probability, statistics and random processes, respectively. The periodogram-based and Blackman-Tukey approaches to spectral estimation, also called the classical approaches, are presented in Chapter 4. Chapter 5 introduces the ARMA models for stationary processes with rational spectral densities and describes their second-order properties. Chapters 6 and 7 present methods and algorithms for estimating the parameters in AR models and study their performance. Model order selection procedures are also briefly discussed.
The estimation of parameters in MA and ARMA models (including order estimation) is discussed in Chapters 8 to 10. Chapter 11 discusses the minimum variance (Capon) spectral estimation. Chapter 12 presents a summary of the spectral estimators derived in the previous chapters and a comparison of their performance on a test case data set.
Chapters 13 to 16 address topics considered to be more advanced. The subject of Chapter 13 is sinusoidal parameter estimation. Chapter 14 discusses multichannel spectral estimation, and Chapter 15 two- dimensional spectral estimation. Finally, Chapter 16 contains a brief presentation of various applications of spectral estimation methods.
The book includes a floppy disk which contains computer program subroutine implementations of all the methods discussed in Chapters 2 to 12. The (FORTRAN 77) source codes of the computer programs are also given in the book. The book contains numerous computer simulations which aim at helping the reader to choose a specific spectral estimation method for a specific application.
The book is written at a reasonable level of mathematical sophistication and may therefore prove useful for the applications-oriented reader. For the expert, the book may serve as an overview of methods, algorithms and theoretical performance results developed prior to 1985. The reviewer recommends this book to anyone engaged in the field of spectral estimation.