Algorithms for Statistical Signal Processing

This authoritative, coherent presentation of computational algorithms for statistical signal processing focuses on advanced topics ignored by other texts on the subject—e.g., algorithms for adaptive filtering, least squares methods, power spectrum estimation, and high-order spectral estimation.

Features

• Basic treatment of algorithms for computing convolutions and the DFT.
  Exposes readers to computationally efficient methods for signal processing involving convolutions and the discrete Fourier transform.
• Treatment of linear prediction and the design of optimum linear filters.
  Introduces readers to the notions of forward and backward linear prediction and to the Levinson-Durbin and Schur algorithms for efficiently computing the inverse of a Toeplitz matrix; also to the design methodology for Wiener filters.
• Exploration of adaptive filtering algorithms with applications to signal processing and telecommunications.
  Shows readers several practical applications of adaptive filtering and the methodology for designing adaptive filtering algorithms and analyzing their performance (including the LMS and RLS algorithms).
• In-depth treatment of recursive least squares algorithms for array signal processing.
  Provides readers with a firm foundation in numerically and computationally efficient recursive least squares algorithms based on the QR decomposition, including the modified Gram-Schmidt algorithm, the Givens algorithm, the Householder algorithm, and several order-recursive lattice algorithms. Shows them how to design QR-type computationally efficient, fast adaptive filter algorithms using signal flow graphs.
• Thorough treatment of power spectrum estimation methods.
  Introduces readers to both nonparametric and parametric methods for power spectrum estimation—including such algorithms as the Bartlett, Welch and Blackman and Tukey nonparametric methods, and the Pisarenko, minimum-variance, MUSIC and ESPRIT algorithms for parametric spectrum estimation based on linear system models (autoregressive, moving average, and autoregressive-moving average).
• Treatment of signal analysis based on higher-order spectra.
  Familiarizes readers with higher-order spectral methods for the analysis of linear and nonlinear random signals, including cepstra of higher-order spectra, the bispectrum, and some applications.
• Basic DSP review—Ch. 1.
  Provides readers with an easy transition into the material by summarizing material typically found in a first-level DSP book (e.g., the z-transform, the analysis and characterization of discrete-time systems, the Fourier transform, the discrete Fourier transform DFT, and the design of FIR and IIR digital filters) and establishing the notation used throughout.

1. Introduction.

2. Algorithms for Convolution and DFT.

3. Linear Prediction and Optimum Linear Filters.

4. Least-Squares Methods for System Modeling and Filter Design.

5. Adaptive Filters.

6. Recursive Least-Squares Algorithms for Array Signal Processing.

7. QRD-Based Fast Adaptive Filter Algorithms.

8. Power Spectrum Estimation.

9. Signal Analysis with Higher-Order Spectra.

References and Biology.

Index.