Adaptive System Identification and Signal Processing Algorithms

An account of an important class of algorithmic families for adaptive system identification and signal processing. The LMS family and R&S and its fast versions, as well as the back propagation algorithms for neural networks, are examined in the context of algorithmic efficiency; that is, the issues of convergence, tracking, computational complexity, parallelism, roundoff error robustness and stability. The chapters are presented in such as way as to provide the basic derivation guidelines as well as the necessary performance analysis without sacrificing rigour. Typical applications areas such as channel equalization, echo concealment, spread spectrum RAKE receiver design and spectral analysis are presented and discussed.