Adaptive filters viewed as iterative linear equation solvers

  • Authors:
  • John Håkon Husøy

  • Affiliations:
  • Department of Electrical and Computer Engineering, University of Stavanger, Stavanger, Norway

  • Venue:
  • NAA'04 Proceedings of the Third international conference on Numerical Analysis and its Applications
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Adaptive filtering is an important subfield of digital signal processing having been actively researched for more than four decades and having important applications such as noise cancellation, system identification, and telecommunications channel equalization. In this paper we provide a novel framework for adaptive filtering based on the theory stationary iterative linear equation solvers. We show that a large number of established, and some quite novel, adaptive filtering algorithms can be interpreted as special cases of a generic update equation forming the cornerstone of our framework. The novelty of our contribution is the casting of the adaptive filtering problem as a problem in numerical linear algebra facilitating a clearer understanding and a unified analysis of the various algorithms.