A unified framework for least square and mean square based adaptive filtering algorithms

Author

Zhang, Zhongkai ; Bose, Tamal ; Gunther, Jacob

Author_Institution

Dept. of Electr. & Comput. Eng., Utah State Univ., Logan, UT, USA

fYear

2005

fDate

23-26 May 2005

Firstpage

4325

Abstract

This paper presents a unified framework for adaptive filters based on a line search method. Expressions for this unified framework are derived. Based on this framework new algorithms are derived, namely, diagonal Q-correlation matrix least square algorithm (DQLS), block diagonal Q-correlation matrix least square algorithm (BDQLS) and their reduced complexity variants. It is shown that both DQLS and BDQLS have less computational complexity compared to EDS and RLS, and better performance than LMS.

Keywords

adaptive filters; correlation methods; least mean squares methods; least squares approximations; BDQLS; DQLS; EDS; LMS; RLS; adaptive filtering algorithms; adaptive filters; block diagonal Q-correlation matrix least square algorithm; computational complexity reduction; diagonal Q-correlation matrix least square algorithm; line search method; Adaptive algorithm; Adaptive filters; Computational complexity; Convergence; Filtering algorithms; Jacobian matrices; Least squares approximation; Least squares methods; Resonance light scattering; Search methods;

fLanguage

English

Publisher

ieee

Conference_Titel

Circuits and Systems, 2005. ISCAS 2005. IEEE International Symposium on

Print_ISBN

0-7803-8834-8

Type

conf

DOI

10.1109/ISCAS.2005.1465588

Filename

1465588