A new class of gradient-based algorithm by using variable step-size technique

Phuvasitkul, Sittisak; Benjangkaprasert, Chawalit; Anantrasirichai, Noppin

doi:10.1109/ICOSP.2002.1180990

A new class of gradient-based algorithm by using variable step-size technique

Author(s)

Phuvasitkul, Sittisak

Benjangkaprasert, Chawalit

Anantrasirichai, Noppin

Date Issued

January 1, 2002

Type

Conference Paper

DOI

10.1109/ICOSP.2002.1180990

Abstract

This paper presents a new class of gradient-based algorithm for second-order IIR adaptive notch filter (ANF), the algorithm base on least mean p - Power (LMP) error criterion. The variable step-size technique is used to detect high convergence speed and high impulse noise robustness. It is proved here that the proposed algorithm has good properties in both cases, when compare with the previous one [4]. Simulation results are presented to demonstrate the effectiveness of the proposed algorithm.

Citation

International Conference on Signal Processing Proceedings ICSP, 1, 85-88, 2002