Mean Square Performance of a family of Adaptive Algorithms for colored noise

TL;DR

This paper analyzes the mean square performance of LMS and affine projection adaptive algorithms in the presence of white and colored noise, highlighting their convergence behaviors and computational considerations.

Contribution

It provides a comparative analysis of LMS and AP algorithms' performance under colored noise conditions, which is less explored in existing literature.

Findings

LMS convergence speed decreases with colored noise

AP algorithms improve convergence but are computationally intensive

Performance metrics vary between white and colored noise environments

Abstract

In real-time applications the characteristics and properties of a signal vary inconsistently. So, to maintain the integrity of such signals there is a need for effective adaptive filters. The conventional Least Mean Squared(LMS) algorithm is widely used because of its computational simplicity and ease of implementation. But, its convergence speed rapidly reduces when colored noise is present in the signal. Affine projection(AP) algorithms are used to speed up the convergence but have high computational costs. In this paper, the mean square performance of LMS and AP algorithms is analyzed when subject to white noise and colored noise.