Robust M-estimate adaptive equalizer for impulse noise suppression

Abstract

In this paper, a FIR adaptive equalizer for impulse noise suppression is proposed. It is based on the minimization of an M-estimate objective function which has the ability to ignore or down-weight large error signal when it exceeds certain thresholds. An advantage of the proposed method is that its solution is governed by a system of linear equation, called the M-estimate normal equation. Therefore, traditional fast algorithms like the recursive least squares algorithm can be applied. Using a robust estimation of the thresholds and the recursive least square algorithm, an M-estimate RLS (M-RLS) algorithm is developed. Simulation results show that the proposed algorithm has better convergence performance than the N-RLS and MN-LMS algorithms when the input signal of the equalizer is corrupted by individually or consecutive impulse noises. It also shares the low steady state error of the traditional RLS algorithm.