New Variable Step-Size NLMS Algorithm for Adaptive Noise Cancellation

Abstract

Numerous Variable Step-Size Normalized Least Mean Square (VSS-NLMS) algorithms have been proposed to solve the problems of fast convergence rate and low value of steady-state misadjustment in the past three decades. In this paper we propose a new VSS-NLMS algorithm that employs the estimated power ratio of the adaptive filter to control the step-size update. The algorithm has a large step size in the initial stages to speed up the convergence rate, and then the step size is adjusted to maintain a low value of steady-state misadjustment. The performance of the proposed algorithm is evaluated for Adaptive Noise Canceller (ANC) using different speech signal with stationary and non-stationary noise added to it. The computer simulation highlights that the proposed algorithm has very fast convergence rate while maintaining a very low value of steady-state misadjustment.