Variable Step-size LMS Adaptive Algorithm Controlled by Summation of Squared Gradient Accumulation

Abstract

This paper presents a variable step-size least mean square (LMS) algorithm for adapting an adaptive finite impulse response (FIR) filter. The proposed algorithm performs to adjust the step-size parameter based on summation function of squared gradient accumulation. The algorithm is intended to produce high convergent rate in transient state and low misadjustment in steady state. The simulation results have shown that the proposed algorithm can provide the performance for such intention. Furthermore, it is very well robust in noise signal compared with other algorithms.