Convergence and Tracking Behavior of the New SPVS-NLMS Algorithm

In this paper, the segment proportionate variable-step-size normalized least mean square (SPVS-NLMS) algorithm is proposed. Using computer simulations, we show that the proposed SPVS-NLMS algorithm performs a faster convergence compared to the segment proportionate normalized least mean square algorithm by Hongyang and Doroslovacki (2005) with a higher tracking ability for speech signals