New Variable Selected Coefficients Adaptive Sparse Algorithm for Acoustic System Identification

Abstract

In communication systems as public switched telephone networks and tele-and-visionconferencing system, addressing the challenge of sparse acoustic echo is of paramount importance.The sparse impulse response identification is very essential in acoustic echo cancellation systems (AEC) exactly in sparse acoustic environments.This paper introduces an enhanced improved proportionate normalized-least-mean-square (IP-NLMS) algorithm, utilizing efficient variable step-size parameters and adapting only the active coefficients based on selection bloc for reducing the computational complexity.The proposed Variable Selection Coefficients IP-NLMS algorithm (VSC-IP-NLMS) focuses on adapting the selected active coefficients of the sparse impulse response (SIR), in order to both accuracy and convergence speed.Extensive simulations conducted under various sparse environments confirm the efficacy of the proposed algorithm.As important characteristic of this proposed VSC-IP-NLMS, it achieves these remarkable results with significantly reduced computational complexity compared to sparse and variable adaptive filtering algorithms, offering a promising avenue for improving the quality of communication systems.