A Convex-Combined Step-Size-Based Normalized Modified Filtered-x Least Mean Square Algorithm for Impulsive Active Noise Control Systems

摘要

The celebrated filtered-x least mean square (FxLMS) algorithm does not work well for active noise control (ANC) of impulsive source. In previous attempts, the robustness of FxLMS algorithm has been improved by thresholding the reference and/or error signals used in the ANC system. However, estimating these thresholds is not any easy task in most of the practical scenarios. The need for appropriate thresholds is avoided in a previously proposed improved normalized-step-size FxLMS (INSS-FxLMS) algorithm, however, there is a tradeoff situation between the convergence speed and steady-state performance as a fixed step-size needs to be selected properly. In this paper, we propose a novel algorithm for impulsive ANC (IANC) systems. The proposed algorithm is based on the previously proposed INSS-FxLMS. The main idea to employ a convex-combined step-size which automatically converges to a large value to improve the convergence speed during the transient state, and to a small value as the IANC system converges at the steady-state. Extensive simulation results are presented to demonstrate the effective performance of the proposed algorithm.