A robust approach for online feedback path modeling in single-channel narrow-band active noise control systems using two distinct variable step size methods

Abstract

Online feedback path modeling and neutralization (FBPMN) is vital in active noise control (ANC) systems to ensure stability and elevate performance in non-stationary environments. In this paper, we propose a modified variable step size (VSS) method for online FBPMN in single-channel ANC systems which incorporates three adaptive filters; a noise control filter, a FBPMN filter, and a linear predictor for disturbance removal. In the proposed method, distinct VSS LMS algorithms are used for FBPMN filter and disturbance removal filter to achieve exceptionally fast convergence and accurate modeling of the feedback path transfer function. Simulation results show that the proposed method yields a significantly more robust ANC system by attaining a higher gain margin at a much faster rate in comparison with the existing methods.