ANC system with random noise injection for online feedback path modeling

Abstract

In active noise control (ANC) systems, an acoustic feedback signal may degrade the ANC performance and/or cause system instability. To solve such a problem, various online feedback path modeling (OFPM) methods using additive random noise have been reported. However, the additive random noise signal may contribute to the residual output noise. In this paper, a new random noise injection control is proposed for ANC systems with OFPM. For that purpose, the additive random noise being generated is scheduled to be stopped when the output error power of an OFPM filter is lower than the estimated steady-state mean square error (MSE) of the filter. Furthermore, a variable step-size normalized mean square (VSS-NLMS) algorithm is employed to achieve faster convergence of the OFPM filter. Finally, simulation results are provided to demonstrate the effectiveness of the proposed approach.