A recursive least M-estimate adaptive controller for active impulsive noise control

The filtered-x recursive square (FxRLS) algorithm has been proven to be efficient for active noise control (ANC) systems. Unfortunately, its performance deteriorates when the ANC systems are corrupted by impulsive noises. To reduce this drawback, a novel filter-x recursive least M-estimate adaptive (FxRLM) algorithm for ANC is proposed in this paper, which can overcome the adverse effect of impulsive noise on the adaptation of controller. Simulations in the context of ANC system show that the proposed FxRLM algorithm outperforms the conventional FxRLS, filtered-x least mean p-power (FxLMP) and robust FxLMS (RFxLMS) algorithms.