Filtered-x RLS algorithm based active noise control of impulsive noise

This paper presents active noise control (ANC) of impulsive noise modeled by non-Gaussian stable processes. The famous filtered-x least mean square (FxLMS) algorithm for ANC systems becomes unstable for impulsive noise as it is based on minimization of variance of error signal. As the family of RLS algorithms give better convergence than that of LMS algorithm, a new solution employing Filtered-x recursive least squares (FxRLS) algorithm is proposed for ANC of impulsive noise. Simulation results illustrate that the proposed solution has faster convergence and better stability than the existing techniques, for highly impulsive environments.