Adaptive Combination of Filtered-X NLMS and Affine Projection Algorithms for Active Noise Control

Abstract

AbstractThe filtered-x algorithm is the most popular technology used in active noise control (ANC) system to update the controllers. This paper proposes an adaptive combined normalized filtered-x least mean square (FxNLMS) algorithm and filtered-x affine projection (FxAP) algorithm to balance the convergence speed and noise reduction performance in ANC. The new algorithm requires only single ANC system, which uses a sigmoid function to adaptively combine FxNLMS algorithm and FxAP algorithm, and a coupling factor designed by gradient descent is used to update the filter weights. The simulation experiment results in stationary and nonstationary scenarios demonstrate the better performance of the proposed algorithm as compared with the conventional algorithms.KeywordsActive noise controlGradient descentAdaptive combination