Robust Filtering of Affine-Projection-Like Maximum Correntropy Algorithm with Bias-Compensated

Abstract

In this paper, a robust adaptive filtering algorithm of the affine-projection-like maximum correntropy based on the bias compensation (BC) is proposed. The proposed bias-compensated affine-projection-like maximum correntropy (BC-APLMC) algorithm is derived by using the cost function based on the maximum correntropy criterion (MCC) and the BC method, which can effectively reduce the adverse effects of impulse noise and input noise on the filter weight updating. Besides, the weight updating formula of the BC-APLMC algorithm is derived. Finally, the simulation results show that the BC-APLMC algorithm is robust in the presence of input noise and impulse noise.