Bias Compensated Least Mean Mixed-norm Adaptive Filtering Algorithm Robust to Impulsive Noises

摘要

This paper proposes the bias-compensated normalized least mean mixed-norm (BC-LMMN) algorithm robust to the impulsive noises. This algorithm is derived from the cost function that uses mixed-norm of the l2 norm and l4 norm, where a mixing parameter is used to combine them. To reduce the bad effects of impulsive noises, two methods step-size scaler (SSS) and modified Huber function (MHF) are utilized. The SSS is derived from the modified log type cost function and is applied to the proposed algorithm to make the algorithm robust against the impulsive noises. MHF is a piece-wise linear function and effective in the impulsive noise environment. An unbiasedness criterion is adopted to eliminate the bias caused by the input noises and derive the bias compensation vector. Simulation results show that the proposed algorithm outperforms the traditional algorithms in the aspect of the convergence rate and the steady-state misalignment.