Robust Maximum Correntropy Criterion Subband Adaptive Filter Algorithm for Impulsive Noise and Noisy Input
Haiquan Zhao, Dongxu Liu, Shaohui Lv
Abstract
In this brief, an improved maximum correntropy criterion subband adaptive filter (MCC-SAF) algorithm is presented, which has excellent performance for alleviating the effect of impulsive noise and noisy input. Though the MCC-SAF algorithm performs well in the face of impulsive interference due to the advantage of the correntropy-based cost function, it yields estimated error when the system input is disturbed by noise. Profiting from the property of unbiased criterion, we propose a bias-compensated MCC-SAF algorithm (BC-MCC-SAF) by introducing a biased compensated term into the MCC-SAF algorithm. Besides, the computational complexity of the BC-MCC-SAF is investigated. Simulation results for system identification under various input signals have illustrated that the presented algorithm both retains robust achievement for the impulsive interference circumstance and achieves lower stable state error and brilliant tracking ability for noisy input.