Bias-Compensated MCCC Algorithm for Widely Linear Adaptive Filtering With Noisy Data

Abstract

Recently, widely linear adaptive filtering with noisy input and output has attracted much attention. As an unbiased estimation method, the widely linear complex-valued estimated input least mean square (WLC-EILMS) algorithm has shown its superiority in the noisy inputs and outputs. However, WLC-EILMS is not robust to impulsive noises. To overcome this defect, this brief uses the maximum complex correntropy criterion (MCCC) to generate a novel widely linear complex-valued estimated-input MCCC (WLC-EIMCCC) algorithm. The proposed WLC-EIMCCC algorithm exhibits superior performance especially in dealing with impulsive noises, and its performance analysis is also given for theoretical analysis. Simulations on system identification and channel equalization validate the correctness of the theoretical steady-state excess mean square error (EMSE) and the superior performance of WLC- EIMCCC.