A robust adaptive FIR filter based on M-estimation

Abstract

This paper suggests a novel approach for design of robust adaptive FIR filter. From the point of view of robustness, a M-estimation regression-based is presented to improve the noise-immune capability of least square (LS) algorithm. Two sorts of algorithms for robust estimation are introduced. A robust design method of adaptive FIR filters is proposed based on the M-estimation. Finally, the results of simulation show that it is feasible and effective.