Adaptive Blind Equalization with Variable Step-Size Modified by Attenuation Function

Abstract

Step-size is one of the important parameters which influence the performance of blind equalization, a variable step-size method control by attenuation function according to instantaneous gradient and iterative times is proposed in this paper. Initially, set larger step-size to obtain faster convergence, with iterative times increasing, step-size would decrease control by attenuation function for instantaneous gradient reducing. Step-size changes according to this method consistent with the demand variable step-size adaptive algorithm that can obtain combination of convergence rate and convergence precision to improve the performance of blind equalization. Simulations show that this method performs better than conventional blind equalization with constant step-size.