Wavelet Vector Machines Blind Equalization Algorithm Based on Variable Segmentation Error Function

Abstract

To overcome the disadvantage of constant modulus algorithm's slow convergence and local minimum, this paper presents a wavelet vector machine blind equalization algorithm based on variable segmentation error function. This proposed algorithm uses support vector machine to optimize the initial weight vector, then, it switches to Wavelet Constant Modulus blind equalization Algorithm(WCMA) with odd symmetry variable segmentation error function. The computer simulation with underwater acoustic channel demonstrates that the proposed algorithm has fast convergence rate and small mean square error.