A Novel Blind Equalization Algorithm for QPSK Signals

Abstract

This paper presents a novel SVM-SCA algorithm combining support vector machines(SVM) with square contour algorithm(SCA). The proposed algorithm sets the initial weights of the blind equalizer by means of SVM and then switches to SCA algorithm which has a lower computational complexity. The simulation results demonstrate that the SVM-SCA algorithm converges much faster, gains better ISI performance and is more steady compared with the traditional blind equalization based on SCA.