Projection minimization algorithm for blind channel equalizer

Abstract

This paper presents a new direct blind channel equalizer by using spatial adaptive filtering technique. Two correlation matrices with time delay parameters d and d+1 are incorporated into design to guarantee the extraction of a d-delay signal. To reduce the system computation complexity of the algorithm, a projection minimization criterion is introduced. It is shown that the optimal equalizer weights can be found by searching the least dominant eigenvector of a generalized eigenvalue problem. The proposed algorithm has benefits of extracting the signals with delay not restricted to the first or the last of a received data. This paper also develops an optimal delay determination method for selecting the delay to reach the best performance. Several simulation examples demonstrate that the proposed algorithm has fast convergence speed and is robust to the channel order selection.