Blind equalization using cumulant based MIMO inverse filter criteria for multiuser DS/CDMA systems in multipath

Abstract

Chi and Chen (see IEEE Trans. Signal Processing, July 2001) previously reported a blind equalization algorithm using cumulant based multi-input multi-output inverse filter criteria (MIMO-IFC) for mulituser DS/CDMA systems in multipath. Assuming that the user of interest is the weak user with signal power /spl epsi//sub 1/ and signal powers of all the interferers are identical, denoted /spl epsi/, the performance of Chi and Chen's algorithm is superior to that of Tsatsanis and Xu's(see IEEE Trans. Signal Processing, vol.46, no.11, p.3014-22, 1998) blind minimum variance (MV) equalizer for low near-far ratio (NFR) (=/spl epsi///spl epsi//sub 1//spl ges/1). In this paper, two blind equalization algorithms, called Algorithms 2 and 3, also using cumulant based MIMO-IFC are proposed. The former (Algorithm 2) can improve the performance of the MV equalizer. The latter (Algorithm 3) based on the former performs as well as Chi and Chen's algorithm for low NFR and outperforms Chi and Chen's algorithm and the MV equalizer for high NFR. Some simulation results are presented to support the efficacy of the proposed algorithms.