Robust Muliuser DFEs Based on Hybrid MMSE and ML for Multipath Fading Channels

Abstract

In this paper, we present robust decision-feedback equalizers (DFEs) for direct-sequence code division multiple-access (DS-CDMA) systems with non-Gaussian channels. One is developed based on minimum mean-square errors (MMSE), another is derived based on maximum likelihood (ML), and the other is based on hybrid MMSE-ML. Nonlinear clippers with suitable thresholds are incorporated into the proposed DFEs to combat non-Gaussian noise. We consider a receiver consisting of a robust linear- trend (LT) channel tracker and a robust DFE following tracking. Simulations show that the robust LT channel tracker together with a robust hybrid MMSE-ML DFE can outperform the robust LT channel tracker along with other types of robust DFEs based on either the MMSE or the ML or the Kalman-based or the minimax scheme.