Turbo multiuser detection with integrated channel estimation for differentially coded asynchronous CDMA systems

Abstract

The problem of joint iterative multiuser detection and channel estimation for the uplink of differentially coded asynchronous direct-sequence code-division multiple-access (DS- CDMA) systems with periodic spreading sequences is considered. The proposed receiver consists of a first stage of channel esti- mation, soft interference cancellation and soft-in-soft-out (SISO) multiuser filtering, followed by a second stage of single-user iterative decoders. The single-user iterative decoder for each user consists of a powerful combination of a recursive systematic convolutional (RSC) decoder and a differential decoder which incorporate their soft information in an iterative fashion. In terms of channel estimation for the first iteration, the multipath channel is estimated blindly by exploiting orthogonality between the signal and noise subspaces. In the next iterations, the channel estimates are improved by using the soft estimates of the coded bits of each user provided by the single-user iterative decoders in conjunction with the output of multiuser detector. By exchanging soft information between the two stages, the receiver performance is improved through iteration. Simulation results demonstrate that the proposed Turbo multiuser receiver offers performance approaching the single-user bound. Keywords-Code-division multiple-access (CDMA) systems, Iterative channel estimation, Differential encoding, Minimum- mean-square-error (MMSE) filtering.