Multiuser detection of synchronous code-division multiple-access signals by perfect sampling

Code-division multiple-access (CDMA) is a multiplexing technique that shows significant advantages over analog and conventional time-division multiple access (TDMA) systems. This technology has become a driving force behind the rapidly advancing communications industry. In order to recover the transmitted signal at the receiver in a CDMA system, demodulating techniques are engaged, where a prominent role is played by the multiuser detector. In this paper, we introduce a class of novel Bayesian multiuser detectors that are constructed by employing perfect sampling algorithms: the sandwiched CFTP and the Gibbs coupler. We show that the detector based on the sandwiched CFTP can be applied to systems with negative cross-correlations, whereas the Gibbs coupler detector can be used without restrictions. A salient feature of the proposed detectors is the use of exact (perfect) samples from posterior distributions. This feature provides them with several advantages over detectors based on the Gibbs sampler. Simulation results on systems with and without near–far resistance demonstrate improved performance of the proposed detectors over some other popular detectors. In the end, we also discuss some important computational issues of the proposed detectors.