Near-MAP Detectors Based on Probabilistic Data Association for Asynchronous MC-CDMA Systems

In this paper, we apply probabilistic data association (PDA) to the detection of asynchronous users in the uplink of a MC-CDMA system. It is shown that the detection can be performed by applying the PDA algorithm to two consecutive MC-CDMA symbols. Moreover, the PDA algorithm definition has been revisited in order to provide a more general and flexible implementation. Several versions of the PDA receiver have been proposed, which have been compared in terms of complexity and achievable bit error rate. The results show that the PDA approach achieves near optimal performance, while the algorithm flexibility can be exploited in order to obtain a reasonable complexity reduction.