Hybrid Turbo Multiuser Detection for OFDM Transmission with Spatially-Correlated Channels

This letter proposes a novel hybrid turbo multi-user (MU) detection technique based on group-wise soft interference canceling minimum mean-square error filtering (SC-MMSE) combined with maximum a posteriori (MAP) signal detection for OFDM MU systems. A new user grouping algorithm is developed that exploits the knowledge of the pairwise spatial correlation among the channels of each group, of which aim is to reduce noise enhancement due to the MMSE interference suppression of highly correlated user signals. It is shown that the proposed detector can achieve significant performance improvements over SC-MMSE in the presence of high spatial channel correlation