Investigations on likelihood function for QRM-MLD combined with MMSE-based multipath interference canceller suitable to soft-decision turbo decoding in broadband CDMA MIMO multiplexing

This paper investigates the effects of soft multipath interference (MPI)-replica generation in a multipath interference canceller (MPIC) using two-dimensional minimum mean squared error (MMSE) and optimum likelihood function of a posteriori probability (APP) for soft-decision turbo decoding in maximum likelihood detection with QR decomposition and the M-algorithm (QRMMLD), in the previously proposed combination of a three-stage MPIC and QRM-MLD for MIMO multiplexing in broadband DS-CDMA. Simulation results show that soft-replica generation with a hyperbolic tangent function reduces the required average received signal energy per bit-to-noise power spectrum density ratio (E/sub b//N/sub 0/) by approximately 0.5 dB compared to hard-replica generation in a three-stage MPIC, and that the proposed likelihood function of the log likelihood ratio (LLR) in QRM-MLD followed by soft-decision turbo decoding, when no surviving symbol replicas remain that contain either "1" or "-1" significantly improves the block error rate (BLER) performance. Consequently, it is shown that by employing QRM-MLD combined with the three-stage MPIC using two-dimensional MMSE for 4-by-4 MIMO multiplexing, the throughput values of 200, 250, 300, and 350 Mbit/s are achieved at the average received E/sub b//N/sub 0/ of approximately 6.0, 9.0, 12.0, and 17.0 dB using QPSK with the coding rate of R = 8/9, 8PSK with R = 3/4, 8PSK with R = 8/9, and 16-QAM with R = 3/4, respectively, for broadband DS-CDMA access assuming a 40-MHz bandwidth in six-path Rayleigh fading channels.