A Novel Low-Complexity MAP SISO Detector for Iterative Detection in Precoded OFDM Systems

In this paper, we propose a novel low-complexity MAP soft-in-soft-out (SISO) detector used for iterative detection in precoded OFDM systems. First, we decompose the precoder into serially concatenated sub-precoders separated by interleavers. Based on this decomposed precoder, our proposed low-complexity MAP SISO detector is accomplished by concatenated SISO sub-detectors separated by deinterleavers or interleavers. Three evaluating algorithms of SISO sub-detectors are developed. The first one does not use approximation. The other two use max-log and Gaussian approximation. Based on the Gaussian approximation algorithm, complexity of the proposed MAP SISO detector with a precoder of size NtimesN can be as low as O(Nldrlog₂N), much lower than that for conventional SISO detectors. Simulation results show that the performance of iterative detection using the proposed MAP SISO detector is almost near the optimal performance for both uncoded and coded systems. EXIT analysis is also presented to predict the BER performance of the iterative detection using our proposed low complexity MAP SISO detector.