Bit-interleaved space-time trellis coding and reduced complexity soft detection in MIMO systems

It is known that a multiple-input multiple-output (MIMO) architecture provides both diversity and multiplexing gains, which makes it attractive to wireless applications aiming at increasing the link reliability or data rate.; A diversity-oriented application, the concatenation of space-time trellis coding (STTC) with an outer code through a bit interleaver (BISTTC) is investigated. BISTTC is shown to outperform the concatenation of space-time block coding (STBC) with bit-interleaved coded modulation (BICM), BICM-STBC, in diversity-rich environments with practical constraints. BISTTC is particularly more attractive than BICM-STBC in the cases with more than two transmit antennas, where BICM-STBC suffers from rate loss.; The pair-wise error probability (PEP) and error bound of BISTTC for general block fading channels are also analyzed. Some techniques, including limitation before integration, virtual trellis, and a novel enumeration method, have been employed to deal with the challenges in block fading channels. The bound is shown to be reasonably tight.; In rate-oriented applications, spatial multiplexing (SM) together with outer BICM (BICM-SM) is preferred. However, the optimal maximum likelihood (ML) detection requires high implementation complexity. A suboptimal reduced-complexity soft detection (RCSD) algorithm is proposed to conduct constrained-delay (in spatial domain) detection while relying on soft decision feedback (SDF). RCSD achieves excellent performance/complexity tradeoff compared to existing algorithms.; If channel state information (CSI) is available at both transmitter and receiver, transmitter precoding can be performed to assist detection. A linear precoding method, based on the complex Householder transform of the channel matrix into a multiple-diagonal, partial-response-like matrix, is proposed for the application of RCSD. Precoded-RCSD (PRCSD) alleviates error propagation of RCSD, and achieves full diversity as ML detection in some scenarios.