基於空時區塊編碼(STBC-SM)的不同累積區塊中的時間排列(TP-DAB)方案以及多層區塊編碼廣義空間調變(MBC-GSM)方案皆為具高頻譜效率之空間調變技術。然而，若直接採用原始的對數最大後驗機率(Log-MAP)解碼器計算軟輸出，其計算複雜度將隨搜尋空間快速增加，使實際實現變得困難。 因此，本文分別針對TP-DAB與MBC-GSM提出低複雜度之簡化Log-MAP解碼器。對於TP-DAB，透過預先儲存小區塊最小度量值並重複利用中間計算結果降低搜尋複雜度；對於MBC-GSM，則透過預先儲存分支度量值並結合BCJR解碼器建立低複雜度軟輸出架構。透過複雜度比較，所提出之簡化Log-MAP解碼器皆能大幅降低解碼複雜度。 此外，本文以低密度奇偶校驗碼(LDPC)作為外通道碼進行模擬。然而，由於系統主要於較低訊雜比區域進行比較，因此尚未觀察到文獻中TP-DAB與MBC-GSM相較於STBC-SM的錯誤率曲線可能產生交會的區域。;Temporal Permutations in Distinct Accumulated Blocks (TP-DAB) and Multi-Level Block-Coded Generalized Spatial Modulation (MBC-GSM) are two high spectral-efficiency spatial modulation schemes. However， the conventional logarithmic maximum a posteriori (Log-MAP) algorithm requires prohibitive computational complexity for soft-output detection. This thesis proposes simplified Log-MAP algorithms for both TP-DAB and MBC-GSM. For TP-DAB， the complexity is reduced by reusing precomputed sub-block metrics. For MBC-GSM， a low-complexity soft-output framework is developed by combining precomputed branch metrics with the BCJR algorithm. Complexity analysis shows that the proposed methods significantly reduce the decoding complexity compared with the original Log-MAP algorithm. In addition， Low-Density Parity-Check (LDPC) codes are employed as the outer channel code. Simulation results are evaluated in low signal-to-noise ratio (SNR) regions; therefore， the crossover point where TP-DAB and MBC-GSM outperform STBC-SM， as reported in the literature， is not observed.
