林茂昭臺灣大學：電信工程學研究所賴建汀LAI, CHIEN-TINGCHIEN-TINGLAI2007-11-272018-07-052007-11-272018-07-052006http://ntur.lib.ntu.edu.tw//handle/246246/58774隨著無線通訊頻寬需求的快速增加，因為跟單輸入單輸出(SISO)系統比較之下，多輸出多輸入(MIMO)系統能夠提供更高的傳輸容量，所以受到愈來愈多的注視。 Turbo principle 已經成功的應用在多輸入輸出系統，Turbo BLAST系統是其中之一。在Turbo BLAST系統，我們需要一個demapper負責對每個編碼後的位元計算事後機率（posteriori-probabilities）。採用暴力解法所導致的高解碼複雜度將使得這個系統變得不切實際。 LISS(LiSt-Sequential) demapper能夠以較低的解碼複雜度，以接收到的事前機率計算每個位元的事後機率。解碼複雜度的降低是因為LISS demapper在事後的機率的計算上應用了循序解碼(sequential decoding)的技巧。 我們考慮對接收到的信號做樹狀搜尋時的先後順序，並對原本的LISS 演算法做了一些修改。模擬結果顯示，修改後的LISS演算法的效能確實比原本的來得更好。With the rapid increasing demand for wireless communication bandwidth, Multiple Input Multiple Output (MIMO) systems draw much attention because it’s able to provide higher capacity than single input single output (SISO) system. Turbo-processing principle has been successfully applied to MIMO systems, one of them is called Turbo BLAST. We need a BLAST dempapper delivering a-posteriori probabilities (APP) about the coded bits. Full-APP detection would lead to prohibitive complexity. LISS(LiSt-Sequential ) demapper accepts a priori probability and produce posteriori-information with reduced complexity. The decoding complexity reduction of Turbo BLAST can be achieved by applying sequential decoding technique on calculation of posteriori probability of demapper. We made modification on the LISS algorithm by considering the ordering of the tree search in the LISS. Simulation results show that the performance of LISS with proper ordering will be better than that of LISS without proper ordering.1 Introduction 1 2 Fundamental 5 2.1 MIMO Channel Model………………………………5 2.2 Decoding Algorithms for BLAST…………………………7 2.2.1 Maximum Likelihood Decoding (MLD)…………………7 2.2.2 The Zero Forcing Algorithm………………………..7 2.2.3 Zero Forcing with QR Decomposition…………………8 2.2.4 The Minimum Mean Square Error (MMSE) Solution……9 2.2.5 Decision Feedback…………………………………10 2.2.6 Simulation Results……………………………………12 2.3 Turbo Decoding……………………………………………13 2.3.1 The MAP Decoding Algorithm…………………………13 2.3.2 The Iterative Decoding Algorithm…………………19 2.4 Sequential Decoding - The ZJ (Stack) Algorithm…24 3 Turbo BLAST 27 3.1 Introduction………………………………………27 3.2 System Model of Turbo BLAST………………………28 3.3 Soft-input/Soft-output Maximum Likelihood Demapper…30 3.4 Simulation Results………………………33 4 LISS Demapper for Iterative Detection…37 4.1 Introduction……………………………………37 4.2 Derivation of APP L-value with a priori information…38 4.3 Rewrite the a posterior probability as an additive metric…41 4.4 Apply stack algorithm on calculating APP L-values…42 4.5 LISS Demapper with Detection Ordering…………………47 5 Simulation Results 49 5.1 (NT, NR) = (4, 4) with QPSK modulation………………50 5.2 (NT, NR) = (6, 6) with QPSK modulation…………52 5.3 (NT, NR) = (4, 4) with 16 QAM modulation…………54 6 Conclusions 59 Reference 61598289 bytesapplication/pdfen-US多重收發天線系統MIMOTurbo MIMOTurbo BLASTLISSthesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/58774/1/ntu-95-R91942096-1.pdf