馮世邁臺灣大學：電信工程學研究所蔡昇甫Tsai, Shen-FuShen-FuTsai2007-11-272018-07-052007-11-272018-07-052004http://ntur.lib.ntu.edu.tw//handle/246246/58781在這篇論文中，我們首先提出適用於多載波分碼多工（multi-carrier code division multiple access, MC-CDMA）系統中多用戶偵測（multiuser detection, MUD）之相對門檻削樹（relative threshold tree pruning, RTTP）演算法. 相對門檻削樹較傳統的M-演算法（M-algorithm）俱淘選力和效率，並能在高訊雜比下達到較低的錯誤率。針對二維矩形傳輸符號，我們提出用戶分離（user partitioning, UP）來和好的用戶排序（user order sorting, UOS）結合以產生沒有用戶分離時無法達到的細緻的用戶順序。這當中，我們提出的DLUPS和DLUP-RMOPS適用於下傳通道及實數展頻碼，而UP-RMOPS則適用於所有的環境，其中RMOPS是一個前人提出的優良的用戶排序演算法，它也可以是任何其他有效的用戶排序。此外，在下傳系統中當各用戶都使用一維傳輸符號時，我們提出了一個簡單的下傳相位旋轉（downlink phase rotation, DLPHR），可以把任何實數展頻碼轉換成複數展頻碼以大幅降低多用戶干擾（multiple access interference, MAI）。下傳相位旋轉可視為一種在同樣有限的資源下增加傳輸維度的方法。 當正交分頻多工（orthogonal frequency division multiplexing, OFDM）系統因為載波頻率誤差而存在載波相互干擾（inter-carrier interference, ICI）時，所有我們提出的演算法都依然適用。模擬結果顯示這些演算法能獲致較低的錯誤率，也因此降低了多用戶偵測的複雜度。In this thesis, we rst proposed the relative threshold tree pruning (RTTP) al- gorithm suitable for pre-whitening tree pruning (PWTP) multiuser detection (MUD) for multi-carrier code-division multiple access (MC-CDMA). RTTP is more e cient and selective than the conventional M-algorithm and achieves relatively lower error rate at high signal-to-noise ratio (SNR). For rectangular 2-D symbols, user partitioning (UP) is proposed and, when combined with good user order sorting (UOS) produces a softer UOS not obtainable with- out UP. The proposed DLUPS and DLUP-RMOPS are applicable with the downlink channel and real-valued spreading code, while the proposed UP- RMOPS is suitable in the general cases; ROMPS is a previously proposed UOS that can be replaced by any other good UOS. A simple downlink phase rotation (DLPHR) is proposed to convert any real-valued spreading code into complex-valued and to signi cantly reduce multiple access interference (MAI) when 1-D symbols are used with downlink channel. DLPHR can be viewed as an increase of transmit diversity with the same limited resource. All proposed algorithms apply in the presence of carrier frequency o set which causes inter-carrier interference (ICI) of orthogonal frequency division multiplexing (OFDM). Simulation results show that the proposed algorithms indeed achieve lower error rate and thus enable low complexity MUD ap- proaching optimal performance.1 Introduction 1 2 Signal Model of MC-CDMA Systems 3 2.1 Signal Model of OFDM System . . . . . . . . . . . . . . . . . 3 2.2 Block Diagram and Signal Model of MC-CDMA . . . . . . . . 8 2.2.1 Viewpoint of Single User . . . . . . . . . . . . . . . . . 8 2.2.2 Uplink Quasi-Synchronous (ULQS) MC-CDMA . . . . 9 2.2.3 Downlink Synchronous (DL) MC-CDMA . . . . . . . . 10 3 Pre-Whitening Tree Pruning Multiuser Detection with User Order Sorting and Real Field Operation 13 3.1 Pre-Whitening Tree Pruning Multiuser Detection . . . . . . . 13 3.2 Performance Enhancement By User Order Sorting (UOS) . . . 18 3.2.1 Reverse Minimum Orthogonal Power Sorting . . . . . . 19 3.3 Real Field Operation . . . . . . . . . . . . . . . . . . . . . . . 20 4 Threshold-Based Tree Pruning 25 4.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 4.2 Absolute Threshold Tree Pruning (ATTP) . . . . . . . . . . . 26 4.3 Relative Threshold Tree Pruning (RTTP) . . . . . . . . . . . 33 5 Achieving Transmit Diversity For 1-Dimensional Constella- tion in Downlink Channel 41 5.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 5.2 Downlink Phase Rotation (DLPHR) . . . . . . . . . . . . . . . 42 5.3 Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 5.4 Optimal Rotation in Special Case . . . . . . . . . . . . . . . . 45 5.5 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 6 User Partitioning 55 6.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 6.2 Preliminary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 6.3 User Partitioning with Reverse Minimum Orthogonal Power Sorting (UP-RMOPS) . . . . . . . . . . . . . . . . . . . . . . 59 6.4 Downlink User Partitioning Sorting (DLUPS) . . . . . . . . . 62 6.5 Downlink User Partitioning With Reverse Minimum Orthog- onal Power Sorting (DLUP-RMOPS) . . . . . . . . . . . . . . 71 7 Simulation Results 75 7.1 Signal Model and Parameters . . . . . . . . . . . . . . . . . . 75 7.2 Threshold-Base Tree Pruning . . . . . . . . . . . . . . . . . . 76 7.2.1 Absolute Threshold Tree Pruning (ATTP) . . . . . . . 76 7.2.2 Relative Threshold Tree Pruning (RTTP) . . . . . . . 80 7.3 Downlink Phase Rotation (DLPHR) . . . . . . . . . . . . . . . 85 7.4 User Partition (UP) . . . . . . . . . . . . . . . . . . . . . . . . 88 7.4.1 User Partition with Reverse Minimum Orthogonal Power Sorting (UP-RMOPS) . . . . . . . . . . . . . . . . . . 88 7.4.2 Downlink User Partition Sorting (DLUPS) and Down- link User Partition With Reverse Minimum Orthogo- nal Power Sorting (DLUP-RMOPS) . . . . . . . . . . . 91 7.5 Joint Frequency O set Compensation and Multiuser Detection 95 8 Conclusions and Future Works 101 8.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 8.2 Future Works . . . . . . . . . . . . . . . . . . . . . . . . . . . 102758400 bytesapplication/pdfen-US削樹用戶分離多用戶偵測tree pruningmultiuser detectionuser partitioningMC-CDMAthesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/58781/1/ntu-93-R91942018-1.pdf