用於多輸入多輸出系統具高效率搜尋設計之複數K-Best球體解碼器

Abstract

在無線通訊系統中，多輸入多輸出(Multiple-Input Multiple-Output, MIMO)技術不需要增加額外的頻寬及傳輸功率便能提高傳輸速率及改善傳輸品質。然而，在多輸入多輸出系統中要設計出具高性能且低複雜度之接收機是一項艱難的挑戰。使用最大可能偵測法能得到最佳的效能，然而其所需的運算複雜度會隨著傳送天線個數的增加呈指數的成長。球體解碼演算法能以較低的複雜度達到與最大可能偵測法相同之效能。然而，傳統的球體解碼演算法會有資料吞吐量不穩定之問題。K-best球體解碼演算法在每一層的節點搜尋當中只保留K個最佳的候選點當作下一次搜尋的依據，因此具有穩定的資料吞吐量。然而，K-best球體解碼演算法需要取相當大的K值才能達到近似最大可能偵測法之效能。除此之外，在每一層的節點搜尋中將候選點作排序取出K個最佳的候選點會耗費大量的記憶體存取。在本論文中，吾人提出ㄧ具高效率搜尋架構之複數K-best球體解碼器。此解碼器能夠大幅降低排序時所花費的運算量。針對所提出之複數候選點搜尋方法，吾人亦設計出相對應之電路架構圖。經由分析與模擬的驗證，此解碼器僅需選取較小的K值即可達到近似最大可能偵測法之效能。

In wireless communication systems, multiple-input and multiple-output (MIMO) technology offers significant increases in data rate and link range without additional bandwidth or transmit power. However, the design of high performance and low complexity receivers for MIMO systems is a challenging task. The maximum-likelihood (ML) detection is the optimal detection scheme but its complexity grows exponentially with the number of transmit antennas. The sphere decoding algorithm (SDA) achieves the ML performance with reduced complexity. Nevertheless, the throughput of the conventional SDA is not stable. The K-best SDA which keeps only K-best candidates at each layer for the search of next layer is guaranteed to have a stable throughput. However, to achieve a near-ML performance, the value of K should be sufficiently large. Besides, applying a sorting algorithm to find K-best candidates at each layer requires a large amount of memory access. In this thesis, we propose a complex K-best sphere decoder with an efficient search architecture. The proposed K-best sphere decoder significantly reduces the sorting complexity. We also provide the hardware architecture of the proposed complex candidate search method. It is demonstrated through analysis and simulations that the proposed K-best sphere decoder achieves a near-ML performance without requiring a large value of K.