在單一載波分頻多工存取系統中，當循環字首的持續時間短於通道響應的持續時間時，符號間干擾與載波間干擾會降低系統效能。在先前有一個盲目波束成型法在考慮了載波頻率偏移的情況下利用以特徵值分解為基礎的架構解決這樣的問題。在這篇文章中，提出用可適性演算法找波束成型的加權向量不利用參考訊號。提出一個新的判斷空間相關性矩陣的概念利用之前的資料，並且這樣空間相關性矩陣可以利用比較少的空間向量得到。提出的演算法可以得到與之前提出的演算法競爭的表現在有不足夠循環字首、載波頻率偏移及同頻干擾的情況下。模擬結果更發現提出的可適性演算法比之前提出的演算法有更好的表現。When the duration of the cyclic prefix is shorter than that of the channel impulse response in single carrier-frequency division multiple access (SC-FDMA) systems, inter-symbol interference (ISI) and inter-carrier interference (ICI) will degrade the system performance. Previously, this problem was solved by a blind beamforming scheme with considering the effects of carrier frequency offsets (CFOs) based on the eigenanalysis in a batch mode [5]. In this thesis, adaptive algorithms are proposed to find the beamforming weight vectors without using reference signals. We propose a new conception for estimating the spatial correlation matrices that are calculated by the data of previous blocks, and we calculate the spatial correlation matrices by less spatial vectors. The proposed adaptive algorithms show competitive performance compared to the previous proposed algorithm under the insufficient cyclic prefix (CP), the effects of CFOs, and the co-channel interference scenarios. The simulation results reveal that the proposed adaptive algorithms provide better performance than the previously proposed algorithm.
