最近,祁忠勇教授與彭俊賢博士等人提出了一個於後置快速傅立葉轉換(post-FFT)下基於子載波平均的多級盲波束成型演算法(MSBFA)應用於正交分頻多工(OFDM)系統。而且也已經知道,多級盲波束成型演算法可有效的工作於存在空間相關路徑(Spatially Correlated Paths)與通道衰落(channel fading)的環境。然而,此演算法在檢測與估計時是對接收的路徑信號採逐路徑信號分離,因此會有較長的時間延遲。有鑑於此,在本篇論文中我們提出了逐用戶盲波束成型演算法(UUBFA);相對比於多級盲波束成型演算法,逐用戶盲波束成型演算法在結構上採用平行處理的想法,使我們可以在一級之中,從接收的路徑信號中分離出同一用戶的所有路徑信號。這不僅使檢測與估計的時間延遲縮短,更適用於軟硬體的實作;論文第六章透過電腦模擬可以發現逐用戶盲波束成型演算法除了擁有以上優點外還保有了多級盲波束成型演算法優良的性能,如:在一個符號區間(Symbol block)內完成所有的解調;性能接近非盲的最小化方均誤差波束成型器(Nonblind MMSE beamformer)。另外在論文末段,針對環境中存在時間相關路徑(Temporally Correlated Paths),使用了快速峰度最大化演算法(FKMA)進行補償,並提出了渦輪峰度最大化波束成型器(TKMBF)於多級盲波束成型演算法可以有效對抗嚴苛的環境(不論是在環境中是否存在時間相關路徑信號或是空間路徑信號或是兩者同時存在的情況)。
Recently, Peng and Chi et al. proposed a block-by-block blind post-FFT multistage beamforming algorithm (MSBFA) based on subcarrier averaging for a multiuser orthogonal frequency division multiplexing (OFDM) system, which works well under multipath block fading channels even in the presence of correlated paths. However, the MSBFA is a sequential path-by-path detection algorithm, thereby resulting in long detection latency as the total number of paths of all the active users is large. In this paper, a blind multistage user-by-user beamforming algorithm (UUBFA) is proposed, which, in contrast to the MSBFA, detects all the path signals for one individual user and meanwhile achieves the maximum diversity gain through a blind maximum ratio combining processing at each stage. In addition to achieving the same performance as the MSBFA, the parallel signal processing structure of the proposed UUBFA is well suited to software and hardware implementations. At the end of thesis, we used FKMA [7] to find the frequency domain equalizer. Finally we proposed the MSBFA (TKMBF) which perfoms well in the presence of both temporally and/or spatially correlated paths.