本文提出在感知無線電中應用降維多階威能濾波器為基礎之固定假警報機率之頻譜偵測。頻譜偵測為感知無線電中極為重要之一環，於多重路徑衰減之環境下，次要使用者利用多天線技術提高對主要使用者訊號之偵測效能；次要使用者進行偵測決策前，廣義旁瓣對消器用以抑制干擾及雜訊之影響，爾後利用自適應匹配濾波器固定假警報機率偵測法進行決策；在短暫偵測時間所得有限取樣數之條件下，降維處理法可達減少計算複雜度及可靠效能之目的。如模擬結果所示，以降維多階威能濾波器為基礎之自適應匹配濾波器固定假警報機率偵測器，於有限取樣數下比滿秩濾波器為基礎之偵測器效能較佳。In this thesis, a reduced-rank multistage Wiener filter based CFAR detector is applied for spectrum sensing in cognitive radio. Spectrum sensing is an essential component in cognitive radio. In order to enhance the performance of primary user signal detection under the multipath fading environments, a multiple antenna technique is employed at the secondary user. In addition, the general sidelode canceller (GSC) is utilized prior to the detection to suppress the effect of interference and noise. And then the AMF CFAR test is utilized for the secondary user. For the purpose of reduction of computational complexity and reliable performance, reduced-rank processing is a well-known technique under the condition of finite sample support in short sensing time. The numerical results show that the reduced-rank multistage Wiener filter based AMF CFAR detector outperforms the full rank filter-based detector in a finite number of samples.
