協力式感知無線網路之研究---子計畫五：協力式感知無線網路之頻譜偵測及接收機設計

Abstract

結合用戶間協力合作之感知無線網路乃目前無線通訊領域中一個非常重要的研究主 題，並且已被公認為一實現下世代高效能無線通訊系統的關鍵技術。而為了達成協力式 感知無線網路所能提供的各種優點，諸如允許主要及次要用戶並存以提升頻譜使用效 益，快速而可靠的基頻信號處理演算法將扮演一個不可或缺的關鍵角色。本子計畫將從 信號處理的觀點出發，針對協力式感知無線網路的兩個重要基本問題-頻譜偵測以及接 收機設計-做一系列深入的研究。首先，我們將針對具正交分頻多工(Orthogonal Frequency Division Multiplexing)信號傳輸實體層機制的感知無線網路系統發展出一套 新式快速頻譜偵測法則，有別於現有的逐一子載波信號能量偵測方式，我們將利用稀疏 信號回復（Sparse Signal Recovery）理論來做共同頻帶檢測以減少整個頻譜偵測時程。 其次，我們將探討如何藉由感知用戶間的協力合作來提升頻譜偵測的準確度，此一研究 重點在於發展出新式協力法則以對抗感知用戶間嚴重通道衰減，而進一步增加頻譜資訊 交換的準確性和整體頻譜偵測結果的可靠度。最後，我們將考慮消除同頻帶用戶干擾的 接收機設計問題，信號接收將藉由能夠同時抑制多用戶干擾及通道估計誤差的強健式波 束形成機制，我們將利用矩陣擾動理論（Matrix Perturbation Theory）來分析通道估計誤 差下的信號及干擾的統計特性進而推導出強健式波束形成器的解析公式。

Cognitive radio wireless networks combined with the user cooperation mechanism is an active yet imporatant research topic in wireless communications, and has been widely acknowledged as a key technique toward high performance next-generation wireless systems. To realize the various advantages of cooperative cognitve radio networks, such as the coexistence of primary and secondary users for enhancing the oveall spectrum efficiency, fast and reliable baseband signal processing algorithms are rather cruical. From a signal processing perspective, this subproject aims to investigate two fundamanetal problems in cooperative cognitive radio networks, namely, spectrum sensing and receiver designs. We consider the wideband scenario with OFDM based transceivers. Firstly, by formulating multi-band spectrum sensing as a sparse signal recovery (SSR) problem we will develop a new fast sensing algorithm based on the SSR theory. The proposed approach enables joint tone signal recovery for spectral hole identification, and can be more efficient (in terms of reduced detection time) than existing energy-detection based tone-by-tone search schemes. Secondly, we will exploit user cooperation for enhancing the spectrum sensing performance. The focus is on devising new cooperative mechanisms to combat severe channel fading among cooperative users. This will guarantee reliabile reception of the exchanged spectrum information, and in turn improve the overall detection performance. Finally, we will address the problem of robust receiver design. Specifically, the in-band interference cuased by spectrum sensing error will be suppressed via receive beamforming. The design of beamforming vector will also take into account the mitigation of channel estimation errors. We will leverage the matrix perturbation theory to derive a closed-form solution of the proposed robust beamformer.