| 摘要: | 本計畫將研究應用於分散式寬頻感知無線電的寬頻頻譜壓縮感測技術，藉由克服下面兩個問題改善頻譜感知能力與傳輸利用機會。首先寬頻無線通訊發展提供主用戶高資料傳輸服務，亦提供感知用戶更多動態接取頻譜機會，然而在滿足奈奎斯特取樣率下，寬頻頻譜感知需使用高速類比數位轉換器與高速數位信號處理方能達到好效能，複雜度也因此增加；再者無線通道效應造成隱藏節點問題是頻譜感知效能差的原因。我們將使用壓縮感測與合作感知方式，以寬帶中的頻譜稀疏特性與空間中的分集特性設計可靠且具低複雜度的頻譜感知技術。在第一年計畫我們研究寬頻頻譜壓縮感知技術，包括：具抗雜訊能力、適應性、強健型等感知技術，以促進在實際通訊環境上應用可行性。在第二年計畫我們針對分散式寬頻感知無線電，以聯合稀疏特性模型與貝式壓縮感測因子圖等方法研究單跳躍合作頻譜感知技術，同時分析分散式合作感知無線電效能。計畫預期成果為聯合稀疏感知模型、因子圖、數種寬頻感知技術、基於聯合稀疏模型或貝式信息傳遞之單跳躍分散式寬頻感知技術、分散式區域合作感知無線電效能分析等，我們將藉由理論分析與電腦模擬來評估所提方法的系統效能。 The goal of this project will be focused on designing distributed wideband spectrum sensing schemes, for which the recently developed compressive sensing techniques are applied in wideband interweaving cognitive radio (CR) networks. The overall objective is to improve the spectrum sensing capability and transmission opportunity exploitation by dealing with the following two critical challenges. On one hand, the growth of heterogeneous broadband wireless communication not only offers high data rate service for primary users, but also provides opportunistic dynamic spectrum access for CR users. The wideband spectrum sensing, however, imposes impractical burden on analogy-to-digital converter and digital signal processing with Nyquist-rate sampling. On the other hand, the hidden terminal problem, which stems from several wireless fading effects, constitutes a major performance degradation factor for monitoring the vacant spectrum hole. To provide reliable spectrum sensing at affordable complexity, in this project, we will resort to compressive sensing and cooperative sensing for wideband CR networks by utilizing the inherent sparsity and diversity nature in spectral and spatial domains, respectively. The first half of this two-year effort will be concentrated on designing several wideband compressive spectrum sensing techniques, including anti-noise schemes, adaptive schemes for sensing matrices, and robust schemes for quantized sensing matrices, to facilitate the future applications in practical communication environments. The focus of the second-year effort is to investigate one-hop cooperative spectrum sensing techniques for distributed wideband CR networks. Two potential methodologies, joint sparsity models and Bayesian compressive sensing with factor graphs, will be utilized to design the cooperative wideband spectrum sensing schemes. Meanwhile, we will analyze the performance of the distributed cooperative CR networks and the required cooperative radius to achieve sufficiently good performance. The outcomes of this project will be the basic cognitive system models, joint sparsity cognitive models, factor graphs for distributed CR networks, and several wideband spectrum sensing schemes, which are based on compressive sensing, with anti-noise, adaptive and robust capabilities. Furthermore, two essential distributed wideband spectrum sensing schemes, derived from joint sparsity models and Bayesian probabilities, will be provided, along with a theoretical analysis on the performance of the distributed CR networks. Computer simulation and theoretical analysis will be conducted to evaluate the overall system performance. 研究期間：10008 ~ 10107 |
