標題: 在多對使用者下放大轉發式中繼器系統之強健分散式波束成型設計
Robust Distributed Beamforming Design in Amplify-and-forward Relay Systems with Multiple User Pairs
作者: 謝沛彣
Hsieh, Pei-Wen
陳紹基
Chen, Sau-Gee
電子工程學系 電子研究所
關鍵字: 強健分散式波束成型;實體層網路編碼;雙向中繼通道;通道不確定性;凸面最佳化;正定規劃;二次錐規劃;Robust Distributed Beamforming;Physical Layer Network Coding;Two-way Relay Channel;Channel Uncertainty;Convex Optimization;Semidefinite Programming;Second-order Cone Programming
公開日期: 2012
摘要: 分散式波束成型是一個用於多使用者中繼器系統有效的空間多工技術。分散式波束成型的最佳化需要通道資訊。但是在一般狀況下我們無法取得完美的通道資訊。所以我們需要強健設計來減少通道資訊錯誤所造成的影響。服務質量(Quality-of-service)是一個很常用的設計準則,而其目的為在服務質量的限制下把傳輸功率最小化。在這篇論文中,我們提出了用訊號對干擾加雜訊比(SINR)作為服務質量下在放大轉發式(amplify- and-forward)單向多使用者中繼器系統的強健分散式波束成型設計。我們推導出新的最糟接收訊號功率及干擾訊號功率的上下界。這些上下界比在[18]中用科西不等式推出來的上下界還要緊。這些上下界有一種二次式的結構。而這種二次式的結構可以幫助我們把本來的非凸面強健最佳化問題近似成一個正定規劃 (semidefinite programming) 問題及二次錐規劃 (second-order cone programming) 問題。模擬結果顯示我們所提出的設計效能接比前人的方法好。除此之外,我們所提出的用二次錐規劃之設計的運算複雜度會比前人的方法少。另外我們研究強健最大最小訊號對干擾加雜訊比(robust max-min SINR)的設計。強健最大最小訊號對干擾加雜訊比的設計主要是要在傳輸功率的限制下最大化訊號對干擾加雜訊比最糟的使用者的訊號對干擾加雜訊比。強健最大最小訊號對干擾加雜訊比的設計可以利用二分逼近法轉化成一連串的強健服務質量的設計問題。而這一連串的一連串的強健服務質量的設計問題可以利用我們提出的方法來解。最後,我們把我們在單向中繼器系統的設計推廣到雙向中繼器系統。我們把在單向中繼器系統提出的方法套用在有使用和沒使用實體層網路編碼的雙向中繼器系統。模擬結果顯示我們的方法再加上實體層網路編碼可以提升系統的吞吐量。
Distributed beamforming is an effective technique for spatial multiplexing in relay systems with multiple user pairs. The optimization of distributed beamforming requires the channel state information (CSI). However, perfect CSI is not available in practice. Therefore, a robust design is required to reduce the effects of CSI errors. Quality-of-service (QoS) is a popular design criterion for distributed beamforming which aims to minimize transmit power while satisfying the QoS criterion. In this thesis, we propose new approaches for signal-to-interference- plus-noise ratio (SINR)-based QoS robust distributed beamforming in amplify- and-forward (AF) one-way relay systems with multiple user pairs. We derive new bounds for the worst-case received signal power and interference power. The proposed new bounds are tighter than the bounds derived by the Cauchy-Swartz inequality in [18]. Besides, the proposed new bounds have a specific quadratic structure, which enables the originally non-convex robust optimization problem to be approximated as a convex semidefinite programming (SDP) problem or a convex second-order cone (SOCP) programming problem. The simulation results show that the proposed SDP-based method and the proposed SOCP method both outperform the existing SDP-based methods in terms of transmit power. Besides, the proposed SOCP-based method has lower computational complexity than the existing SDP-based methods. In addition, we investigate the robust max-min SINR design which optimizes the max-min SINR under the power constraint. The robust max-min SINR design problem is converted to a series of robust SINR-based QoS design problem using the bisection searching technique. The robust SINR-based QoS design problem in each bisection iteration is solved by the proposed SDP-based method or the proposed SOCP-based method. Finally, we generalize the distributed beamforming design to two-way AF relay systems. We apply the proposed methods to solve the design problems with and without physical layer network coding (PLNC). The simulation results show that our methods with PLNC improve the system throughput greatly.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT070050262
http://hdl.handle.net/11536/72815
Appears in Collections:Thesis