合作式放大傳遞多輸入多輸出中繼系統之線性與非線性傳收機設計

Abstract

三節點放大傳遞多輸入多輸出中繼系統之傳收機設計包含兩種通訊鏈結 - 直接鏈結(direct link)與中繼鏈結(relay link)，及兩種前置編碼器 – 來源端前置編碼(source precoder)與中繼端前置編碼(relay precoder)。在這種系統中，大部分的傳收機設計都只設計中繼前置編碼器，有些設計甚至只考量中繼鏈結以簡化最佳化過程。在本論文中，我們提出新穎的線性與非線性傳收機設計，其中來源端前置編碼與中繼端前置編碼是根據直接鏈結與中繼鏈結的通道資訊，合併最佳化設計。在本論文所探討的傳收機中，中繼端前置編碼為線性，傳送端前置編碼與接收機可為線性或非線性。具體而言，我們考量四種傳收機設計，第一種為線性來源端前置編碼、線性中繼端前置編碼與最小均方錯誤(minimum mean-squared error)接收機。第二種考量非線性來源端前置編碼、線性中繼端前置編碼與最小均方錯誤接收機之傳收設計，第三種為線性來源端前置編碼、線性中繼端前置編 碼與非線性QR 干擾消除接收機(successive-interference-cancellation)傳收設計，最後一種為線性來源端前置編碼、線性中繼端前置編碼與非線性最小均方錯誤干擾消除接收機傳收設計。在所有考量的傳收機設計中，不管是線性或非線性，困難點在於其最佳化的成本函數為來源端前置編碼與中繼端前置編碼的非線性函數，並且為非凹曲線最佳化(convex optimization)，為了克服這個設計上的困難，我們提出不同以往的前置編碼結構與設計方法，使得原本的傳收機設計可轉換為凹曲線最佳化問題，因此可推導出解析解。最後，本論文進一步探討以服務品質(quality-of-service)觀點的線性來源端前置編碼、線性中繼端前置編碼與線性最小均方錯誤接收機之傳收設計，並延伸前述所提及之設計方式，提出此問題的解析解，由模擬結果得知，相較於其他現有方法，所提出的傳收機架構有較好的效能 表現。

The transceiver design in three-node amplify-and-forward (AF) multiple-input multiple-output (MIMO) relay systems involve two links, the direct and relay links, and two precoders, the source and relay precoders. Most existing methods only consider the design with a relay precoder, and some even ignore the direct link. In this dissertation, we propose new linear/nonlinear transceiver design methods taking the direct and relay links into account, and jointly optimizing the source and relay precoders. In our designs, the relay precoder is linear, and the source precoder and the receiver can be linear or nonlinear. Specifically, four scenarios are considered. The first is the design with a linear source precoder and a liner minimum-mean-square-error (MMSE) receiver, the second a nonlinear Tomlinson-Harashima source precoder and a linear MMSE receiver, the third a linear source precoder and a nonlinear QR successive-interference-cancelation (QR-SIC)receiver, and the fourth a linear source precoder and a nonlinear MMSE-SIC receiver. All the designs, either the linear or nonlinear precoded systems, are difficult since the cost functions to be optimized are highly nonlinear functions of the source and relay precoders. Yet, the corresponding optimization problems are not convex. To overcome the difficulties, we propose new precoder architectures and methods such that the design problems can be translated into scalar-valued and convex optimization problems. And, the closed-form solutions can be obtained by the corresponding Karush-Kuhn Tucker (KKT) conditions. Finally, we consider the precoders design with quality-of-service (QoS) constraints. In the scenario, the linear precoder is used at the source and the MMSE receiver at the destination. Again, this problem is difficult and the optimization problem is not convex. We then extend the method proposed for the systems mentioned above to derive a closed-form solution. Simulation results show that the performance of the proposed transceiver design methods is significantly better than that of existing methods.