於全雙工無線系統之傳輸波束成型與功率控制聯合設計

Abstract

在蜂巢系統中，頻譜分工被採用於避免上行與下行之間的干擾。然而下行對頻譜的需求遠遠大於上行。為了緩減下行流量的堵塞狀況並充分利用上行頻帶，本論文提出將上行頻帶提供下行傳輸用。故於上行頻帶成為一個全雙工系統。同時，為保證上行的訊雜比，每個用戶上行的通訊品質均需要被保證。根據此情境，吾人可建立將此情境建立一個最佳化問題：最大化下行吞吐量、受限於上行的通訊品質受到保證與各種功率的限制。因著此最佳化問題並不是凹向上，吾人可將此問題經過適當地轉換與化簡成為凹向上。再者，利用專門解最佳化問題的軟體CVX，求出此最佳化問題的數值解。從數值解的模擬結果，可得知：全雙工的表現並不是在所有條件狀況皆勝過半雙工。唯有當下行用戶半徑小、傳輸天線多，全雙工的表現方能勝過半雙工。最後，吾人根據雙重分解，提出此最佳化問題的演算法，並附上其收斂結果。

In a cellular system, the technique of Frequency-division duplexing (FDD) is adopted to avoid interferences between uplink signals and downlink signals. The downlink spectrum is full of traffic, while the uplink band is not. To alleviate the traffic on the downlink, the spectrum of uplink can be utilized for downlink transmission, i.e. the downlink signals are delivered both in the uplink and downlink spectrum. In the uplink band, the downlink and uplink signals exist simultaneously, which constitutes a full-duplex system. The goal is to maximize the downlink throughputs. However, the interferences between uplink and downlink transmission occurs, and the uplink rate degrades. Therefore, QoS constraints for uplink are imposed to guarantee the SINR of uplink users. A downlink throughput maximization problem subject to some uplink QoS constraints is formulated. Since the formulated problem is too complicated, some simplifications are applied to the problem, including fixing the uplink receive beamforming and downlink received power. The weights are added to each downlink user according to their priorities. The Karush-Kuhn-Tucker condition and the feasibility analysis of the problem are also provided. At first, the optimization is solved by software CVX numerically. From the numeric results, it is inferred that the full-duplex outperforms the half-duplex only when the downlink cell radius is small and the number of transmit antennas is big. Last but not least, an iterative algorithm based on the dual-decomposition method is proposed to solve the problem.