多天線非正交多工接取之低複雜度排程演算法研究

Abstract

在本篇著作中，我們將半正交使用者選擇(SUS)排程演算法實作在多天線非正交多工多重接取(NOMA-MIMO)系統中，並考慮迫零預編碼(Zero Forcing Beamforming)和假設傳送端完美(Perfect CIST)的知道整個通道的狀況。而多天線非正交多工的基本概念為傳送端將多個使用者的訊號於一個束波中重疊，並在接收端使以疊代干擾消除(SIC)技術正確的資料解調。而在多天線正交多工系統中，由於不同束波之間存在干擾，因此束波間的相關性是個相當重要的議題。此外，在零迫預編碼的情境中，雖然可藉由向量投影消除使用者干擾，若將兩個非正交的使用者排程在一起，仍會因投影而造成有效通道增益的降低。因此，半正交使用者排程演算法會考慮束波相關性之下排程使用者，並且只需要相當低的運算複雜度。就公平性而言，我們採用比例公平調度的模型作為最佳化的目標函數來評估系統效能。我們首先在多天線正交多工系統中實作半正交使用者選擇排程，並且提出傳輸率限制方法來改善半正交使用者排程考慮公平性模式。而後，我 們將半正交使用者排程考慮公平性模式實作在多天線非正交多工接取系統中。然而，原先的半正交使用者排程考慮公平性模式在此系統中的效能增益並不佳，因此我們提出三個使用者選擇方法來改善演算法的排程效能。

這篇著作主要的貢獻為將低複雜度演算法實作在作在多天線非正交多工多重接取系統中，並考慮公平性進行只用者排程的最佳化。

In this work, we investigate the non-orthogonal multiple access multiple-input multiple-output (NOMA-MIMO) based on the semi-orthogonal selection (SUS) scheduling scheme with zero-forcing (ZF) beamforming in the assumption of perfect CSIT. The basic concept of NOMA-MIMO is to superpose multiple user signals within a beam at the transmitter and apply successive interference cancellation (SIC) at the user terminal. The correlation among beams is a critical issue in the orthogonal multiple access (OMA) system due to the loose constraint of the correlation would cause the serious inter-beam interference. Besides, in the ZF beamforming scenario, selecting a non-orthogonal user degrades the e↵ective channel gains. Therefore, the SUS is a scheduling scheme which selects the user set considering the correlation among beams with low computation complexity. In terms of fairness and sum rate, the proportional fair (PF) scheduling model is adopted to evaluate system performance. We first implement the SUS-MS and SUS-PF algorithm into the OMA-MIMO system, and the rate constraint method are proposed to solve the performance decreasing phenomenon as the orthogonality factor increasing especially in SUS-PF. After that, the SUS-PF are implemented into the NOMA-MIMO system. However, the original SUS has poor performance in NOMA-MIMO system. Therefore, three user selection metric are proposed to improve this problem. The main contribution of this work is the improvement of the SUS-PF method in both OMA-MIMO system and the NOMA-MIMO system under the objective to maximize the proportional fairness(PF). Simulation results show that the proposed method can reduce the computational complexity without lost too much performance compared to the optimal solution.