隨著射頻元件以及天線技術的突破，低軌道衛星通訊在B5G大頻寬時代逐漸盛行，且因為可以支持高資料傳輸率及大幅改善硬體成本高昂的問題，受到人們的高度關注，被視為是未來蜂巢式網絡中非常有潛力的技術之一。本論文基於多用戶 (MU) 大規模 MIMO 低軌道衛星通訊系統設計混合波束成型，目標是最小化上行鏈路的總系統傳輸功率，同時必須確保在一定的服務質量（QoS）如每個用戶的位元/塊錯誤率和資料傳輸率的情況下。同時，我們還提出一種用戶選擇的方案，根據每個使用者在每個子通道做的奇異值分解之結果，在眾多使用者中優先挑選出通道條件較佳之使用者作訊息傳輸。在傳輸過程中，不同頻率上的信號可能會受到不同的衰減、多路徑干擾、衰減失真等影響，從而導致通道的頻率響應有所差異，所以正交分頻多工通常被用來對抗通道的頻率選擇性。除此之外，現有的研究大多僅關注混合波束成型演算法的設計，關於傳輸數據時應該如何選擇調變或編碼的問題幾乎沒有探討過。基於這個想法，對於用戶選擇配置和波束成型設計中的傳輸問題，我們設計的演算法能提供合適的調變與編碼方案。
綜上所述，我們探討了所提出的方案和其他基準方案在各種參數下的性能差異，以及提出的用戶選擇演算法與隨機選擇用戶方法的比較。模擬結果顯示，所提出的混合波束成型設計和用戶選擇方案均可以實現非常有競爭力且優於全數位設計的性能。 ;With breakthroughs in RF components and antenna technology, low earth orbit (LEO) satellite communications are gradually becoming prevalent in the era of B5G with its large bandwidth. It has garnered significant attention and is regarded as one of the highly promising technologies in future cellular networks due to its ability to support high data transmission rates and significantly alleviate the issue of high hardware costs. In this paper, a design of hybrid beamforming is proposed for a multi-user (MU) massive multiple-input multiple-output (MIMO) LEO satellite communication system. The objective is to minimize the total system transmission power in the uplink while ensuring a certain quality-of-service (QoS) such as bit/block error rate and data rate for every user. Additionally, a user selection method is introduced, prioritizing users with better channel conditions for information transmission among many users, based on the singular value decomposition results for each user in each subchannel. During transmission, signals at different frequencies may experience different fading, multipath interference, and distortion, resulting in frequency-dependent channel responses. So, orthogonal frequency-division multiplexing (OFDM) is commonly adopted to combat the frequency selectivity of the channel. Furthermore, existing research mainly focuses on the design of hybrid beamforming algorithms, while the problem of how to select modulation and coding schemes during data transmission has hardly not been addressed. In light of this, the algorithm we design provide the most suitable modulation and coding scheme for transmission in user selection configurations and beamforming designs.
In summary, we investigate the performance difference between the proposed scheme and other benchmark schemes under various parameters, and the comparison of the proposed user selection algorithm with random selection of users. Simulation results show that both the proposed hybrid beamforming design and user selection scheme can achieve very competitive and better performance than the fully-digital design.
