第五代通訊系統(5G)擁有巨大的潛力可實現低功耗、覆蓋範圍廣、低延遲、大頻寬和可靠傳輸。 但其容易受到過大的路徑損耗、延遲擴展與同通道干擾(Co-Channel Interference, CCI)等問題。 因此通過到達角度(Direction of Angle, DoA)估測與波束成型(beamforming)可以有效提高陣列天線接收之訊雜比， 並且得到較佳的通訊品質進而解決以上問題。其中又以MUSIC演算法為到達角估測的經典方法之一， 本文利用此方法為基礎，進行改良即延伸，並結合OFDM設計接收機。 本論文提出在正交分頻多工(Orthogonal Frequency Division Multiplexing, OFDM) 系統接收端使用均勻平面陣列天線(Uniform Planar Array, UPA)為基礎， 在時域上藉由到達角度的估測來當作無跡卡爾曼濾波的DOA初始預測。 我們假設在目標物為移動訊號源的情況下所產生的角度變化量為基礎， 提出無跡卡爾曼濾波器適應性演算法在每次傳輸進行高效且精確的角度追蹤， 以提高OFDM系統中容量(capacity)，並利用模擬結果進行性能分析與討論。 ;Nowadays, the ever-growing demand for mobile communications is constantly increasing the need for improved capacity, better coverage, and higher-quality service. Three major disabilities limit the capacity and reliability of wireless communication systems including multipath fading, delay spread, and co-channel interference. With the high speed of communication in 5G, Direction-of-Arrival (DOA) estimation and fast beamforming techniques need to be adopted. The MUltiple SIgnal Classification (MUSIC) algorithm has obvious advantage in high resolution signal source estimation scenarios. The training time required to form and steer the main lobes toward 5G users must be short. In this thesis we evaluate an OFDM receiver with Uniform Planar Array (UPA). This work proposes tracking the Direction of Arrival (DOA) through the Unscented Kalman Filter (UKF) algorithm based on a motion model governing the moving source. DOA of the moving source is estimated using MUltiple SIgnal Classification (MUSIC) and later the estimated DOA is used as an initial value and provided to the UKF algorithm to track the moving source. Computer simulation is used to evaluate the performance of this work with MATLAB. Details drawing the process of the proposed scheme are presented in this thesis.
