近年來隨著科技的進步，無人機系統也逐漸發展完善， 各產業也爭相將此投入各個產業以及軍事方面，因此勢必需要對 此進行研究和追蹤。此篇論文將著重以輔助波束對(Auxiliary Beam Pair) 演算法追蹤無人機與基站之間的角度，但演算法在均勻平面 天線陣列下只能得到近似角度，因此對這部分提出以前一個取樣 時間，用演算法算出的角度當作下一個取樣點的視軸(Boresight) 來降低近似誤差，也希望可以利用其他方式再降低誤差，所以會 加入擴展卡曼濾波器修正近似的問題，也使用機器學習修正輔助 波束對誤差的參數，再代回演算法求出角度，並比較這些方法在 不同環境下的效果，最後再針對結果分析各方法的優缺點。;In recent years, with the advancement of technology, unmanned aerial vehicle (UAV) systems have become increasingly sophisticated. Various industries, including military applications, have been eager to adopt UAVs. Consequently, there is a growing need for research and tracking of UAVs. This paper focuses on using the Auxiliary Beam Pair algorithm to track the angle between a UAV and a base station. However, when using a uniform planar array, the algorithm can only obtain an approximate angle. To mitigate this approximation error, we propose using the angle calculated by the algorithm at the previous sampling time as the boresight for the next sampling point. Additionally, we aim to further reduce the error by incorporating a Extended Kalman filter to correct the approximation and by using machine learning to adjust the parameters of the Auxiliary Beam Pair algorithm. The angles are then recalculated using the algorithm, and the performance of these methods under different environments is compared. Finally, the advantages and disadvantages of each method are analyzed based on the results.
