以虛擬基地台為輔助的無線位置估測

Abstract

隨著商業應用的潮流，無線位置估測的技術在近年來被廣泛的研究著。很多種形式的無線訊號都可以作為位置估測之用。在這篇論文之中，由傳播時間轉換過來的量測距離首先被線性化，以便利用最小平方 (Least Square) 估計的方法。論文裡考量兩個實務上的問題，分別是〝非直接路徑 (NLOS) 誤差〞與〝幾何衰減 (GDOP) 效應〞。在量測時間資訊時，非直接路徑誤差會引發ㄧ個正值而且很大的偏差。另ㄧ方面，基地台的幾何分佈不佳時，幾何衰減效應的值會增加，進而降低定位的準確度。這篇論文所提出的〝虛擬基地台 (VBS) 輔助演算法〞藉由加入一些幾何限制以降低非直接路徑誤差的影響；也藉由加入一些虛擬基地台來減少幾何衰減效應。所提出的虛擬基地台輔助演算法在模擬時與一些現有的定位演算法相比較。模擬結果顯示虛擬基地台輔助演算法的表現較好，特別是在非直接路徑誤差很大與基地台幾何位置不佳時的錯誤率表現更佳。

The wireless location estimation techniques have been wildly investigated with the trend of commercial applications in recent years. Various types of radio signals can be used to develop the location estimation algorithms. In this thesis, the range measurements transferred from the received time-based information are first linearized to utilize the Least Square (LS) method. The practical issues such as the Non-Line-of-Sight (NLOS) errors and the Geometric-Dilution-of- Precision (GDOP) effect are concerned. The NLOS errors will cause a large positive bias while measuring the time information. On the other hand, the poor geometric layout will raise the GDOP value and reduce the accuracy of a location algorithm. The proposed VBS (Virtual Base Stations) algorithm mitigates the influence of the NLOS errors by adding some geometric constraints and reduces the GDOP effect by joining the assisted virtual base stations. The proposed VBS algorithm is compared with several existing location algorithms via simulations. The performance is comparably better than other methods, especially in the environments with large NLOS errors and poor geometric layout.