具幾何輔助之無線網路定位與追蹤技術

Abstract

在過去的幾十年內，定位和追蹤技術已經成為許多通訊系統或應用的重要輔助功能。隨著位基服務在實際應用上的價值提升，定位演算法在諸如大型蜂巢，微型蜂巢與感測網路等無線網路的發展得到很大的鼓舞。當通訊系統一代一代更迭的同時，定位追蹤技術的精進以達到更好的精準度仍是需要的。本論文的研究目的即是在不同的限制因素和無線網路環境中，提升定位追蹤的效能。 無線訊號的傳播路徑對於訊號到達接收端的時間或訊號強度有很大的影響。因此對定位演算法而言，在多重路徑環境中的混合式直線路徑和非直線路徑是一個影響精準度的重要因素。在進行定位之前，距離的估測是一個前置的工作。為了降低用戶端的製造成本，接收訊號強度是被用來估測距離的資訊。在給定信賴水準和誤差限制的需求下，統計推論式距離估測 (Statistical inference distance estimation, SIDE) 演算法可以推論出理論上最少的取樣數，並且使距離的估測結果趨近於正確值。另一方面，粒子式距離估測 (Particle-based distance estimation, PDE) 則是考量在LTE-A網路中，在混和式直線路徑和非直線路徑下利用接收端訊號強度來估測距離。基於粒子分佈所帶來的多樣性，PDE演算法對於訊號衰減模型的誤差具有顯著的抵抗力。這些距離估測的結果將被應用於不同無線網路中的定位與追蹤演算法。 為了探討幾何效應在不同無線網路下對於定位問題的影響，基地台和用戶之間的位置關係是一個重要考量因素。不良的幾何位置關係會造成較大的誤差區域，並且使定位精準度變差。由接收端訊號強度所估測的距離資訊，可以進一步得到距離差的量測；這兩種量測類型分別被運用在大型蜂巢和微型蜂巢網路中。不同的量測類型在空間中會有不同形狀的交集，而這些交集即是用來規範用戶位置估測的幾何限制。距離差量測在空間中具有雙曲線的特性，幾何改良式定位 (Geometric improved location estimation, GILE) 定位演算法即是藉此產生一多邊形的幾何限制區域來規範用戶的位置估測。另一方面，在虛擬基地台定位演算法(Location estimation algorithms with the virtual base stations, VBS) 中則是探討具有圓形特性的距離量測來產生幾何上的多邊弧形限制區域。這些幾何限制被整合到傳統的二階式最小平方法中以對抗不良的幾何位置關係和訊號不足的問題。 幾何輔助的概念也運用在有限電力的無線感測網路中去追蹤用戶的移動，以期使用最少的感測節點並且延長整個感測網路的生命期。在排程式省電追蹤 (Scheduled power-saving tracking, SPT) 演算法中，利用追蹤用戶信標發送的時刻，在一個用戶與網路不同步時間的情況之下，先告知特定的感測結點，以卡爾曼濾波器來追蹤用戶的移動軌跡。此外，幾何輔助式省電追蹤 (Geometry-assisted power-saving tracking, GPT) 演算法根據最小化幾何稀疏精度，只使用三個感測節點來追蹤用戶並且更進一步達到省電的目的。 總而言之，諸如混合式路徑和幾何效應等實際的環境因素，在配合不同距離量測和無線網路的考量之下，本論文提出相對應的定位追蹤演算法，以期增進位置估測和追蹤的精準度。

Acknowledgement

Foremost, I would like to express my sincere gratitude to my advisor Prof. Kai-Ten Feng for his continuous guidance and supports to my Ph.D. study and research. Undoubtedly, Dr. Feng always shows patience and enthusiasm with immense knowledge and wise insights to help me through the cultivation of my Ph.D. program. It gains rich harvest to me by giving reasonable explanations to elaborate on my motivations and defend my work when facing the argumentative inquiries from Dr. Feng. The most valuable guidance that Dr. Feng gives to me is to discover problems, to clarify ideas, to propose solutions, and finally to express with logical statements. This will be an important asset in my lifetime. Special thanks to the committee of my dissertation, including Prof. Yu-Ted Su, Prof Kuo-Chen Wang, Prof Ching-Yao Huang, Prof. Yuh-Ren Tsai and Prof. Ding-Bing Lin for their criticisms, insightful comments and encouragements. In addition, I want to thank the professors from whom I get theoretic knowledge and practical experiences for the advanced researches in the future. I want to thank many lab mates in Mobile Intelligent Network Technology laboratory. As a post doctor for past three years, Dr. Tain-Sao Chang is always willing to give valuable insights with respect to physical aspects in communication systems. Being a paradigm of making good use of time, Dr. Wen-Juinn Liu always acts with high efficiency. Besides, Dr. Po-Hsuan Tseng always reveals his capability to hit the points of problems and deduce many things from one case. Moreover, Dr. Jia-Shi Lin has an agile mind to catch the essentials quickly and propose feasible countermeasures. Many other lab mates continue bringing boundless vitality to this family. Yu-Lu Lin is a trustable partner for team works, while Ke-Ting Lee is a polite and studious girl brimming with vibrancy. Yu-Chiun Lin, Lin-Chih Chu, and Cheng-Tse Chiang also accompany me through the ups and downs in prior researches. I want to thank my parents, Chin-Kuan Chen and Hsiu-Pei Chien, for their endless love and tolerance to support me to pursuit knowledge and build confidence. I want to share with them all of my achievements. Thanks to my in-laws for their daily cares. Finally, special thanks to my wife, Yin-Ying Hsieh, for her supports with no regrets to handle home affairs and maintain family life with my lovely girls, Yuri and Candy.

Chien-Hua Chen in HsinChu June 2013