基於接收訊號強度之分散式室內定位系統

近年來，全球定位系統(Global Position System, GPS)為大眾所熟知的定位技術。然而，GPS定位技術的應用必須讓目標與衛星之間擁有直視距離(Line of Sight, LoS)。因此，GPS定位技術無法滿足室內定位應用上的需求。在室內定位技術的相關研究中，由於無線感測器網路(Wireless Sensor Network, WSN)技術在室內應用上，擁有體積小、方便佈建以及成本低廉等優點。同時，無線感測器節點佈建在室內時，感測器節點就像環繞在地球的衛星一樣，能夠提供目標節點進行定位計算時所需的資訊。因此，無線感測器網路技術是室內定位研究中常被使用的技術之一。本研究利用接收訊號強度設計並實現一套分散式室內定位系統。接收訊號強度定位法擁有方便使用、無須額外硬體成本等優勢，是WSN定位技術中常被使用的一個方法。透過分散式定位法的優勢，目標節點能夠快速計算出自身的位置。無須透過閘道器或是特定節點做運算。本系統先在室內建立訊號衰減曲線，作為測量距離的依據。當目標節點藉由周遭錨節點取得足夠的距離資訊時，目標節點本身即可進行分散式定位演算法的計算。本研究考量到接收訊號強度在室內的穩定度隨距離增加而減少，因此定位演算法的部分是以三角定位法為基礎。同時，利用定位圓的伸縮與擴張的概念，修正三角定位法定位時，因為接收訊號強度受到環境影響所產生的定位誤差。本研究以MATLAB軟體模擬此分散式室內定位系統的結果。同時，本研究克服實作上的困難，將演算法實作於無線感測器節點Octopus II上，以驗證模擬結果。經過實地實驗證實，此分散式室內定位系統的實作結果與模擬結果相符合，定位誤差分別為0.87 m與0.71 m。同時，在不同環境下的定位結果分別為0.87 m與0.67 m，證實本研究提出的定位方法能夠修正不同環境造成的定位誤差。

關鍵字

分散式室內定位；接收訊號強度；無線感測器網路

並列摘要

In recent years, Global Position System (GPS) is the most common technology of localization. However, the applications of the GPS must include Line of Sight (LoS) between user and the satellites. For this reason, the GPS is not conformed to applications of indoor localization. Wireless Sensor Network (WSN) is one of the technologies often used in the studies of indoor localization for the advantages of small sizes, convenience of set up, and low cost. The nodes, like the satellites, can provide information for localization algorithm which set in the indoor environment. In this study, a distributed indoor localization system based on the received signal strength indicator (RSSI) was designed and implemented. RSSI is often used in localization of WSN due to no additional hardware being needed and convenient workability. In this study, the target node can calculate its position through the distributed algorithm. In the beginning of the localization system, the RSSI is detected and the RSSI-distance curve is built between two nodes in the indoor environment. The target node in the indoor environment could start the distributed localization algorithm when it gathers enough information from neighboring anchor nodes. The distributed localization algorithm is based on triangulation location. Considering the RSSI is more unstable with increasing distance, this algorithm used the expanded local circle for reducing the located error from unstable RSSI. This study simulates the distributed indoor localization system by the MATLAB and implements this system on the Octopus II. The location error from the simulation is 0.71 m. The location errors from the implementation in various environments are 0.87 m and 0.67 m. The simulation and the implementation have similar results that confirm this system worked. The similar implementation results from different environments show that the algorithm can reduce location error from different environment.