本論文提出一個利用低耗能藍芽(Bluetooth Low Energy, BLE) iBeacon技術來完成室內二維以及三維非測量距離定位(indoor 2D and 3D range-free positioning)的新方法，稱為wClub。在wClub中佈置多個位置已知之錨節點(anchor node)於室內環境，而錨節點會以固定訊號強度，週期性地送出夾帶錨節點位置資訊的BLE iBeacon廣告(advertisement)封包。位置未知之目標節點(target node)則可透過附近錨節點所送的廣告封包的接收訊號強度(Received Signal Strength Indicator, RSSI)計算每個錨節點的加權值，再利用加權質心定位演算(weighted centroid localization algorithm)計算出目標節點位置。本研究並根據BLE iBeacon設備的印刷電路板天線輻射方向圖(antenna radiation pattern)，利用天線對於YZ和XZ平面最符合全向輻射(isotropically radiated)的有效角度，提出長方體錨節點佈置模型。透過此模型能夠於典型教室和走廊等室內環境適當地佈置錨節點，降低天線方位對定位所產生的誤差，達到滿意的定位精準度。本研究以長為9公尺寬為5公尺的室內教室為實驗場景，透過密集實驗資料來調整出較佳的加權值計算參數，以進一步提升定位精準度。在上述的環境透過安裝四個錨節點可以執行二維定位，而80%二維定位實驗結果的誤差在50公分以內，其所需要定位時間介於400 ms 至1500ms之間。透過額外增加一個錨節點可以執行三維定位，而90%三維定位實驗結果的誤差在80公分以內，其所需要定位時間介於400 ms至1900 ms之間。;This study proposes a new indoor localization scheme, which uses the Bluetooth Low Energy (BLE) iBeacon technology. The proposed scheme, called wClub, can achieve indoor two dimensional (2D) and three dimensional (3D) range-free positioning. In wClub, several anchor nodes with known positions are deployed, and every anchor node periodically broadcasts BLE iBeacon advertisement packets containing its position information with fixed signal strength. A target node, whose position is unknown and is to be determined, receives advertisement packets from nearby anchor nodes and calculates the weight of every anchor node by the received signal strength indication (RSSI) values of advertisement packets. The target node can then use the weighted centroid localization algorithm to estimate its position. According to the printed circuit board (PCB) antenna radiation pattern of iBeacon devices, the study derives the effective angles of the YZ and XZ plane that are close to the isotropically radiated pattern. The study then proposes a cuboid anchor deployment model suitable for indoor environments of typical classrooms and the corridors to reduce the positioning error caused by the antenna orientation. Some experiments are performed to derive better parameter settings for further improving the positioning accuracy. The experimental environment for 2D positioning is an indoor 5 m by 9 m classroom in which 4 anchor nodes are deployed. The 80% of the 2D positioning experiment results are of less than 50 cm positioning errors, and the positioning takes the time between 400 ms and 1500 ms. The experiments of 3D positioning are performed by adding one extra anchor node. The 90% of the 3D positioning experiment results are of less than 80 cm positioning errors, and the positioning takes the time between 400 ms and 1900 ms.
