利用蜂巢式拓樸建置大面積之ZigBee定位系統

A New ZigBee Positioning System with Cellular Topology Used for Large Size Areas

鄭崇廷(Chong-Ting Jheng)

指導教授 ： 黃國鼎

國立虎尾科技大學/電機資訊學院/電機工程系碩士班/碩士(2016年)

摘要

由於ZigBee具有低功耗、低成本、支援大量節點的特性，在建置室內定位環境時有低成本與簡單的優勢。ZigBee之節點進行定位時利用接收訊號強度指標(Received Signal Strength Indicator, RSSI)判斷待測點與參考點間的距離，與事先建立的通道模型曲線即可進行距離估算。本文將透過通道模型曲線在不同距離等級的變化，區分有效距離範圍與無效距離範圍，並以此提出定位面積幾何之優化後再進行三邊定位法，藉此提高定位可靠性。優化方式首先將定位幾何的邊長設為有效距離的最大值，讓定位幾何形成正三角形。除了能讓待測點接收到每個參考點的RSSI準確性高，也能增加可定位面積以提高硬體的使用成本效率(C/P值, Cost/Performance)。為確實了解其效率差異，將進行實驗比較相同面積之兩種不同三角形定位幾何，分別為邊長都為有效距離最大值的正三角形定位幾何和其中一邊長含有無效距離範圍的等腰三角形定位幾何的定位效果差異。從實驗結果顯示邊長都為有效距離最大值的定位幾何的總平均誤差為27公分，標準差為12公分。另外未經優化的定位幾何的總平均誤差為61公分，標準差為57公分。實驗結果得知正三角形之優化定位幾何的定位效果和穩定性明顯提升，本文後續研究將採用此正三角形定位幾何。
為了建置更大有效面積的定位範圍，本文將六個正三角形組合成為一個正六邊形定位幾何，作為本文提出新的優化定位方法時之基本定位範圍。由於在此六邊形之基本定位範圍內進行定位時，待測點需從六個正三角形中判斷其所在之三角形方能進行正確三邉定位，但在任意相鄰兩正三角形附近之待測點存在邊界模糊區域，容易錯選定位三角形，造成進行三邊定位時誤差增大。為了改善邊界模糊區域問題，本文將提出一新的優化方法，稱為動態混合三邊定位法。在六邊形之基本定位範圍中，除原來六個正三角形外將額外規劃出兩個大面積之正三角形，藉由不同參考點之RSSI值強弱判斷出合適之正三角形進行定位，此方法除了可降低邊界模糊區域問題外也可優化定位誤差。實驗中，平均擺放待測點於固定範圍內，原本正六邊形定位幾何整體的平均誤差211公分，標準差為237公分，大幅縮小至平均誤差29公分，標準差為13公分，證明了此方法之可行性。因此本文提出動態混合三邊定位法用於正六邊形定位幾何具有較佳的定位效果和穩定性，未來在大於正六邊形面積之範圍需要進行定位之應用時，可組合不同正六邊形面積擴展成蜂巢式定位拓樸涵蓋整個需要定位之範圍，因此本研究亦可在大面積範圍達到室內定位功能。

關鍵字

ZigBee ； 訊號強度指標 ； 室內定位 ； 三邊定位法

並列摘要

We proposed a new positioning scheme with ZigBee technology for the indoor environments. The received signal strength Indicator (RSSI) of reference nodes measured of the target are used to estimate the distances between the target and reference nodes. For each specific environment, we need to establish its channel model (the relation between the distance and RSSI). Because of the nonlinear relation, the estimation error of distance becomes larger as the corresponding RSSI is smaller.
Therefore, we classify the error to two levels: valid and invalid distance ranges. Accordingly a trilateral positioning method with an optimized positioning geometry is proposed by exploiting the valid distance range. The optimized geometry is a regular triangle with the side length to be set up to the maximum valid distance range. As a result, a target node could receive more accurate the RSSI from each reference node, and thus the positioning area is enlarged. In order to verify our method, we perform a positioning experiment by comparing two different triangle geometries of same area. In addition we also consider the isosceles triangle with a longest side length up to the invalid distance range. The experimental results reveal that the average estimation error of distances is 27cm and its standard deviation is 12cm for the regular triangle. The isosceles triangle produces an average error of 61cm and standard deviation of 57cm. Thus, the trilateral positioning method with an optimized regular triangle geometry has an excellent positioning performance.
In order to enlarge the positioning area, we further propose a hexagonal positioning geometry by combining six regular triangles as mentioned above. However a new problem occurs when a target node located around the boundary area of any two adjacent regular triangles. The target node is likely to choose a wrong triangle in which the target node is not located. This results in the increase of estimation error of distance. Therefore, we propose a new dynamic hybrid trilateral positioning method for the hexagonal positioning geometry to reduce the estimated error. According to our experiments, by the proposed method, the average estimation error of distance is reduced from 211cm to 29cm and the standard deviation from 237cm to 13cm. Moreover, if the considered area is much larger than that of the hexagonal geometry, we can combine the different hexagonal geometries to form a cellular topology, which produces an extended area. The target node receives the 7 highest RSSI values of reference nodes to create a hexagonal positioning geometry. It is shown that the proposed new ZigBee positioning system with cellular topology that is suitable to be used for any size of positioning coverage.

並列關鍵字

ZigBee ； RSSI ； Indoor positioning ； trilateral positioning method

參考文獻

[33]黃俊元，”用ZigBee無線感測網路做為室內自我調整誤差定位方法之研究”，國立虎尾科技大學電機工程研究所在職專班碩士論文，100年7月。

[1]K. Maneerat, and C. Prommak, (2015), “Low Complexity Wireless Indoor Positioning Approaches based on fingerprinting Techniques,” Telecommunication Networks and Applications Conference (ITNAC), pp.244-249.

[2]R. Faragher, and R. Harle, (2015), “Location Fingerprinting with Bluetooth Low Energy Beacons,” IEEE Journal on Selected Areas in Communications, pp2418-2428.

[3]L. Pei, R. Chen, J. Lin, T. Tenhunen, and H. Kuusniemi, (2010), “Inquiry based Bluetooth Indoor Positioning via RSSI Probability Distributions,” Advances in Satellite and Space Communications (SPACOMM),pp151-156.

[4]A. K. Nabih, H. Osman, M. Gomaa, and G.M. Aly, (2012), “New Fuzzy based Indoor Positioning Scheme using ZigBee Wireless Protocol,” Computer Engineering & Systems (ICCES), pp.16-20.