基于 Zigbee 室内定位系统的指纹库优化算法

摘要

Zigbee 无线传感器网络的接收信号强度指示(RSSI)可以用来提供位置服务，使用 RSSI 建立指纹数据库设计定位算法能够得到良好的定位效果，但指纹数据库的采集和建立精度会直接影响到最终的定位精度，而一般方法建立的指纹数据库误差较大。为此，通过对 Zigbee 网络指纹数据库采集和建立过程进行研究，提出一种 Zigbee 网络中指纹数据库采集、建立及优化的算法。在定位区域将各采样点采集到的指纹数据库源数据进行滤波处理，建立高精度的指纹数据库。利用 Zigbee 平台组建无线网络进行实验，分别使用针对该定位系统提出的限定区域最邻近算法、限定区域加权最邻近算法、限定区域贝叶斯算法进行定位计算。实验结果表明，运用优化后的指纹数据库在短距离范围内定位平均误差可限制在1.5 m 以内。%Received Signal Strength Indication(RSSI) of the Zigbee Wireless Sensor Network(WSN) can be used for providing location service. Using RSSI establishing fingerprint database to design positioning algorithm can get good positioning result. But fingerprint database collection and establishment accuracy directly affect the positioning accuracy. The general method to establish the fingerprint database has larger error, and after researching for the process of Zigbee network fingerprint database collection and establishment, the fingerprint database collection, establishment and optimization method is proposed. In the positioning area, all the collected fingerprint database source data from every sampling point are filtered, and high precision fingerprint database is established. Using the defined region adjacent algorithm, defined region weighted adjacent algorithm and defined region Bayesian algorithm are proposed for this positioning system calculates user’s location, respectively. Experimental result shows that the average error using the optimized fingerprint database is limited within 1.5 m in short range positioning system.