基于改进WKNN的蓝牙室内定位方法  

Method of Bluetooth indoor positioning based on improved WKNN

作　　者：沈天盛 陈文莹 朱彬斌 王燕 周勇良 艾青 SHEN Tiansheng;CHEN Wenying;ZHU Binbin;WANG Yan;ZHOU Yongliang;AI Qing(State Grid Shanghai Municipal Electric Power Company,Shanghai 200122,China;Shanghai University of Electric Power,Shanghai 201306,China)

机构地区：[1]国网上海市电力公司,上海200122 [2]上海电力大学,上海201306

出　　处：《现代电子技术》2023年第10期11-16,共6页Modern Electronics Technique

摘　　要：大型商超、地下停车场等公共场所对室内定位的需求迫切,而GPS定位技术无法实现室内定位。针对传统定位算法定位精度低的问题,文中提出一种基于改进WKNN的蓝牙室内定位方法。首先构建欧氏距离与曼哈顿距离融合的相似度度量函数,改进权值的计算方法;在此基础上,针对偏差较大的跳跃点以及长时间连续定位采集数据量庞大的特点,提出一种基于粒子群的滤波算法,使预测轨迹更贴近真实轨迹。实验结果表明:与传统WKNN算法相比,所提方法定位精度提高约40 cm;相比传统的卡尔曼滤波,基于粒子群的滤波算法可以进一步将定位精度提高约4 cm。Since there is an urgent need for indoor positioning in public places such as large supermarkets and underground parking lots,GPS positioning technology cannot achieve indoor positioning,and the traditional positioning algorithms has low positioning accuracy,a Bluetooth indoor positioning method based on improved WKNN is proposed.The similarity measure function fusing Euclidean distance and Manhattan distance is constructed,and the calculation method of weight is improved.On this basis,in view of the large deviation of jump points and the large amount of data collected in long⁃term continuous positioning,a filtering algorithm based on particle swarm is proposed to make the predicted trajectory closer to the real trajectory.The experimental results show that in comparison with the traditional WKNN algorithm,the positioning accuracy of the proposed method is increased by about 40 cm;in comparison with the traditional Kalman filtering,the particle swarm based filtering algorithm can further increase the positioning accuracy by about 4 cm.

关 键 词：蓝牙室内定位 WKNN算法 权值计算 粒子群滤波 算法流程 实验分析 

分 类 号：TN915-34[电子电信—通信与信息系统] U666.1[电子电信—信息与通信工程] TP212[交通运输工程—船舶及航道工程]