一种基于WiFi信号特征的聚类过滤定位算法研究  

On Location Algorithm of Cluster Filtering Based on WiFi Signal Features

作　　者：蔡炯炯 沈涵生[1] 张文辉 王子辉 袁琳[1] CAI Jiong-jiong;SHEN Han-sheng;ZHANG Wen-hui;WANG Zi-hui;YUAN Lin(School of Automation and Electrical Engineering,Zhejiang University of Science and Technology,Hangzhou310023,China;School of Electrical Engineering,Zhejiang University,Hangzhou 310013,China;Suzhou Industrial Technology Research Institute of Zhejiang University,Suzhou 215010,China)

机构地区：[1]浙江科技学院自动化与电气工程学院,浙江杭州310023 [2]浙江大学电气工程学院,浙江杭州310013 [3]浙江大学苏州工业技术研究院,江苏苏州215010

出　　处：《浙江水利水电学院学报》2018年第5期63-67,共5页Journal of Zhejiang University of Water Resources and Electric Power

基　　金：浙江省自然科学基金(LY17E070002);江苏省博士后资助计划(169491)

摘　　要：提出一种用于室内车库定位的基于Wi Fi信号强度特征的快速算法.根据信号强度(RSSI)波动变化,提出信号波动聚类算法,可在指纹库中标定初选区域,缩小后续计算范围;根据在该区域内不同RSSI区间的信号波动变化规律设置过滤阈值,可实现动态选择过滤区间,将初选后的指纹数据组进行多次过滤,可进一步减少待选数据组;剩余少量数据组可由最小欧式距离快速计算,完成目标精确定位.算法可大幅度减少计算量和复杂度,提高定位速度.A fast algorithm based on WiFi signal strength features for indoor garage location is proposed. According to thefluctuation of signal strength （RSSI）, a signal fluctuation clustering algorithm is proposed, which can calibrate the primaryring region in the fingerprint database and reduce the subsequent calculation range. Multiple filtering can further reduce thenumber of data groups to be selected, and the remaining small number of data groups can be quickly calculated by the mini-mum Euclidean distance to achieve accurate target location. The algorithm can greatly reduce computation complexity andimprove positioning speed.

关 键 词：室内车库定位 过滤 信号波动特征 聚类 

分 类 号：TN925[电子电信—通信与信息系统]