摘要
针对目前利用WiFi信号进行室内定位实时精度较低的问题,该文提出了一种改进的K最近邻算法.由于室内人体走动对于WiFi信号的不规律干扰,使得室内实时定位的精度带有很大的不确定性.为了实时地消除外界干扰带来的误差,改进的K最近邻算法增加了外部节点来监测周围WiFi信号的强度变化,通过将获取的信号强度与指纹数据库中对应节点的信号强度比对,获取差值,并应用于节点周围的客户端,来实时地校正客户端的定位结果.利用此算法在Android平台上的实验表明,该算法定位简单,可以较为明显地改善节点周围2.4m范围内的实时定位精度,使平均精度能提高0.8-1m左右.
Aiming at the problem of low accuracy of indoor real-time positioning using WiFi signal, this paper presented an improved KNN algorithm. The signal interference caused by humans walking could make indoor real-time positioning accuracy uncertain. To avoid the error caused by interference in real-time, the improved KNN algorithm monitored the change of WiFi signal strength around the nodes by increasing some external nodes. The signal strength was used to compare with the strength in the fingerprinter database, and the difference could be used to correct the client＇s positioning results. According to experiments based on Android platform, this algorithm was simple, besides, it could more obviously improve the client＇s positioning accuracy within the scope of 2.4m from the node and the average precision was improved by 0.8 m-1 m or so.
出处
《测绘科学》
CSCD
北大核心
2015年第6期125-128,155,共5页
Science of Surveying and Mapping
