一种基于压缩感知与最近邻的联合定位方法  

Joint location method based on compressed sensing and nearest neighbor

作　　者：胡久松[1] 刘宏立[1] 肖郭璇 徐琨[1] 

机构地区：[1]湖南大学电气与信息工程学院 [2]国家电网浙江乐清市供电公司

出　　处：《电子测量与仪器学报》2018年第6期72-78,共7页Journal of Electronic Measurement and Instrumentation

基　　金：中央国有资本经营预算项目(财企[2013]470号);国家自然科学基金(61771191);中央高校基本科研项目(1053214004);湖南省自然科学基金(2017JJ2052);教育部产学合作协同育人项目(201601004010,201701056026);湖南省普通高校教学改革研究项目(湘教通[2016]400号);湖南省研究生创新项目(CX2017B112)资助

摘　　要：针对基于压缩感知的Wi-Fi室内定位算法中,由于信号强度的时变性和多径效应等原因,算法恢复的结果有时并不理想而降低了定位精度的问题,提出了一种基于压缩感知与最近邻的联合定位方法。在线定位时,通过条件判断决定是选择压缩感知的定位结果还是最近邻的定位结果。这种联合定位的方式可以使得两种定位方法相互补充各自的不足,定位结果更佳。采集了大量的真实数据进行了实验。实验结果表明,提出的联合定位算法比使用压缩感知和最近邻单项算法分别提高了5%和11%。In order to solve the problem of decrease of positioning accuracy using compressed sensing( CS) based Wi Fi indoor location algorithm,caused by non-ideal restore results,due to the time-varying signal strength and the multipath effect,this paper proposes a joint location method based on CS and nearest neighbor. According to the conditional judgment,the joint location method determines whether to select the positioning results of compressed sensing or nearest neighbor. The joint location method can make the two positioning methods complement each other which can make the positioning results much better. The experiment collected a lot of real data for testing. The experimental results show that the proposed joint location method is 5% and 11% higher than that of compressed sensing and nearest neighbor single algorithm.

关 键 词：室内定位 指纹识别 压缩感知 加权K邻近 无线电地图 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]