Aiming at the problem that the current indoor fingerprint positioning techniques are of passive positioning based on multiple access points (APs),which limits the use of a single AP in indoor scenarios,the paper proposed an indoor fingerprint positioning method based on channel state information (CSI) and joint angle of arrival (AOA):it was pointed out that the key technique is to use multidimensional signal parameters on a single link to construct a fingerprint model;the spatial diversity in multiple input multiple output (MIMO) systems was utilized,and a CSI quotient was established to obtain more robust CSI fingerprint signals;then,a multi carrier AOA fingerprint representation method was designed based on the principle of multiple signal classification (MUSIC) algorithm,which has discriminability compared to the original CSI fingerprint;finally,in order to solve the symmetry problem of AOA fingerprints for a single AP,a new fingerprint was gained by combining CSI quotient with AOA fingerprints,and target position matching was performed through machine learning ways.Experimental results showed that the proposed method could achieve localization accuracy of 98.96％ and 97.08％,with average localization errors of 0.46 m and 0.68 m in empty classroom and laboratory environments,respectively,demonstrating high localization performance.
关键词
无源定位/到达角(AOA)估计/信道状态信息(CSI)/位置指纹/指纹库
Key words
passive localization/angle of arrival (AOA) estimation/channel state information (CSI)/location fingerprint/fingerprint database
维普
万方数据