Hierarchical CSI-fingerprints Classification for Passive Multi-person Localization.

Indoor human target localization is an important enabling technology for many intelligent applications. The attractive advantage of passive localization algorithm is that it can estimate target location without needs of carrying electronic devices by human target. Due to complex signal radiation in indoor environment, most localization algorithm adopt fingerprints approach for indoor passive localization. Fingerprints can achieve good performance in one person indoor passive localization. However, when there are more than two people in target area, the system performance may degrade due to high complex of the matching task. In this paper, a hierarchical CSI-fingerprints classification system is proposed for passive indoor multi-people localization. In training phase of coarse classification, fingerprints with similar CSI are first grouped into coarse classes. Then, a coarse classifier is trained for coarse fingerprints matching. Fingerprints belonging to the same coarse class are then feed into fine classifier for fine fingerprints matching. Experimental results reveal that the proposed approach can achieve acceptable accuracy in 37 configurations including 0 to 2 people. Furthermore, CSI grouping shows that the similarity of CSI depends on LOS and number of people.