Fingerprint Localization Algorithm Based on Occlusion Judgment

With the development of Device-free localization technology, indoor fingerprint localization has attracted more and more attention. However, in indoor environment, the degree of occlusion at different locations is obviously different, the positioning accuracy based on Channel State Information (CSI) system decreases. In this paper, we propose a fingerprint localization algorithm based on occlusion judgment (FLOJ algorithm). In the fingerprint processing phase, we use principal components analysis (PCA) to reduce the dimension of fingerprint database to remove redundant data. And k-nearest neighbor (KNN) position estimation algorithm is used to get the matching results. In the phase of multi-position fusion, we use the algorithm based on CSI propagation model as the criterion to measure the occlusion degree of indoor environment. And we fuse the matching positions obtained by multi-receivers at the same time. Experiment results show that compared to KNN localization methods, our proposed algorithm can effectively reduce positioning error.