Generating indoor Wi-Fi fingerprint map based on crowdsourcing

Now indoor fingerprint maps are mostly generated and maintained by professional companies, where the generation efficiency and update frequency are relatively slow. To address such issues, the paper devises a crowdsourcing method, which uses the sensed data from personal smart devices to build a Wi-Fi fingerprint based map for indoor navigation. Our method uses the positions with high centrality as landmarks. Moreover, based on the improved Euclidean Distance method and Access Points (APs) rank, we devise a Coefficient Weighting algorithm to match the Wi-Fi fingerprints among the Wi-Fi Received Signal Strength (RSS) sequences, which also applied together with a scoring matrix to decrease the redundant data brought by trajectories overlapping. The experimental results show that: (1) our proposed method not only reducing the redundant data brought by crowdsourcing, but also updating the map in a near-real-time manner; (2) the mean positioning error of the generated map is <2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$< 2$$\end{document} m, which obviously outperforms the known methods such as LiFs, CALL, etc.