An on-demand approach for indoor localization based on crowdsourced Wi-Fi fingerprints

Indoor localization based on crowdsourced Wi-Fi fingerprints is an emerging technique that allows to tackle site survey problem by utilizing contributions from common people. In order to support flexible location accuracy in terms of fruitful application or user requirements, in this paper, we analyze the radio map constructed by crowdsourced Wi-Fi fingerprints and propose an on-demand approach to optimally select representative fingerprints from the radio map for user indoor localization. The evaluated results demonstrate that our method achieves high confidence level of location accuracy under a wide-range of requirements.