Autonomous wireless positioning system using crowdsourced Wi-Fi fingerprinting and self-detected FTM stations

Wi-Fi positioning system (WPS) has been proven as an effective way to realize universal indoor navigation for smart city-related applications. The localization ability of the existing WPS is affected by the low precision of the crowdsourced database and the unknown location of local wireless stations. This paper develops an integrated indoor localization framework using crowdsourced Wi-Fi fingerprinting and self-detected Wi-Fi Fine Time Measurement (FTM) stations (IL-CFSW). A novel crowdsourced mobile sensors data modeling algorithm with self-calibrated parameters is proposed, and modeled trajectories are further segmented and matched with the existing pedestrian indoor network to enhance the performance of the final database generation. Furthermore, the iteration unscented Kalman filter (iUKF) is adopted to recognize the position and calibrate the bias of existing Wi-Fi FTM anchors combined with the hybrid distance measurement model. Finally, an enhanced particle filter with error ellipse constraint is developed to fuse different location sources and indoor network information. Comprehensive experimental results present that the developed IL-CFSW can achieve autonomous 3D indoor localization performance in multi-floor contained scenes and meter-level positioning accuracy is achieved with the consideration of pedestrian indoor network information.