A calibrated-RSSI/PDR/Map intergrated system based on a novel particle filter for indoor navigation

A novel positioning system integrating Fingerprinting/PDR/map is proposed for indoor location-based service (LBS). The proposed system is used to achieve a superior indoor localization accuracy. The particle filter with multi-directions and adaptive weights, also named MDAW-PF, is proposed to fuse the fingerprinting, PDR and map. In the MDAW-PF method, the particle weights fully consider the noise of fingerprinting, the direction of the particles is to uniformly sample the direction of PDR estimation into multiple directions to ensure particle diversity, and the map information provides constrains for the particles. It is well known that the positioning accuracy of fingerprinting using RSSI is rather susceptible to the changing environment. Thus, the improved fingerprinting, also named MCRT, which calibrates online-RSSI by the matching model between offline-RSSI and online-TDOA, is proposed to mitigate the effect of RSSI variation. All measurements are conducted in indoor public areas, and the results show that the proposed system is featured with better localization performance, with the average error 0.68 m.