A Large Indoor Localization System based on WiFi and Smartphone Inertial Sensors

The accuracy of existing WiFi-based indoor localization methods with smartphones generally suffers from WiFi signals perception bias given by different smartphones. To this problem, we present a high-accuracy smartphone localization method that combines WiFi and Pedestrian Dead Reckoning (PDR). Specially, we design a novel Received Signal Strength (RSS) ranked vocabulary suited to the FAB-MAP algorithm to enhance fingerprint matching accuracy with negligible computation. The PDR component of the proposed method estimates the user heading and step length using data from inertial sensors embedded in smartphones. We then integrate all the available information in a pose graph optimization framework to iteratively refine the localization results. The experiments on the real and public datasets show that our method achieves an average localization error of 1.88 m and 1.84m, respectively.