Flexible indoor localization and tracking system based on mobile phone

As the WIFI access points are widely deployed, the received WIFI signal strength is commonly adopted as a positioning characteristic for mobile phone based indoor localization systems. Although WIFI based localization has achieved great development, there are still several key challenges in tracking applications, such as how to modify irregular trajectory obtained from the sequential positioning results. To tackle those challenges, this paper integrates the typical WIFI indoor positioning system with a Pedestrian Dead Reckoning (PDR) system based on the sensors in the mobile phone as many newly emerged systems proposed. The Maximum Likelihood (ML) algorithm is proposed to retrieve the user's initial location and moving direction without any intervention from the user. During the tracking process, a filtering algorithm can revise the moving direction indicated by the sensors only if a straight walking is detected. To obtain more accuracy and efficiency, a combination of Kalman Filter (KF) and auto-adaptive dynamic grid filter (GF) named KAGF is proposed for the fusion of the results from WIFI and PDR system. Experiments in the real scenarios show that our fusion system achieves better results than the widely adopted one, in which the particle filter is used, both in accuracy and computational complexity. Furthermore, the system's effectiveness is improved largely with longer WIFI updating period and larger reference points' interval to achieve the same encouraging results.