Smartphone PDR positioning in large environments employing WiFi, particle filter, and backward optimization

The evolution of an existing offline algorithm for smartphone based indoor localization is outlined, discussed and evaluated. The existing algorithm performs dead reckoning using accelerometer-, compass- and gyroscope data. Accelerometer data is used to detect step events. Compass and gyroscope are used for heading estimation. Drift compensation and step length estimation is performed by a particle filter using map information. Floor detection is based on RSSI (Radio signal strength indicator) evaluation. In an offline scenario where positions are determined on recorded datasets, a backtracking algorithm allows a considerable increase in accuracy by elimination of unfeasible traces. The evolution comprises a radio map generation algorithm, the WiFi RSSI position determination, the inclusion of WiFi- and floor change information into the particle filter of the dead reckoning algorithm, as well as outlier identification in the RSSI position- and floor estimation. The modifications are described and discussed. The performance is evaluated against a publicly available database in the frame of a competition.