Compass and WLAN integration for indoor tracking on mobile phones

Indoor positioning with smartphones in ubiquituos computing scenarios still poses some problems with respect to accuracy and precision as well the need for a calibrated infrastructure and map data. This paper presents a method for indoor positioning based on the combination of 802.11 WLAN fingerprinting using weighted kNN and a simple indoor movement model based on digital compass information. We introduce a novel computation scheme for the distance in WLAN signal space, additionally considering the Euclidean real world distance between each fingerprint and the position predicted by the movement model. A detailed evaluation in a test environment at our site demonstrates a performance gain of more than 10% as compared to a classical Kalman filter. Moreover, we also show that the Kalman filter offers a slightly better capability to correct the accumulation of errors over time when accurate movement information in form of step detection is available.