A Hybrid Modeling for WLAN Positioning System

The wireless-LAN positioning system (WPS) is an indoor positioning system, which uses the wireless local area networks infrastructure to determine the position of the user with mobile terminal. There are essentially two kinds of signal-strength-based modeling for WPS: propagation modeling and fingerprinting modeling. Since the indoor environment is very complicated and changeful, it is very difficult to fit the signal strength distribution nicely using mathematical formula in propagation modeling. The problem with fingerprinting modeling is that it takes much effort and time to build the coordinate-signal strengths maps for each WLAN. The proposed hybrid modeling has three stages. Firstly, establish a database that contains a distance-signal strength map of the interesting area. Secondly, use the database to obtain an estimate of the distances between mobile terminal and base stations. Thirdly, apply the trilateration to compute the mobile terminal position. Experimental results show that the mean position error is much better than propagation modeling and database establishment is simpler than fingerprint modeling.