A novel method for metropolitan-scale Wi-Fi localization based on public telephone booths

The ability of low-cost mobile devices to quickly and accurately localize their geolocation is extremely important for pedestrian navigation and location-based services (LBS). Traditional approaches typically rely on global information provided by Global Navigation Satellite System (GNSS), beacons, or the high-cost tracking system. Weak signals and slow time to first fix are two limitations of GNSS, other sensors such as inertial measurement unit (IMU) devices will increase the hardware cost and need a substantially more complicated hybrid algorithm. Wi-Fi localization now regarded as a cost-effective way for indoor and outdoor positioning, it's mainly based on received signal strength indicator (RSSI) measurements, and there are many commercial products like Skyhook's Wi-Fi positioning technology. But the Wi-Fi database is collected through wardriving and it only provides a coarse location with 30-50m accuracy. In this paper, we present a novel method for metropolitan-scale Wi-Fi localization in Hong Kong. Our PCCW telephone booth Wi-Fi database is gathered through precise measurement based on the public telephone booths information, including the media access control address (MAC address), service set identifier (SSID) and coordinates of the Wi-Fi devices installed in the public telephone booths. We also propose a path loss model in the typical urban environment for improved accuracy in the relationship between RSSI and distance. Experimental results demonstrate that the Wi-Fi installed in the public telephone booths can be effectively used for localization in Hong Kong. The cell identification (Cell-ID) rank approach can generally achieve better than 40m accuracy. The triangulation approach following the Cell-ID rank approach can improve the Wi-Fi positioning accuracy to better than 10m level, if the effective range for path loss model is less than 30m, and under good triangulation geometry.