WiFi fingerprint indoor positioning system using probability distribution comparison

Positioning services are increasingly used for applications such as navigation, advertising and social media. While outdoor navigation based on GPS and/or cellular systems works well, indoor navigation is a much tougher challenge. This paper presents a new indoor positioning method based on Wi-Fi fingerprints, i.e. RSSI measurements from multiple Wi-Fi access points. During an offline phase, fingerprints are collected at known positions in the building. This database of locations and the associated fingerprints are called the radio map. In the online mode, the current Wi-Fi fingerprint probability distributions are compared with those of the radio map. The user location is estimated by calculating a weighted average of the three offline positions that best match the online measurements. Experiments show that our technique is superior to other proposed methods and reaches a median error of 2.4m.