Integrating Wi-Fi and Magnetic Field for Fingerprinting Based Indoor Positioning System

Smartphone based personal tracking is very important for people to find their destination in large complex buildings (e.g. shopping malls, airports and museums). Such applications are highly demanded in both industries and research organizations. One critical issue for these applications is lack of mature technologies for highly accurate indoor location tracking. In this paper, a new Wi-Fi and magnetic field based smartphone tracking system, named WMLoc, was introduced. The system is a part of a collaborative project between the RMIT University and a famous Australian software company. A number of tracking algorithms such as K nearest neighbor (KNN), artificial neural network (ANN) and back tracing (BT) have been developed or adopted for the system in order to obtain a real-time precise location of the smartphone user. The integration of Wi-Fi and magnetic field includes physical floor analysis, label pattern creation using ANN and BT for enhancing the tracking reliability and improving positioning accuracy. The WMLoc system was tested in two buildings at the RMIT University, Australia. The preliminary results showed that the average root-mean-square (RMS) error of the WMLoc system was less than 2.6 m.