Abstract: In this paper, we examine the results of the EvAAL-2013 localization competition. We present a comprehensive description of RealTrac™ technology, the winner of this competition. Focused on real-time location technology, RealTrac hardware includes a server, access points (gateways and repeaters), and mobile devices (voice intercoms and tags). A location calculation engine is based on Bayesian algorithms fusing data from time-of-flight and received signal strength measurements, inertial measurement units, building structure, and optionally, from context events. The peculiarity of the embedded navigation system is that it works effectively independent of the place where the mobile tag is actually attached, i.e., it could be mounted on one’s foot, carried freely in a jacket pocket, held in one’s hand, or even hung on a shoestring. The RealTrac system can easily be combined with third-party systems via the Real Time Location System Communication Protocol that consists of an open API and uses a common Keyhole Markup Language format for geo-data presentation. Furthermore, in this paper, we show that the scores obtained in the competition regarding positioning accuracy might be significantly higher when the applied particle filter is appropriately configured and a probability-based approach for area-of-interest determination is introduced.
Keywords: Local positioning system, indoor navigation, Bayesian filter, time-of-flight, received signal strength
