A Bluetooth-based minimum infrastructure home localisation system

Indoor location tracking is a function best suited to wireless LAN devices. This generally precludes it from home use in isolated rural areas, where WLAN is a rare commodity. We propose an affordable localisation system which can be implemented using a variety of Bluetooth enabled mobile phones. This permits the incorporation of cellular network signal measurements as well as Bluetooth link measurements into the localisation framework. This paper presents a Hidden Markov Model localisation method, utilising the Viterbi algorithm, which enables single Bluetooth access point localisation. The improvement of accuracy this presents over a Naive Bayes classifier is illustrated, along with the optimal method of obtaining training data.