Experimental evaluation of a single access point Bluetooth localisation system

Location tracking is a wide field of research with many different solutions. One common trait of many of these solutions is the necessity for a large amount of hardware infrastructure. This work attempts to address this deficiency of typical solutions by developing a localisation system which requires only one piece of infrastructure to be installed, the basestation computer. The proposed localisation system uses classifiers to translate the radio signals coming from a Bluetooth enabled mobile phone to predictions of the currently inhabited room. This paper investigates the applicability of Linear Discriminant Analysis and Support Vector Machines to such a system. Furthermore, experimental results for an in-home environment are used to evaluate the accuracy and repeatability of the proposed localisation system.