Pattern Recognition Comparative Analysis Applied to Fingerprint Indoor Mobile Sensors Localization

Location Fingerprinting methods are an alternative to accurate localization of mobile sensors and actuators in indoor environments, which learn a radio map for a given scenario and use this information for recognizing the position of a given node at a particular moment. In fact, when using other conventional methods in complex scenarios that may present irregular geometries and materials, fingerprinting techniques can be a very good alternative. Moreover, although they need a previous training of a knowledge database for each scenario, once this is done the method runs in a quite stable and accurate manner without needing any sophisticated hardware. In this paper we present a comparation between the CC2431 Texas Instruments textit{Location Engine} and two pattern recognition methods applied to fingerprinting localization in several indoor scenarios in our university. Results show that this kind of technique is very interesting for known environments such as warehouses, universities and so on.