An analysis of radio fingerprints behavior in the context of RSS-based location fingerprinting systems

Location Fingerprinting is a localization technique which uses a radio database as the key element for the positioning process. Performance of location fingerprinting systems has been analyzed empirically in some previous works. However, there is a lack of theoretical models that analyze and predict the fingerprints behavior under different physical situations. In this work we present an analysis to evaluate the distinctiveness of fingerprints in a local neighborhood, in the context of cellular radio fingerprinting. The model is developed assuming the presence of correlated log-normal shadowing. Based on the presented analysis, we conclude some general guidelines for design and implementation of radio databases in fingerprinting systems.