Indoor Localisation Based on Wi-Fi Infrastructure

Over the last decade, indoor positioning has played an increasing role in the navigation market. In this work we present an indoor attempt for localization based on radio-environment properties. Our research concerns implementation of machine learning algorithms for a Wi-Fi fingerprints-based positioning system. The algorithms we chose are kNN, NB and RF. The method was evaluated using several different mobile devices, with samples collected in different locations of a school building. The results we achieved are very promising.