Hierarchical Approach to Enhancing Topology-based WiFi Indoor Localization in Large Environments

Traditionally, WiFi has been used for indoors localization purposes due to its important advantages. There areWiFi access points in most buildings and measuring WiFi signal is free of charge even for private WiFi networks. Unfortunately, it also has some disadvantages: when extending WiFi-based localization systems to large environments their accuracy decreases. This has been previously solved by manually dividing the environment into zones. In this paper, an automatic partition of the environment is proposed to increase the localization accuracy in large environments. To do so, a hierarchical partition of the environment is performed using K-Means and the Calinski-Harabasz Index. Then, different classification techniques have been compared to achieve high localization rates. The new approach is tested in a real environment with more than 200 access points and 133 topological positions, obtaining an overall increase in the accuracy of approximately 10% and reducing the mean error to 2.45 metres.

keywords: WiFi indoor localization, large environments, learning algorithms, clustering, classification