GraphLoc: a graph-based method for indoor subarea localization with zero-configuration

Abstract

Indoor subarea localization can facilitate numerous location-based services, such as indoor navigation, indoor POI recommendation and mobile advertising. Most existing subarea localization approaches suffer from two bottlenecks, one is fingerprint-based methods require time-consuming site survey and another is triangulation-based methods are lack of scalability. In this paper, we propose a graph-based method for indoor subarea localization with zero-configuration. Zero-configuration means the proposed method can be directly employed in indoor environment without time-consuming site survey or pre-installing additional infrastructure. To accomplish this, we first utilize two unexploited characteristics of WiFi radio signal strength to generate logical floor graph and then formulate the problem of constructing fingerprint map as a graph isomorphism problem between logical floor graph and physical floor graph. In online localization phase, a Bayesian-based approach is utilized to estimate the unknown subarea. The proposed method has been implemented in a real-world shopping mall, and extensive experimental results show that the proposed method can achieve competitive performance comparing with existing methods.