Abstract: Smart environments require context information about their inhabitants in order to dynamically adapt their functionality. The location of persons or entities in an environment is an important piece of the context information. To this end, several types of indoor localization systems have been developed, with fingerprinting-based systems being the most common. Fingerprinting-based indoor localization systems tend to achieve higher accuracy compared to other approaches such as signal propagation modeling. However, they also tend to have a higher effort/cost for deployment and maintenance. Changes in the configuration of the indoor space like moving of furniture, or defective signal sources can cause the signal characteristic distribution in the environment to change significantly. This renders the fingerprint radio map (used for training the system) outdated, and causes a corresponding drop in localization performance over time. This paper proposes an approach for environment self-monitoring and autonomous recalibration using the system infrastructure, and demonstrates that it can reliably detect changes in signal distribution and recalibrate the radio map of the localization system. The proposed approach achieves localization performance of up to 93% of the optimum achievable through manual system recalibration.
