Fault detection and mitigation in WLAN RSS fingerprint-based positioning

Abstract

Fingerprint-based positioning methods that utilise WLAN received signal strength (RSS) measurements are becoming increasingly popular for inferring terminal location in indoor environments. They are very attractive mainly because of the low deployment cost, as they exploit the ubiquity of WLAN access points (AP) and the worldwide availability of mobile devices featuring WLAN adapters for wireless connectivity. They provide reliable and accurate location estimates, however their performance can be compromised when faults occur in the overall positioning system. Such faults are injected in different ways, including the removal or relocation of one or several APs, and are caused either accidentally or deliberately due to malicious attacks. These faults corrupt the RSS fingerprints and may lead to accuracy degradation, unless countermeasures are taken. In this work, we focus on positioning under faults and construct detection schemes using fault indicators that are applicable to standard fingerprint-based methods and achieve high fault detection rates. In addition, we incorporate our detection mechanisms into hybrid positioning algorithms that greatly improve fault tolerance in the presence of different types of faults. Finally, we present experimental results to verify the effectiveness of the proposed approach