Indoor localization using Wi-Fi fingerprinting

Nowadays the widespread availability of wireless networks has created an interest
in using them for other purposes, such as localization of mobile devices in indoor
environments because of the lack of GPS signal reception indoors. Indoor localization
has received great interest recently for the many context-aware applications it could make possible. We designed and implemented an indoor localization platform for Wi-Fi nodes (such as smartphones and laptops) that identifies the building name, floor number, and room number where the user is located based on a Wi-Fi access point signal fingerprint pattern matching. We designed and evaluated a new machine learning algorithm, KRedpin, and developed a new web-services architecture for indoor localization based on J2EE technology with the Apache Tomcat web server for managing Wi-Fi signal data from the FAU WLAN. The prototype localization client application runs on Android cellphones and operates in the East Engineering building at FAU. More sophisticated classifiers have also been used to improve the localization accuracy using the Weka data mining tool.