Device Diversity Effects on RF Fingerprinting Based 3D Positioning System

This paper presents the device diversity effects on Radio Frequency (RF) fingerprinting framework in the indoor environment for the three-dimensional positioning system. RF Fingerprinting positioning was studied inside of a multistory university building by devices from multiple manufacturers at different times. The goal was to study the effects of equipment heterogeneity on positioning accuracy in Wireless Local Area Network (WLAN). The performance evaluation demonstrates that RF fingerprints from homogeneous devices maintain better similarity in a crowdsourced database. By keeping the reference of user device information, better positioning accuracy can be achieved with less amount of training data.