An ANN based approach for wireless device fingerprinting

Abstract

Device fingerprinting is a technique to characterize or identify a device by using collected information about the device. Wireless device fingerprinting plays an important role in finding counterfeit devices in a wireless network. Fingerprinting methods can be active by sending some information to the target device; or by observing the information sent by the target device which is known as passive fingerprinting. In this paper, a technique of fingerprinting wireless devices based on Artificial Neural Networks has been presented. The parameters used in our work are transmission time and frame inter-arrival time. Our technique classifies unique devices from GTID and Sigcomm2008 datasets using frame inter-arrival time and transmission time and then the results are presented. Using these technique accuracies of 92.3% and 95.8% have been achieved by considering inter-arrival time and transmission time respectively.