RF Fingerprinting Using Artificial Neural Networks

Abstract

Wireless communications systems are highly susceptible to spoofing attacks where attackers attempt to impersonate other stations. This project applies the technique of Radio Frequency (RF) fingerprinting to verify the identity of wireless devices at the physical layer. While prior research into RF fingerprinting has looked at hand-crafted features, this project takes a deep-learning approach by training artificial neural networks to perform classifications of raw I/Q samples captured by a software defined radio. A proof-of-concept system using commercial off-the-shelf hardware for automatic identification of spoofed beacon packets in a realistic wireless environment yields greater than 99% accuracy over an AWGN channel with -10dB SNR. These results show substantial improvement over prior fingerprinting schemes relying on hand-crafted features and open the way for applying RF fingerprinting to a diverse range of wireless security and cognitive radio applications.