Identification of Wireless Devices of Users Who Actively Fake Their RF Fingerprints With Artificial Data Distortion

Variations in the RF chain of radio transmitters caused by imperfections of manufacturing processes can be used as a signature to uniquely associate wireless devices with a given transmission. In our previous work, we proposed a model-based approach that allows for identification of wireless devices based on signatures obtained with time domain analysis of a pair of received and decoded signals. Here, we consider strong adversaries who intentionally introduce distortions to the data symbols before the symbols are exposed to the transmitter's inherent nonlinearities, with the intention of faking the signatures of their devices while still allowing for proper data decoding. The method proposed in this work is based on spectral analysis and on the observation that nonlinear components cause in-band distortion and spectral regrowth of the signal that is dependent on the parameters of the nonlinearity. Hence, by analysis of the in-band distortion of the spectrum as well as the spectral regrowth, we show that wireless devices can be successfully identified even when the users are digitally modifying their data symbols. The utility of the proposed identification approach is demonstrated with simulations based on parameters obtained from the measurements of commercially employed WLAN RF transmitters.