Physical-Layer Security via Radio Frequency Fingerprint Disguise: A Physics-Inspired Anti-Identification Framework

With the rapid advancement of wireless communication technology, critical devices are increasingly at risk of malicious attacks and identity compromise. To protect these critical devices from potential threats effectively, this paper introduces a novel physical radio frequency fingerprint disguise (RFFD) and anti-identification method designed to preserve these vital devices from malicious radio frequency fingerprint identification (RFFI) attacks. Our approach begins with nonlinear modeling of the transmitter’s (Tx) radio frequency fingerprint (RFF), followed by applying inverse and forward models to the digital baseband signal. This strategy effectively masks the RFF of critical Tx while simulating the RFF of non-essential Tx. Additionally, we develop a multi-channel convolutional long short-term memory auto-encoder (MCLAE)-based RFFD technique. We simulated twelve different types of Tx using MATLAB Simulink and validated the effectiveness of the disguise methods on a real-world testing platform. Experimental results within the RFFI framework indicate that our RFFD method achieves a higher target success rate (TSR) and maintains lower bit error rates (BER) and normalized mean square error (NMSE) compared to conventional model-based methods. Notably, the MCLAE-based approach performs better in disguising capabilities and preserving stealthiness under various scenarios, including different disguise targets and channel conditions, making it the most effective among the tested RFFD methods.