New Frontier of Communication Security on Radio Frequency Fingerprints Concealment

Due to device-specific defects introduced during the hardware manufacturing process, the radio frequency fingerprint (RFF) can be extracted to identify wireless devices and further avoid spoofing attacks. Many effective RFF identification methods have been proposed based on either machine learning or deep learning. However, from the perspective of communication security, if the RFF of the transmitter can be easily extracted and identified, attackers can disguise themselves as legitimate transmitters by impersonating RFF and other means, thereby undermining the security of wireless communications. Therefore, concealing the RFF of legitimate transmitters from detection and camouflage attacks has become a highly challenging issue in the field of wireless communications. This article presents an active RFF concealment (RFFC) method, which removes the nonlinear features of the transmitter system, thereby preventing attackers from obtaining the transmitter's RFF and ensuring the identity security of the transmitter. To evaluate the performance of RFF concealing technology, we simulate seven types of RFFC systems, and collect datasets without and with RFFC technology. The simulation results show that the performance of traditional transmitter identification methods decreases sharply after RFFC. Especially in low signal-to-noise ratio environments and complex multipath channel conditions, the proposed RFFC technology makes the RFF features chaotic and difficult to detect, leading to dramatically reduced effectiveness of existing transmitter identification methods.