Domain Adaptation-Based Radio Frequency Fingerprint Identification for Industrial Cyber-Physical Systems

With the widespread adoption of cloud and fog platforms, security issues facing Industrial Cyber-Physical Systems (ICPSs) are increasingly prominent. Radio Frequency Fingerprinting (RFFI) emerges as a promising physical layer security technique for user authentication, effectively enhancing ICPS security over wireless networks by exploiting inherent hardware imperfections in transmitters. However, as technologies evolve and user environments become denser and more complex, existing RFFI techniques encounter significant domain shift challenges, leading to performance degradation. Therefore, developing domain-adaptive RFFI techniques is crucial for the effectiveness of ICPSs. In this article, we first provide a comprehensive review of the RFFI framework, emphasizing its applications in ICPSs and the key processes involved. Next, we focus on the domain shift issues faced by RFFI in ICPSs and present a comprehensive survey of domain adaptation methods. To tackle these challenges, we propose a novel feature-alignment-based domain adaptation (FADA) framework for RFFI, which is designed to mitigate the performance degradation caused by cross-domain shifts by aligning the feature representations of the source and target domain. Finally, we discuss the inherent challenges and future directions for developing domain adaptation-based RFFI to enhance the security of ICPSs.