RFF Identification with CGAN-based Channel Estimation in Time-Varying Channels

Radio Frequency Fingerprinting (RFF) has emerged as an effective solution for enhancing device security. However, its performance is often degraded by interference from time-varying multipath channels. To mitigate this issue, we propose a Conditional Generative Adversarial Network (CGAN)-based channel equalization method, which improves channel estimation by simulating extended preambles. Furthermore, we integrate a multi-channel attention network for feature extraction, utilizing four distinct signal representations, with estimated channel responses incorporated to adapt to varying channel conditions. Experimental results demonstrate that our model is robust against time-varying channels and achieves high classification accuracy.