An RF Fingerprint Data Enhancement Method Based on WGAN

RF fingerprinting is an emerging technology in the field of IoT security and is widely used in many areas, such as management, wireless device authentication, and interference source determination (Hall et al. in IEEE Trans Dependable Secure Comput 201–206, 2004 [1]; Wu et al. in Sci China Inf Sci 65(7):170304, 2022 [2]; Lin et al. in Sci China Inf Sci 2023 [3]). Most of these application scenarios rely on recognition methods for devices. Most of the mainstream recognition methods are based on a large amount of data for training. In case of insufficient sample size, the mainstream recognition methods are not applicable. Generative adversarial networks (GANs), with their adversarial properties, are well-suited and effective for applications in scenarios where the amount of data is insufficient. In this paper, we propose an RF fingerprint data enhancement method based on Wasserstein Generative Adversarial Network (WGAN). The experimental results show that the method can effectively improve the accuracy of RF fingerprint recognition in the same and limited data set.