Radio frequency fingerprint (RFF) identification is an emerging device authentication technique that leverages hardware imperfections arising from the manufacturing process. Due to the strong time-varying characteristics of the UAV channels, it is difficult to apply the existing RFF identification methods to the UAV. In this paper, we propose a data augmentation method on the raw IQ signal, which use to enhance the robustness of the pre-trained model when extracting RFF in the presence of severe fast fading in the real flight environment. By employing this technique, we are able to extract robust features on the UAV signals. We evaluated the performance of our method in diverse environments, including UAV hovering state and moving state. In the hovering and moving state of UAV, using data augmentation can improve the average recognition accuracy by 8.2%.
Robust Radio Frequency Fingerprint Identification for UAVs During Fast Fading Channels
22.03.2024
1959349 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch