@article{ZHENG2025, 
author = {Yunfei ZHENG and Xuejun ZHANG and Yuanhao TAN and Xueyuan LI},
title = {Transformer-based identification for ADS-B transmitters in open–time sets},
year = {2025},
journal = {Chinese Journal of Aeronautics},
volume = {38},
number = {8},
keywords = {Automatic Dependent Surveillance-Broadcast, Radio frequency fingerprinting, Identification, Open-time set, Time-frequency feature diagram, Swin Transformer},
url = {https://www.sciopen.com/article/10.1016/j.cja.2025.103418},
doi = {10.1016/j.cja.2025.103418},
abstract = {Radio Frequency Fingerprint Identification (RFFI) technology provides a means of identifying spurious signals. This technology has been widely used in solving Automatic Dependent Surveillance–Broadcast (ADS-B) signal spoofing problems. However, the effects of circuit changes over time often lead to a decline in identification accuracy within open-time set. This paper proposes an ADS-B transmitter identification method to solve the degradation of identification accuracy. First, a real-time data processing system is established to receive and store ADS-B signals to meet the conditions for open-time set. The system possesses the following functionalities: data collection, data parsing, feature extraction, and identity recognition. Subsequently, a two-dimensional Time-Frequency Feature Diagram (TFFD) is proposed as a signal pre-processing method. The TFFD is constructed from the received ADS-B signal and the reconstructed signal for input to the recognition model. Finally, incorporating a frequency offset layer into the Swin Transformer architecture, a novel recognition network framework is proposed. This integration can enhance the network recognition accuracy and robustness by tailoring to the specific characteristics of ADS-B signals. Experimental results indicate that the proposed recognition architecture achieves recognition accuracy of 95.86% in closed-time set and 84.33% in open-time set, surpassing other algorithms.}
}