Unsupervised Identification Method of Radio-Frequency Fingerprint Based on Deep Clustering

The development of radio frequency fingerprint recognition technology is becoming increasingly mature, with abundant research and applications confirming the reliability of supervised radio frequency fingerprint technology. However, in real electromagnetic environments, particularly in areas such as intrusion detection and military confrontation, it is not always feasible to collect and analyze target signals in advance, manually annotate them, and enter the annotated signals into a fingerprint database, as obtaining authentic information about the target devices is often challenging. Additionally, unsupervised clustering of radio frequency fingerprints can identify the number of source devices, effectively detecting impersonation and identity-simulation attacks in wireless network security. In summary, research on unsupervised electromagnetic target blind recognition technology based on radio frequency fingerprints holds significant practical application significance. The method of electromagnetic target identification based on deep clustering aims to extract more effective radio frequency fingerprint features, considering the temporal characteristics of I/Q signals. This approach significantly enhances the precision of target recognition without requiring prior labeling, demonstrating superior performance in practical applications such as wireless network security, intrusion detection, and military operations. This method significantly improves the classification accuracy of radio frequency signals in real-world environments compared to previous studies.