Lightweight Multi-Modal Few-Shot RF Fingerprint Recognition for Drones

With the increasing applications of drones in public safety, military reconnaissance, and commercial logistics, efficient and reliable identification of their identities has become a pressing issue. Addressing the challenges of sparse samples and diverse interference in complex electromagnetic environments, this paper proposes a drone RF fingerprint recognition method based on dual-modal few-shot learning. Researchers extract both timedomain and frequency-domain features by performing timefrequency transformation and statistical analysis on the original I/Q signals,. Experimental results demonstrate that the proposed method significantly outperforms baseline approaches in accuracy and maintains a high recognition rate even in mixed interference scenarios and low signal-to-noise ratios. This provides an effective and feasible technical solution for drone identity authentication in complex electromagnetic environments.