Open-Set RF Fingerprinting via Improved Prototype Learning
This work addresses a domain-specific challenge in RF fingerprinting by moving beyond closed-set assumptions, though it appears incremental as it builds on existing prototype learning methods.
The paper tackles the problem of open-set radio frequency fingerprinting, where signals from unknown devices not seen during training must be recognized, and proposes improvements to prototype learning that achieve promising performance on a real-world dataset.
Deep learning has been widely used in radio frequency (RF) fingerprinting. Despite its excellent performance, most existing methods only consider a closed-set assumption, which cannot effectively tackle signals emitted from those unknown devices that have never been seen during training. In this letter, we exploit prototype learning for open-set RF fingerprinting and propose two improvements, including consistency-based regularization and online label smoothing, which aim to learn a more robust feature space. Experimental results on a real-world RF dataset demonstrate that our proposed measures can significantly improve prototype learning to achieve promising open-set recognition performance for RF fingerprinting.