Radar Emitter Signal Pulse Repetition Interval Pattern Recognition Based on Physics-Informed Neural Networks and Spectrogram

Radar emitter signal identification is a critical task in electronic countermeasure reconnaissance. With the emergence of integrated sensing and communication (ISAC) systems, electronic warfare reconnaissance is evolving toward dual-function radar-communication signal processing. To address the challenge of identifying radar emitters in complex electromagnetic environments, this paper proposes a method that leverages spectrogram generated by a communications receiver and a physics-informed neural networks (PINN) to recognize pulse repetition interval (PRI) patterns. Simulation results demonstrate that the proposed approach achieves over 90% recognition accuracy under small-sample learning conditions. Experimental validation confirms its effectiveness, generalization capability, and practical applicability.