Radar Signal Identification Using AI

RAIS identifies different radar submodes characterized by pulse width, pulse repetition interval, and intrapulse modulation type using AI-based techniques. The system operates in the 13–43 dB SNR range and classifies five submodes including constant RF, LFM chirps, and BPSK. A CNN and an MLP are deployed on an AMD Artix-7 FPGA via UART communication. The CNN achieves 100% accuracy and MLP achieves 99.5% on hardware, with the MLP satisfying the 150 µs latency requirement at 120 µs inference time, making it the primary candidate for EHSIM's ESM system integration.