 |
Recent research classifies radio signals between drones and controllers using RF signals processed with SDR. The signals are converted into spectrograms and analyzed with CNNs to categorize them. One study [16] achieved 98% accuracy in drone signal classification with CNN and power-based spectrograms at an SNR of -15 dB. The top right picture shows distinct signal patterns for each drone model, enabling effective classification. Below is the result when noise is added to simulate a real-world environment. |
IEEE/ICACT20250020 Slide.11
[Big Slide]