Machine learning-based radar waveform classification for cognitive EW

In this paper, we propose a waveform classification approach for cognitive electronic warfare applications in which a supervised classification method is presented in an efficient framework. In this manner, we introduce an end-to-end framework for detection and classification of radar pulses. Our approach is complete, i.e., we provide raw radar signal at the input side and produce categorical output in the end. We use short-time Fourier transform to obtain time-frequency image (TFI) of the signal. Hough transform is used to detect pulses in TFIs. Convolutional neural networks (CNN) are used for intra-pulse modulation classification. In experiments, we provide supervised classification results at different signal-to-noise ratio (SNR) levels and achieve 98.08% classification accuracy for 10 dB SNR on a diverse set of both frequency- and phase-modulated signals. The method sustains high classification accuracy levels as [93.9%;85.83%] for 0 and -10 dB SNR, respectively, that signifies the robustness of the method against noise for low-power intercepted signals. Simulation results also show that CNN outperforms artificial neural networks in intra-pulse modulation classification task.

Automated summary

The paper proposes a waveform classification approach for cognitive electronic warfare applications, achieving high classification accuracy in experiments and demonstrating strong robustness against noise for low-power intercepted signals. Additionally, simulation results indicate that CNN outperforms artificial neural networks in intra-pulse modulation classification task.