Specific Radar Recognition Based on Characteristics of Emitted Radio Waveforms Using Convolutional Neural Networks

Area of Science:

• Electronic Warfare
• Signal Processing
• Artificial Intelligence

Background:

• Modern battlefields feature increasingly complex electromagnetic environments with advanced radar systems.
• Traditional radar signal recognition methods struggle with agile waveforms, posing challenges for electronic warfare (EW) systems.

Purpose of the Study:

• To propose and evaluate a novel recognition method for emitted radar signals with agile waveforms.
• To address the limitations of traditional models in identifying complex radar signals in EW.

Main Methods:

• Development of a radar signal recognition method utilizing Convolutional Neural Networks (CNNs).
• Signals were captured by electronic recognition receivers, digitized, and processed.
• A simulation environment with a signal generator was used to test the CNN's performance.

Main Results:

• The proposed CNN method demonstrated effective recognition of raw radar signals with agile time waveforms.
• High recognition probabilities, ranging from 92% to 99%, were achieved in simulation.
• The method proved effective in noisy conditions and environments with multiple interfering radar signals.

Conclusions:

• The CNN-based approach offers a robust solution for identifying agile waveform radar signals in challenging EW scenarios.
• The study validates the effectiveness of CNNs for real-time radar signal recognition and processing.
• Further development of learning and processing algorithms is possible for enhanced performance.