Radar PRI modulation recognition based on ResNet-LSTM model

The radar pulse repetition interval (PRI) is a key parameter of the radar system, and its modulation pattern recognition is an important means of analyzing the working status and tasks of the radar. However, traditional PRI modulation recognition methods face difficulties in manual feature extraction and low recognition success rates in noisy environments. To solve this problem, a radar PRI modulation recognition method based on a combination of residual network (ResNet) and long short-term memory network (LSTM) is proposed, which can effectively identify 7 common PRI modulation types including complex modulation. The experimental results under different degrees of pulse loss and interference show that the method proposed in this paper has better accuracy and robustness compared to existing algorithms.