Research on Intelligent SEI Method for Navigation Radar Based on Metric

Specific emitter identification (SEI) of radars is an important function of electronic reconnaissance. Aiming at the individual identification of navigation radars, this paper provides a method combining metric learning and deep learning. That method extracts the front & back edge and instantaneous frequency of pulse envelop based on original pulse waveform data as initial input features, then combines the central loss function and cross entropy loss function to training the deep neural network so as to extract the individual features, at last makes identification to radar individuals through the optimized fully connected classification network. The test results show that the proposed method can effectively improve the performance of critical parameters including cluster effect of radar individual features and correct individual identification rate etc.