Airborne radar operating status recognition based on the CIB-ResNet2

The traditional way to identify the working state of radar is to extract pulse descriptors from the received electromagnetic signals, and the input vectors are identified by machine learning or deep learning models. This paper proposes a CIB-ResNet2 recognition method, which directly performs spectrum change and recognition of electromagnetic signals based on electromagnetic spectrum sensing technology, without extracting pulse descriptors, extracting the characteristic curve by the circular integral bispectrum (CIB), and input ResNet2 model for identification, the simulation results show that the proposed method can also achieve high recognition accuracy under low signal-to-noise ratio conditions, and the method has broad industrial application prospects.