Radar Signal Recognition Based on Squeeze-and-Excitation Networks

Classification and recognition of radar signals is an important part of modern warfare. Traditional radar signal recognition method based on the pulse description word may be not suitable for the complex electromagnetic environment. In this paper, a novel radar signal recognition method based on Squeeze-and-Excitation networks is proposed. Firstly, time-domain, frequency-domain and time-frequency-domain features of five kinds of radar signals are transformed into images. Then the images are classified by convolutional neural networks and output the result of each domain. Finally, the results of the three domains are combined to obtain the correct identification results. And the recognition effects of three different modules are compared. The experiments prove the efficiency of our method, especially on different SNR condition and center frequency. And confusion matrixes demonstrate that under the low SNR condition, our method still has a good recognition effect and are better than traditional radar signal recognition method based on PDW.