Radar Waveform Recognition Based on Time-frequency Images Under Deteriorating Channel Conditions

In the process of electronic countermeasures, most studies of radar signal recognition focus on Additive White Gaussian Noise (AGWN) channel scenarios, while pay less attention to the influence of channel deterioration on radar signal recognition. Therefore, we propose a low interception rate radar signal recognition method based on time-frequency images under channel impairment. First, we obtain shallow features of LPI radar signals in fading channels by Choi-Williams distribution (CWD) time-frequency transform that converts the signals to time-frequency images. Second, we design a parallel residual network model for feature learning and signal classification to realize the automatic modulation recognition of radar waveforms under channel fading conditions. Simulation results show that the proposed method effectively achieves the recognition of 12 kinds of radar waveforms both in AGWN channels and in multipath channels, which demonstrates the superiority of this model.