Deep Learning Approach to LPI Radar Recognition

In this study, an advanced automatic low probability of intercept (LPI) radar recognition technique (LWRT) that includes both LPI radar signal classification and parameter extraction is proposed. It is shown with Monte Carlo simulation that, even without the unrealistic assumptions used in the previous studies, the proposed LWRT achieves classification performance similar to that of the state-of-the-art LWRT for pulse wave (PW) LPI radar waveforms. And by the combination of the `single shot multi-box detector' (SSD) or `you only look once version 3' (YOLOv3) and a supplementary classifier, the proposed LWRT achieves an extraordinary classification performance for continuous (CW) LPI radar waveforms for all the twelve modulation schemes considered in the literature (i.e., BPSK, Costas, LFM, Frank, P1, P2, P3, P4, T1, T2, T3 and T4). Moreover, the proposed LWRT summarizes the existing and proposed new parameter extraction functions, which can help to design the countermeasure in electronic warfare.