A Deinterleaving Method for Mechanical-Scanning Radar Signals Based on Deep Learning

Radar signal deinterleaving has been deeply studied in the field of electronic reconnaissance. In this work, a deinterleaving method for mechanical-scanning radar signals based on deep learning is proposed. In this method, the pulse repetition interval (PRI) and pulse amplitude (PA) are used to deinterleave mechanical-scanning radar signals. A bidirectional gated recurrent unit (BGRU) is employed, and the difference of time of arrival (DTOA) and PA of the pulse stream are input into the BGRU. Based on the PA variation features of different radars, each pulse in the obtained pulse stream is classified, and the radar signals are deinterleaved. Compared to the PRI-based deinterleaving methods, the proposed method utilizes the two-dimensional information of radar signals. As a result, signal deinterleaving of different radars with the same PRI is achieved. Compared to other existing radar signal multiparameter-based deinterleaving methods, the proposed method is applicable to radar signals with complex features and to complex environments, and can utilize multiparameter in one step.