A New Deinterleaving Approach Based On Clustering and PRI Type Recognition

In an electronic warfare environment, numerous radars operate, each designed with distinct signal waveforms tailored to fulfill specific missions. The deinterleaving of radars is a fundamental function of an electronic warfare system. Following deinterleaving, identifying the Pulse Repetition Interval (PRI) modulation type becomes essential for enhanced radar recognition and understanding of its function. In this study, a new clustering and rule-based method is proposed to deinterleave radar pulses and recognize the PRI modulations. Ordering Points to Identify Clustering Structure (OPTICS) method is employed to cluster the radar signals. Difference of Time of Arrival (DTOA)-based method is employed to find the PRIs in the clustered data. A rule-based method is used to determine the PRI type from the obtained PRI values. In the experiments, the clustering and PRI type recognition phases were analyzed separately. The performance of OPTICS was tested under different conditions: (i) high cluster counts, (ii) close proximity of clusters, (iii) different cluster densities and forms, and showed good results in clustering. The PRI type detection performance was also tested on a simulation dataset consisting of 4 different PRI types (constant, agile, stagger and dwell&switch). The results indicate that the new method is effective in determining the PRI modulations.