Pulse Deinterleaving for Multifunction Radars With Hierarchical Deep Neural Networks

Multi-function radars (MFRs) work with pulse groups, and pulses in different groups are weakly correlated, which greatly increases the difficulty for MFR deinterleaving, especially cases of significant data noises. At present, no relevant research results have been reported to address this problem. In article, a hierarchical deep learning model will be established describe sequential patterns trains from MFRs. The bottom layer represents discrete continuous vectors, describes temporal pattern makes them machine readable. top uses recursive neural network mine between consecutive groups. two layers then synthesized form model, is able semantic correlations within trains, parameters subsequent their inner can predicted based on preceding pulses. Based article proposes an iterative deinterleaving method parallel one Simulation demonstrate that, proposed methods perform satisfyingly separating interleaved same or types, they robust