Online Trajectory Generator for Synthetic Aperture Radar Imaging Based on MOEA-TNets Framework

The synthetic aperture radar (SAR) imaging system offers high-resolution two-dimensional imaging and has wide applications. The imaging performance of SAR is linked to the observation configuration, which depends on the platform trajectory. During platform flight, the changing relative positions between the platform and observed targets result in varying observation configurations, affecting the imaging resolution. Therefore, real-time planning of platform trajectories is essential. To address the above issues, this paper models the problem as a multi-objective optimization problem that integrates comprehensive ballistic control and imaging performance. It proposes a multi-objective online search framework based on multi-objective evolutionary algorithms (MOEA) and temporal networks (TNets). This framework employs speed-constrained multi-objective particle swarm optimization (SMPSO) to plan trajectories, forming a dataset for regularized long short-term memory (R-LSTM) network training and enabling real-time trajectory planning. Experimental studies affirm the efficiency of the proposed method.