Online Trajectory Planning for Hypersonic Glide Vehicle under Multiple No-Fly Zones: An Attention Mechanism-Based BiGRU Framework

In trajectory planning for hypersonic glide vehicles under multiple no-fly-zone constraints, existing methods struggle to achieve effective avoidance due to limitations in the vehicle's maneuverability. In addition, the computational burden of iterative planning hinders their applicability for online implementation. To address the avoidance of dense no-fly zones, a waypoint generation method based on no-fly zone tangents is designed to produce feasible paths. Subsequently, the threat-aware spatial ivy algorithm is proposed for offline feasible path decision-making. Then, to mitigate computational delays while ensuring prediction accuracy, a bidirectional gated recurrent unit network incorporating a dual attention mechanism is developed for online control decision-making. Numerical simulations validate the reliability of the proposed method for online no-fly zone avoidance and the accuracy of bank angle prediction.