Coordinated Decision-Making of Heading and Morphing for a Morphing Reusable Launch Vehicle via Multi-Agent Reinforcement Learning

The application of single agent reinforcement learning faces significant challenges in coordinated decision-making of heading and morphing for morphing reusable launch vehicles. By applying mutli-agent reinforcement learning, this paper proposes a coordinated decision-making method that is capable of determining heading and morphing command simultaneously. Firstly, the coordinated decision-making framework is established specifically for the morphing reusable launch vehicle. The heading and morphing decision-making agents are designed to execute the mission of reaching designated location and avoiding no-fly zones along trajectory. A multi-stage training strategy is developed to accelerate the training process. Simulation results show the effectiveness and superiority of the proposed method over the coordinated decision-making method based on single agent reinforcement learning. Compared with the fixed-wing reusable launch vehicle, the morphing reusable launch vehicle with the proposed method performs less number of bank angle reversal.