Reinforcement Learning-Based Trajectory Planning for Tracking and Rendezvous of Small Spacecraft

To address the problem of multi-mission conflicts in trajectory planning for small spacecraft, a reinforcement learning-based trajectory planning method for tracking and rendezvous has been proposed. First, a reinforcement learning mission su-pervisor (RLMS) was designed based on a null-space behavioral control framework. Then, during the offline training phase, the mission supervisor learns an optimal mission priority assignment strategy, while in the online phase, only a lightweight pre-trained neural network is required to be invoked, enabling adaptive trajectory planning. The proposed RLMS can intelligently adjust task priorities according to the environment, overcoming the limitations of fixed-rule priority schemes. Finally, comparative numerical simulations with a finite state automate mission supervision demonstrate the effectiveness and superiority of the proposed method.