Low-orbit Space Debris Warning and Autonomous Collision Avoidance for Space Environment Governance

Abstract With the increasing frequency of human space activities, the issue of space debris has become progressively more severe, posing a significant threat to spacecraft in orbit and highlighting the need for effective space environmental governance. This study addresses the challenges of debris warning and autonomous collision avoidance in low earth orbit by analyzing the limitations of current technologies and developing a more comprehensive problem model. In the context of space environmental governance, we design multiple optimization scenarios tailored to various avoidance situations, employing a range of reinforcement learning methods to solve and optimize these issues. These methods include the Monte Carlo Tree Search Algorithm, the Cross Entropy Method, and Proximal Policy Optimization in continuous action spaces. Our research not only provides an effective debris avoidance strategy for spacecraft but also contributes to the advancement of technological solutions for the sustainable management and protection of the space environment.