Safe Hierarchical Reinforcement Learning for CubeSat Task Scheduling Based on Energy Consumption

• DOI: 10.48550/arxiv.2309.12004
• Publication Date: 2023-01-01
• Authors: Mahya Ramezani; M. Amin Alandihallaj; José Luis Sánchez...; Andreas Hein
• Keywords: CubeSat

This paper presents a Hierarchical Reinforcement Learning methodology tailored for optimizing CubeSat task scheduling in Low Earth Orbits (LEO). Incorporating high-level policy global distribution and low-level real-time adaptations as safety mechanism, our approach integrates the Similarity Attention-based Encoder (SABE) prioritization an MLP estimator energy consumption forecasting. Integrating this mechanism creates safe fault-tolerant system scheduling. Simulation results validate superior convergence success rate, outperforming both MADDPG model traditional random across multiple configurations.