A potential game-based distributed task planning method for revisit period optimization in MEO remote sensing constellation

To address the complex coordination challenges of meeting revisit period requirements within medium Earth orbit (MEO) remote sensing satellite constellations under distributed architectures, a distributed task planning method is proposed. While revisit period is a traditional metric, achieving its global optimization through decentralized collaboration in MEO systems remains a critical and under-explored problem. The proposed method transforms the revisit period optimization into an observation time coverage problem and constructs a corresponding objective function to evaluate requirement satisfaction. At the single-satellite level, a Hybrid Coding Differential Evolution (HCDE) algorithm is developed to optimize task selection and imaging timing simultaneously. By integrating discrete and continuous decision variables into a unified evolutionary framework, the HCDE algorithm effectively handles the mixed-integer search space to ensure high-quality local planning. For multi-satellite collaboration, a task coordination strategy based on the improvement path idea is designed within a potential game framework. This strategy allows the constellation to reach a global Nash equilibrium through finite iterations, ensuring that individual satellite utility improvements align with the global optimization direction. Simulation results demonstrate that the proposed distributed method outperforms centralized planning in optimization effectiveness and offers a significant advantage in terms of computation time.