Optimizing Micro Agile Earth Observation Satellite Mission Planning with MOPSO-GA and hp-RPM

Abstract

As a cost-effective microsatellite platform characterized by rapid attitude maneuverability, mission planning for micro agile Earth observation satellite (MAEOS) requires efficient scheduling of high-resolution observation sequences under resource constraints. Metaheuristics are suitable for solving discrete combinatorial optimization problems in satellite mission planning. However, they remain constrained by computational inefficiency, premature convergence, and parameter sensitivity, particularly in handling the fundamental multiobjective trade-off between maximizing imaging quality and minimizing energy consumption. To effectively mitigate these combinatorial challenges and convergence limitations, a hybrid multiobjective particle swarm–genetic algorithm (MOPSO-GA) is proposed in this study through a bidirectional coevolutionary mechanism that integrates the velocity-driven global exploration of particle swarm optimization (PSO) with the discrete sequence optimization of the genetic algorithm (GA). Additionally, an adaptive hp-Radau pseudospectral method (hp-RPM), where h denotes mesh-interval refinement and p denotes polynomial-order refinement, is embedded to refine mesh density dynamically and polynomial order for high-fidelity attitude trajectory optimization to ensure physical feasibility. Rigorous simulations demonstrated that MOPSO-GA significantly outperforms state-of-the-art methods in Pareto solution quality, convergence speed, and robustness across diverse MAEOS scenarios, establishing a novel paradigm for discrete–continuous cooperative planning in agile satellites.