Intelligent Perturbation-Driven control methodology for Low-Thrust satellite constellation systems

This study proposes a closed-form dynamical modeling and intelligent control methodology for low-thrust RGT-Walker satellite constellations. By analytically incorporating the first-order secular J 2 perturbations, the presented dynamic framework enables unified ground track maintenance and precise constellation-wide control based on the reference trajectory. The hybrid Intelligent Perturbation-Driven (IPD) Control system, combining Proportional-Derivative (PD) feedback and real-time environmental disturbance learning via Long Short-Term Memory (LSTM) networks, maximizes adaptive thrust scheduling and fuel efficiency. Numerical simulations under both nominal and worst-case disturbance scenarios validate stable orbit maintenance, 50 % reduction in propellant consumption, and a twofold extension of mission lifetime. This methodology offers a robust and scalable academic foundation for advancing the autonomy, efficiency, and precision of next-generation low-thrust electric propulsion satellite constellation operations.