Trajectory Design and Robust Guidance for Solar Sail Spacecraft under Complex Uncertain Conditions
In response to the complex and uncertain conditions faced by solar sail spacecraft during in-orbit flight,a deep reinforcement learning-based integrated algorithm for trajectory design and robust guidance is proposed.The uncertainties conditions acted as solar radiation pressure model uncertainty,navigation er-ror,control execution error and randomly triggered safety events are incorporated into the Markov decision process modeling of solar sail spacecraft in-orbit flight,based on the algorithm proposed regarding orbital dynamics of solar sail spacecraft.A reward function refllecting the optimization of solar sail energy supply is designed by the minimum solar phase angle,and training is conducted by using the proximal policy optimi-zation algorithm to achieve the optimization design of solar sail spacecraft trajectories and robust guidance under complex and uncertain conditions.This algorithm is applied to the heliocentric transfer mission of a solar sail spacecraft exploring the near-Earth asteroid 2019 GF1.Simulation results show that the terminal arrival precision of nominal trajectory tracking fllight under uncertain conditions can be decreased and the solar phase angle along the trajectory is reduced by using this new algorithm.
Solar sail spacecraftTrajectory designRobust guidanceUncertaintiesDeep reinforcement learning
NETL
NSTL
万方数据
