探索了利用深度强化学习算法训练智能体,以代替人类工程师进行火箭姿态控制器参数的离线设计方案。建立了多特征秒的火箭频域分析模型,选定了设计参数。选择深度强化学习算法中的双深度Q学习(Double Deep Q Network,DDQN)算法,通过记忆...探索了利用深度强化学习算法训练智能体,以代替人类工程师进行火箭姿态控制器参数的离线设计方案。建立了多特征秒的火箭频域分析模型,选定了设计参数。选择深度强化学习算法中的双深度Q学习(Double Deep Q Network,DDQN)算法,通过记忆回放和时间差分迭代的方式让智能体在与环境交互过程中不断学习。设计了对应的马尔科夫决策过程模型,进行了智能体的训练和前向测试。结果说明该方法对于运载火箭姿控设计具有一定参考价值。展开更多
This paper investigates the attitude and orbit control for the combined spacecraft formed after a target spacecraft without the autonomous control ability is captured by a service spacecraft.The optimal controller of ...This paper investigates the attitude and orbit control for the combined spacecraft formed after a target spacecraft without the autonomous control ability is captured by a service spacecraft.The optimal controller of fully-actuated system is proposed to realize the attitude and orbit stabilization control of combined spacecraft.The stability of the system is proved by introducing Lyapunov function.Numerical simulation of the combined spacecraft and physical experiment based on the combined spacecraft simulator(CSS)are completed.Both simulation and experiment results demonstrate the effectiveness and practicability of the optimal controller of fully-actuated system.展开更多
文摘探索了利用深度强化学习算法训练智能体,以代替人类工程师进行火箭姿态控制器参数的离线设计方案。建立了多特征秒的火箭频域分析模型,选定了设计参数。选择深度强化学习算法中的双深度Q学习(Double Deep Q Network,DDQN)算法,通过记忆回放和时间差分迭代的方式让智能体在与环境交互过程中不断学习。设计了对应的马尔科夫决策过程模型,进行了智能体的训练和前向测试。结果说明该方法对于运载火箭姿控设计具有一定参考价值。
基金This paper was supported in part by the National Natural Science Foundation of China under Grant Nos.62173255 and 62188101.
文摘This paper investigates the attitude and orbit control for the combined spacecraft formed after a target spacecraft without the autonomous control ability is captured by a service spacecraft.The optimal controller of fully-actuated system is proposed to realize the attitude and orbit stabilization control of combined spacecraft.The stability of the system is proved by introducing Lyapunov function.Numerical simulation of the combined spacecraft and physical experiment based on the combined spacecraft simulator(CSS)are completed.Both simulation and experiment results demonstrate the effectiveness and practicability of the optimal controller of fully-actuated system.
