Neural network adaptive robust attitude control of spacecraft
Neural network adaptive robust attitude control of spacecraft
复制标题
航天器神经网络自适应鲁棒姿态控制
DOI：
10.1109/icicisys.2009.5358286
复制
发表时间：
2009
期刊：
影响因子：
--
通讯作者：
Chaoyang Dong
中科院分区：
文献类型：
--
作者：
Xiyuan Huang;Qing Wang;Chaoyang Dong
设计了一种用于航天器姿态跟踪的神经网络自适应鲁棒控制器。在存在参数不确定性和时变的情况下，考虑了执行器的动态特性和外部干扰，实现了带有变速恒模控制器的刚体航天器的姿态跟踪。它由一个非线性反馈补偿器、一个神经网络估值器和一个鲁棒控制器构成。利用三层神经网络对含有不确定参数的未知函数进行估计和消除。该控制器用于增强控制器对近似误差和外部干扰的鲁棒性。李亚普诺夫稳定性理论被认为是用控制律来保证闭环系统动态稳定的。结果表明，所提出的控制器保证了跟踪误差的一致最终有界性和对干扰的鲁棒性。
A neural network adaptive robust controller for spacecraft attitude tracking is developed in this paper. The proposed controller achieves attitude tracking for a rigid spacecraft with a VSCMG cluster in the presence of parametric uncertainties and time-variation, taking into account dynamics of actuator and external disturbance. It is constructed by a nonlinear feedback compensator, a neural network estimator and a robust controller. The three-layered neural network is used to estimate and eliminate the unknown function with uncertain parameters. The robust controller is used to enhance the robustness of controller to approximate error and external disturbance. Lyapunov stability theory is considered to guarantee the stability of the closed-loop dynamics with the control law. As a result, the proposed controller ensures the uniform ultimate boundedness in tracking error and robustness to disturbance.