摘要
为提高传统卫星姿态控制系统精度,提出了一种基于小脑模型(CMAC)神经网络的比例-积分-微分(PID)的复合控制器。给出了具在线学习功能的复合控制器结构,并证明了神经网络学习收敛条件与最终控制目标的一致性。仿真结果表明,设计的控制器具有较好的自适应性和鲁棒性。与传统控制器相比,进入稳定状态的速度更快,指向精度更高。
To improve the control accuracy of traditional attitude controller for satellite, a CMAC neural network PIE） hybrid attitude controller was put forward in this paper. The construction of the controller with the online learning ability was given out. The consistency of learning convergence condition of neural network and final control target was proved. The simulation results showed that this controller had good adaptability and robustness, and was faster to realize the stable state with higher accuracy than traditional controller.
出处
《上海航天》
北大核心
2007年第1期38-41,共4页
Aerospace Shanghai
关键词
卫星姿态控制
神经网络
比例-积分-微分控制器
控制精度
动态特性
Satellite attitude control
Neural network
Proportion-integral-differential controller
Control accuracy
Dynamic characteristic
