摘要
针对不确定非线性混沌系统,提出了一种基于动态神经网络辨识器的自适应跟踪控制新方法.通过滑模控制技术在线调整动态神经网络辨识器权值,并在获取动态神经网络模型的基础上设计出优化控制器,实现混沌系统的轨道跟踪.对辨识误差和轨道跟踪误差进行分析并证明了它们的有界性.Lorenz混沌系统的仿真实验结果表明了控制策略的有效性.
An adaptive tracking controller based on dynamical neural network identifier for uncertain nonlinear chaos systems is presented. The weights of the dynamic neural networks used as neuro-identifier can be on-line (adjusted) through the usage of the sliding mode technique. An optimal controller via dynamic neural network model is (presented) to perform reference trajectory following control for chaotic system.The identification error and the (trajectory) tracking error are analyzed and guaranteed to be bounded. The experiment results of the chaotic system given by Lorenz equation show the effectiveness of the method.
出处
《控制与决策》
EI
CSCD
北大核心
2004年第4期455-458,共4页
Control and Decision
基金
国家自然科学基金资助项目(60075008
60102010)
湖南省自然科学基金资助项目(02JJY2095
03JJY3107).
关键词
混沌
动态神经网络
滑模控制
非线性系统
稳定性
chaos
dynamic neural networks(DNN)
sliding mode control
nonlinear system
stability