时变非线性系统直接自适应迭代学习控制  

Direct adaptive iterative learning control of time-varying nonlinear systems

作　　者：李静[1,2] 胡云安[1] 

机构地区：[1]海军航空工程学院控制工程系 [2]中国人民解放军91055部队

出　　处：《系统工程与电子技术》2012年第1期154-159,共6页Systems Engineering and Electronics

基　　金：国家自然科学基金(61004002)资助课题

摘　　要：针对一类含有非周期时变参数化不确定性的非线性系统,设计了一种新的迭代神经网络估计器,解决了非周期时变不确定性带来的设计难题。用迭代神经网络直接对期望控制量进行整体逼近,利用Lyapunov稳定性理论和自适应迭代学习控制技术设计了控制器,并进行稳定性分析,证明了系统所有状态量有界,且输出量将收敛至期望轨迹的一个邻域内。仿真结果验证了控制器设计方案的正确性。For a class of time-varying nonlinear systems with non-periodically parameterized uncertainties,a new iterative learning neural network approximator（ILNNA） is designed to solve the design difficulties brought by the non-periodical time-varying uncertainties.Subsequently,the ILNNA is directly applied to approximating the desired controller.At the same time,the controller is designed by the adaptive iterative learning control technique.Corresponding stability theorem is obtained according to Lyapunov stability theory.Then the conclusion that all state variables are bounded and the output will converge to a neighborhood of the desired trajectory can be obtained.Simulation results verify the correctness of the proposed scheme.

关 键 词：时变不确定性 迭代学习控制 神经网络 自适应控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]