摘要
对一类不确定非线性系统,将反推控制和神经网络相结合,研究了其鲁棒渐近镇定控制问题。与通常研究中被控对象仅局限于严格反馈形式相比较,研究对象更具一般性。基于反推控制方法来构造镇定控制器,利用神经网络来逼近控制器构造过程中产生的不确定项,并提出一种新的自适应算法来在线调节神经网络权值。通过一步步适当地选取虚拟控制器的参数和神经网络权值的自适应律,最终得到的控制器使得整个闭环系统是全局渐近稳定的。
The problem of robust stabilization control for a class of uncertain nonlinear systems is considered. Compared with the conventional controlled objects with strict feedback form, the controlled systems considered in this paper are more general. The robust stabilization controller is designed based on Back- stepping mechanism, which employs neural networks to approach the uncertainty incurred during the process of controller design. The weights of neural networks are updated on-line with a new adaptive algorithm. By choosing appropriate parameters of the virtual controllers and the adaptive law of neural networks step-by-step, a stabilization controller is achieved making the closed-loop systems globally asymptotically stable.
出处
《南京邮电大学学报（自然科学版）》
2010年第2期77-80,共4页
Journal of Nanjing University of Posts and Telecommunications：Natural Science Edition
基金
江苏省高校自然科学基础研究项目(KJD510150)资助项目
关键词
非线性系统
神经网络
鲁棒控制
反推
nonlinear systems
neural networks
robust control
backstepping
