基于在线逼近的不确定非线性系统自适应控制研究

项目名称： 基于在线逼近的不确定非线性系统自适应控制研究

项目编号： No.61463019

项目类型： 地区科学基金项目

立项/批准年度： 2015

项目学科： 自动化技术、计算机技术

项目作者： 孙刚

作者单位： 江西理工大学

项目金额： 45万元

中文摘要： 在线逼近自适应控制是利用通用逼近器的学习性能消除系统不确定性对控制器的影响，进而确保被控系统稳定。然而，不确定非线性系统自适应控制中普遍存在设计复杂、结构循环依赖和限制条件过多等缺陷，这对研究高度不确定非线性系统的自适应控制起到阻碍作用。本项目将基于这些缺欠展开讨论。首先，针对不确定下三角非线性系统，采用单一在线逼近与动态面控制技术简化自适应控制器的复杂性，使其结构简单且容易实现。其次，对不满足下三角形式的不确定非线性系统，依据动态面控制技术探究自适应反步控制器设计中的结构循环依赖问题。此外，针对含有死区、扰动及时滞的不确定非线性系统设计鲁棒自适应控制器，使其能有效地应用于实际系统控制中。最后，为保证控制器的可靠性，通过Lyapunov稳定性、输入状态稳定性及小增益定理等方法对闭环系统进行稳定性分析，并利用数值仿真技术对控制系统进行模拟验证。

中文关键词： 在线逼近；自适应控制；鲁棒控制；不确定非线性系统

英文摘要： In order to ensure the stability of controlled systems, online approximation technique is incorporated into adaptive control design framework by using the learning performance of universal approximator to eliminate the influence of system uncertainties. There are some drawbacks exist in adaptive control design of uncertain nonlinear systems, such as complex in design, circular construction of controller, too many constraints, etc., which limit the development of adaptive control of nonlinear systems with high uncertainties. This project aims to investigate these issues. Firstly, for uncertain lower triangular nonlinear systems, techniques of single online approximation and dynamic surface control are used to simplify the complexity of adaptive controllers, such that the designed controllers are simple in structure and easy in implementation. Secondly, based on dynamic surface control technique, the problem of circular structure of adaptive backstepping controller is investigated for uncertain nonlinear systems with non-lower triangular form. In addition, in order to make the designed controller effectively applied to actual systems, robust adaptive control approaches are developed for uncertain nonlinear systems with dead zone, disturbances, and time delays. Finally, to guarantee the reliability of the controllers, Lyapunov stability, input state stability, and small gain theorem are employed to analyze stability of closed-loop systems, and numerical simulation technology is used to model the control systems.

英文关键词： Online Approximation;Adaptive Control;Robust Control;Uncertain Nonlinear Systems