非线性时滞系统的神经网络稳定控制方法研究

项目名称： 非线性时滞系统的神经网络稳定控制方法研究

项目编号： No.60864001

项目类型： 地区科学基金项目

立项/批准年度： 2009

项目学科： 金属学与金属工艺

项目作者： 王汝凉

作者单位： 广西师范学院

项目金额： 24万元

中文摘要： 本项目主要成果：（1）对一类含未知变时滞的非线性系统，引人新的基于自适应的神经网络控制方法，用RBF神经网络来逼近未知的连续函数，用积分型Lyapunov-Krasovskii 函数来消去未知的变时滞，在非线性闭环系统中实现所有信号的有界性，并证明系统的输出收敛到预期轨迹一个小的邻域内；(2) 研究一类非线性外部系统作用下的不确定扩展输出反馈系统的全局干扰抑制问题。利用系统输出变量信息和外部系统结构信息，构造一个新的内模方程以实现控制目标, 并且可以完全抑制非线性干扰输入；（3）通过Lyapunov函数和Razumikhin技术，获得一类带变时滞的脉冲神经网络系统一致有界和一致渐近稳定性的新判据。（4）利用拓扑理论和Lyapunov泛函研究了S—#20998;布时滞反应扩散Hopfield 神经网络平衡点的存在性及全局鲁棒指数稳定性。研究成果具有重要的理论价值和潜在的应用价值。 本项目撰写学术论文19篇，其中SCI待收录2篇, EI收录12篇，核心期刊6篇；邀请了国内外多个著名专家来校交流；项目组成员参加了数十次国际、国内学术会议交流；为研究生提供了一个很好的研究平台，共完成硕士学位论文6篇.

中文关键词： 非线性；时滞系统；稳定性；神经网络；鲁棒控制

英文摘要： The major project results are ：(1) A new adaptive neural control is proposed for a class of nonlinear unknown time-varying delay systems. Radial basis function (RBF) neural networks are utilized to estimate the unknown continuous functions. The unknown time-varying delays are compensated for using integral-type Lyapunov-Krasovskii functionals in the design. Boundedness of all the signals in the closed-loop of nonlinear systems is achieved, and the outputs of the systems are proven to converge to a small neighborhood of the desired trajectories. (2) The study proposes an asymptotic rejection algorithm for rejecting exotic disturbances in nonlinear systems. A new internal model is proposed to deal with these disturbances, and it can be completely rejected. (3) By employing Lyapunov functions and Razumikhin techniques, some new criteria are obtained to ensure the uniform boundedness and uniform asymptotic stability of the impulsive neural networks. (4) By means of the topological degree theory and Lyapunov function method, the sufficient conditions to guarantee the exponential robust stability of the an equilibrium solution are obtained. There are important research value and potential practice value in the project results. Nineteen articles of academic papers are published, two of which are SCI featured, twelve of which are EI featured and six of which are published in core journals；Famous experts were invited home and abroad to our university to exchange who are artificial wisdom；Team members participated in dozens of international and internal academic exchanges; This project provides a very good research platform for postgraduates, and 6 essays of master's degree are finished.

英文关键词： non-linear;time-delay system; stability;neural network;robust control