【作者】 李杨；

【导师】 吴学礼；

【作者基本信息】 燕山大学 ， 控制科学与工程， 2013， 博士

【摘要】 在实际的工业过程中，系统通常具有非线性、时变性、不确定性等其它复杂特性，时滞现象也是普遍存在于实际的工业生产过程中，这使得非线性系统的控制问题变得具有挑战性。因此，带有不确定性和时滞的非线性系统的鲁棒滑模控制方法研究越来越受到重视和关注。论文针对带有不确定性和时滞的非线性系统，基于线性矩阵不等式技术、自适应技术和神经网络技术、Backstepping技术、鲁棒控制方法、有限时间控制方法，研究非线性系统控制问题。论文的主要内容分以下几部分：针对不确定一致混沌系统，基于有限时间控制方法和Lyapunov稳定性理论，设计鲁棒反馈同步控制器，使得主从系统有限时间同步；针对未知参数不确定混沌系统，基于有限时间控制方法，给出带积分项的滑模面，并利用自适应参数辨识技术估计未知参数，进而基于Lyapunov稳定性理论，设计自适应滑模控制器，使得主从系统有限时间同步。针对带有外部扰动的不确定非线性中立系统，基于Lyapunov稳定性理论，通过线性矩阵不等式形式给出鲁棒稳定性新判据，并利用矩阵不等式技巧处理系统的不确定；针对带有外部扰动的不确定非线性中立系统，基于Lyapunov稳定性理论，通过线性矩阵不等式形式给出鲁棒稳定性新判据，并利用矩阵不等式技巧处理系统的不确定，给出不带时滞项的滑模面，进而基于Lyapunov稳定性理论，设计了鲁棒滑模控制器，使得闭环系统一致渐近稳定。针对时变时滞细胞神经网络，基于Lyapunov稳定性理论，通过线性矩阵不等式形式给出鲁棒稳定性新判据，并利用矩阵不等式技巧处理系统的不确定；针对非仿射非线性系统，利用神经网络对非线性系统进行辨识，并利用自适应参数辨识估计神经网络权值系数，进而基于Lyapunov稳定性理论，设计神经网络自适应滑模控制器，从而实现闭环系统一致渐近有界稳定；针对带有时滞的非仿射非线性系统，利用神经网络对非线性系统进行辨识，并利用自适应算法调整神经网络的权值系数，进而基于Lyapunov稳定性理论，设计基于Backstepping控制策略的神经网络自适应滑模跟踪器，使系统状态能够跟踪上预设轨迹。针对带有匹配扰动的离散非线性系统，给出滑模函数，并基于Lyapunov稳定性理论，设计非线性滑模控制器，进而利用矩阵技巧解出控制器参数，进一步利用Newton-Raphson算法解出控制器输入，从而使系统状态稳定在很小的临域内，并消除了抖振；针对不确定时变时滞离散非线性切换系统，基于Lyapunov稳定性理论，通过线性矩阵不等式形式给出鲁棒稳定性新判据，并利用矩阵不等式技巧处理系统的不确定，给出不带时滞项的滑模面，进而基于Lyapunov稳定性理论，设计非线性鲁棒滑模控制器，进一步利用Newton-Raphson算法解出控制器输入，从而实现闭环系统一致渐近稳定。更多还原

【Abstract】 Man-made systems with advanced functionalities in practical industrial processesusually exhibit nonlinearities, time-varying, unpredictable parameters and othercomplexities from design limitation, operation conditions, integrated inherent dynamics,even internal and external disturbances. In addition, delay is also common phenomenonencountered in the actual industrial production process, which makes the control systemdesign even challenging. Therefore, the robust sliding mode control method for uncertainnonlinear systems with time delay has attracted greater attention. In the thesis, theproblems of robust sliding mode control method for uncertain nonlinear systems withtime delay are studied by using linear matrix inequality, adaptive，backstepping andneural network, robust control and finite time control methods. A brief introduction of themajor research contents and achievement is outlined below.Firstly, it studies unified chaotic systems with uncertain parameters, and the robustfeedback synchronization controller is designed based on finite time control method andLyapunov stability theory, in order to achieve finite time synchronization between themaster and slave systems. Secondly, it studies chaotic systems with unknown parametersand uncertainties, the integral sliding mode surface proposed based on finite time controlmethod, the unknown parameters estimated by adaptive technique, the adaptive slidingmode synchronization controller is designed based on Lyapunov stability theory, in orderto achieve finite time synchronization between the master and slave systems.Firstly, it studies nonlinear neutral systems with uncertain parameters and externaldisturbance, the new robust stability criterion is provided in the form of linear matrixinequality based on Lyapunov stability theory, the system uncertainties are disposed bymatrix inequality technique. Secondly, it studies nonlinear neutral systems with uncertainparameters and external disturbance, the new robust stability criterion is provided in theform of linear matrix inequality based on Lyapunov stability theory, the systemsuncertainties were disposed by matrix inequality technique, the new sliding mode surfacewithout time delay term is designed, the robust sliding mode controller designed based on Lyapunov stability theory, in order to achieve uniformly asymptotic stability of theclosed-loop systems.Firstly, it studies cellular neural networks with uncertain parameters andtime-varying delays, the robust stability criterion is provided in the form of linear matrixinequality based on Lyapunov stability theory. Secondly, it studies non-affine nonlinearsystems, the nonlinear systems identified by neural network technique, the weightingcoefficients of neural network are adjusted by adaptive technique, the adaptive neuralnetwork sliding mode controller is designed based on Lyapunov stability theory, in orderto achieved uniformly ultimately boundedness of the closed-loop systems. Thirdly, itstudies non-affine nonlinear systems with time delays, the nonlinear systems identifiedby neural network technique, the weighting coefficients of neural network are adjusted byadaptive technique, the adaptive neural network sliding mode controller is designedbased on Lyapunov stability theory, so that the system trajectories can track the set pointvalues.Firstly, it studies discrete nonlinear systems with matching perturbations, the slidingfunction proposed, the nonlinear sliding mode controller is designed based on Lyapunovstability theory, the controller parameters obtained by matrix technique, the controllerinput obtained by Newton-Raphson Algorithm, in order to achieve uniformly asymptoticstability of the closed-loop system meanwhile eliminating the chattering effect,. Secondly,it studies discrete nonlinear switched systems with uncertain parameters and time-varyingdelays, the new robust stability criterion is provided in the form of linear matrixinequality based on Lyapunov stability theory, the systems uncertainties were disposedby matrix inequality technique, the new sliding mode surface without time delay term isdesigned, the robust sliding mode controller designed based on Lyapunov stability theory,the controller input obtained by Newton-Raphson Algorithm in order to achieveuniformly asymptotic stability of the closed-loop system.更多还原