基于ADP的一类未知非线性系统事件触发输出反馈最优控制  

ADP-BASED EVENT-TRIGGERED OUTPUT FEEDBACK OPTIMAL CONTROL FOR A CLASS OF UNKNOWN NONLINEAR SYSTEMS

作　　者：李琳[1] 潘忠成 李昶志 Li Lin;Pan Zhongcheng;Li Changzhi(School of Information,Guangdong Institute of Communications,Guangzhou 510000,Guangdong,China;School of Materials Science and Chemical Engineering,Harbin Engineering University,Harbin 150001,Heilongjiang,China;Shaanxi McCullough Biotechnology Co.,Ltd.,Weinan 715500,Shaanxi,China;Shaanxi Public Resources Exchange Center,Xi’an 710000,Shaanxi,China)

机构地区：[1]广东交通职业技术学院信息学院,广东广州510000 [2]哈尔滨工程大学材料科学与化学工程学院,黑龙江哈尔滨150001 [3]陕西麦可罗生物科技有限公司,陕西渭南715500 [4]陕西省公共资源交易中心,陕西西安710000

出　　处：《计算机应用与软件》2023年第3期292-297,共6页Computer Applications and Software

基　　金：陕西省科技厅项目(2018ZKC-173,2019-PT-15);国家自然科学基金面上项目(51979064);陕西省科技厅重大项目(S2018-YF-ZDNY-0199)。

摘　　要：针对一类未知非线性系统,提出基于ADP(Adaptive Dynamic Programming)的事件触发输出反馈最优控制策略,此方法只用到了系统的输出量。考虑到系统内部状态量无法测量和系统模型难以获得的问题,设计神经网络状态观测器来估计系统的不可测状态量并通过输出信息重构了系统的内部状态。在获得系统的未知动态和状态量信息后,设计结合事件触发技术的ADP输出反馈最优控制策略。通过Lyapunov理论推导了神经网络观测器和评估网络的权值更新率,并且证明了闭环控制系统的稳定性。通过仿真实验验证了该控制方法的有效性。Aimed at a class of unknown nonlinear systems,an event-triggered output feedback optimal control scheme based on adaptive dynamic programming(ADP)is proposed,which uses only the output information.Considering that the internal state of the system could not be measured and the system model was difficult to obtain,a neural network state observer was designed to estimate the unmeasurable system states and reconstruct the system internal state based the output data.After the information of unknown dynamics and the unmeasurable states information were obtained,an ADP output feedback optimal control scheme combined with event-triggered technique was proposed.Through the Lyapunov theory,the weight update rates of neural network observer and critic network were obtained,and the stability of closed-loop control system was proved.The simulation results demonstrate the effectiveness of the proposed control scheme.

关 键 词：ADP 神经网络观测器 事件触发 输出反馈 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]