基于事件触发的神经网络控制器稳定性分析  

Analysis on Stability of Neural Network Controller Based on Event Triggering

作　　者：郭欣 高燕[1] 蒋琳 张志姝 GUO Xin;GAO Yan;JIANG Lin;ZHANG Zhi-shu(School of Electronic and Electrical Engineering,Shanghai University of Engineering Science,Shanghai 201620,China)

机构地区：[1]上海工程技术大学电子电气工程学院

出　　处：《测控技术》2019年第10期113-117,共5页Measurement & Control Technology

基　　金：国家自然科学基金项目(61503238)

摘　　要：针对非线性系统稳定性问题,设计了基于事件触发的神经网络控制器来稳定非线性系统,控制器只有在触发规则满足的条件下,才更新控制参数,降低了网络传输率。算法开始先建立非线性系统模型,在采样过程中引入事件触发机制,并且设计了神经网络控制器,对于系统中包含的传输时滞,引入系统时滞模型,运用输入延迟法将同步控制器求解问题转化为时滞系统的稳定性问题。再通过构造分段Lyapunov-Krasovskii泛函并结合Jensen不等式,给出了非线性系统稳定条件。与传统数据采样系统相比,本文所提出的方法有效地增大了采样间隔,结尾通过例子仿真验证了所提出方法的有效性。In order to solve the stability problem of nonlinear systems,an event-triggered neural network controller is designed to stabilize the nonlinear system.The controller only updates the control parameters under the condition that the triggering rules are met,which reduces the network transmission rate.At the begining of the algorithm,the nonlinear system model was established firstly,the event trigger mechanism was introduced in the sampling process,and the neural network controller was designed.For the transmission time-delay contained in the system,the system time delay model was introduced,and the input delay method was used to transform the synchronous controller into the stability problem of time-delay system.By constructing the piecewise Lyapunov-Krasovskii functional and combining Jensen’s inequality,the stability conditions of nonlinear systems were given.Compared with the traditional data sampling system,the proposed method effectively increases the sampling interval,and the effectiveness of the proposed method is verified by an example simulation.

关 键 词：神经网络 数据采样控制 事件触发机制 时滞模型 非线性系统 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]