基于ESO的电动伺服系统RBF神经网络滑模控制  

RBF Neural Network Sliding Mode Control of Electric Servo System Based on ESO

作　　者：李晓飞 范元勋[1] 许鹿辉 LI Xiao-fei;FAN Yuan-xun;XU Lu-hui(School of Mechanical Engineering,Nanjing University of Science and Technology,Nanjing 210094,China)

机构地区：[1]南京理工大学机械工程学院,南京210094

出　　处：《组合机床与自动化加工技术》2023年第4期108-111,115,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：航天一院CALT基金资助项目(CALT201512)。

摘　　要：为了提高电动伺服系统的加载力跟踪精度,基于扩张观测器(ESO)和RBF神经网络,设计了一种自适应滑模控制器。ESO可以观测电动伺服系统的状态变量和外界扰动,解决实际工程应用中因对输出力直接微分造成的微分爆炸问题;利用RBF神经网络逼近系统存在的非线性不确定因素,将ESO观测到的状态变量、扰动项和RBF神经网络的估计值引入到滑模控制其中,实现对滑模控制的补偿,并通过构造李雅普诺夫函数对所提出的控制策略进行稳定性证明;在MATLAB/Simulink环境中搭建仿真模型,验证了ESO和RBF神经网络在不同工况下对系统相应量的精准估计,且误差均满足所设定的性能指标,同时与传统滑模控制进行了对比,所提出的控制策略能有效地解决传统滑模控制的抖振问题,加载力跟踪精度更高,鲁棒性能更优。In order to improve the tracking accuracy of loading force of electric servo system,an adaptive sliding mode controller is designed based on extended observer(ESO)and RBF neural network.ESO can observe the state variables and external disturbances of the electric servo system,and solve the problem of differential explosion caused by the direct differentiation of the output force in practical engineering applications.Using RBF neural network to approximate the nonlinear uncertain factors of the system,the state variables,disturbance terms and estimated values of RBF neural network observed by ESO are introduced into sliding mode control to realize the compensation of sliding mode control.The stability of the proposed control strategy is proved by constructing Lyapunov function.The simulation model is built in the MATLAB/Simulink environment to verify the accurate estimation of the corresponding quantity of the system by ESO and RBF neural networks under different operating conditions,and the errors meet the set performance indexes.at the same time,compared with the traditional sliding mode control,the proposed control strategy can effectively solve the chattering problem of traditional sliding mode control,and the loading force tracking accuracy is higher and the robust performance is better.

关 键 词：电动伺服系统 扩张观测器 RBF神经网络 滑模控制 

分 类 号：TH16[机械工程—机械制造及自动化] TG659[金属学及工艺—金属切削加工及机床]