文献类型：期刊文献

中文题名：基于简化模型预测的网络化控制系统设计

英文题名：Networked Control System Design Based on Simplified Model Prediction

作者：佟世文[1,3];钱殿伟[4];于庆林[5];赵叶[1,2];程光[2]

第一作者：佟世文

机构：[1]北京联合大学机器人学院,北京100101;[2]北京联合大学工科中心,北京100101;[3]北京市信息服务工程重点实验室,北京100101;[4]华北电力大学计算机与控制学院,北京102206;[5]中国石油工程建设公司北京设计分公司,北京100085

第一机构：北京联合大学机器人学院

年份：2021

卷号：28

期号：2

起止页码：367-374

中文期刊名：控制工程

外文期刊名：Control Engineering of China

收录：CSTPCD;;北大核心:【北大核心2020】；CSCD:【CSCD_E2021_2022】；

基金：北京市教委科技计划一般项目(KM201811417001);北京市自然科学基金-北京市教委联合项目(KZ201811417048);北京联合大学科研项目(ZK70202001);北京市自然科学基金-轨道交通联合基金项目(L191006)。

语种：中文

中文关键词：模糊聚类建模;网络化控制;模糊控制;神经网络;随机时延;

外文关键词：Fuzzy cluster modeling;networked control;fuzzy control;neural network;random delay;

摘要：针对机理不清、难建模的被控过程,提出了一种基于数据的网络化控制方法。仅根据过程的输入输出数据,先采用模糊聚类建模技术,获得系统的T-S模型,然后利用神经网络对前件隶属度函数的拟合能力,将其转化成显式表达,从而获得一步预测简化模型。再利用递推和反馈校正,获得一系列预测的过程输出,进而将控制系统设计转化成"由未来的偏差及偏差的变化得到未来的控制作用"问题。通过分析典型被控过程的动态响应曲线设计隶属度函数及控制规则,可获得一系列未来的控制作用。再通过选择合适的控制作用即可补偿网络时延和数据丢包的影响。仿真表明,该控制方法在处理随机时延和数据丢包问题上较PID控制有更快的动态响应和更高的精度。
Considering the unclear mechanism of some controlled processes and the difficulty of directly establishing mathematical models,this paper presents a data-based networked control method.First,the fuzzy clustering modeling technique was used to obtain the T-S model of the system only based on the input and output data of the controlled process.Then,the neural network was used to fit the membership functions of the antecedent variables to convert it into an explicit expression.Thus,a one-step prediction simplified model was available.Recursion and feedback correction were used to obtain a series of predicted process outputs,which in turn transformed the control system design into the problem of‘getting a series of future control actions according to deviations and the change of deviations in the future’.Through analyzing the dynamic response curve of a typical controlled process,membership functions and control rules were designed and a series of future control actions were obtained.Thus,the influence of network delay and data packet loss on the performance of the control system could be compensated by selecting the appropriate control action.The simulation results show that the proposed control method has faster dynamic response and higher accuracy than conventional PID control in dealing with random delay and data packet dropout.