Hybrid H_2/H_∞ force/position control based on neural networks  

作　　者：温淑焕 蔡建羡 王洪瑞 

机构地区：[1]Dept. of Electronic Engineering, Yanshan University, Qinhuangdao 066004,China

出　　处：《Journal of Harbin Institute of Technology(New Series)》2004年第5期569-572,共4页哈尔滨工业大学学报（英文版）

基　　金：SponsoredbytheHebeiProvinceScienceTechnologyTacklingKeyProblemItem(GrantNo.A393) .

摘　　要：A neural network control scheme with mixed H2/H∞performance was proposed for robot force/position control under parameter uncertainties and external disturbances. The mixed H2/H∞tracking performance ensures both robust stability under a prescribed attenuation level for external disturbance and H2optimal tracking. The neural network was introduced to adaptively estimate nonlinear uncertainties, improving the system’s performance under parameter uncertainties as well as obtaining the H2/H∞tracking performance. The simulation shows that the control method performs better even when the system is under large modeling uncertainties and external disturbances.A neural network control scheme with mixed H_2/H_∞performance was proposed for robot force/position control under parameter uncertainties and external disturbances. The mixed H_2/H_∞tracking performance ensures both robust stability under a prescribed attenuation level for external disturbance and H_2optimal tracking. The neural network was introduced to adaptively estimate nonlinear uncertainties, improving the system’s performance under parameter uncertainties as well as obtaining the H_2/H_∞tracking performance. The simulation shows that the control method performs better even when the system is under large modeling uncertainties and external disturbances.

关 键 词：ROBOTICS force/position control mixed H_2/H_∞control neural networks 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置] TP273[自动化与计算机技术—控制科学与工程]