摘要
针对含有模型不确定性的机电伺服系统,设计一种基于多层神经网络干扰补偿的控制策略。通过多层神经网络对与状态有关的干扰进行在线估计,以提高基于模型前馈控制输入的补偿精度,然后结合误差符号积分鲁棒(RISE)反馈控制方法,通过RISE的鲁棒增益处理神经网络逼近误差与未估计干扰,从而抑制干扰对伺服性能的不利影响。基于Lya⁃punov稳定性理论,证明了所提出控制器的闭环系统半全局渐近稳定,且系统所有信号有界。仿真结果表明:所提出的控制策略具有很好的干扰抑制能力,可显著提高机电伺服系统的跟踪精度。
Aimed at mechatronic servo systems with model uncertainty,a control strategy based on multilayer neural network in⁃terference compensation was designed.Multi⁃layer neural network was used to estimate state⁃related disturbances online to improve the compensation accuracy based on the model feedforward control input.Then combined with the robust integral of the sign of the error(RISE)feedback control method,the neural network approximation error and unestimated interference were processed by the robust gain of RISE so as to suppress the adverse effect of interference on servo performance.Based on Lyapunov stability theory,it was proved that the closed⁃loop system of the proposed controller is semi⁃globally asymptotically stable,and all signals of the system were bounded.The simulation results show that the proposed control strategy has good interference suppression ability and the tracking accu⁃racy of the mechatronic servo system can be significantly improved.
作者
吉珊珊
陈传波
JI Shanshan;CHEN Chuanbo(Department of Computer Engineering,Dongguan Polytechnic,Dongguan Guangdong 523808,China;School of Software Engineering,Huazhong University of Science and Technology,Wuhan Hubei 430074,China)
出处
《机床与液压》
北大核心
2020年第23期142-146,共5页
Machine Tool & Hydraulics
基金
2017广东省教育厅青年创新人才类项目(2017GkQNCX119)
东莞市社会科技发展项目(2017507156388)。
关键词
机电伺服系统
建模不确定性
鲁棒控制
多层神经网络
Mechatronic servo system
Modeling uncertainty
Robust control
Multi⁃layer neural network
