基于RBF网络最小参数学习法的机械手终端滑模控制  

Terminal Sliding Mode Control of Manipulator Based on RBF Network Minimum Parameter Learning Method

作　　者：刘昕明[1] 吕东东[1] LIU Xin-ming;LV Dong-dong(College of Electrical and Control Engineering, Liaoning Technical University, Huludao 125105, China)

机构地区：[1]辽宁工程技术大学电气工程与控制工程学院,葫芦岛125105

出　　处：《自动化与仪表》2018年第6期28-33,共6页Automation & Instrumentation

基　　金：辽宁省创新团队基金项目(LT2010047)

摘　　要：针对机械手轨迹跟踪控制算法的问题研究,该文提出了基于RBF神经网络最小参数学习法的终端滑模控制(TSMC)方案。终端滑模控制算法解决了线性滑模控制算法不能在有限时间收敛到系统滑模面的问题,并保持了其对被控系统不确定性的鲁棒性。采用RBF神经网络逼近系统中的不确定项,用单个参数代替神经网络中的权值,从而简化自适应算法,增强了实时控制的要求。同时,用一个鲁棒控制项来抑制神经网络的建模误差和估计误差。Lyapunov理论保证闭环系统的有限时间收敛性和稳定性。最后,以两关节机械手作为被控对象,实验结果证实该控制方案的有效性。A terminal sliding mode control（TSMC） scheme based on the minimum parameter learning method is pro- posed for the trajectory tracking control algorithm of manipulator . The terminal SMC algorithm solves the problem that the linear siding mode control（LSMC） algorithm can not converge to the sliding surface of the system in finite time while maintains its robustness to uncertainties of the controlled system. By using the RBF neural network to ap- proximate the uncertainties in the system,a single parameter is used to replace the weights in the neural network, thus the adaptive algorithm is simplified and the requirement of real-time control is enhanced. Meanwhile,a robust control is used to suppress the modeling error and estimation error of the neural network. The Lyapunov theory guar- antees the finite time convergence and stability of the closed-loop system. Finally,the two joint manipulator is taken as a controlled object.and the experimental results confirm the effectiveness of the proposed control scheme.

关 键 词：滑模控制 机械手 轨迹跟踪 RBF神经网络 终端滑模 Matlab 

分 类 号：TP241.3[自动化与计算机技术—检测技术与自动化装置]