考虑驱动系统动态的机械手神经网络控制及应用  

Neural Network Control and Application of Robotic Manipulators Including Actuator Dynamics

作　　者：王良勇[1,2] 柴天佑[1,2] 方正[1,2] 

机构地区：[1]东北大学流程工业综合自动化重点实验室,沈阳110004 [2]东北大学自动化研究中心,沈阳110004

出　　处：《自动化学报》2009年第5期622-626,共5页Acta Automatica Sinica

基　　金：国家重点基础研究发展计划(973计划)(2009CB320601);国家自然科学基金(60534010);国家创新研究群体科学基金(60521003);高等学校学科创新引智计划(B08015)资助~~

摘　　要：针对结构和参数均未知的机械手控制问题,提出了考虑驱动系统动态的机械手神经网络控制方法,采用稳定的径向基(Radial basis function,RBF)神经网络辨识机械手未知动态,而附加的鲁棒控制可以保证存在神经网络的建模误差和外部干扰时系统的稳定性和性能,并且该方法使机械手闭环系统一致最终有界.同时开发了基于半实物仿真技术的机械手控制系统,最后,将本文方法与经典的PD控制器和自适应控制器在同一机械手平台上进行了实验验证与分析,实验结果表明该方法具有良好的控制性能.A neural network control scheme is proposed for the control of robotic manipulator including actuator dynamics in this paper. In the proposed control scheme, the radial basis function （RBF） network is adopted to approximate the nonlinear dynamics of the robotic manipulator. In addition, a robust control is used to eliminate the neural network modelling error and disturbance. Uniformly ultimate boundedness （UUB） stability of the closed-loop system can be guaranteed by Lyapunov theory. Finally, a hardware-in-the-loop simulation technique based control system is developed. Furthermore, the proposed control scheme is applied to the same robotic manipulator together with PD control and adaptive control. Experiment results confirm the validity of the proposed control scheme by comparing it with other control strategies.

关 键 词：机械手 神经网络 鲁棒控制 半实物仿真 

分 类 号：TP241[自动化与计算机技术—检测技术与自动化装置]