摘要
针对模型未知和动力学非线性机器人轨迹跟踪，提出了一种基于分布式动态回归神经网络（ＤＲＮＮ）的自适应控制方法．该方法在ＰＤ动态反馈控制的基础上，引入神经网络辨识器（ＮＮＩ）在线逼近对象的非线性动力学，并设计出神经网络自适应控制器（ＮＮＣ）来补偿机器人动力学非线性造成的误差．仿真结果表明该控制方案具有良好的跟踪性能和较强的鲁棒性．
This paper presents an adaptive coposite control architecture for robot manipulators based on dynamic recurrent neural networks (DRNN). The control objective is to adaptively compensate for the unknown nonlinearity of robot. This method combine PD feedback controller will neuroidentifier and neurocontroller, in order to reduce the tracking error caused by nonlinearity. Simulation results reveal that the proposed control scheme has good tracking performance and robustness.
出处
《应用基础与工程科学学报》
EI
CSCD
1998年第4期111-115,共5页
Journal of Basic Science and Engineering
基金
江苏省自然科学基金
航空科学基金
关键词
机器人控制
动态回归神经网络
自适应控制
robot control, dynamic recurrent neural networks, adaptive control
