具有H_∞性能的机械手神经自适应跟踪控制  

Neural Adaptive Tracking Control of Manipulators with H_∞ Tracking Performance

作　　者：蔡壮 张国良[1] 田琦[1] 

机构地区：[1]第二炮兵工程大学301教研室,西安710025

出　　处：《计算机测量与控制》2014年第3期720-722,共3页Computer Measurement &Control

基　　金：陕西省自然科学基金(2012K06-45)

摘　　要：针对非线性不确定机械手系统的轨迹跟踪控制问题,提出一种具有H∞性能的神经自适应控制算法;该算法为机械手系统分别设计了主控制器和监督控制器;主控制器由神经网络控制为基础,基于滑模控制原理得到神经网络权值的自适应律;基于李亚普诺夫稳定性理论和鲁棒控制设计的监督控制器,补偿自适应神经网络对系统不确定项学习的误差,同时使系统具有H∞性能;通过二自由度机械手模型进行仿真实验,仿真结果验证了方法的有效性。A neural adaptive algorithm with H∞ performance is proposed for trajectory tracking of nonlinear manipulator with uncertain- ties. A main controller and a supervised controller is designed for manipulator system. The main controller is designed based on neural net- work control theory, the adaptive law of neural network weight is achieved based on sliding mode control theory. The supervised controller is proposed to compensate the approaching error of neural network, and ensure the H∞ tracking performance of manipulator system in the sense of Lyapunov function. Finally, the simulation results based on two--link manipulator show the effectiveness of proposed algorithm.

关 键 词：机械手 轨迹跟踪 神经网络 鲁棒控制 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]