摘要
针对目前弧焊机器人的控制算法大多是基于关节空间的算法,而这种算法无法实现对机械臂末端位置的直接控制的问题,提出了基于笛卡尔空间的轨迹跟踪控制算法.首先运用RBF(radical basis function)神经网络技术对实际机械臂数学模型的建模误差和参数不确定性进行补偿,接着定义Lyapunov函数并运用HJI(Hamilton-Jacobi inequality)定理设计基于笛卡尔空间的机器人鲁棒控制器.在此基础上以二自由度机械臂为被控对象进行仿真研究,仿真结果表明,基于笛卡尔空间算法的轨迹跟踪控制算法误差小于基于关节空间的控制算法,在基于笛卡尔空间的控制算法的仿真中末端轨迹跟踪误差小于0.08 mm,神经网络能够有效地在线学习机器人的建模误差和参数不确定性.
Aimed at the problem that most of present control algorithms of arc-welding robots are based on the joint space algorithm and this algorithm can not realize the direct control of the end point of the manipulator of the robot, a trajectory tracking control algorithm is proposed based on Cartesian space. Firstly, the RBF(radical basis function) neural network is used to compensate the modeling error and parameter uncertainty of the mathematical model of a practical manipulator. Then the Lyapunov function is defined and a robust controller is designed for the robot based on the Cartesian space and theorem of Hamilton-Jacobi inequality(HJI). The simulation results show that the error of the trajectory tracking control algorithm based on Cartesian space algorithm will be less than that of joint space-based control algorithm. In the simulation of the control algorithm based on Cartesian space, the tracking error of the robot end will be less than 0.08 mm and the neural network will be able to catch effectively the modeling error and parameter uncertainty of the robot.
作者
王保民
张明亮
WANG Bao-min;ZHANG Ming-liang(College of Mechano-Electronic Engineering, Lanzhou Univ. of Tech., Lanzhou 730050, China)
出处
《兰州理工大学学报》
CAS
北大核心
2019年第3期85-89,共5页
Journal of Lanzhou University of Technology
