基于学习方法的机器人轨迹控制的实现

LEARNNING-BASED CONTROL OF ROBOT TRAJECTORY

本文针对机器人周期性重复劳动的运动轨迹控制,提出了一种基于 PID 调节器型的学习控制方法.本文首先给出了线性时变系统学习控制方法的稳定性条件,并导出了非线性机器人系统的迭代学习方法.此方法结构简单,计算量小,便于实时控制.最后就 PUMA-560的前三关节机械手的轨迹控制作了仿真研究.结果表明,随着学习次数的增加,机械手的实际运动轨迹将收敛于期望轨迹.

A learnning control method based on the principle of PID regulator for robot repeated trajectory control ispresented in this paper.The stability criterion of the learnning control method for linear time-varying systemis established and an iterative learnning method of the nonlinear robot system is deduced in this paper.Theproposed learnning scheme is structurally simple and can be used in Real-Time control.Finally,simulationson the PUMA 560 robot trajectory control show that the actual motion trajectories of the manipulator con-verge to the desired ones as the learnning number increases.