机器人跟踪的迭代学习控制方法

Iterative learning control algorithm for tracking control of robot

针对机器人跟踪控制问题,提出一种改进的D型迭代学习控制律,给出迭代学习控制算法的充分条件和必要条件.该控制律除了利用机器人的关节位置及关节速度信息外,还引入了机器人的关节角加速度.利用控制律中的机器人关节角加速度构成的学习律在二关节操作臂上进行试验并且与D类学习律进行比较.仿真结果表明,该方法能大大提高学习速度,具有良好的控制性能.

Aimed at tracking control of robot, an improved D-type iterative learning control law was presented and the sufficient and necessary conditions of its algorithm were given. In this control law presented, not only the information of angular joint position and velocity of the robot was employed, but the angular joint acceleration information was also introduced. In addition, the experiment on bi-jointed operative arm was conducted by using the learning law formed out of angular joint acceleration of the robot and compared with the traditional D-type algorithm proposed. The results of simulation showed that this method exhibited a higher learning speed and better controlling capability.