Iterative Learning Control with Application to Robotics

Abstract

Many machines and robots working today in factories are programmed to perform the same task repeatedly. By observing the tracking error in each iteration of the same task it becomes clear that it is actually repetitive, even though disturbances from noise and possibly sligthly changing friction dynamics affect the response. The idea of Iterative learning Control D ILC E is to use the knowledge from the previous iterations of the same task to reduce the tracking error the next time the task is performed. ILC utilizes the tracking error knowledge from the previous iteration to change the input signal to the system. In the thesis ILC is applied to an ABB Irb-2000 industrial robot. Using ILC the tracking error on the motor side has been reduced without changing the internal structure or any parameter in the robot controller. Three different ILC algorithms are considered in the thesis. Also some important theorems about ILC stability are taken into account. Two of these three ILC algorithms have been applied to the robot to improve the tracking of desired trajectories. ILC has also been used in order to improve the robot motion of an open container with liquid. The purpose was to shorten the motion time of the package transfer with control of the slosh inside.