Dynamic Compensation - Toward the Next-Generation Industrial Robot
Traditionally it is difficult for industrial robots to achieve high-speed motion with high accuracy due to large dynamical uncertainties. Researchers from the University of Tokyo's Ishikawa Watanabe Laboratory present a solution using dynamic compensation by adopting high-speed vision and compensation actuators to compensate for the uncertainties caused by the system itself as well as the external environment. They present two application tasks: fast and accurate contour-tracking and high-speed peg-and-hole alignment, with a commercial industrial robot. Traditionally, the playback method is the most common approach to control an industrial robot. However, it is time-consuming to teach an accurate path point by point. They propose to perform the 2D contour-tracking task by adding a high-speed robotic module under the dynamic compensation scheme. Through this method, a coarse global path can be easily taught with very few roughly chosen teaching points. The errors between the coarse path and the target path are then dynamically compensated by the high-speed robotic module under 1,000 fps visual feedback. As a result, accurate tracking for a target contour with random pattern is achieved with fast speed.