AUTONOMOUS GLUING BASED ON IMAGE-BASED VISUAL SERVOING

Le D. Hanh∗ and Chyi-Yeu Lin∗∗

Keywords

RGB-D camera, visual servoing, trajectory following, PD control

Abstract

This research presents a practical strategy that can significantly reduce the time for a gluing task in footwear industry by using a 6-DOF industrial robot and RGB-D camera. The manipulator system equipped with a PD image-based visual servoing module, will autonomously learn the 2.5D object’s profile when a soling bond is placed inside the robot working space. After learning, the trajectory based on the raw data will be smoothed and then saved in the trajectory database. Once a random soling bond is positioned inside the robot working area, the manipulator can autonomously execute the trajectory tracking task without re-learning. By its ability that can autonomously update the trajectory of an object during the learning process, this methodology also reduces the errors of the system which occurred during calibration and image processing. The high accuracy and robustness of the system and algorithm are tested and proven in the real experiment with a 6-DOF Denso robot.

Important Links:

Go Back