Department for Automation, Biocybernetics and Robotics





Grasping using ArToolKit library

Research activity duration: 2008 - 2008
Research area: Automation and Intelligent Control of Robots
Activity funding: Research Agency
Activity leader: Bojan Nemec

Abstract

The project deals with grasping of randomly positioned and oriented objects using a single camera.

Activity description

For object location and orientation estimation we used an open source library ArTool Kit. A miniature camera was mounted at the robot gripper and an appropriate gripping pose was achieved using visual servoing, thus eliminating the tedious calibration procedures. The overall control scheme was implemented in Matlab/Simulink using a newly developed robot programming language MatRol (Matlab Robot Language), which is especially convenient for research and educational work in robotics. The proposed visual servoing algorithm can be implemented in the majority of the industrial robots with the ability of the real time trajectory control in the task space. Hence, the algorithm is very convenient also for the low cost industry automation. The figure shows an illustrative example, which consists of sorting randomly chattered cubes on the table using vision controlled robot.

Multimedia

Sorting cubes with a robot (VideoATK.avi, 13490 kb)
Grasping using ArToolKit (Video_short_H264.avi, 3273 kb)

Publications

Article

  • Nemec B., Robot grasping using an ArToolKit library = Robotsko prijemanje objektov z uporabo ArToolKit knjižnice, 2010. [More] [PDF]