Alten Mechatronics Unstructured Box Palletizing with ROS-Industrial

| No Comments | No TrackBacks

From Berend Kupers, Alten Mechatronics cross-posted on ROS-Industrial blog

Alten Mechatronics, in collaboration with CSi Palletising systems, developed a demonstrator using an ABB IRB6640 robot to show that a system based on ROS-Industrial can easily cope with unknown products and uncertainties in the environment. Very little setup time was required for the actual hardware.

Two pallets were placed in the workspace of the robot. On one of these pallets three boxes of unknown size were randomly placed. The size, position and orientation of the boxes were determined using a Kinect, the Openni package and the PCL library. Based on this information, a path was calculated by MoveIt! and was then sent to the robot controller using the ABB ROS-Industrial package.

To make offline testing of the connection between the ROS-pc and the ABB motion controller possible, Alten Mechatronics developed a setup in which a ROS-pc was connected to a pc running the official ABB ABB RobotStudio software. This reduced the hardware set up time to less than one day. The current implementation reached a grasping accuracy of approximately 1cm, which indicates the accuracy of the vision system can be improved. This can be done by using more accurate sensors (e.g., laser triangulation, Time of Flight cameras) and more advanced filtering algorithms. For this demonstrator, the existing ABB drivers had to be adapted to allow controlling the gripper using the I/O interface of the IRC5 controller, which indicates a the need for a standardized Robot I/O interface in ROS-Industrial, to make implementation even easier.

No TrackBacks

TrackBack URL:

Leave a comment

Find this blog and more at

Monthly Archives

About this Entry

This page contains a single entry by Ugo Cupcic published on December 11, 2013 2:20 AM.

Shadow Robot's RoNeX available for ROS Hydro was the previous entry in this blog.

New Repository: kth-ros-pkg is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.