Jul 2, 2016 @ 23:24 |
A team of engineers from Delft has won the Amazon Picking Challenge 2016. Team Delft won the two separate picking and stowing finals, making them a double champion.
The Amazon Picking Challenge 2016 was held during RoboCup 2016, the famous international robot competition which is held from 29 June to 3 July in Leipzig, Germany. During this challenge, robots autonomously retrieved a wide range of products from a container and put them on the shelves.
About Amazon Picking Challenge 2016 and Scores
This competition was to build their own robot hardware and software that can attempt simplified versions of the general task of picking items from shelves. The challenge combines object recognition, pose recognition, grasp planning, compliant manipulation, motion planning, task planning, task execution, error detection and recovery.
This was the second annual Amazon Picking Challenge, first one was held at ICRA 2015 and Team RBO was the winner.
The challenge 2016 was divided into two separate finals – picking and stowing finals: during the ‘stow task’ the robots, equipped with grippers, had to autonomously retrieve a wide range of products from a container and put them on the shelves. With the ‘pick task’, it was the other way around: the robot had to retrieve items from the shelves and put them in a container. The full rule of challenge can be read here.
Team Delft won the stow task finals by collecting 214 points. Second came NimbRo Picking (186 points) and the team from MIT ended third (164 points).
The pick task finals proved to be nerve wracking. Both Team Delft and the Japanese team PFN collected 105 points. What followed was a ‘photo finish’, whereby the jury analysed the first pick of the two teams. The fastest team would win. With a difference of about half a minute, that proved to be Team Delft. NimbRo Picking ended second.
The team built a flexible robot system based on industry standards. The system is equipped with a robot arm with seven degrees of freedom, high-quality 3D cameras and an in-house developed gripper.
To control the robot, the team integrated advanced software components based on state of the art artificial intelligent techniques and robotics. The components are developed with the Robot Operating System for industry (ROS-Industrial), and will be released as open software.
To get the job done, the robot needs to deal with unsolved problems for automation. “The robot needs to be able to handle variety and operate in an unstructured environment. We are really happy that we have been able to develop this successful system”, says Carlos Hernández Corbato from TU Delft Robotics Institute.
Kanter van Deurzen from the company Delft Robotics points out some of the key success factors of the team’s preparations, “We built a very robust system that hardly makes any mistakes in picking the products, thanks to our expertise in bin picking”. In addition, the robot chooses the maximum of points that can be scored for each product to pick.
Team Delft was one of 16 finalists for the Amazon Picking Challenge, and is a collaboration between TU Delft Robotics Institute and the company Delft Robotics. Team Delft is supported by RoboValley and the EU project Factory-in-a-Day.
|Team Delft||First Place|
|Team NimbRo||Second place|
|Team MIT||Third place|