Need a Pick-Me-Up?

Amazon awards $50,000 prize in robotics competition’s 2nd year

Amazon held its second annual Picking Challenge last week, in conjunction with Robocup 2016 in Leipzig, Germany from June 30 to July 3. A robotic arm, designed by a team of engineers from Delft Robotics and TU Delft Robotics Institute, took home the $50,000 grand prize.

The robotics competition is designed to address the challenge of commercially viable automated picking in unstructured environments (or—as humans might call them—human environments). For example, finding seven specific items of merchandise on Amazon’s hundreds of shelves. The competition’s stated goal is to “strengthen the ties between the industrial and academic robotic communities and promote shared and open solutions to some of the big problems in unstructured automation,” per the Amazon Picking Challenge website.

A robotic arm is shown picking and stowing a pack of Oreos at last year's competition in Seattle.

A robotic arm is shown picking and stowing a pack of Oreos at last year’s competition in Seattle.

The contest was divided into two parallel competitions this year, introducing a new task. The Pick Task, essentially a more difficult version of the 2015 competition, made up the first section, and the new Stow Task introduced the challenge of taking items out of a tote and putting them into the shelving unit. Results are shown below:

Day 1: Stow Results

Competition on the first day was fierce, with some excellent performances. First place was won by team Delft who stowed 11 items into densely packed bins to earn 214 points. NimbRo Picking placed second and MIT took third place.

Team Name Score
Delft 214
NimbRo Picking 186
MIT 164
PFN 161
Duke 66
Team K 52
KTH 40
microRecycler 20
C^2M 16
Applied Robotics 15
AA-Team 3

Day 2: Pick Results

This year’s picking task was significantly more difficult than the 2015 task, with denser bins, occluded items, and products that are more difficult to see and grasp. The teams rose to the challenge and demonstrated much more sophisticated perception and manipulation solutions. The competition was so close that the judges had to resort to the second tie-breaker using video replay to determine the winner. Both Team Delft and PFN finished with 105 points, but Delft achieved their first pick in a mere 30 seconds, beating PFN’s 1:07 time.

Team Name Score
Delft 105 (0:30 first pick)
PFN 105 (1:07 first pick)
NimbRo Picking 97
MIT 67
Team K 49
C^2M 21
Dataspeed-Grizzly 21
AA-Team 16
Applied Robotics 10
Duke 0
microRecycler 0
Rutgers ARM N/A

For a video of the competition, make sure you follow the link in this week’s Direct and Current e-Newsletter!

No comments yet... Be the first to leave a reply!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: