right_division Green SCM Distribution
Bookmark us
SCDigest Logo

Focus: Distribution/Materials Handling

Feature Article from Our Distribution and Materials Handling Subject Area - See All

From SCDigest's On-Target E-Magazine

- June 1, 2015 -


Supply Chain News: Amazon Comes Up with New Way to Innovate - Let the Grad Students Figure It Out for Almost Free

First Amazon Picking Challenge Sees 28 Robot Entries from Around the Globe, but Many Say Prize Money Needs Jacked to Come Back Next Year


SCDigest Editorial Staff


The innovation machine that is came up with a new way to find innovated ideas, sponsoring what it called the "Amazon Picking Challenge," a contest it announced last year that just completed its inaugural event in Seattle last week.

With existing its Kiva Systems robots (in place through an acquisition of the company in 2012), Amazon says that its "automated warehouses are successful at removing much of the walking and searching for items within a warehouse."

SCDigest Says:

For humans, determining how to best grab items is second nature. But at the moment it's a big challenge for a robot.
What Do You Say?
Click Here to Send Us Your Comments
Click Here to Post or See Reader Feedback

However, "commercially viable automated picking in unstructured environments still remains a difficult challenge," the company says on the contest web site, noting its contest was designed to spur improved capabilities in this area.

The Amazon contest was held in conjunction with IEEE's International Conference on Robotics and Automation in Amazon's home town of Seattle.

The competition asked contestants to build their own robot hardware and software to attempt simplified versions of the general task of picking items from shelves. The robots were presented with a stationary shelf holding items including a rubber duck, a pack of Oreo cookies, dog toys, pencils and a paperback book. The robots were asked to pick a subset of the products on the shelf and put them on a nearby table, as shown in the photo below.

The challenge combined the need for object recognition, pose recognition, grasp planning, compliant manipulation, motion planning, task planning, task execution, and error detection and recovery, Amazon said. In the end, each robot was asked to pick the same 12 items from the shelf and place them into a tote on the table.

Contestants were allowed to either build their own hardware and software systems, or acquire at no charge a "platform" to work with from leading robotics vendors such as Rethink Robotics, Clearpath, Fanuc and others (the number of free platforms was limited). Amazon hasn't released information on which teams used what hardware platforms.

For humans, determining how to best grab items is second nature. But at the moment it's a big challenge for a robot, and they can easily become confused. In general, the issues as much of a software problem as a hardware one, though there are some hardware challenges too.

Teams received full points for successfully moving each item from shelf to tote, and partial points if they picked up an item but subsequently dropped it or misplaced it on the table.

A team from Technical University of Berlin won the contest using a vacuum-powered robot successfully grabbing 10 out of the 12 items - in about 20 minutes.

(Distribution/Materials Handling Story Continues Below )


Download Longbow Advantage

Business Briefs



The Keys to WMS Success,

Maximizing JDA WMS

Performance and More






The web site reported that the winning entry was comprised of two main components: a robotic arm by Barrett Technologies installed on top of an mobile base robot from Nomadic Technologies.

The robotic arm used can mimic a human arm with its seven degrees of freedom. It grabbed items using a suction cup that can conform to the shape of different objects. Of course, the machine uses a number of sensors, including one for object recognition, another to find the base's position in relation to the shelf and a third one to make sure the arm doesn't exert too much force when handling items.

Key to its success, however, was an object recognition algorithm the team developed, which the team leader told engadget was "simple but robust."

A team from MIT came in second place, successfully moving 7 out of 12 items using a robotic arm with a flat scoop and a suction gripper.

For its win, the team from Berlin claimed a $20,000 prize, but some say that the total prize pool of $26,000 isn't nearly enough. For example, it cost $6,000 to ship the Berlin team's $250,000 robot, which got damaged on the way to Seattle, threatening the school's ability to use the equipment in other research.

Amazon also provided about $60,000 in travel assistance funds to get the teams to Seattle, but some teams said that fund needs a boost as well.

Only about half the 28 teams told Amazon Chief Technology Officer Peter Wurman they would return next year with the same basic levels of funding.

"If you really want strong teams, you need a bigger investment," said Alberto Rodriguez, who led the MIT team. "This probably needs 10 times the money."

"I brought the executives who funded this by and they were all very impressed," Wurman said. "So we'll work on them and see what we can do."

Amazon hopes to make its challenge a regular event that encourages innovation in robotics and steers academic research toward e-commerce automation.

What do you think of Amazon's robotic picking challenge? Let us know your thoughts at the Feedback button (email) or section (web form) below.

Recent Feedback


No Feedback on this article yet