A robot designed by German engineering students was the winner of Amazon’s first-ever “Picking Challenge” competition, which was held during a robotics conference in Seattle last month.
The contest challenged students to design robots that could grab a series of small objects from a shelf according to a pre-determined sequence in the shortest possible time. The scenario is similar to the process that occurs in Amazon’s warehouses. as the world’s largest online retailer fulfills millions of customer orders per day.
Mimics Amazon’s Daily Operations
The world’s largest online retailer fulfills millions of customer orders per day and nearly all of those orders are composed of unique sets of individual items. Robots assemble orders so that they can be shipped to Amazon customers within a few minutes after the order has been placed.
The winning team was composed of students from the Robotics and Biology Laboratory at the Technische Universitat Berlin.
The event was part of the Institute of Electrical and Electronics Engineers’ annual International Conference on Robotics and Automation, which was held May 26 to 30. The event is the IEEE’s flagship conference and is one of the world’s most important forums for robotics researchers to present their work.
Other participants included teams from some of the world’s leading robotics companies, including ABB, Fanuc, Rethink Robotics, Universal Robots, Willow Garage, and Yaskawa-Motoman.
Robots of Every Size and Shape
Teams use a variety of grippers, robotic hands, suction cups and scoops to collect the diverse items — which include a joke book, a box of Cheez-Its, a lint brush, a duck toy, a box of crayons, and a package of Oreos — in order and without causing any damage.
The German engineering students used a WAM arm developed by Barrett Technology, a highly dexterous back-drivable manipulator that looks like a metallic human arm that has only three fingers. They were awarded 148 points, beating the second-place team from the Massachusets Institute of Technology, which used a robotic arm developed by ABB Robotics that had a gripper with creative finger geometries.
Other top finishers included teams from UC Berkeley and Georgia Tech.
A Team Effort
Oliver Brock, a professor at TU-Berlin, said that the win was the result of a team effort and that every single member of the team contributed with enthusiasm and ingenuity, enabling the team to produce a “compelling showcase for mobile manipulation as a winning approach to industrial manipulation”.
The two-day competitions required teams to combine robotic hardware and software that could perform simplified versions of the general task of picking items from shelves. The robots were set up in front of a stationary, lightly-populated inventory shelf and teams were given a set of products and had to pick them and place them on a table in a particular sequence.
The challenge combined object recognition, pose recognition, grasp planning, compliant manipulation, motion planning, task planning, task execution, and error detection and recovery. Their robots were scored by how many items they picked and how fast they went.
Teams were also required to explain their approach to future challenge results and industrial implementation.