Presented by:

Fetch It challenge mobile manipulation

May 14, 2019      

The concept of mobile manipulation for robots – combining a robotic arm that can grab objects and then move them to another location – has often been considered a “holy grail” in the space. Several companies have limited applications of mobile manipulation, but most companies we speak with say this concept and technology are in the very early stages.

Fetch Robotics, one of the leading companies in the autonomous mobile robot sector, is sponsoring a competition for teams to showcase their work within mobile manipulation. The “Fetch it! The Mobile Manipulation Challenge” will be held from May 20-22, 2019, at the International Conference on Robotics and Automation, in Montreal.

At least four teams are competing in the event, including members from Case Western Reserve University, Columbia University, the Georgia Institute of Technology, and the University of Massachusetts Lowell. The grand prize is a Fetch Mobile Manipulation Research Robot, along with 7,000 Schunk Bucks. Other prizes include LiDAR sensors and scanners, and additional Schunk Bucks. In addition to Fetch, the competition includes sponsorships from SICK, Schunk, EandM, and The Construct.

How it works

Fetch Mobile manipulator

Fetch Mobile Manipulator

Teams will be required to autonomously navigate to stations in a work cell, pick up items, operate basic machinery, place items into kits, and transport the finished kits to a drop-off location. The team that assembles the most kits within 45 minutes will win the competition. The work cell will be approximately 10 feet by 10 feet in size, with items and machinery that the robot needs to complete the tasks. Objects used for manipulation will be announced ahead of times, and will be detectable and able to be manipulated by the Fetch robot.

The official competition robot will be a Stock Fetch Mobile Manipulator, with Ubuntu 18.04 and ROS Melodic installed. Official competition robots will be provided for teams, but they can also use their own robot as long as no major hardware modifications were used, Fetch said.

“The combination of mobility and manipulation is an intriguing field of study,” said Russell Toris, director of Robotics at Fetch, and the director of the challenge. “My hope for this competition is to push the boundaries of manipulation, but grounded in a real-world application.”

Holly Yanco, professor of computer science at UMass Lowell, said the challenge is a great way to have students working with two Fetch robots that the school recently purchased. “Additionally, the tasks in the challenge represent a good set of skills that can be used in our ONR MURI on robot self-assessment, a collaborative project with Carnegie Mellon University (Aaron Steinfeld, Reid Simmons and Henny Admoni), Brigham Young University (Michael Goodrich and Jacob Crandall) and Tufts University (Matthias Scheutz),” Yanco said.