Article from Fetch Robotics
Fetch Robotics announced the results of the inaugural FetchIt! Mobile Manipulation Challenge. The competition was designed to advance the state of technology for applying mobile manipulators, which are autonomous mobile robots (AMRs) fitted with robotic arms, for use in manufacturing and related applications.
The FetchIt! Challenge attracted teams from leading universities, who were tasked with using a Fetch Mobile Manipulator robot to navigate to stations in a work cell where they picked up items with the arm, inserted them into a machining tool, placed the machined items into kits, transported the finished kits to an inspection station and then to a drop-off location. The competition encompasses the full range of activities that are commonly found in manufacturing environments.
Georgia Tech was awarded first prize for successfully assembling three kits in 39 minutes, earning a prize package that included a Fetch Mobile Manipulation Research Robot - a $100K value - along with additional prizes from co-sponsors EandM, SCHUNK and SICK AG.
FetchIt! Challenge participants included:
• Team Columbia: Columbia University, led by Professor Peter K. Allen, PhD, and Neil Chen
• Team DeRAILers: The Georgia Institute of Technology, led by Associate Professor Sonia Chernova, Ph.D. and David Kent
• Team RoboHawks: The University of Massachusetts Lowell, led by Professor Holly Yanco, Ph.D., Assistant Professor Reza Ahmadzadeh, Ph.D. and Zhao Han
• Team Fido: Independent Competitors, Thomas Butterworth and Ben Jarvhi
Autonomous mobile robots are seeing rapid adoption in multiple applications, particularly for material movement tasks in warehouses and factories. This has fueled growing interest in using AMRs equipped with robotic arms - so called mobile manipulators - for applications that combine autonomous transport with robotic arms that can grasp and manipulate objects. While conceptually simple, combining autonomous mobility with robotic arm activity is extremely challenging, requiring complex interaction between the robot navigation, machine vision, arm operation and safety systems.
The FetchIt! competition focused on teams autonomously completing combined manipulation and navigation tasks. The goal was to assemble a kit from six objects obtained from stations around the designated arena. Designed to mimic a common machine-tending process in manufacturing, the objects were picked from bins and then placed in a kit and transported to a drop-off location. All teams used a stock Fetch Mobile Manipulator, which is the most widely deployed mobile manipulation robot, deployed at over fifty of the world's leading academic and commercial robotic research centers.
“I’d like to congratulate all the teams for their accomplishments during the course of this challenge,” said Russell Toris, director of robotics at Fetch Robotics. “When setting out to create this challenge, we knew we wanted to keep it grounded to a real-word scenario. Interacting with machinery that is designed to be used by humans is no easy task. Piece-picking, kitting and countless other tasks are going to require state-of-the-art perception, motion planning, navigation and safety all seamlessly working together. The teams’ performance this week indicate that they represent some of the world’s leading experts in these skills.”
“We’re very excited to have won the FetchIt! challenge,” said Georgia Tech’s Sonia Chernova. “It has allowed us to validate our research code in a complex domain. We can’t wait to continue our work with our newest Fetch robot.”
Added Holly Yanco from UMass Lowell: “Everyone from Fetch has been helpful and very encouraging. This has been an amazing experience and the tasks used for the competition form a great basis for our ONR MURI research.”
In addition to Fetch Robotics, additional sponsors included EandM, Schunk, SICK and The Construct.
First Place: Georgia Tech
Fetch Mobile Manipulation Research Robot, MRS1000 4-layer LiDAR sensor (1) provided by SICK and EandM, along with 7,000 Schunk Bucks
Second Place: University of Massachusetts Lowell
MRS1000 4-layer LiDAR sensor (1) and TiM561 LiDAR laser scanner (1) provided by SICK and EandM, along with 5,000 Schunk Bucks
Additional Prizes: TiM561 LiDAR laser scanner (1) provided by SICK and EandM, for all entrants
Related Glossary Terms
Cone-shaped pins that support a workpiece by one or two ends during machining. The centers fit into holes drilled in the workpiece ends. Centers that turn with the workpiece are called “live” centers; those that do not are called “dead” centers.
Arm or basic object-transferring device. Hands or gripping devices vary according to application, as do arm design and number of joints (axes or degrees of freedom). See degrees of freedom; effectuating device.
Discipline involving self-actuating and self-operating devices. Robots frequently imitate human capabilities, including the ability to manipulate physical objects while evaluating and reacting appropriately to various stimuli. See industrial robot; robot.