An untethered robot, invented at the Oregon State University College of Engineering, and produced by OSU spinout company Agility Robotics, has established a Guinness world record for the fastest 100 metres by a bipedal robot. Named Cassie by its inventors, the robot set a time of 24,73 seconds, starting and finishing the sprint from a standing position without falling.
Unlike a human sprinter, Cassie has bird type legs like an ostrich, with knees that bend backwards. The robot does not have a vision system and operates without cameras or external sensors, essentially as if blind. To learn how to sprint, the OSU researchers say the robot’s programming was conducted in a week-long simulation. The simulation compressed a year’s worth of training experiences by computing numerous calculations simultaneously.
The 100 metre dash is Cassie’s second record setting performance. In 2021, the robot completed a 5 km run in just over 53 minutes on a single battery charge, making it the first untethered bipedal robot to use machine learning to control a running gait on outdoor terrain, the researchers say.
Cassie was developed under the direction of OSU robotics professor, Jonathan Hurst, with a 16 month, $1 million grant from the Defense Advanced Research Projects Agency (DARPA). Introduced in 2017, the robot became Agility Robotics’ first commercial robot, and has been used by top universities and robotics labs in the U.S. as a platform for exploring machine learning.
Since Cassie’s introduction in 2017, in collaboration with artificial intelligence professor Alan Fern, OSU students funded by the National Science Foundation and the DARPA Machine Common Sense programme have been exploring machine learning options in Oregon State’s Dynamic Robotics and AI Lab. “We have been building the understanding to achieve this world record over the past several years, running a 5 km stretch and also going up and down stairs,” says graduate student Devin Crowley, who led the Guinness effort. “Machine learning approaches have long been used for pattern recognition, such as image recognition, but generating control behaviours for robots is new and different.”
Fern says that the Dynamic Robotics and AI Lab melds physics with AI approaches more commonly used with data and simulation to generate novel results in robot control. Students and researchers come from a range of backgrounds including mechanical engineering, robotics and computer science.
“Cassie has been a platform for pioneering research in robot learning for locomotion,” Crowley adds. “Completing a 5 km run was about reliability and endurance, which left open the question of how fast Cassie can run. That led the research team to shift its focus to speed.”
Cassie was trained for the equivalent of a full year in a simulation environment, compressed to a week through a computing technique known as parallelisation – multiple processes and calculations happening at the same time − allowing Cassie to go through a range of training experiences simultaneously.
“Cassie can perform a spectrum of different gaits but, as we specialised for speed, we began to wonder which gaits are most efficient at each speed,” Crowley explains. “This led to Cassie’s first optimised running gait and resulted in behaviour that was strikingly similar to human biomechanics.”
The remaining challenge, a “deceptively difficult” one, is to get Cassie to start reliably from a free-standing position, run, and then return to the free-standing position without falling.
“Starting and stopping in a standing position are more difficult than the running part, similar to how taking off and landing are harder than actually flying a plane,” Fern continues. “This 100 metre result was achieved by a deep collaboration between mechanical hardware design and advanced artificial intelligence for the control of that hardware.” Hurst, also chief technology officer at Agility Robotics, calls the Guinness-recognised accomplishment “a big watershed moment”.
“This may be the first bipedal robot to learn to run, but it won’t be the last,” he says. “I believe control approaches like this are going to be a huge part of the future of robotics. The exciting part of this race is the potential. Using learned policies for robot control is a very new field, and this 100 metre dash is showing better performance than other control methods. I think progress is going to accelerate from here.”
For more information contact Steve Lundeberg, Oregon State University, [email protected], www.oregonstate.edu
© Technews Publishing (Pty) Ltd | All Rights Reserved
printer friendly version