A bipedal robot breaks the Guinness World Record for 100 meters

Cassie the robot sets the 100 meter record, photo by Kegan Sims.

By Steve Lundenberg

Cassie the robot, invented at Oregon State University College of Engineering and produced by OSU spin-off company Agility Robotics, set a Guinness World Record for the fastest 100 meters by a bipedal robot.

Cassie clocked the historic time of 24.73 seconds at OSU’s Whyte Track and Field Center, starting from a standing position and returning to that position after the sprint, without a fall.

The 100-meter record builds on the robot’s previous achievements, including running five kilometers in 2021 in just over 53 minutes. Cassie, the first bipedal robot to use machine learning to control a running gait on outdoor terrain, completed the 5K on the Oregon State campus untethered and on a single battery charge.

Cassie was developed under the guidance of Oregon State robotics professor Jonathan Hurst. The robot has knees that bend like those of an ostrich and operates without a camera or external sensor, essentially as if it were blind.

Since Cassie’s introduction in 2017, in collaboration with artificial intelligence professor Alan Fern, OSU students have explored machine learning options in Oregon State’s Dynamic Robotics and AI Lab.

“We’ve built the understanding to achieve this world record over the past few years, running a 5k and going up and down stairs,” said 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 behaviors for robots is new and different.”

The Dynamic Robotics and AI Lab merges physics with more commonly used AI approaches with data and simulation to generate new results in robot control, Fern said. Students and researchers come from a variety of backgrounds, including mechanical engineering, robotics, and computer science.

“Cassie was a platform for pioneering research in robot learning for locomotion,” Crowley said. “Finishing a 5K was all about reliability and endurance, which left open the question of how fast can Cassie run? This led the research team to focus on 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 parallelization – multiple processes and calculations happening at the same time, allowing Cassie to spend through a range of training experiences simultaneously.

“Cassie can perform a range of different gaits, but as we specialized her for speed, we started to wonder which gaits are most effective at each speed?” said Crowley. “This led to Cassie’s first optimized running gait and resulted in behavior strikingly similar to human biomechanics.”

The remaining challenge, a “deceptively difficult” challenge, was to get Cassie to reliably start from a free-standing position, run, and then return to the free-standing position without falling.

“Starting and stopping while standing is harder than running, just like taking off and landing is harder than flying an airplane,” Fern said. “This 100-meter result was achieved through close collaboration between mechanical hardware design and advanced artificial intelligence for hardware control.”

Hurst, chief technology officer at Agility Robotics and professor of robotics at Oregon State, said, “It may be the first bipedal robot to learn to run, but it won’t be the last. I think control approaches like this will play an important role in the future of robotics. The exciting part of this race is the potential. Using learned policies for bot control is a very new area, and this 100-meter sprint shows better performance than other control methods. I think progress will accelerate from here.