Bipedal robot from Oregon State University completes 5-km run using machine learning

Bipedal robot first to complete five-kilometre run A robot invented and built at Oregon State University made history by completing a five-kilometre course using machine learning.

TORONTO — Oregon State University students took one small step for robot marathon runners when their history-making bipedal bot went for a long-distance jog.

The robot, nicknamed "Cassie," completed a five-kilometre course in just over 53 minutes, a time comparable to the walking average set by humans. The university said in a news release last week that Cassie is the first bipedal robot to "use machine learning to control a running gait on outdoor terrain."

That means it taught itself to run – on legs that bend backwards like a bird's pronounced ankles – using what the school called "a deep reinforcement learning algorithm." It learned to make the subtle adjustments required to balance and stay upright while moving.

"The Dynamic Robotics Laboratory students in the OSU College of Engineering combined expertise from biomechanics and existing robot control approaches with new machine learning tools,” said robotics professor Jonathan Hurst in the release. "This type of holistic approach will enable animal-like levels of performance. It’s incredibly exciting."

Hurst led the project to develop Cassie under a 16-month, US$1-million grant from the U.S.'s Defense Advanced Research Projects Agency starting in 2017. Since then, students at the university have explored how machine learning could change a robot’s performance.

“Deep reinforcement learning is a powerful method in AI that opens up skills like running, skipping and walking up and down stairs,” Dynamic Robotics Laboratory undergraduate Yesh Godse said in the release.

Hurst said the five-kilometre run is just the start, and bipedal walking robots will one day be common, whether that's working in warehouses, delivering packages or helping people at home.

“In the not very distant future, everyone will see and interact with robots in many places in their everyday lives, robots that work alongside us and improve our quality of life,” Hurst said.

The university said during the 53-minute run, the robot had to stop for about six minutes of reset time after it fell, once for an overheated computer and another time because it had to take a turn too fast.