It was slow but it still made history.
Cassie, a robot invented at Oregon State University in Corvallis, has set a record as the first two-legged robot to use machine learning to control its running gait on outdoor terrain, the university said.
Using a single charge and untethered, Cassie ran 5 km or 3.1 miles outdoors in a leisurely 53 minutes. That equates to about a 17-minute mile and compares to world-record time of about 12-1/2 minutes for the entire distance.
The robot fell twice as a pack of students trailed it past buildings, homes and country fields. Making distinct clomping sounds as its metal plates hit the ground, Cassie consists of two red "legs" with a motor and controls where a person's abdomen would be, and nothing above.
Such robots could be used in the future to deliver packages or help people in their homes. Part of the breakthrough is that Cassie learned how to balance dynamically, performing subtle adjustments to stay upright while moving, the university said.
Cassie was developed with a US$1 million grant from the U.S. Department of Defense. The robot was produced by Agility Robotics, a company spun out of the university.