Students at Oregon State University created a bipedal robot that used machine learning to teach itself how to successfully complete a 5K racecourse, the school announced.
The history-making bot, dubbed Cassie, traversed 5 kilometers, or 3.1 miles, of gravel, sidewalk, and fields and completed the course on a single charge in about 53 minutes.
The successful experiment showed the robot can use its two, human-like legs to keep a steady but varied pace across multiple outdoor terrains, paving the way for a future filled with automatons working alongside humans, the school said.
“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,” robotics professor Jonathan Hurst, who led the project, said.
Prior to this successful run, scientists hadn’t been able to figure out legged locomotion, which makes Cassie’s 5K that much more historic because it’s the first bipedal robot to teach itself to run on outdoor terrain, the school said.
“Running requires dynamic balancing – the ability to maintain balance while switching positions or otherwise being in motion – and Cassie has learned to make infinite subtle adjustments to stay upright while moving,” OSU explained in a news release.
Hurst called the breakthrough “incredibly exciting.”
“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,” the professor said.
“This type of holistic approach will enable animal-like levels of performance.”
The research was funded by a $1 million grant from the Defense Advanced Research Projects Agency, which is housed under the Department of Defense.
While Hurst said bots like Cassie will be used in the future for mundane tasks like package delivery and other logistics work, DARPA’s primary goal is to fund emerging technology for use by the military, which means they’ll also be used for security purposes.