Oregon State University's machine athlete only fell twice while on the 5K run.
Cassie was invented at Oregon State University and created by OSU's spinout company Agility Robotics that's been working on manufacturing bipedal robots since 2015. The running robot was developed under the supervision of Agility Robotics' CTO and robotics professor Jonathan Hurst. The team received a 16-month $1 million grant from the Advanced Research Projects Agency of the U.S. Department of Defense.
Bipedal robots are a bigger technical challenge than four-legged robots as naturally they are less stable and fall down easily. Cassie's knees are bent resembling ostrich's anatomy which allows for more stability and smoother running overall. The robot learned how to run using a deep reinforcement learning algorithm which basically means she is a self-trained athlete. Thanks to Cassie's learning mechanisms, the robot learned how to perform dynamic balancing (i.e. make a number of tiny adjustments to its posture to stay upright) to avoid falling while on the go.
Cassie only fell twice during the 5K run: the first collapse happened due to an overheated computer, the second one occurred because the athlete had to make a turn at a speed that would not allow for a tricky maneuver like this. The resets took around 6.5 minutes and that means that Cassie completed the run in even less than 53 minutes.
"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," commented Jonathan Hurst, Agility Robotics CTO.
You can read more about Cassie here. Don't forget to join our new Reddit page, our new Telegram channel, our Discord, follow us on Instagram and Twitter, where we are sharing breakdowns, the latest news, awesome artworks, and more.