OSU engineers used a deep reinforcement learning algorithm to train Cassie to run 5km. (Image Credit: Oregon State University)
Cassie, a bipedal robot developed by Oregon State University and Agility Robotics, ran a five-kilometer route in over 53 minutes on a single battery charge while untethered. In the past, OSU engineers trained Cassie in a simulator to climb up and down a flight of stairs without using cameras or LIDAR.
“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 Hurst, who co-founded Agility in 2017. “This type of holistic approach will enable animal-like levels of performance. It’s incredibly exciting.”
Cassie used a deep reinforcement learning algorithm to teach itself how to run. Such an approach allowed the robot to stay upright by shifting its balance while running. However, Cassie needed to make adjustments to run while maintaining its balance.
“Deep reinforcement learning is a powerful method in AI that opens up skills like running, skipping and walking up and downstairs,” added Yesh Godse, an undergraduate in the lab.
To test Cassie’s running ability, the team made it run on turns for five kilometers, which took 43:49. However, the robot finished its run across the OSU campus in 53 minutes and 3 seconds. This took a bit longer because Cassie had 6 ½ minutes worth of resets following two falls, one of which was the result of an overheated computer. It fell a second time after executing a turn too quickly. “Cassie is a very efficient robot because of how it has been designed and built, and we were really able to reach the limits of the hardware and show what it can do,” said Jeremy Dao, a Ph.D. student in the Dynamic Robotics Laboratory.
“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. Along with logistics work like package delivery, bipedal robots could also have intelligent and safety capabilities to help people in their homes.
Have a story tip? Message me at: http://twitter.com/Cabe_Atwell