Overview
Feeling a bit tired of taking the dog for a walk? Well, your worries might be over soon once Cassie the robot takes a break from sprinting and takes on more domestic duties.
Cassie is the most recent robot to make headlines for a fantastic feat. Developed by a joint team of Oregon State University (OSU) and Agility Robotics, the robot registered its name in the Guinness Book of World Records by running a 100-meter (328-foot) dash in 24.73 seconds to become the fastest bipedal robot in the world.
The science and other stuff to know
Cassie’s achievement is a watershed moment in robotics. Existing technology has relied on cameras and external sensors to map terrain and control the machine’s movement. However, the world’s fastest bipedal robot uses none of that. As a matter of fact, it is completely blind.
Its movements are purely a result of machine learning, which performs hundreds of nanosecond calculations to control and adjust the machine’s gait to the terrain. Cassie’s latest achievement builds on earlier successes in which the robot successfully traversed 5 kilometers (3.1 miles) in a little over 53 minutes, according to a press release.
Cassie was developed under the direction of Oregon State robotics Professor Jonathan Hurst, who is also the chief technology officer at Agility Robotics. Devin Crowley, part of the team that led the effort, told OSU News and Research Communications that machine learning approaches “have long been used for pattern recognition, such as image recognition, but generating control behaviors for robots is new and different”.
So what?
The bipedal robot’s ability to first traverse the OSU campus and now run the 100-meter course without tripping defines a “watershed moment” in the development of robots, as stated by Hurst. It not only marks the achievement of a major scientific milestone, but it also provides hope that machines could eventually be made intelligent enough to mimic human biomechanics and perform a variety of redundant tasks and even improve their delivery. For example, robots like Cassie could be a useful feature in the healthcare industry. They could provide a reliable, failsafe resource for tasks that require only monitoring of vital stats, such as in ICUs, freeing up human resources for other critical tasks.
Moving ahead, improvement in tech and machine learning could eventually make robots like Cassie a key feature of our homes as well, efficiently tending to tasks like taking the trash out. A Rosie from the Jetsons of sorts.
What’s next?
While it might be several years before scientists turn Cassie into an independent robot worthy of full-scale service, the developments appear to be on the right track. Alan Fern, AI professor at Oregon State University who helped train Cassie, said Cassie’s feat brought engineers “one step closer to humanlike robots that could one day move around packages in warehouses, build homes or provide care for the elderly in homes”.
But it is yet to be known that, like their intelligent human counterparts, whether or not the robots love the jobs envisioned for them by their creators.