Cassie’s 40 or so supporters were delighted, cheering as she crossed the finish line. Its success was a defining moment in the history of robotics, they said. Cassie’s speed and agility, honed by training in artificial intelligence, showed that bipedal robots could maneuver through difficult real-world situations while maintaining balance, a problem that has plagued designers in the past.
The race built on Cassie’s 2021 achievement of a 5k in around 53 minutes, which initially showed Cassie could stand for long periods of time. It was also the cornerstone of about five years of work by engineering and machine learning researchers at Oregon State University and a spin-off company, Agility Robotics, paving the way for more advanced designs.
“This is the first big step for humanoid robots to do real work in the real world,” said Alan Fern, a professor of artificial intelligence at Oregon State University who helped train Cassie. “Because [now]we can make robots move robustly around the world on two legs. ยป
For decades, scientists, entrepreneurs, and engineers have been clamoring for two-legged robots. In the 1960s, Japanese researchers created rudimentary prototypes of bipedal machines. Over the past decade, engineers at MIT and the California Institute of Technology have attempted to do the same. Last week, Tesla CEO Elon Musk launched a two-legged humanoid robot, Optimus.
But two-legged robots have always faced problems, researchers say, including losing their balance and falling.
To solve this problem, Fern teamed up with Jonathan Hunt, a professor at Oregon State University and co-founder of Agility Robotics, to train bipedal robots using machine learning and neural networks, which are algorithms that mimic the functioning of a human brain.
The research is funded in part by the National Science Foundation and the Defense Advanced Research Projects Agency (DARPA), a secret government organization behind creations such as the Internet.
Since 2017, the team has been teaching Cassie to walk properly, using algorithms to reward the robot for moving appropriately. “It’s all inspired by Pavlovian psychology,” Fern said. “He just learns to anticipate those rewards and do the right thing.”
Once the team had successfully run the remote-controlled robot in simulation, the next step was to see how it would handle real-world environments, where surfaces are uneven, friction can change, and a robot’s mass can move.
In 2021, when the team asked Cassie to run a 5k, she learned a few things. The robot was “too stomping,” Fern said, and researchers began rewarding the robot when it smoothed out its gait. With the success of this year’s 100 meters, the team is taking the next step: putting a chest and a head on Cassie. (Agility Robotics is working on one called Digit.)
Fern said it will bring engineers closer to human robots who could one day move packages through warehouses, build houses or provide care for the elderly in their homes.
But such advances come with their own challenges.
Humanoid robots with heads resting on Cassie’s leg design will need peripheral vision to navigate difficult terrain. “Now Cassie has to look around the world,” he said, “to figure out what objects are there and not bump into them.”
The robot will also need to identify an object as something to pick up and then be smart enough to pick it up like a human would. (For example, Fern said, if a robot were to put boxes in a room, it would have to load the boxes from back to front.)
More importantly, these robots must walk with intention. “When you’re in the real world,” Fern said, “sometimes you have to be careful where you step.”
Still, engineering experts said it would be a tough climb to replace humans with robots.
Nancy J. Cooke, professor of human systems engineering at Arizona State University, noted that robots are getting very good at doing things like running or kicking a soccer ball. The most difficult thing is to create a machine that interacts with humans in a natural way.
“What they lack is really complex cognition,” Cooke said. “There is still a deep understanding of humans that is needed to interact with humans that they don’t have.”
Cooke also said it’s laudable that robots like Cassie are advancing the robotics industry, but it seems pointless to build machines that simply replicate what humans do. It might be more interesting, she said, to create robots that can do things humans can’t.
“Why [do] we want to rebuild ourselves? she asked. “I think it’s sci-fi stuff, but other than the entertainment value, I think it’s over the top.”
#Meet #Cassie #Usain #Bolt #robots