The squishy side of robots

Caitrin E. Eaton, Ph.D. thinks robots have a bad image—especially when they move. She should know. Robots are her specialty.

According to Eaton, who is an assistant professor of computer science at New College, the typical science fiction movie portrays robots as stiff and mechanical. Terrifying, like the hulking Terminator. Or comic, like the keg-shaped R2-D2.

Now that robots are an emerging reality, Eaton thinks science can improve their moves. She and other scientists are looking to the animal world for better kinetic models.

The secret ingredient?

“Squishiness,” Eaton said.

Eaton explains that the archetypal robot has a rigid metal structure. But that’s not nature’s way. There are few rigid, moving structures in the biological world.

“The stuff that produces movement—things like cartilage, tendons and muscle—is squishy,” Eaton said. “If you talk to a biologist about that, they’d say ‘duh.’ Roboticists are just now figuring out that screwing a couple of pieces of metal together isn’t the ideal way to create movement.”

Roboticists like Eaton have begun looking to nature to find a better way. In her current research, she draws on principles of animal physiology to design more agile, energy-efficient robots. She also applies those principles to physics-based computer simulations that can test physiological hypotheses without endangering animals—or robots, for that matter.

According to Eaton, everything starts with real-world observations. That includes observations of animals interacting with their environments, as well as studies of animal tissue in the lab.

Eaton participated in such a study. It happened during her post-doctorate laboratory studies in muscle physiology at the University of California, Irvine. That’s when she had her first aha moment. She was working with a biologist in the physiology lab.

“It was totally eye-opening every day,” she said.

According to Eaton, principles that were self-evident to biologists came as a revelation to her. The revelations and aha moments continued throughout her studies. The value of interdisciplinary collaboration was one of them.

Different fields of science and engineering tend to be walled off.  And game-changing discoveries happen when the walls come down. For Eaton, it was a transformative insight. She shares that lesson every day in her classroom work. It informs her New College research as well.

Applying animal kinesthetics to robots is an exciting field of research—one that defies academic specialization by its very nature. How does Eaton define it?

“I’m really interested in using robots as testbeds for building a better understanding of how animals move,” she said. “It’s two sides of the same coin—like using robots to understand animals and learning from animals how to build better robots.”

Speaking of interdisciplinary collaboration, Eaton welcomes classroom students from a range of disciplines besides computer science. That includes New College students interested in physiology, physics, mechanical engineering, and even kinetic sculpture.

“These are all great fits for the lab,” Eaton said. “I’m always looking for undergrads out there who are interested in the way these fields overlap. Even if somebody is really into physiology and maybe they’ve never thought of robotics before, that student might be a great fit for my lab.”

What about humanities majors? Absolutely.

Eaton stresses how crucial the humanities are to the future of robotics, especially when it comes to the question of preventing that technology from being misused. To put it in simple terms: How do we avoid building the Terminator?

“The terrifying thing is that even the engineers who are building those machines can’t answer that question. And the people who build these machines aren’t the people who decide how to use them,” Eaton said. “Philosophers and political scientists need to be involved in this conversation. If not, we’re going to end up with the Terminator.”

Su Byron is the communications specialist for the New College Foundation.