Stronger Artificial Muscles Will Make Robots Seem More Human
Motors are so 20th century.
The Repliee Q2 robot. Image: BradBeattie/Wikipedia
Ray Baughman, a professor of chemistry and director of the NanoTech Institute at the University of Texas at Dallas, recalls seeing a beautiful woman across the room while at a conference in South Korea. But the closer he got, the more disturbed he became.
"That was EveR, a humanoid robot," he told me. EveR's creators had done a pretty good job of emulating the human form, he said, but "she didn't have enough muscles in her face to smile naturally." That's just the sort of subtle imperfection that makes robots more creepy than friendly.
Artificial muscles have the capacity to transform robots and prosthetic devices, but they're surprisingly hard to make. Researchers have been trying for years to have them contract and release quickly enough while still being flexible and not brittle. They've tried lots of different structures, from fishing wire that stretches to a rubber cylinder that can be electrically stimulated to contract.
But one of the most futuristic-feeling concepts is the idea of a gel-based muscle, which could completely reshape the way we think about muscles packaging—if anyone can ever build one strong enough to last.
Japanese researchers published a study earlier this month in the journal Nature Communications about a new hydrogel, or a network of synthetic molecules that can expand by absorbing water, that they say could be used as an artificial muscle, even if they're not yet ready to be implemented in the field.
"Traditionally, gels have been used by researchers in many fields because the preparation of the gels is very easy," Yukikazu Takeoka, a professor of molecular engineering at Nagoya University in Japan and an author of the Nature study, wrote in an email. But he and his team have found a way to strengthen the gels, he said, which should inspire other researchers to find different applications.
They note in their paper that while hydrogels "have the potential … to be employed in artificial muscles, their brittleness must be reduced before they can be used in these applications," adding that their new method of developing a gel can lead to "significant improvements" in their durability.
"We've done work with gel artificial muscles in the past, but ours were too weak and too slow," Paul Calvert, a professor of bioengineering at the University of Massachusetts in Dartmouth, said in an interview. "They were too slow because stuff has to diffuse in and out, and that just takes time if you're talking about microscopic things."
And they were weak because the strings of material are pulled tight when they were filled with water, which makes them brittle. "Think of Jello—when you put a nick in it, it just splits," Calvert said.
Over the past 15 years, Calvert said, small innovations have brought engineers incrementally closer to making gels with the right qualities for artificial muscles. That process is still going, and has a fair bit to go, even with the latest contributions from the Japanese research team.
A gel is like a net of fibers, Calvert said, and they intersect at a number of points that have been fixed in the past. The authors of the new study added a "donut" to the end of the fibers so that the intersections are no longer fixed in space.
"It gives you more stretchability because everything can slide to one end before it pulls tight," Calvert said. This innovation isn't brand new, but "no one had put [the cross link and the donut] together before," he added.
Even though the authors of the Nature study mention artificial muscles as a possible application for their hydrogel, Baughman thinks that gels are "generally too stretchy" to work as such. They're better designed for drug delivery because they can release or absorb materials in the presence of particular chemicals, which means that they could release drugs in the body only in the presence of the appropriate cells like antigens.
For now, Calvert agrees; the most promising developments are coming out of SRI International, a research-based company in Silicon Valley that has come closest by putting greased metal diodes on either side of a rubber cylinder.
"For the moment, the gels are paralyzed," Calvert said, because no one has been able to figure out how to make them strong and sensitive enough.
But they point at a future in which robots have far more flexibility in packaging. "If you look at all the robots at the moment, including prosthetic arms, they're all driven by electric motors and gear wheels. It's obvious we need an artificial muscle and a humanoid kind of arm," Calvert said.
One exciting potential application is the growing field of soft robotics. Calvert envisions that the first soft robots will look like "dogs with a battery inside," suitable to look for landmines. But they'll quickly progress beyond those practical applications. Artificial muscles are key to that progression, even if people aren't really sure why they need humanoid robots in the first place.
"As people start to interact more with robots, they're going to need things that are a lot more like people," Calvert said. "For the moment, the only question is, 'Can we do it?'"