August 10, 2012      

Taking his inspiration from organisms without hard internal skeletons like squid, starfish, and worms, Harvard University Professor George Whitesides and his team have fashioned a new type of flexible, soft-bodied robot that could one day assist in dangerous search and rescue missions or serve as a surgeon’s assistant.

Adaptable robot ideal for search and rescue operations that require both crawling through tight, dangerous spaces and moving across flat ground.”

“If you look at most of the robots that are widely deployed, they either have the characteristics of, or they sort of look like, people or animals — a model that has a rigid structure with actuators that move around joints,” Whitesides said.

Multigait soft robot

But Whitesides’ pneumatically actuated robots are capable of sophisticated locomotion — i.e., fluid movement of limbs and multiple gaits –according to his paper on the subject.

Made mostly from elastomeric polymers, these robots are capable of two gaits: crawling and undulating. That means they can navigate difficult obstacles and go where no hard robot could ever go.


“Unlike hard robots that are fabricated from metals and are often heavy, these robots will do some things that hard robots have a hard time doing, like picking up an egg or crawling under door jambs — although they’re not very fast,” Whiteside said. “Since they’re intrinsically cheap, lightweight, and easy to make, we expect them to be stable in unstable environments.”

To get around this issue of speed — or lack of it –Tufts researchers have developed a robot that mimics the behavior of caterpillars, a terrestrial soft-bodied creature, so they can better understand the mechanics of “ballistic rolling.”

In a report published in Bioinspiration & Biomimetics, researchers including biology professor Barry Trimmer report that their GoQBot mimics the way a caterpillar moves: inching along like a worm or ballistically rolling at relatively high speeds. That means it’s able to quickly curl itself into a wheel and propel itself away from its enemies.

Whitesides said that because soft-bodied robots are simple and cheap to make, they’re made for simple, cheap applications. One such application would be in search and rescue.

“Imagine an earthquake, and the ceiling has half fallen down, and there are piles of rubble, and you want to crawl over the unstable rubble — something hard robots can’t do,” Whitesides said. One of these objects could go in there and climb around. Of course, there’s the chance that the roof will fall in, and the rubble will collapse, but you [only] lose a $25,000 robot or a $50,000 robot.”

These flexible robots could also be used as surgeons’ assistants in the operating room.

“Where you might use fingers [such as to push aside tissue], you might build a robotic system that’s fingerlike and has the ability to push aside or grip but it wouldn’t have a large hand or a large arm to go with it,? he says.

Whitesides doesn’t think of these soft-boded robots as being competitive with conventional hard robots, but he regards them as things that stand alone or that in collaboration with hard systems can open new opportunities for robotics.

As for Trimmer, he envisions creating an entirely new capability for the robots.

“There’s a bunch of capabilities they don’t have — things that you and I don’t think of as particularly extraordinary, like being able to walk though the forest,” Trimmer said. “We think that by exploiting soft materials for completely soft robots, you’re creating something that’s really, really adaptable. So beyond crawling into holes to rescue people or crawling inside bodies, I like to think of this as trying to reinvent some aspect of robots and their control.”

Turning everything on its head

However, the problem is that researchers have no idea how to control the movements of something that’s very deformable. That’s why Trimmer and his team are trying to understand how the nervous system and the body work together to create those complex movements.

“So we turn everything on its head,” Trimmer said. “We’ve determined that the complexity of the material is the answer, so the solution is that the calculations are being done between the nervous system and the mechanics.”

“You have to design the whole thing as a system,” he said. “You can’t build a robot and stick a computer chip in and control it. You have to design it from the ground up as a complete system.”

Trimmer said this means all the machines of the future will look more like animals and be more organic (i.e., robots will be grown) so people can do lots of things they couldn’t do previously.


“So rather than having to fold your complicated, hard antenna up and put it into a box to send it into space, you could crumple it and collapse it because it’s soft,” Trimmer said. “Then when you unfurl it, it just inflates. It’s truly amazing. And it’s biodegradable, so once you decide your machine isn’t valuable anymore, you can either dissolve it or eat it.”

It may sound like science fiction, but Trimmer said not so much.

“This whole idea of building robots by growing them is feasible within a decade, but within the next few years, we’ll have small devices, although they won’t be commercially available,” Trimmer said. “We’re working on those right now.”

Whitesides said his team is working on prototypes that they hope to take out of the university lab and put into the hands of real-world engineers who really know what do with them and who can figure out how much it would cost to turn them into commercially viable products.

“We’re putting together a company to take them out,” he said.