Somewhere in a laboratory, something is squirming. It’s an eel robot, swimming silently through an enormous darkened saltwater tank; its rhythmic, ribbon-like motions echoing those of its natural counterpart. Or is it an octopus-inspired robot, looking like the kind of thing H.P. Lovecraft might have dreamed up had he been born 100 years later and become a roboticist instead of a horror writer?
Maybe it’s neither of those. Maybe it’s a soft rubbery sleeve, wrapped around a human heart, and giving it regular reassuring squeezes to allow it to continue beating in the face of heart failure. Or a hydrogel gripper which can reach out and grab a swimming fish, securing it without hurting it. Or. Or. Or.
The reality is that there is no shortage of examples from the burgeoning field of soft robotics. One of the most exciting, fast-developing fields in robotics research, these robots don’t resemble the hard, metallic machines that science fiction promised us. They’re made of rubbery materials. They’re also much cheaper to build, weigh considerably less, and offer far more flexibility and (perhaps counterintuitively, given their soft materials) durability. In the process, they’re changing what we think of as a robot — and proving themselves immensely useful in the process.
THE ROBOT OF A THOUSAND USES
The current surge of soft robots into a world previously dominated by metal doesn’t take anything away from the more traditional robots being built by companies like Boston Dynamics. Advances in traditional hard robotics have, in the past few years alone, given us all manner of versatile machines capable of doing everything from fine assembly work on production lines to performing dances, parkour displays or even Olympics-worthy backflips. But these traditional robots aren’t good for everything.
They’re great at carrying out the same task many times in a row, which is exactly what’s required if, for example, they’re helping to put together iPhones on a conveyor belt in a Foxconn factory. But remove them from the structured domain that they’re used to working in and suddenly their astonishing precision can disappear in an instant.
This is problematic for all sorts of reasons, not least the fact that, increasingly, robots are going to be working alongside people and other living people. This could mean directly working with humans as colleagues. It could also mean an even closer level of interaction, such as the aforementioned robot intended to keep a person’s heart beating in the face of possible cardiovascular failure. It’s these scenarios which have led, in part, to the increasing popularity of soft robots.
“[One other] exciting application field for soft robotics is in the search and rescue sector,” Giada Gerboni, a postdoctoral fellow at Stanford University, who last year gave a TED Talk titled “The incredible potential of flexible, soft robots,” told Digital Trends. “[In the lab I’m working in], there is a very motivated team of colleagues that is exploring how a soft robot that grows from the tip — like a plant, or more specifically a vine — would be able to navigate archeological sites inaccessible by humans and full of very delicate artifacts, or to explore the delicate underground habitats of some endangered animal species.”
Gerboni’s own research has medical applications. She is presently working on flexible robots which could be used as surgical instruments to access hard-to-reach parts of the body, just as her colleagues’ soft robots could help access remote sites.
“My current work in Stanford uses a flexible needle that can reach completely different parts of the liver from a single insertion point and can burn a tumor — [destroying] tumor cells by heat — with its tip,” she continued. A steerable needle isn’t what we might think of as a robot, but it’s one of the possibilities opened up by soft robots.
The latest development in soft robotics is particularly exciting. Recently, researchers figured out how to eliminate the last hard, metal components in soft robots. Where previous soft robots have still required components like metal valves, this latest soft robot can function using only rubber and air — with pressurized air replacing the need for electronic innards. In doing so, it integrates memory and decision-making directly into its soft materials, using a kind of digital logic-based soft computer.
“These states of high pressure and low pressure are analogous to a digital signal, with 1 equal to high pressure, and 0 equal to low pressure,” Philipp Rothemund, one of the researchers on the project, explained to us. “Typically, the control of soft robots is done with electronic controllers and using hard valves resulting in hard/soft hybrid robots. Our soft computer allows integration of complex control directly into the structure of a soft robot.”
All of this is still at a relatively early stage, but it demonstrates the fast-moving world of soft robotics: lagging a half century behind its harder counterpart, but hastily doing its best to catch up. It seems to be working, too. In addition to the applications already mentioned, soft robot grippers are now being used on assembly lines, due to their ability to interact with objects without the potential of damaging them.
As Giada Gerboni notes, however, it’s a mistake to view soft robotics as being in conflict with traditional hard robots.
“I would not say that soft robots are better, but it is only a class of robots — or a way to do robotics — that we cannot avoid considering anymore,” she said. “Soft robots have already demonstrated to have great potential in navigation tasks because they can articulate their body easily, and their navigation is not compromised by abruptly contacts with unknown objects. But when it comes to another very common task for robots like manipulation tasks, then having non-rigid parts is very bad because the robot can [exert very little force, meaning that it cannot lift many objects.] Also, given the high body flexibility, degrees of freedom, and limited sensor integration, they cannot be precisely controlled so they cannot perform a very precise manipulation task.”
BOTH HARD AND SOFT
Ultimately, she says, the best way to truly make robots that will be useful in our lives will be to combine both hard and soft: just as we see with the materials found in nature. “Just look at humans,” she said. “They are perfect robotic machines with soft components, but also stiff structures and standard joints. [They] can carry out tasks in different contexts, but can also be precise and exert forces.”
Some researchers are even busy exploring materials which allow soft robots to stiffen on demand, or else combine different rigidity materials for achieving both precision and dexterity. The term “soft robot” has barely been around for a decade — but it’s already carving out a niche for itself. (And finding its way into pop culture, too: check out Baymax, the lovable robot from Disney’s Big Hero 6.)
As we move toward a world in which robots are part of our everyday life, soft robots are going to take on more importance than ever. One thing’s for sure: here in 2019, the word “robot” is no longer synonymous with “metal machine.”