soft robot

The robots look somewhere between ridiculous and creepy as they shimmy up walls, bop up and down, and sort crockery. Ridiculous because their motions seem like the machine equivalent of a drunk uncle on a wedding dance floor, all extraneous motion and unearned confidence, but creepy because there’s something deeper, something more fundamentally alive in strange, silly movements.

These robots — soft robots, to be a hair more technical — may look like a mere novelty, but there’s a reason the research behind them is featured Wednesday in the journal Science Robotics. Jamie Paik, one of their creators and the director of the Reconfigurable Robotics Laboratory at the Swiss Institute of Technology, tells Inverse these machines have a chance to revolutionize how humans and robots interact. It’s a shift that’s really begun to take hold over the past three years.

“Our focus in robotics and design of robotics has changed from making robots faster and more intelligent, more accurate than humans, to how we bring robots into our everyday lives,” she says. “And in that case what’s most important is not so much the mechanical performance but the safety, the intuitive usage, the interactivity, and the adaptiveness to certain tasks that are not necessarily known at the time of construction of the robot.”

When thinking of the robot of the future, Paik’s descriptions don’t sound much like C-3PO, or even a real-life military robot like the terrifying “dog” Spot. Rather, she imagines robots that are part of our lives in much the same way that, say, smartphones are. An iPhone is fundamentally a tool, or really dozens of tools in one: phone, camera, web browser, music player, instant messenger, flashlight, among many others.

soft robot
Shimmying up the wall.

Soft robots like the ones Paik has developed aspire to that same kind of multifaceted usefulness. They would have apps just like your phone does, meaning you wouldn’t need to be an expert roboticist to use them anymore than you need to be a programmer to navigate all a smartphone can do. Where Paik’s team has made the biggest advance over what was previously possible lies in the robots’ versatility.

“There’s a lot of roboticists who talk about modularity of the robot, and the plug and play-ness of modules, but they’re never really truly plug-and-play,” she says. “You have to spend a day assembling or disassembling, spend another day designing, spend another day programming. So that’s the extent of modularity. This one you can take it apart and put it back in and it will still be functional.”

For now, these robots still are not too much more than toys, but it’s early days. The most basic commercially available version of these robots would live in your toolbox, providing simple, preprogrammed assistance for household tasks.

“When you need to have something hold your nails for you, and you don’t have a free hand or arm, these are the robots you can call out,” she says.

soft robots
Or sort tupperware, as the case may be.

But that’s just the beginning. Wearables are a particularly exciting application for these robots, she says. These robots’ programmed movements and ability to take in sensory data from its environment and adjust — the crucial bit that makes a robot a robot — could help them guide humans through physical therapy. Attach the robot to, say, a sprained ankle, and it would help you rehab it with precisely the right movements to make your recovery go smoothly. If you’re looking to up your sports game, it could help you perfect the mechanics of the swing of a bat or golf club, or help you build the muscle memory to throw a tight spiral.

Building in this kind of versatility could also make a big difference when the stakes are highest. Robots already play a crucial role in modern surgeries, but no two patients are ever the same, which means as much as half the procedure is taken up with simply reprogramming the robot to do its job correctly. Soft robots like these could cut down on that time, intuitively adjusting to the patient’s body.

These soft robots don’t yet incorporate artificial intelligence into their programming, and Paik says she’s not an expert on the subject. Researchers focused on A.I. are working on robots’ minds, whereas she’s looking much more at the body. But improving the latter will ease the task of the former.

“It really reduces what A.I. experts need to do,” she says. “So if you have a very dumb body, there’s only so much a 70-year-old body can do with a really, really smart head. It’s much nicer to have a novice, 16-year-old body that may not have as much knowledge, but the body is already so much more agile that it’s less prone to break its hip when it falls down. I’d like to think these types of prototypes and hardwares give a much better platform for A.I. engineers and roboticists to give more tasks that are more exciting.”

Photos via Robertson and Paik, Sci. Robot. 2, eaan6357 (2017)