Humans Are Soft. Robots Can Be, Too.

ECE Associate Professor Kris Dorsey is designing soft robots that are safer and more user-friendly than traditional robots that could assist with delicate situations without causing harm.


This article originally appeared on Research at Northeastern. It was published by Marigo Farr. Main photo: Origami robot during a test. The black hexagon on the front of the robot is used for optical tracking. Photo credit: Sonia Roberts

Picture the iconic robot duo from Star Wars: R2-D2 with its dome-like swivel head and C-3PO with its polished gold plating, jerky movements and clumsiness. While the two may have cute robot personalities, they’re not exactly the best huggers. And if one of them rammed into you, its metal and plastic exterior would be enough to knock you out.

Now, imagine those robots helping you bend your injured leg, or guide an elderly person out of bed. By the mere fact that a machine is imperfect, it could get something wrong. It could bump you or hit something.

This is the risk of bringing robots into our everyday lives, says Kris Dorsey, Associate Professor of Electrical and Computer Engineering as well as Physical Therapy, Movement and Rehabilitation Sciences. She knows that the potential for robots to help people is immense, especially in healthcare settings, but safe implementation is key. And that’s what soft robotics offers — robots made from materials like soft plastic or rubber that move more organically and are less likely to cause harm.

Or, as Dorsey puts it, “if they hit you, you’d be kind of annoyed rather than injured.”

“If we want to bring robots into our daily lives — and that’s a fair conversation to have about whether we do or not — we can’t keep making robots in the same way,” says Dorsey. “We need to make them soft. We need to make them friendly.” And friendly doesn’t mean they’re loyal to their jedi owners, or even that they look like humans. In fact, Dorsey says that a soft robot arm looks more like a balloon animal at a birthday party.

But despite their gentleness, the tricky thing about soft robots is that their movement can be hard to predict. And that’s where Dorsey’s research comes in. She’s trying to understand exactly how flexible materials behave.

“One of the challenges with soft robots is understanding how that balloon is going to deform…when it brushes up against a wall or someone squeezes it in a particular way.” If you push on the balloon and it bends in a way you didn’t expect, that could affect the next command you give it.

For a human, it would be no big deal, because lucky us, we have something called proprioception, which is a sense that lets us perceive the location and movement of our bodies. Basically, we can tell where we are in space. Robots don’t, so we need their movements to be predictable and precise.

With a 2019 grant from the National Science Foundation, Dorsey did a deep dive into the mechanical and electrical properties of a sensor specifically designed to measure the position of soft robots within a space. This work prepared her for a forthcoming study submitted to a robotics conference, in which the goal was to see how a flexible, 3D-printed material called NinjaFlex, behaved as a component of a robot.

Her team assembled origami block-shaped materials out of the material and stacked them one on top of each other to make an arm. Then, they checked to see if the arm ended up where it was supposed to be after it contracted and expanded.

“The analogy I might use…is asking [a person] to raise their hand off a table by 10 centimeters versus measuring the distance from their hand to the table with a ruler — the ‘truth.’”
When they used sensors to compare the true location of the robot to where the robot “thought” it was, they had a pretty good match — within a millimeter and a half. Dorsey said that the results were “good news.”

Origami block with conductive fabric sensors that sense its position. Photo credit: Kris Dorsey

One unique aspect of this study is the way Dorsey’s team took these measurements. Typically, scientists use cameras to track reflective dots placed on the material. But that method requires a lot of set-up and it’s expensive. In Dorsey’s study, they attached sensors directly to the robot arm and got the readings on its position that way.

While much of Dorsey’s work focuses on the technical aspects of soft robots, the bigger picture matters to her. “I always like keeping whomever is going to use the thing that we’re making in mind,” she says. “I think in order to make sure I’m designing the right thing, and my students and I are working on the right technical problems, we have to understand not only the technical problems, but why is it important that we’re doing what we’re doing?”

Beyond the wide range of healthcare and potentially lifesaving applications for robots, Dorsey imagines a future in which robots can make anyone’s life easier in everyday situations: Like an extra arm that could be stored in a backpack or on a belt and inflate as needed.

“When I’m putting up a picture on my wall, and I really want something that will hold the nail in place that’s not my hand, can I have a little robot arm that comes and holds the nail for me?” Dorsey says this isn’t a five years down the road kind of thing, but a look into a future that offers people more ease and convenience.

And her vision is not without nuance. She doesn’t shy from either the practical or ethical questions that robots introduce. On the practical front — how do we design robots to be recyclable? Or biodegradable? On the ethical front, in what ways could robots replace humans to the social detriment of society?

“I certainly have labor questions about how we design robots to make people’s jobs easier but not replace them,” she says. “Or if our robots do end up replacing people’s jobs, what avenues do people have available to them for new jobs?”

The intention behind the robot’s use is also a key ethical concern for Dorsey, because she anticipates that certain kinds of robots will be a goldmine for corporations looking to invade people’s privacy.

“People will tell cute, inanimate objects or cute avatars different things than they would tell another person,” she says. “But then if that cute little thing can record your entire conversation and share it with a company for commercial purposes, or share it with someone else for other nefarious purposes, we should probably think about what that means for society.”

Related Faculty: Kris Dorsey

Related Departments:Electrical & Computer Engineering