The shape of things to come: Assisting humans with soft robotics

A woman using a robotic glove and hologram of an assistive dog looks up information on a computer kiost, artist's rendering

As engineers look to solve harder and harder problems in robotics, they are pushing the boundaries of what a robot looks like. Nowhere is this more evident than in the emerging field of soft robotics.

"One need for soft robotics is for human-machine interaction," says Wanliang Shan, assistant professor of mechanical engineering. "Think about a patient in a hospital that needs care. You cannot ask a very rigid and cold metal robot to do that. It makes the interaction much harder. Another one is for machine-environment interactions. After a huge earthquake you want to send the robot into the wreckage of a building to find people. If you have a conventional rigid robot, it might get stuck anywhere. But if you have a soft robot, it might be able to squeeze through."

Shan is working toward a future in which soft robots are capable of the kind of autonomy and interaction existing rigid robotic platforms have achieved. But before soft robotics researchers can tackle the kind of control and reasoning problems traditional roboticists are working on, they have to develop soft materials capable of the sensing and movement metallic machines have.

"Right now for soft robotics we are proposing that we actually use soft materials, like elastomer, rubber or gel to achieve the functionalities that traditionally have been achieved using metal, using silicon," Shan says. "People design all kinds of composite materials right now to achieve different kinds of functionalities."

Materials of the future

Shan's lab is exploring multifunctional composite materials and the mechanics of soft materials for soft robotics. Here's a look at two promising technologies:

Flexible dry adhesion

Designed to help robots handle thin or fragile materials, Shan's group is developing a soft material that will stick to an object - imagine a high-tech suction cup-enabling a robot to pick it up, move it and then release it by softening a core inside the material. Shan's group has an NSF grant dedicated to exploring the fundamental research underlying this approach, which Shan hopes will be safer, more cost-effective and more energy efficient than existing methods.

Robotic needle

Shan is working on a flexible robotic needle that would be capable of bending significantly more than current needles on the market. In surgical applications, this could dramatically reduce the invasiveness of certain procedures, but the fundamental mechanics of how the needle bends and buckles need to be studied in depth, so precise control algorithms can be developed.

"I envision that in 10 years the emphasis of soft robotics research will shift from materials development to the integration between hardware and software," Shan says. "We'll have the capability for what the rigid robots can do right now and even more."

Co-robotic systems

As experts look forward, many envision a blurring of the line between humans and robots. Some robots will work alongside humans, but other kinds of robotic devices will be designed to augment humans.

"I think the trend is to work together," says Yantao Shen, associate professor of electrical and biomedical engineering. "Currently humans are humans, and devices are devices, but later those devices could become very organic; they could meld with your body. The human eye can see a very narrow spectrum, from 400 nanometers to 700 nanometers, but sensors can extend the spectrum and help us see, for example, infrared. So that's expanding the capabilities of the human body."

That's a research direction Shen is very excited about. Much of his research has focused on developing smart assistive technologies to aid users with visual impairments.

Fingereye

This technology, under development by Shen and collaborators at the University of Arkansas, Little Rock, combines vision, tactile, force, temperature and audio sensors and actuators to help the wearer pre-sense an object - telling its location, feeling its shape and size - and then grasp it.

"The miniaturized system will contribute to the lives of visually impaired people by enabling them to identify and move objects," Shen said. "We will build a lightweight form-fitting device that attaches to the hand using key locations for cameras and mechanical and electrical sensors. It will be simpler than a glove, and less obtrusive."

E-Braille

With funding from the National Science Foundation CAREER program, Shen is developing a wearable fingertip scanner that captures text and converts it to braille signals, transmitted via tactile sensations on the wearer's finger.

"This technology will allow the blind and visually impaired to conveniently access more documents, books and libraries, anytime and anywhere," Shen said. "It's based on electrical stimulation with online skin bioimpedance sensing and tactile-preference rendering functions."

According to Shen, robotics - in particular the kind of devices that are being developed with biomedical applications in mind - are trending toward smarter, softer, less invasive technologies. Shen imagines our current trend toward wearables proliferating as the devices themselves practically disappear.

"Currently wearable systems are actually pretty huge," says Shen. "When you wear them, you feel them. Later if biomedical robotic systems are like part of your body, you will not feel them, but they still can do similar functions. That's the trend."

The devices themselves will become smarter as well, Shen says. Shen imagines prosthetic legs that don't need to be controlled by human wearers but instead learn how to generate force and motion so that they are capable of naturally controlling themselves as a user walks.

Of course, there are a number of technical challenges still to be solved. Chief among them are developing biocompatible materials; integrating multiple functionalities on small, thin, flexible materials; and one that Shen is particularly passionate about solving: endowing these devices with sensation and feeling.

"One important issue is currently all those prostheses cannot generate feeling to your body," he says. "They can generate actuation. Actuation is no problem. I move my arm, I can control it and grasp with it. But you touch with it; your brain, your body cannot feel the temperature, cannot feel the roughness and softness, cannot feel those kind of things. So that's another challenge. How can you, based on the physical sensor measurement, generate a signal to the human body or neurosystem to help you feel it."

It's a tall order, but Shen sees progress on the horizon.

"In 10 years we will have small, flexible devices that can measure a lot," Shen says. "You can still feel them, maybe, but it's a big step."