'Optical Nerves' Give This Prosthetic Hand a Super Sensitive Touch

The Gentle Bot's touch is so delicate, it's able to run its fingers over several tomatoes and pick out the ripest one.

Each finger looks like a fake rubbery Vienna sausage, but don't be fooled. This new soft robotic prosthetic hand developed at Cornell University has an extremely sensitive touch.

Called, naturally, Gentle Bot, the hand boasts complex, sophisticated components. The bot can handle challenges like grasping a coffee mug, holding an extremely ripe tomato without crushing it, and shaking a real human hand. Nice to meet you, too, Gentle Bot.

Robert Shepherd, an assistant professor of mechanical and aerospace engineering and principal investigator in Cornell's Organic Robotics Lab, headed up the bot's creation. He and his group sought to get as close as possible to human touch, ditching the motors, rigid parts, and bulkiness usually required for making a responsive prosthetic. They published a paper about the bot this week in the first issue of Science Robotics.

Their approach drew on advancements made in optics, particularly in the manufacturing of snaking optical cords called waveguides. Fiber optics are a type of optical waveguide, if that helps. Although that fabrication was once limited to large corporations and national laboratories, it's now possible for academic researchers, like Shepherd, to do their own custom printing using a multi-step soft lithography process.

RELATED: Top Prosthetic Limbs Bring Hope to Amputees

Gentle Bot's innards are quite clever. Stretchy optical waveguides containing LEDs are built right into the pneumatic fingers, allowing them to "sense" the surroundings. When the soft fingers mounted on a rigid palm flex, even a tiny bit, that affects how much the light goes into the device. Those changes are measured by a light detector, or photodiode. The internal optical cords act kind of like nerves.

"Most robots today have sensors on the outside of the body that detect things from the surface," doctoral student Huichan Zhao, the lead author on their paper, said in a university statement. "Our sensors are integrated within the body, so they can actually detect forces being transmitted through the thickness of the robot, a lot like we and all organisms do when we feel pain."

The group's tests bore that out. Gentle Bot was able to check objects for shape and texture, running its fingers over the surface of three tomatoes in a row and then correctly picking out the ripest one. Using one finger, Gentle Bot also measured the softness of various objects like a sponge, silicone rubber, and acrylic.

However, there are limitations. The hand couldn't tell the difference between the acrylic and an unripe tomato. Plus, it's currently hard to localize where a touch comes from, Shepherd told the university. Next steps could include improvements like 3D printing more complex waveguide sensors, increasing the pressure range on the soft actuators, and adding machine learning. I'm hoping we'll see this bot graduate to separating eggs. Then we can truly wave goodbye to robotic butter fingers.

WATCH VIDEO: What Is the Future of Robotic Limbs?