In a breakthrough that could give robots and medical devices a finer sense of touch, researchers have developed DIGIT Pinki - a miniaturized artificial fingertip that can detect tactile sensations with human-like sensitivity. The project is a collaboration between Meta AI, Stanford University, Technische Universität Dresden, and the German Cancer Research Center (DFKZ). Their innovative sensor, described in a recent paper, combines cutting-edge machine learning with a biomimetic design to digitize the faculty of touch. 

"We are primarily interested in the science of digitizing touch for robots," explains Julia Di, a researcher at Meta AI and co-author of the paper. "With DIGIT Pinki, we wanted to investigate how to build sensors that have touch capabilities resembling those of humans but in the form factor of a human finger."

Inspired by Human Fingertips 

At just 15 millimeters wide, DIGIT Pinki replicates the slim dimensions of a human fingertip. It consists of an optically clear silicone "skin" that deforms when pressed against an object or surface. Embedded fiber optic bundles relay images of these deformations to an external camera and machine learning model.

By training on visual data of how the skin distorts under different forces and textures, DIGIT Pinki can accurately interpret tactile signals like pressure, texture, and shape - similar to real human nerves and brain processing.

"DIGIT Pinki is a type of vision-based tactile sensor that uses images from a miniature camera to get a sense of touch," says Di. "When these gel fingertips touch an object, an internal camera will take images of the resulting deformations in the gel."

A Key Advantage: Safety for Medical Use 

One of DIGIT Pinki's most promising applications is in the medical field, where its slim, electronics-free design allows it to sense inside the human body safely. Cancerous growths, for instance, are often slightly stiffer than healthy tissue - a characteristic DIGIT Pinki could detect through careful palpation.

In initial tests on tissue phantoms and ex vivo prostate samples, the sensor could discriminate between healthy and unhealthy tissue with reasonable accuracy. This proof-of-concept hints at DIGIT Pinki's potential for cancer screening and other medical diagnostics that rely on the clinical "finger test."

"One real-world application is cancer diagnostics in constrained spaces - for example, digital palpation in routine rectal or cervical exams," says Di. "Clinicians may feel cancerous growths with their fingertips as a diagnostic method."

A Robotic Sense of Touch 

While medical uses are an exciting frontier, DIGIT Pinki could also bestow a profound new capability on robots: dexterity and tactile awareness akin to human hands.

Current robotic grippers largely operate through brute force, grasping objects with stronger-than-needed pressure. With artificial fingertips that can perceive texture, stiffness, shape, and grip force, robots could learn to grasp and manipulate objects with human-like gentleness and precision.

"It would be exciting to have these kinds of highly sensitive fingertips on prosthetic hands too," Di says. "We are interested in how to use AI to interpret touch signals for manipulation."

From surgical instruments that can "feel" their way around organs to robotic hands that pluck fragile items, a realistic robotic sense of touch could open up new worlds of capability. As Di and her colleagues continue refining the software and shrinking the hardware, DIGIT Pinki may bring the human faculty of tactile perception to machines in ways previously unimaginable.