Robotic Finger Can ‘Feel’ With 90% Accuracy

Researchers have developed a system that gives robotic fingers the sense of touch. 

The team from the Chinese Academy of Sciences applied triboelectric sensors to the artificial digit’s fingertips to give it sensation, with these sensors responding differently to objects it was brought into contact with. Results showed the system could identify materials on average 96.8% of the time, with a minimum accuracy of 90% for all of the surfaces. Results were published in the journal Science Advances.

 Previously, scientists have given robotic fingers the ability to discern different temperatures and changes in pressure, however, the Chinese team took this research further by adding the ability to identify the material being touched. 

“To date, several artificial haptic-based sensing techniques can accurately measure physical stimuli,” the team wrote. “However, quantifying the psychological parameters of tactile perception to achieve texture and roughness identification remains challenging.” 

Each fingertip sensor was connected to an individual processor, which was also connected allowing researchers to compare results between the sensors. Small LCD screens were also added to each finger to display the results.

Not only was the system capable of correctly identifying different materials, but it also demonstrated its long-term efficacy, with the team repeating the test thousands of times. 

Further testing is expected, though potential use cases are already anticipated to include improving the capabilities of prosthetic devices and assistive robots, or in the product testing phase of manufacturing.