In a recent achievement, a team of engineers created an artificial fingertip that reacts to texture and forms in the same way as a human fingertip does.
The latest finding, according to experts at the University of Bristol, might make robots far more agile and perhaps aid enhance prosthetics in the long run.
The artificial fingertip was created by Nathan Lepora, a robotics and artificial intelligence (AI) professor at the University of Bristol, and his colleagues using a 3D-printed mesh. The movement observed between the inner and outer layers of human skin is imitated by this mesh.
The researchers started working on an artificial fingertip in 2009, using human skin as a basis.
Their first fingertip, which they had made by hand, was roughly the size of a Coke can. Then, by 2018, they’d moved on to 3D printing, which allowed them to produce the tip and all of its components around the size of an adult’s big toe and develop a succession of layers that resembled the stratified composition of human skin.
Scientists have recently put neural networks onto the fingertip, which they have dubbed TacTip. A neural network’s large network of neurons sharpens a robot’s senses and pushes it to react appropriately, much like an actual finger.
When the skin on a human fingertip comes into touch with an object, a layer of nerve endings contorts and transmits a signal to the brain.
These nerves send “rapid” impulses to assist us in avoiding dropping something and “slow” signals to relay information about the shape of an object.
How It Works
TacTip’s comparable signals are generated by an array of pin-like projections beneath a rubbery surface layer that moves when touched.
The array’s pins are rigid but malleable, like hairbrush bristles. A tiny camera beneath the array detects when and how the pins move.
The amount of bending of the pins produces a sluggish signal, but the speed of bending produces a quick signal.
Following that, the neural network interprets those impulses into fingertip actions, such as tightening the grasp on the object or altering the angle of the fingertip.
This discovery could lead to a better imitation of human hands in robotics, with enhanced dexterity, to better grip items and understand the forms that they are touching.
Lepora and his colleagues are now looking ahead, hoping to develop this new artificial skin as excellent as human skin.
To read our blog on “Humans don’t make as good bipedal robots as birds,” click here.
