Ů��С��Ƶ

An artificial skin is even better than human skin at sensing objects, because it can detect and identify items that it hasn’t touched yet.

“Human skin has to touch something to tell it what is there,” says at Nanyang Technological University in Singapore. “Human skin can only tell the softness or hardness of an object. We wanted our artificial skin to have more functions.”

Even without touching an object, Wang and his colleagues’ artificial skin can sense if it is close by and can also discern some clues about the type of material it is made of. “We can tell whether it’s a piece of metal, plastic… or some biological material,” he says.

The skin is made up of two outer layers of conductive fabric coated with nickel to serve as electrodes. These surround a porous sponge soaked in ionic liquid, which is a salt in a liquid state that acts as a conduit for electricity. The two layers act as a capacitor, storing electrical energy in an electric field.

The ions in the sponge boost the performance of the capacitor, which effectively measures how much the distance between the two layers of electrodes changes. That ability to detect tiny shifts is behind how the artificial skin is able to detect that it has touched something.

The sensing performance of the capacitor, which Wang claims is between 10 and 100 times more sensitive than a standard capacitor, means it is also able to detect very small changes in the electric field around the skin, allowing it to sense when objects are near. What’s more, those subtle changes can help it identify what type of material a nearby object is made of.

In tests, the skin managed to detect and successful classify a series of objects brought near it as being either polymer, metal or skin, indicated by specific changes in the capacitor’s measurements.

“The process is relatively simple. As the component comes close to contact, it enters the edges of the electric field of the capacitive structure,” says at the University of Sheffield, UK. “There are several interesting future routes,” he says, but at present the skin relies on machine learning techniques to identify how the object it detects compares with data on known materials.

Wang thinks the skin could work on a robotic finger to allow factory robots to better understand which objects to pick up and which to leave without having to grasp them, as well as being useful for prostheses.