Artificial skin that can sense potentially harmful forces could help robots detect danger, in a similar way to how pain helps people avoid hazards.
In humans, painful sensations come from a complex interplay between peripheral nerve signals and the brain鈥檚 interpretation of them. Electrical spikes sent by pain sensors in the skin, called nociceptors, are relayed through nerves to several parts of the brain, which can lead to self-preserving actions, such as moving a hand away from a sharp knife.
Advertisement
at Hunan University in China and his colleagues developed an artificial pain-sensing system using a zinc and gallium crystal. When a strong force is applied to the crystal, electrons are released, creating an electrical signal. The force also produces flashes of light in the crystal, which can be monitored by a camera to indicate where the pain is.
鈥淏y using the electrical and optical signals at the same time, the intensity and the location of the pain can be simultaneously perceived,鈥� says Tan. 鈥淚t functions like a natural nociceptor.鈥�
Tan and his team also developed an artificial intelligence algorithm that can interpret when a stimulus should be avoided. They trained the algorithm on 100 different electrical and optical readings each for a knife, a rod and a cotton ball pressing on the sensor, so it could learn to differentiate between harmful and non-harmful objects that produce similar readings.
A robotic hand equipped with the skin and algorithm successfully differentiated between safe and harmful objects 97.5 per cent of the time. This meant it could decide to hold soft objects, like tofu or a boiled egg, and drop hard ones, such as a spiky ball.
The researchers also attached the skin to the heart, liver and lungs in a rat while a surgical robot carried out a biopsy. The signals guided the robot so that it didn鈥檛 accidentally damage these organs.
at Northeastern University in Massachusetts says the artificial skin could be incorporated into surgical tools to give surgeons the ability to feel things remotely.
Advanced Materials
