ReSkin works thanks to the use of particles that generate magnetic fields. When it touches a surface, the skin deforms. A circuit board monitors it and sends information to artificial intelligence, which tries to understand the changes in force and touch that have been made to transmit a response.
"ReSkin's touch skin technology will provide a source of rich contact data that could be useful in powering AI in a wide range of touch tasks, including object sorting and robotic grasping," stated Abhinav Gupta, the head of this research, to NewScientist.
To create the synthetic skin, robots were trained by using a key to open a door and ordering them to grasp delicate objects, such as grapes and blueberries. To grab the fruits, the researchers coated a robotic gripper with ReSkin. The research team shared some images of the robots collecting blueberries with the right pressure from a surface on the blog of the artificial intelligence division. They picked up the fruit without causing any harm to it. "The caliper's built-in force sensor is unable to complete the task, but with ReSkin, it can detect force feedback well enough to control its grip," Gupta stated.
Other tests, in which ReSkin was used to monitor the movements of a dog, the force used by humans performing manual tasks, and locating contacts in larger surface areas, such as the body of a human being, were also carried out.