New robotic skin responds to touch

Scientists from Singapore have introduced artificial skin, which determines touch with almost 100% accuracy. Development will improve the design of robots that care for patients.

Using Intel’s neuromorphic chip, scientists from the National University of Singapore (NUS) have developed artificial skin that allows robots to detect touch a thousand times faster than the human sensory system. She is also able to determine the shape, texture, and hardness of objects ten times faster than analogs. Researchers believe that their development can improve human-robot interaction, and will improve aspects such as robotic surgery or patient care.

To prove effectiveness, the NUS team first taught the robot arm equipped with artificial leather to read Braille. Then the robotic arm transmitted the tactile data and translated it. The accuracy was 92%, the device used 20 times less energy than analogs.

Scientists went even further and combined tactile and visual data. They taught the hand to classify objects using both artificial leather and a camera. They sent data to the processor and proved that the combination of event vision and touch using a spiked neural network made the system 10% more accurate than using only visual data.

They also found that the device processed touch data 21% faster than the best processor, while consuming 45 times less energy. Researchers presented their results in the scientific journal Robotics: Science and Systems.