Artificial nervous system for robots will allow them to feel touch

A team of researchers from the Korea Advanced Institute of Technology (KAIST) has created an artificial sensory nervous system that allows robots to respond differently to safe and potentially dangerous touches.

The development is based on a neuromorphic semiconductor – a new type of memristor that can process external stimuli without the use of complex software and high computing power.

In living organisms, similar reactions are regulated by the processes of gibbitation (weakening of the response to repeated neutral stimuli, such as noise or the touch of clothing) and

The key component is an improved memristor, which, unlike previous versions, can change conductivity not only in one direction, but also in the opposite direction. This allows the device to gradually reduce sensitivity to repeated stimuli, but quickly activate in the presence of a pain signal.

The system was tested on an artificial hand. Initially, it responded to every touch, but with the repetition of safe stimuli, the sensitivity decreased. When a painful stimulus (electrical impulse) was added, the reaction sharply increased.

According to the authors, this is proof that compact and energy-efficient sensor systems for robots can function without powerful computing resources, which opens up new prospects in the development of robotics and neuroengineering.