A group of researchers from the Universities of California, Yale, Stanford, Cambridge and the Seoul National University have carried out a study in which they examine how far science has advanced in the design of what are known as e-skins or electronic skins.
The document has been published in Science Robotics and exposes how these electronic skins could help to create robots with a texture similar to that of human skin (“soft” robots) and with the capacity to feel, similar to our own sense of touch . In addition, from the collection of these tactile data, the machine learning of artificial intelligence would be promoted, which allows robots to learn by themselves.
According to the researchers, one of the main challenges still facing the field of e-skins is the integration of cabling into structures. Multiplexing is a technique that can substantially reduce the number of cables required to implement these electronic skins, but even with it, each sensor would need to have at least one connection via two cables.
In addition, two specific robotic functions that could significantly benefit from the combination of electronic skins and machine learning from artificial intelligence are shape detection by robots and control of their feedback . Both would promote more advanced interactions between a robot and its environment, while allowing it to explore what is around it, providing them with valuable information that helps them respond effectively to different stimuli .
Walking towards “affective touch”
Finally, the researchers have considered that the next step is to ensure that the robots also acquire a sense that goes beyond, the affective touch . This would promote their ability to communicate with people, something that scientists consider of vital importance since, in the future, they believe that soft robots will be more than integrated into our lives and will coexist with us in our homes.