Artificial Intelligence (AI) has become very good at seeing, talking, and making things. However, it has not been great at sensing touch or feeling how rough or smooth something is.

At Stevens Institute of Technology, researchers have now found a way to teach AI how to feel surfaces. They do this by using a laser that fires light at an object.

The researchers describe the methods and results of this study in a paper published in Applied Optics.

When laser beams hit a surface and bounce back, they create a pattern known as speckle noise, which is usually seen as interference in images. However, the team at Stevens uses this noise to learn about the surface. They’ve trained their AI on how these patterns change with different textures.

A Stevens press release describes this setup is a blend of AI and quantum science, because quantum interactions provides critical information to the AI. “This is a marriage of AI and quantum,” says researcher Daniel Tafone.

The researchers tested their method on sandpapers with different roughness levels. The roughness they measured was in microns (one micron is a thousandth of a millimeter). The researchers found that their system could tell very fine differences in the surface texture, down to about 4 microns after some improvements, which is as good as the best tools currently used in industry.

Applications to medicine, manufacturing, and robotics

This technology could have lots of uses. For instance, in medicine, it might help doctors tell if a mole on the skin is just a harmless mark or something dangerous like melanoma. AI could discriminate by sensing tiny differences in the mole’s surface that are too small for the eye to see.

In manufacturing, where even the smallest flaw can cause big problems, this technology could check the quality of parts by measuring how smooth or rough they are.

This could be integrated into devices like self-driving cars or robots that already use similar light-based technologies for navigation and mapping.

AI with touch-like sensory capabilities would better understand the world around them.

Let us know your thoughts! Sign up for a Mindplex account now, join our Telegram, or follow us on Twitter.