Robotic ‘skin’ with flexible magnetic film could improve artificial touch

Robots are clumsy but pliable magnetized “skin” could help them mimic touch better.

A human finger is seen reaching a robotic finger.

A team of researchers from the City University of Hong Kong, South University of Science and Technology in Shenzhen, and Carnegie Mellon University of Pittsburgh, have published a Science Robotics study on robotic tactile sensing, which Wired reports could improve the sense of touch in machines.

Researchers say that touch sensitivity can be achieved to a reliable degree in robots by using a pliable magnetized film on a robotic fingertip to essentially replicate the way human skin works. To give the robot skeletal support of the sort bones give fingers the team used a circuit board covered in magnetic sensors. If you press down on the robotic fingertip, the sensors note a depression in the “skin” of the tip, register the location, and relay that tactile feedback to the computer. It’s not identical to the sensory loop we humans have for touch but it’s surprisingly close to it.

We’ve got a long way to go — A robot will never be able to emulate the ultra-sensitivity of human touch, at least not 100 percent correctly. Our skin has bundles of sensory nerves that then connect to neurons, which are found in our spinal cords. Mechanoreceptors like thermoreceptors and nociceptors in our skin sense temperature and pain respectively. Touch is processed in humans by feeling something — a hand, table, leaf, hot stove, what-have-you — and those signals travel to the thalamus, which then relays sensory signals to the rest of your brain.

Your somatosensory cortex then processes signals related to touch. All of this is to say that a robot does not have any of this. So for engineers to emulate even a fraction of touch effectively within robotics, they have to find alternatives or close approximations that can achieve similar results.

In this case, that’s meant creating magnetic fields on the robot's fingertips, as Wired points out. If successfully done, such sensitivity could help to create prosthetic hands and other parts of the body that grasp, reach, hold, and feel like an actual human does. Another example is that an effective prosthetic hand could help someone read Braille more accurately or hold a slippery dish tighter.

Slippery when wet — Researchers want to create touch-sensitive robotics that not only know how to manipulate touch and exert pressure if needed, they also want to improve shearing force among bots. This means that if a robotic hand can feel an object (like a cup) slipping out of its grip, it knows to hold the object more firmly. Of course, all of this trial and error is limited to laboratories at the moment. But if engineers can get machine sensitivity right and mainstream it to the public, robots will become a lot more dexterous.