Innovative Wristband Uses Sound Waves to Track Every Hand Motion and Direct Robotic Hands Wirelessly

MIT engineers have created an innovative wearable wristband that can measure hand movements with super-high accuracy, even minor shifts in between. Dian Li, a graduate student, demonstrated the technology by moving her hands around as if she were in real life, and a robot hand on the opposite side of the room could duplicate every finger bend and palm tilt.

This little band employs tiny ultrasonic stickers that sit flat against the skin, just like a watch, and compact electronics around the size of a phone manage the processing, all of which sits snugly on the band itself. Sound waves enter the wrist and bounce off the muscles, tendons, and ligaments, creating a vivid black and white image of what is going on inside your wrist. And those images demonstrate how much the tissues stretch and glide as you curl or extend a finger.

LEGO Disney & Pixar Wall-E & EVE Building Set for Adults, Ages 18+ – Home Office, Book Shelf, or Room…

• LEGO SET FOR ADULTS – The WALL-E and EVE (43279) building set offers adults 18 years old and up an immersive construction challenge featuring…
• 4 DISNEY PIXAR CHARACTERS – Builders can create iconic robots WALL-E, EVE, M-O and Hal from the hit movie—each with authentic functionality like…
• MINDFUL BUILDING EXPERIENCE – This detailed construction set lets builders practice advanced construction techniques for an immersive and relaxing…

The team has compared tendons to puppet strings, and one of their team members, Gengxi Lu, stated that being able to take a snapshot of those strings at any one time provides a very precise picture of where your hand is positioned. Your fingers can move in a variety of ways, from basic bends to many various angles, and ultrasound images reveal every single one of these changes in crystal clear clarity. An AI algorithm then takes this information and explores all of the patterns in the photographs, learning how to match them up with the real motions, all with the help of some training from volunteers who have provided the program with a plethora of labeled samples.

So, during these recording sessions, volunteers sat down with cameras tracking their hand movements as the band collected ultrasound data. The AI then went through and studied the matched pairs until it was able to figure out the movements for itself using only a fresh image. Then, when they tried it on eight people with various hand and wrist shapes, they discovered that it recognized every single move, whether it was spelling out the 26 letters of American Sign Language, picking up a tennis ball, a plastic bottle, a pair of scissors, or even a pencil. And the forecasts came in quickly enough for them to apply it in real time without any issues.

It’s fair to say that other approaches are rather constrained in their own ways, yet this band manages to overcome all of them. Cameras tend to lose track if there is an obstruction in the path or if the lighting changes even slightly. Sensory gloves tend to get in the way and reduce touch sensitivity. Sensors that detect electrical impulses from the forearm may receive a lot of background noise and miss the subtle distinction between open and closed postures. The ultrasonic approach from the wrist avoids all of this by simply looking at the movement source directly, eliminating the need for any specific views or covers.
[Source]