Design a smart bracelet

MADRID (EFE).— A scientific team has developed a wrist-worn device that allows users to interact with computers using hand gestures, such as typing.

Details of the bracelet were published yesterday in the journal Nature, in an article focusing on efforts to build a generic neuromotor interface that would enable interactions with computers.

Researchers claim their prototype, which has been the subject of nearly ten years of study, can translate electrical signals generated by wrist muscle movements into computer commands without the need for laborious, customized calibration or invasive procedures.

The findings, they emphasize, could help make human-computer communications more fluid and accessible on a large scale. Behind this development is a team from Meta's Reality Labs, led by Patrick Kaifosh and Thomas Reardon.

Traditional methods of human interaction with technology, such as computers and smartphones, require direct contact through input devices like keyboards, mice, and touchscreens, the magazine notes, adding that this can be limiting for some people.

The new research proposes a highly sensitive, slip-on cuff that can detect electrical signals from wrist muscles—via surface electromyography—and translate them into computer signals using training data from thousands of participants.

This wrist-based electromyography technology represents an innovative way to monitor devices throughout the day, Meta describes in a summary.

According to the company, creating a bracelet of this type is an extremely challenging task, as it must take into account differences in human anatomy, physiology, and behavior: the shape and function of muscles and how each person moves.

“For the first time, we have revealed a formula for creating a surface electromyography neuromotor interface that works immediately for a variety of human movements and computer controls, without requiring laborious and cumbersome calibration.”

For its implementation, the team used deep learning to create generic decoding models.

By Bluetooth

The device, which communicates with the computer via a Bluetooth receiver, recognizes gestures in real time to enable a wide range of interactions with the device.

The tests involved virtual navigation and selection tasks, as well as handwriting text entry at a speed of 20.9 words per minute (the average typing speed on a cell phone keypad is about 36 words per minute).

Engineers also found that personalizing writing models with minimal adjustment for each participant can result in a 30% improvement in performance.

The authors suggest that their neuromotor bracelet provides a portable method of computer communication for people with reduced mobility, muscle weakness, paralysis, and other conditions.

In addition to detailing the key technical and conceptual advancements, the paper offers the community a blueprint—a set of key design rules and best practices in hardware, experimental design, data requirements, and modeling—for what it takes to create a generic neuromotor interface, which Meta hopes will be widely used.