Publications
- Researchers have created an AI device that can make different hand movements
- The device is in the form of an armband and uses brain signaling to make gestures
- Although similar inventions already exist, this invention has a lower power budget
In the near future, making prosthetic hand movements won’t necessarily require any form of physical effort – all it would take is sending signals from the brain, triggered by thoughts.
Engineers at the University of California (UC) collaborated to develop a device that uses electrical signals to make hand signals. This system could be used to control prosthetics or interact with almost any type of electronic device.
A paper with the full description of the device was published recently in the journal Nature Electronics. The structure relies on artificial intelligence (AI) combined with wearable biosensors.
A breakthrough for prosthetics
Ali Moin – a doctoral student at the university who helped with designing the device – stated: “Prosthetics are one important application of this technology, but besides that, also offers a very intuitive way of communicating with computers.
“Reading hand gestures is one way of improving human-computer interaction. And, while there are other ways of doing that, by, for instance, using cameras and computer vision, this is a good solution that also maintains an individual's privacy.”
Creating a system for gesture classification
Designing the device involved creating a flexible armband that can pick up electrical signals at 64 points on the forearm.
These signals are then loaded onto a chip programmed with an algorithm that can link the signals in the arm with specific hand gestures.
The algorithm can successfully make 21 different hand gestures, some of which include a fist and a thumbs up.
With artificial intelligence software in general and the current device, programming of each gesture needs to be done first in order for the algorithm to “learn” them.
This means a user needs to wear the device and make every gesture one by one for the algorithm to remember it.
A device that updates itself
The researchers stated that “[their] system implements a neuro-inspired hyperdimensional computing algorithm locally for real-time gesture classification, as well as model training and updating under variable conditions such as different arm positions and sensor replacement.”
Through its hyperdimensional computing algorithm, the device is able to update itself with new information.
An added advantage of using the chip in the device is that it does not send any data on the chip to nearby devices, which means personal usage data stays private.
“Most of these technologies already exist elsewhere, but what's unique about this device is that it integrates the biosensing, signal processing and interpretation, and artificial intelligence into one system that is relatively small and flexible and has a low power budget,” said senior author of the paper, Jan M. Rabaey.
READ | Is your smartphone or tablet an injury waiting to happen?
READ | Welcome to the 'smart toilet' that can spot disease
READ | Engineers develop cheap biosensor 'tattoos' made with pencil and paper
