One of the most troublesome facets of using a wheelchair is maneuvering it through crowded or congested places.
Cashing in on their scientific know-how, leading manufacturers have come up with devices that claim to facilitate this process; however, they have succeeded to a certain extent only.
While people using wheelchairs continue waiting with bated breath for a reliable and dependable invention that would make going from one place to another a little less arduous for them, an international team of researchers has developed a highly portable wireless wheelchair control platform that might do exactly that.
The BMI (wearable brain-machine interface) is an upgraded version of traditional EEG (electroencephalography) tools, which come in handy for measuring signals in the human brain. A vital part of this radical overhaul, BMI enables people with serious motor disabilities, chronic stroke, or ALS to control prosthetic accessories.
Combining deep learning algorithm, flexible electronics, and nanomembrane electrodes, scientists enabled people with disability to wirelessly control an electric wheelchair, mini robotic vehicles, and even a computer. Wheelchair users do not need to put on a heavy hair-electrode cap with all sorts of wires hanging behind.
All they need to do is to slip a headband, which has three elastomeric scalp electrodes, paired with a printed electrode, and an ultrathin wireless electronics patch. Recorded EEG data from the brain gets refined in the flexible circuitry when it is wirelessly transmitted via Bluetooth to a tablet up to 5ft away.
In a statement, Chee Siang Ang, senior lecturer at the University of Kent said, “Deep learning methods, commonly used to classify pictures of everyday things such as cats and dogs, are used to analyze the EEG signals.”
Just like pictures of a dog that can have numerous variations, Chee Siang Ang says EEG signals have a similar challenge of high variability. It is no secret that deep learning methods work extremely well with pictures; Ang says they have shown that the same strategy is effective on EEG signals as well.
Deep learning models were also used to determine the most useful electrodes for accumulating information. Ang says they discovered that the model accurately identifies the appropriate locations in the brain for BMI, thus minimizing the number of sensors needed, while cutting cost and improving portability at the same time.
Until now, the device has been tested on six able-bodied human subjects. The group is leaving no stone unturned in a bid to improve its electrodes and make the system more advantageous for motor-impaired individuals.
The full study carried out by researchers from the University of Kent, Wichita State University, and Georgia Tech was originally published in the journal Nature Machine Intelligence earlier this month.
Image credit: reidy68 / Pixabay