Infants at risk of cerebral palsy are getting a bit of support from a skateboard, a robot, and a machine-learning algorithm at the University of Oklahoma (OU) Health Science Center, where this year a team of researchers have embarked on the third version of an innovative effort, IEEE Spectrum reports.
As infants mature, they often learn to crawl in an effort to reach some goal, such as a toy or another interesting object. In this case, learning to crawl is reward-based. But, when an infant suffers from cerebral palsy, her movements and muscle coordination can be severely stunted, and the motivation to keep crawling may be reduced since the task is either too difficult or doesn’t result in a reward.
With this motivation gone, as the infant focuses energy elsewhere, the brain stops building and reinforcing vital spatial cognition and motor connections, leading to further issues later in life.
The Self-Initiated Prone Progression Crawler (SIPPC) — invented by physical therapist and researchers Thubi Kolobe and Peter Pidcoe — sees infants lie on a padded skateboard, while strapped to a robot and wearing a cap packed with dozens of electrodes to monitor brain activity. The cap transmits the infant’s movement to a 3D screen as a camera on the robot records the movements of her limbs. This data is finally relayed to a machine-learning algorithm that interprets what actions the infant is attempting to perform and informs the robot to move slightly in accordance with the infant’s desires.
The comprehensive effort rewards infants for simply attempting to crawl by assisting in their movements.
“As soon as you start to crawl, the world seems like a much bigger place,” OU engineering professor Andrew Fagg told IEEE Spectrum. “We hope, with the crawling, we’ll set them up to build other capabilities that will be really important later on in life.”
In the trials, Fagg is joined by Kolobe and engineering professors David Miller and Lei Ding. They realize their cause is honorable but Fagg admitted that fatigue is a factor. “It’s wearing everybody down,” he said, after 1,000 sessions, 10 gigabytes of data, and another six to nine more months of research ahead of them.
And although Fagg insists that more research needs to be done before anything definitive is concluded, results from this year’s study have been in line with what the researchers found in their pilot study and — perhaps just as promising — the parents of the infant participants are already eager to take a device of their own home.
- The DroNet algorithm teaches drones to navigate city streets like cars
- Bad habits? This tech predicts impulsive behavior, zaps your brain to stop it
- Rise of the machines: Here are the best robots we saw at CES 2018
- This bug-like robot is learning to improvise on the go
- Spirit animals: 9 revolutionary robots inspired by real-world creatures