Robotics and AI often get a bad rap for the whole destroyer of the human race thing. But when you watch a motorized machine help a baby crawl, you can’t help put feel like robots aren’t so bad after all. And that’s exactly the kind of machine that researchers at the University of Oklahoma built.


Specifically, the Self-Initiated Prone Progression Crawler (SIPPC) is designed to mitigate neurological damage caused by cerebral palsy at an early age. Cerebral palsy refers to a number of neurological disorders that occur during pregnancy, infancy, or early childhood. Infants at risk can suffer from severe loss of motor skills and, sometimes, intellectual capabilities. Although children usually aren’t diagnosed with cerebral palsy until their first birthday, aiding movement in those crucial early months can help children at risk to develop motor and cognitive skills.

So researchers designed a motorized scooter for infants around two to eight months that helps them crawl. Additionally, an EEG cap monitors brain activity during these exercises, while mounted cameras capture movement 20 times a second to create a 3D graph of the child’s crawling.

The centerpiece to this whole robotic operation, however, is the machine learning algorithm which analyzes the infant’s movements and anticipates what the child is trying to do. The crawler then kicks in some motorized assistance to help the kiddo go.


The device featured in the above video is actually third iteration of the robot’s design. Since receiving funding from the National Science Foundation in 2012, the project has seen a series of successes leading to the current study of 56 newborns. Unfortunately, the device is still in its early stages and can’t be used by families at home. But the researchers hope that won’t be the case for long.

[IEEE Spectrum]