Alexa could soon be the ultimate baby monitor.
Researchers at the University of Washington created a speaker system that is capable of monitoring the breathing and movement of sleeping children. The work paves the way for a baby monitor that can track how much sleep kids are getting and detect problems like sleep apnea.
Parents already use Amazon’s Alexa and other smart speakers to play soothing noise and as a baby monitor, but the researchers took things to the next step by using microphones on the speakers and algorithms to decode breathing and movement patterns.
The system works by playing white noise, detecting how the sound reflects off of the bodies of infants, and turning that information into a breathing rate.
“By tracking the specific distance and figuring out exactly where the chest is moving, we can figure out the exact breathing pattern of the infant. And we can do it in a completely contactless manner,” Shyam Gollakota, a professor at UW and co-author on the study, told GeekWire.
Locating the direction of the child helped the team to overcome the challenge of detecting the subtle movements of babies’ breathing. Due to limitations on how developers are able to use microphones on consumer smart speakers like Alexa and Google Home, the researchers created their own speaker system for the study.
Gollakota and his collaborators plan to create a public version of the speaker skill through Sound Life Sciences, a startup that spun out of UW earlier this year. The team initially created Sound with the idea of developing a smartphone app that can monitor opioid users for signs of an overdose. They are also working on a consumer version of a smart speaker app that listens for signs of a heart attack.
The researchers tested the speaker on five babies in a Seattle-area hospital’s neonatal intensive care unit and found that the system could detect breathing patterns that mirrored readings from the hospital’s own monitors.
There are a number of devices currently available that measure heart rate and breathing on babies, from smart socks to smart cameras. One advantage of the researchers’ system, called BreathJunior, is that it might work on smart speakers that people already own.
Going forward, the team wants to advance the system so that it can play lullabies in addition to white noise. Gollakota worked on the project, which was funded by the National Science Foundation, with doctoral student Anran Wang and Dr. Jacob Sunshine.