The staff properly trained their creation’s neural network and tested its precision by collecting 200 several hours of sleep data from 26 topics who experienced to put on sensors on their chest and tummy in the beginning. They said that after teaching the unit on a week’s truly worth of data, it predicted the subject’s correct body posture ninety four % of the time.
In the long term, BodyCompass could be paired with other equipment to prod sleepers to adjust positions, these as intelligent mattresses. When that comes about, the unit could alert individuals with epilepsy if they’ve taken a perhaps deadly sleeping placement, cut down sleep apnea activities and notify caregivers to move motionless clients at chance of building bedsores. It could also assist absolutely everyone else get a great night’s sleep, due to the fact we definitely all want it.
Team member Shichao Yue will introduce the program at the UbiComp 2020 conference on September 15th, but you can read through their paper on MIT’s site (PDF).