The Wi-Fi radio waves that our laptops use to communicate with wireless routers are quite sensitive to the environment they pass through. They’re so sensitive that a few years ago researchers at MIT were able to use them to wirelessly detect a child’s movement, breathing, and heart rate. Now the MIT team working with Massachusetts General Hospital have successfully demonstrated that they’re able to repurpose this Wi-Fi data to detect what state (light, deep, or REM) a sleeping person is in.
The technology relies on sophisticated algorithms that are able to tease out, and interpret, the changes in the heart rate, breathing, and movement, and to convert the data into an indicator of the sleeping stage someone is in. The artificial intelligence algorithms that do this essentially learned what is important and what can be ignored in the data coming form the Wi-Fi system. As the irrelevant stuff didn’t have to be analyzed, the remaining useful information could then be processed much easier.
In a study of 25 healthy individuals that were analyzed using EEG as well as the Wi-Fi system, both methods were about 80% accurate in providing sleep specialists the sleeping state of a given subject. This is quite impressive and perhaps a non-intrusive Wi-Fi device will soon be available for people to monitor a few of their vitals and sleep quality at home.
Here’s a short video from MIT about the technology:
Paper presented at International Conference on Machine Learning: Learning Sleep Stages from Radio Signals: A Conditional Adversarial Architecture…
Via: MIT…