For people with sleep disorders, a sleep study can help target the underlying problems behind why they’re tossing and turning all night. But the actual study is obtrusive and uncomfortable, with sensors attached to the person while they sleep. Researchers at MIT have started using radio signals combined with artificial intelligence (AI) to analyze sleep stages without the hassle of wires and sensors.

While the system created by the researchers is just a prototype, using AI algorithms has been producing a high rate of accuracy in tracking sleep. Using radio frequency (RF) signals through a wireless router, data can be collected about a person’s sleep patterns. The MIT team created deep neural networks and AI algorithms to turn the raw data into sleep data. The sleep data was collected without the user having to alter their behavior or be hooked up to any machines.

The researchers plan to use this new technology to study how Parkinson’s disease impacts sleep. Using AI and radio waves can potentially provide answers to the sleep problems that occur in people with other illness, like Alzheimer’s and epilepsy.

Written by IEEE on August 16, 2017