Wi-Fi radio signals detect problematic breathing patterns
For many people, Wi-Fi is a figurative lifesaver. Now, new research from the National Institute of Standards and Technology (NIST) might be able to turn it into a literal one as well. By modifying an off-the-shelf Wi-Fi router with a firmware update and using a deep-learning algorithm, scientists were able to detect breathing patterns that indicate respiratory distress in a medical mannequin.
The idea to try to harvest Wi-Fi signals to monitor the breathing patterns of people in their homes was born during the height of the COVID-19 pandemic. "As everybody's world was turned upside down, several of us at NIST were thinking about what we could do to help out," said NIST researcher Jason Coder. "We didn't have time to develop a new device, so how can we use what we already have?".
The answer came in looking at the radio waves that allow communication between devices like cell phones or tablets and the routers they use to connect to the internet. As these radio waves travel back and forth, they encounter obstacles like furniture or people that alter them slightly.
By examining these alterations, Coder, research associate Susanna Mosleh, and colleagues at the Office of Science and Engineering Labs (in the FDA's Center for Devices and Radiological Health) believed they could detect subtle changes in a person's body that would be indicative of breathing difficulties – much in the same way Wi-Fi signals have been used to count people through walls and monitor sleep patterns.
To test out the idea, the team placed a mannequin that simulates breathing in a radio-wave-absorbing room known as an anechoic chamber. They also set up a commercial Wi-Fi router and a receiver. As the mannequin mimicked a variety of breathing patterns including those that would indicate asthma, COPD, and abnormally slow and fast breathing rates, the disruptions in the radio waves were recorded, with data transmitted about 10 times per second.
That led to a tremendous amount of information, which needed to be analyzed in order to discover which wave disruptions corresponded to simulated breathing difficulties in the mannequin. To sift through it all, Mosleh created a deep-learning algorithm the team dubbed "BreatheSmart." Once the equation was established and that data was fed through it, it was found to be 99.54% effective in correctly classifying the breathing patterns.
The fact that the system can work with existing routers gives the researchers hope that it could someday roll out simply through a smartphone app that would deliver the firmware update. They also say that their work creates a framework into which other types of monitoring algorithms could be fit.
Of course, testing conducted in a sealed room with a medical mannequin will be quite different from real-life applications where people are moving around amongst furniture, pets, and each other, but the research is at least proof-of-concept for a system that could hold promise.
Information about the research has been published in the journal IEEE Access.