Every breath you take, every move you make, wi-fi’s watching you

A new technology that uses wi-fi router signals could soon help health professionals remotely monitor patient breathing and detect any irregularities.

The federal National Institute of Standards and Technology (NIST) and FDA researchers used a deep learning algorithm (which falls under the machine learning subset of AI) to examine how body movements alter wi-fi router signals. They found that changes in the path radio frequencies taken from routers to computers or other devices could be used to detect various respiratory issues.

It’s not the first time scientists have studied wi-fi signals to monitor breathing and other vital signs. One team of researchers examined wi-fi monitoring in a 2020 paper, while others at MIT’s Computer Science and Artificial Intelligence Lab published a study on how “radio-frequency re-identification” could be used to monitor individuals.

But new developments, which were detailed in a December 2022 IEEE Access paper, lay the groundwork for potential future remote respiratory monitoring technologies.

“All the ways we’re gathering the data is done on software on the access point (in this case, the router), which could be done by an app on a phone,” Jason Coder, who leads shared spectrum metrology research at the NIST, said in a statement. “This work tries to lay out how somebody can develop and test their own algorithm. This is a framework to help them get relevant information.”

NIST scientists, working with colleagues at the Office of Science and Engineering Labs in the FDA’s Center for Devices and Radiological Health, used a commercially available wi-fi router and receiver to measure the simulated breathing of a manikin (not to be confused with a mannequin) designed to replicate respiratory conditions—ranging from normal to chronic obstructive pulmonary disease—and recorded the wi-fi signal data.

They analyzed the data with BreatheSmart, a deep learning algorithm that can recognize patterns indicating different breathing problems. It successfully classified several respiratory patterns simulated with the manikin 99.5% of the time, according to NIST.

Susanna Mosleh, an NIST research associate, noted that “most of the work that’s been done before was working with very limited data.”

“We were able to collect data with a lot of simulated respiratory scenarios, which contributes to the diversity of the training set that was available to the algorithm,” she said.