In a day and age when everyone worries about surveillance and privacy, the last thing we want to hear is that technology has cooked up yet another way to identify and track us.

A team at MIT has done exactly that, but the group emphasizes its technology is intended for safety and health-related uses. Researchers at the Computer Science and Artificial Intelligence Lab have found a way to use Wi-Fi signals — the same signals that already permeate the air you're breathing right now in your home or office — like sonar. Their technique allows for locating a person's shape in the Wi-Fi soup, tracking their movements, and even learning to identify them based on their silhouette.

It's both cool and kind of creepy.

The technique, dubbed "RF-Capture," works through walls, too. Right now it's accurate enough to identify a human shape down to the shape and movement of their hand, but the MIT team expects it to return higher-resolution images in the future. At the moment, it can also capture a person's breathing and heart rate. In the future, it could be capable of detecting things like heart problems.

Something they've discovered is that every person has a unique Wi-Fi silhouette, just as everyone has a different fingerprint. With repeated use, the MIT team believes learning algorithms can be employed to allow RF-Capture to recognize and remember a person based on their Wi-Fi silhouette. They postulate that this process could be used to create devices for scanning an elderly person's home to monitor them for falling over, and then calling 911 automatically.

The team is presenting its work on RF-Capture at SIGGRAPH Asia in November. A paper [PDF] was published this morning on its website at MIT.

The group makes it clear that privacy is a top concern, so encryption is being built into the technology from the start. They also intend to utilize blockers that will keep unauthorized users out of an individual's Wi-Fi system.

ⓒ 2021 All rights reserved. Do not reproduce without permission.