Your humble Wi-Fi router (opens in a new tab) Signals can be used to track your movements around a room, Bat-style, a new report says.
Researchers at Carnegie Mellon University recently published a report in which they detail an experiment that used common off-the-shelf Wi-Fi routers to detect the position and posture of people in a room.
The experiment, while not without its flaws, was generally successful, proving that endpoints can be used to track people. It has been described as an ethical and privacy-sensitive way to monitor (mainly elderly and lonely) individuals.
accurate image
In layman’s terms, the Wi-Fi signal transmitted by the router can be used as a kind of sonar, and the AI driven program analyzes the difference in density between the outgoing and incoming signals and returns a wireframe image of the people in the room.
In some cases, the returned images were incomplete or showed people in weird, unnatural poses, suggesting that the method clearly needs improvement. But in many cases, the images created by AI are remarkably accurate. People’s positions in the room are accurate, their dimensions are accurate, their poses are accurate.
Besides the occasional bug in the rendering, another major challenge was being able to keep track of more people. So far, the router has been able to successfully track up to three people.
For the experiments, the researchers used a TP-Link Archer A7 AC1750 device, which costs just $32. Using a Wi-Fi router for this purpose is much cheaper than other tracking technologies like lidar or radar. In some cases, routers may even be a better solution than cameras because they can work even when someone is hiding behind objects like furniture.
It seems that the researchers will continue their work, trying to improve the solution with better public training data based on Wi-Fi perception.
pass: Tom’s Hardware (opens in a new tab)