Your humble Wi-Fi router (opens in a new tab) The signal could be used to track its movements around a room, bat-style, a new report claims.
Researchers at Carnegie Mellon University recently published a report detailing an experiment using standard Wi-Fi routers to detect people’s locations, as well as their poses, in a room.
The experiment, while not without flaws, was an overall success, demonstrating that the endpoints could be used to track people. It is described as an ethical and privacy sensitive way of monitoring people (mostly elderly and lonely).
In simple terms, the Wi-Fi signal transmitted by routers can be used as a kind of sonar, where an AI-powered program analyzes the difference in density between outgoing and incoming signals, and returns with wireframe images of people in The bedroom.
In some cases, the images appeared sketchy or showed people in strange, unnatural poses, showing that the method obviously still needs work. But in many cases, the images created by the AI were quite accurate. The positions of people within a room were accurate, their dimensions were accurate, their poses were accurate.
Apart from the occasional rendering error, another major challenge is being able to track a larger number of people. So far, the routers can successfully track up to three people.
For the experiment, the researchers used TP-Link Archer A7 AC1750 devices, which cost a measly $32. Compared to other tracking technologies, such as lidar or radar, using Wi-Fi routers for this purpose is immensely cheaper. In some cases, routers might even be a better solution compared to cameras, since they work even if people are hidden behind objects like furniture.
It looks like the researchers will continue their work, trying to improve the solution through better public training data for Wi-Fi-based perception.
Via: Tom’s Hardware (opens in a new tab)