Researchers at the Karlsruhe Institute of Technology (KIT) in Germany have published a paper describing a novel — and disturbingly invasive — passive surveillance technique. The system, called BFId, reportedly identifies people moving through a room or past a building with 99.5% accuracy.
What's most unsettling is that you don't need to break into the router, obtain the Wi‑Fi password, or install special radar gear. The person being tracked doesn't even have to carry a phone or any wireless gadget.
- Connected devices (smartphones, laptops, TVs) routinely send back technical reports so routers can steer their signals—these are called Beamforming Feedback Information (BFI).
- Those BFI frames travel unencrypted at the physical (MAC) layer; any ordinary Wi‑Fi adaptor, e.g., a laptop or Raspberry Pi put into monitor mode, can capture them passively.
- A human body walking through a space perturbs and scatters radio waves; those changes are mirrored almost immediately in the BFI records.
- Collecting these traces from several vantage points produces a unique 'radio fingerprint' tied to a person's walking style and body form.
KIT tested BFId on data from 197 participants — reportedly the largest dataset of its kind. The neural network learned to pick out identities within seconds. Even when the researchers intentionally reduced the data to mimic noisy, real-world conditions, BFId still reached about 99.5% accuracy for gait-based recognition; by comparison, the older CSI method managed roughly 82.4%. Frankly, that gap is hard to ignore.