Surveillance in the digital age is no longer limited to cameras and smartphones. From facial recognition to GPS logs, the tools used to monitor people have grown increasingly sophisticated.
Now, researchers in Italy have shown that even ordinary Wi-Fi signals can be used to track people, without needing them to carry any device at all.
A team from La Sapienza University of Rome has developed a system called ‘WhoFi,’ which can generate a unique biometric identifier based on how a person’s body interacts with surrounding Wi-Fi signals.
The approach, described in a preprint paper, uses signal distortions caused by the human body to re-identify individuals as they move across spaces covered by different Wi-Fi networks.
Biometrics through the air
The researchers behind WhoFi, Danilo Avola, Daniele Pannone, Dario Montagnini, and Emad Emam, claim their system can match people with up to 95.5 percent accuracy using the public NTU-Fi dataset.
Their method builds on a technique known as Channel State Information (CSI), which captures how Wi-Fi signals change when they pass through physical environments.
“The core insight is that as a Wi-Fi signal propagates through an environment, its waveform is altered by the presence and physical characteristics of objects and people along its path,” the authors state in the paper. “These alterations, captured in the form of Channel State Information (CSI), contain rich biometric information.”
CSI includes data on both the amplitude and phase of electromagnetic transmissions.
According to the team, these signal changes are specific enough to serve as a kind of digital fingerprint, especially when analyzed by a deep neural network.
In this case, the researchers used a transformer-based architecture, a type of model popular in advanced AI applications.
New angle on old problem
The concept of re-identification, linking the same person across multiple points of observation, isn’t new.
It’s widely used in video surveillance, often by tracking clothing or physical traits. But Wi-Fi presents new advantages.
Wi-Fi signals present a powerful alternative to traditional surveillance tools like cameras.
Unlike visual systems, they can operate regardless of lighting conditions, pass through walls, and avoid capturing identifiable images, making them appear more privacy-conscious on the surface.
The WhoFi technique doesn’t rely on phones or wearable devices. A person’s body alone can create a distinct enough pattern in Wi-Fi signals to enable re-identification.
This raises new concerns about passive tracking, especially as Wi-Fi sensing becomes more widely adopted.
The groundwork for such applications was laid in 2020 with the approval of the IEEE 802.11bf specification.
Since then, the Wi-Fi Alliance has been actively promoting Wi-Fi Sensing, reframing routers and access points as environmental sensors.
A comparable system called ‘EyeFi’ was introduced in 2020, achieving 75 percent accuracy.
WhoFi significantly improves on this with up to 95.5 percent accuracy, highlighting the growing effectiveness of signal-based re-identification tools.
The study has been published as a preprint on arXiv.
Source link
