Portuguese 
Spanish 
3 min of reading 
0 comments 
Italian Researchers Developed a System That Identifies People Using Only Wi-Fi Signals in an Environment. Without Needing Cameras or Cellphones, the Technology Reaches Over 95% Accuracy and Can Transform Any Router Into an Invisible Biometric Tracking Tool.
Digital surveillance has taken another bold step. Now, neither cameras, cell phones, nor wearable devices are needed to identify someone.
Researchers at La Sapienza University in Rome have developed a new technique that allows tracking people using only Wi-Fi signals. The system is named WhoFi.
This innovation turns Wi-Fi signals into a tool capable of recognizing individuals based on how their bodies affect the waves as they move through connected environments.
-
After building a 5-acre lake on his farm, the owner reveals plans for a new artificial lake with a flowing river, waterfalls, islands, and even an indoor version with a retractable roof to control light, temperature, and transform the rural landscape.
-
A giant is disappearing from the map: the Dead Sea, between Israel, Jordan, and the West Bank, loses 1.2 meters per year, has already shrunk by one-third in 50 years, and is 427 meters below sea level.
-
Ghost town of 4,000 residents, abandoned after the collapse of an iron mine, now becomes a drone training field for police, rescue, and emergency teams.
-
Colorectal cancer advances worldwide, but most cases can be prevented.
The accuracy of the method is surprising: up to 95.5%, according to the tests conducted.
Invisible Biometrics
WhoFi uses an approach known as Channel State Information (CSI).
This technique analyzes how Wi-Fi signals behave when crossing environments and encountering obstacles such as walls, furniture — and human bodies.
According to researchers Danilo Avola, Daniele Pannone, Dario Montagnini, and Emad Emam, the distortions caused by the presence of a person are unique enough to generate a kind of fingerprint.
These data, when processed by deep neural networks, become biometric identifiers.
The architecture used in the study was based on transformers, artificial intelligence models known for their performance in complex tasks.
The team applied the technique to the public dataset NTU-Fi, achieving results superior to previous methods, such as the EyeFi system from 2020, which only reached 75% accuracy.
New Form of Reidentification
Reidentification is not a new concept. It is already common in security camera systems, which analyze characteristics such as clothing or movements to track a person from one place to another. The difference lies in the medium used.
With Wi-Fi signals, recognition becomes invisible. There is no need for lighting or direct line of sight. Furthermore, these signals pass through walls, which expands the coverage area of surveillance. This represents an important change in how biometric data can be collected.
Another relevant point is that, at first glance, this method may seem less invasive since it does not involve images or explicit recordings.
However, the fact that it works without the person carrying a cell phone or being in front of a camera raises new debates about privacy.
Impact of Wi-Fi Sensing Technology
The advancement of WhoFi occurs in a scenario of transformation of Wi-Fi itself. Since 2020, with the approval of the IEEE 802.11bf specification, the technology has gained a new function: environmental sensing.
The Wi-Fi Alliance, responsible for promoting wireless network standards, has been encouraging the use of what is called Wi-Fi Sensing.
The idea is that routers and access points can also function as sensors, detecting movements, presence, and now even biometric profiles.
This expands the possibilities of application in areas such as security, home automation, and environmental monitoring. However, it also opens up space for passive tracking without explicit consent.
Surprising Accuracy
WhoFi stands out for its accuracy. With up to 95.5% success in reidentifying individuals, the technology shows a leap over previous systems.
This rate was achieved even with variations in the environment and without the need for any device with the monitored person.
The research was published as a preprint in arXiv, a repository of studies that have not yet been peer-reviewed. Nevertheless, the work is already garnering attention in the scientific community and in sectors related to digital security.
The possibility of identifying someone solely by the effects of their body on Wi-Fi signals opens up new discussions about the boundaries between technological innovation and personal privacy.
-
1 person reacted to this.