Jakarta, INTI – In the era of rapid digital transformation, Wi-Fi is no longer just a tool for connecting devices to the internet. A new breakthrough has emerged: Wi-Fi signals can now detect, recognize, and identify humans without the need for cameras. This innovative capability is made possible through a technology called WhoFi, developed by researchers from Sapienza University of Rome.
From Internet Connector to Human Sensor
Traditionally, Wi-Fi (Wireless Fidelity) has served as a bridge between digital devices and the internet. However, scientists have proven that Wi-Fi signals are capable of more than just data transmission. They can penetrate walls, detect human presence, and even identify individuals through subtle interactions between the human body and wireless signals.
This advancement is driven by the development of Wi-Fi Sensing, a feature introduced in the IEEE 802.11bf Wi-Fi specification promoted by the Wi-Fi Alliance. With this, Wi-Fi becomes more than just a transmitter; it functions as a sensor system, identifying movements and even human identity based solely on signal disturbances in a given area.
WhoFi: Human Identification Without Cameras
The WhoFi system, developed by researchers Danilo Avola, Daniele Pannone, Dario Montagnini, and Emad Emam, works by interpreting how a person’s body interferes with Wi-Fi signals in the surrounding environment. According to their paper, “WhoFi: Deep Person Re-Identification via Wi-Fi Channel Signal Encoding,” Wi-Fi has distinct advantages over traditional cameras because it can interact with the internal structure of the body, not just the outer surface.
Despite the name being previously used by a community internet service in Oklahoma, the academic version of WhoFi has a completely different mission: to identify people based on how their unique body movements and physical presence alter Wi-Fi signal patterns essentially acting as an invisible biometric signature.
The Science Behind WhoFi
Every human body is unique in terms of size, shape, and movement. These differences produce specific interference patterns in Wi-Fi signals, which can be captured and analyzed using Channel State Information (CSI) technical data that shows how Wi-Fi signals behave when passing through an environment.
Even when someone is standing still, their body still affects the surrounding Wi-Fi signal. By processing this CSI data, researchers can create unique "Wi-Fi fingerprints" that represent each individual.
This technology eliminates the need for visual surveillance, offering a more privacy-conscious alternative for identity detection, which is useful in homes, offices, and public spaces especially where cameras might be invasive or restricted.
Real-World Application and Accuracy
Prior to WhoFi, the research team developed a system called EyeFi, which achieved an accuracy rate of about 75%. With the new WhoFi approach, accuracy rose significantly.
When tested using a public dataset called NTU-Fi, WhoFi successfully recognized individuals with an impressive accuracy rate of up to 95.5%, as long as the Wi-Fi signal could be received in the tested locations. This high accuracy is made possible by transformer encoder technology, a deep neural network architecture commonly used in modern artificial intelligence applications.
Conclusion
Wi-Fi sensing and WhoFi mark a new chapter in how wireless technology interacts with human identity. Without any cameras or invasive tools, our mere presence how we move and exist within a space can be uniquely identified by the invisible signals around us. This not only opens new doors in biometric security but also raises important questions about privacy and the future role of smart technology in daily life.
Read More:APJII Recommends Three Key Measures to Support ISPs in Expanding Internet Access to 3T Areas