Researchers train AI to detect concealed handguns using Wi-Fi signals
Researchers have developed WiFireNet, an AI system that detects concealed handguns from Wi-Fi channel-state information, achieving 72.5% average precision in controlled indoor tests without cameras or metal detectors.
Swedish researchers at the FOI have created an AI system called WiFireNet capable of detecting concealed handguns through Wi-Fi signals. The system, outlined in a paper posted to arXiv on October 13, examines how Wi-Fi signals change when obstructed by objects. A Wi-Fi transmitter, receiver, and AI comprise the complete system. When Wi-Fi signals hit an object, they bend and reflect differently, depending on the object’s shape, angle, and material.
This alteration, known as channel state information, can be analyzed to recognize the object with high precision. Previous studies have demonstrated that Wi-Fi routers can identify individuals with 99.5% accuracy. To construct WiFireNet, researchers first gathered a database, the WiFire Dataset, which merged video footage of people moving with Wi-Fi signals reflecting off various objects, encompassing everyday items like phones and handguns.
The researchers then trained a deep learning algorithm using this data, teaching it to correlate the visual of a concealed handgun with its unique Wi-Fi signature. The system was refined by training on recordings of a plastic handgun, a metal water bottle, an actual metal handgun, and a phone. Upon completion, WiFireNet could pinpoint concealed handguns solely based on Wi-Fi CSI, achieving an average precision of 72.5% in controlled trials involving up to four individuals moving around in a small indoor area.
However, precision decreased to 53.8% when more exact localization was needed, such as determining the exact spot on the carrier where the weapon was hidden. The neural network also faced challenges in unfamiliar indoor settings. Despite these limitations, the system remains experimental, with plans for future enhancements. The prototype utilized an older 1999 wireless standard to demonstrate feasibility.
Future improvements will involve using modern Wi-Fi 7 and Wi-Fi 8 networks with higher-resolution CSI for enhanced detection. The researchers have chosen to make the project open source, releasing the code, models, and dataset to inspire further research in the field. Wi-Fi sensing is being seen as a promising surveillance technology that doesn’t require elaborate equipment or complex setups.
However, skeptics warn that the technology’s widespread use and the lack of regulatory guidelines may lead to significant privacy issues.
Written by urgent.news from Tom's Hardware's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
