How's this for X-ray specs? Wi-Fi can read through walls... if the letters are solid objects

Researchers in California have found that Wi-Fi signals can be used to image objects on the far side of a wall, and claim to have demonstrated that such a system can even pick out complex shapes such as letters of the alphabet.

The researchers at the University of California, Santa Barbara used off-the-shelf Wi-Fi transmitters in their experiments, along with a receiver mounted on a moving platform to emulate a receiver grid, as detailed on the UC Santa Barbara site.

However, the key part of the setup is a proposed method, dubbed Wiffract, for interpreting the received signals from the transmitter and receiver apparatus to allow them to recreate an image of still objects on the other side of the wall.

Wiffract relies on a way of interpreting those signals in order to detect the edges of objects and their orientation, with this approach and the team's experimental results appearing in the proceedings of the 2023 IEEE National Conference on Radar (RadarConf) on June 21, 2023.

"Imaging still scenery with Wi-Fi is considerably challenging due to the lack of motion," said Professor Yasamin Mostofi. "We have then taken a completely different approach to tackle this challenging problem by focusing on tracing the edges of the objects instead."

Mostofi, working with two graduate students, developed their approach based on a bit of science known as Keller's Geometrical Theory of Diffraction (GTD), which concerns light rays, or in this case radio waves, reflected when they hit edges, corners or vertices of boundary surfaces.

According to the GTD, when a wave is incident on an edge point, a cone of outgoing rays emerges, referred to as a Keller cone. Wiffract uses a mathematical model developed by the research team to infer edge angles from the signal traces of the Keller cones picked up by the receiver grid.

Once an edge point is identified to a high level of confidence, Wiffract then propagates their inferred angles using Bayesian information propagation and enhances the resulting edge map using advances in the field of computer vision, according to the UC Santa Barbara report.

"Depending on the edge orientation, the cone leaves different footprints (i.e. conic sections) on a given receiver grid. We then develop a mathematical framework that uses these conic footprints as signatures to infer the orientation of the edges, thus creating an edge map of the scene," Mostofi explained.

• How to snoop on passwords with this one weird trick (involving public Wi-Fi signals)
• Whiffy malware stinks after tracking location via Wi-FI
• Intel seems to think Wi-Fi 7 is too cool for old-school Windows 10
• Ford SYNC 3 infotainment vulnerable to drive-by Wi-Fi hijacking

The results appear to demonstrate that Wiffract can indeed pick out the outline of complex shapes such as letters of the alphabet on the opposite side of a wall from reflected Wi-Fi signals.

The research team placed letters spelling out the word "BELIEVE" behind the wall in one experiment, although this actually involved one letter at a time being placed and imaged separately, not all the letters at the same time.

And before anyone gets alarmed, the letters involved were objects in the shape of letters, so while the report is provocatively titled "Wi-fi can read through walls," the UC Santa Barbara team is not claiming that Wi-Fi can be used to read material such as printed pages through solid walls.

It does, however, have a serious purpose. Identifying still objects is important for various applications such as smart homes, smart spaces, structural health monitoring, search and rescue, surveillance, and excavation.

But while sensing with wireless signals has shown promise for applications where there is motion, such as detecting activity, imaging details of still objects has remained a considerably challenging problem, according to the research team.

Mostofi has also published other papers on using Wi-Fi for sensing, such as one proposing a technique to identify an individual on the other side of a wall by analyzing video footage of the same person walking. ®