Researchers from UC Santa Barbara have developed a method of sensing the outlines of still objects that are out of sight. This new Wi-Fi-based technique, dubbed Wiffract, can even sketch out objects in an adjacent room. No AI is used here, just pure physics and math, but perhaps the tech could be enhanced in the future by leveraging Deep Learning.
The research team noted that sensing objects using Wi-Fi has gained traction in recent years, but most of the rival methods are based upon the analysis of movement, or objects in motion, with stationary objects beyond their scope. Behind the new Wiffract technique is the Geometrical Theory of Diffraction and corresponding Keller Cones created by Wi-Fi's interaction with edges, explains the UC Santa Barbara blog.
To test Wiffract RF imaging, the researchers decided to see how accurately they could trace letters of the alphabet that were out of view. With only Wi-Fi and complex math that analyzes the incidental RF waves reflected by object edges, motionless objects that are out of view can be sketched out.
You can see the outlines generated by Wiffract and the corresponding concealed letters in the image above. You can watch the embedded video for more background on the scientific method behind this technique. However, we can explain that the aforementioned Keller Cones are central to the new imaging technique. These cones occur when RF waves, like Wi-Fi signals, hit object edges.
"Depending on the edge orientation, the cone then leaves different footprints (i.e., conic sections) on a given receiver grid," says professor of electrical and computer engineering at UC Santa Barbara, Yasamin Mostofi. "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."
In developing Wiffract, the researchers used a single laptop as the Wi-Fi transmitter, but six antennas were attached to a small moving vehicle to read the reflected RF waves and plot their results on a 2D grid. In effect, the movement of the vehicle and its antenna tower 'scanned' the letters behind the wall.
Finally, it is interesting that a pivotal decision in achieving impressively accurate outlines of still objects was the construction of a moving sensor tower.