Mostofi and her researchers developed a system that allows two drones working together to generate detailed images of objects through walls, using only Wi-Fi RSSI measurements. In a proof-of-concept demonstration, one drone flew outside an enclosed four-sided brick structure with an unknown interior. While flying, it transmitted a Wi-Fi signal, which was then measured by the other drone.
While this is done, some smart algorithms then model the 3D unknown area of interest using some highfalutin technology such as a Markov Random Field to measure its spatial dependencies. Although the image generated isn’t 100 percent perfect, it’s impressively accurate in broadly stating what is behind a particular wall. Making it even more impressive is the fact that the prediction can be achieved with a very small number of Wi-Fi measurements, and uses only off-the-shelf technology such as commercially-available drones, Wi-Fi transceivers, and Tango tablets.
“There are several potential applications, such as emergency response, archeological discovery, and structural health monitoring, that can benefit from obtaining a high-resolution 3D image of an unknown area through walls with everyday Wi-Fi signals,” Mostofi told Digital Trends. “For instance, consider structural health monitoring for bridges. It would be very helpful if unmanned aerial vehicles can fly around a bridge, especially in the hard to access areas, and image the details inside to assess the health of the bridge.”
From here, Mostofi said that the team is working to improve the work and find new possibilities and potential implementations.
“The general area of sensing and learning about our environment with everyday radio frequency signals, such as Wi-Fi, is one of the major research thrusts of my lab,” she said. “From our past experience, there are good commercialization opportunities in this general area.”