With the proliferation of mobile devices around the world, the demand for connectivity is constantly increasing. As a result, we’re bombarded by Wi-Fi, essentially low-frequency radio waves, nearly everywhere we go.
“Can these signals be used not just for communication but for sensing the environment? We started looking at this problem about nine years ago,” said Yasamin Mostofi, associate professor in the department of electrical & computer engineering at UC Santa Barbara.
Seeing through walls
If you’ve ever tried to connect to the internet when there’s distance and walls between you and your router, you know that Wi-Fi signals are affected by the environment. They lose energy as they pass through objects, like light through a lampshade, and they can bounce off of objects, like light off a mirror.
“Our first interest was X-ray vision with Wi-Fi,” said Mostofi. “We basically wanted to see if we can see behind walls with Wi-Fi signals.”
And if that doesn’t sound fantastic enough, Mostofi decided to throw some robots into the mix.
To test the system, Mostofi and her lab constructed a tall, square enclosure out of brick on the UC Santa Barbara campus. In the middle, they placed objects: a barrel, two barrels, a graduate student.
To peek through the wall, two remote vehicles passed around the enclosure.
“One robot transmits, and that transmission goes through the object,” explained Mostofi. “The objects interact with it, depending on their material property and their location.”
The objects leave a signature on the Wi-Fi signal. Once the other robot receives the signal, the trick is to then reconstruct what the signal passed through.
“Can they image that person? Can it image the object that is there? Maybe there are multiple objects – can they image them correctly?” said Mostofi. “We’ve shown that you can actually do this.”
Initially, the goal was to devise a system that could tell you the location of the objects, their geometry, and their shape – and early tests have confirmed that this is possible, albeit with relatively low resolution so far. But Mostofi’s team has found that there’s even more information in the Wi-Fi signature.
“What we extract also tells us something about the material property,” said Mostofi. “In the future, we should be able to say something more about what kind of object is there. Is this wood? Is this human?”
Robots to the rescue
Applications of this technology for search and rescue are immediately apparent.
Mostofi thinks back to the Tōhoku Earthquake in 2011 and the tsumani that caused widescale destruction and triggered the Fukushima nuclear disaster.
“There was a nuclear spill and it was too dangerous for humans to go in,” said Mostofi. “It would be great if we can have unmanned vehicles that from outside can assess the situation inside and figure out what’s going on behind walls.”
Being able to spot people trapped in collapsed buildings from the outside could quickly direct search and rescue resources to areas of need and avoid wasting precious time.
There are other less catastrophic reasons to see through walls. Mostofi envisions Wi-Fi as a tool for archaeology, providing non-destructive ways to peer behind walls. Archaeologists recently located a possible hidden chamber in the tomb of King Tutankhamun, but the Egyptian government is leery that further exploration could damage the tomb. Wi-Fi sensing could prove useful in cases like this.
When what’s behind a wall isn’t a mystery, Wi-Fi sensing could have everyday uses.
“One possible application is optimization of heating, and cooling in a building,” said Mostofi. “Wouldn’t it be great if the existing Wi-Fi network can tell something about the density of the people, which areas are more crowded? Then the heating and cooling can be optimized, accordingly.”
To make this possible, Mostofi’s lab has been hard at work using Wi-Fi to estimate occupancy of spaces.
In an office, laptops and other devices are connected to routers by Wi-Fi. If a person crosses through a space and blocks the direct link from a device to a router, it results in a huge drop in signal strength. If a person is not directly blocking the connection, the signal will still bounce off of them. Both of these effects help Mostofi estimate occupancy in a room.
“One can envision that in a future smart home we would use the existing radio waves, to tell more things about us,” said Mostofi.
Other smart home applications could include security systems (e.g., intruder alert), or health applications like baby monitoring.
The research highlighted in this article has been supported in part by a National Science Foundation CAREER award.