If you already feel like we're living in a surveillance state with all of the cameras on our phones and computers, along with all of the security cameras plastered all over the place, just wait until you hear this one. Researchers at the University of California, Santa Barbara have developed a way to sense a person through a wall using Wi-Fi signals. The technique can be used to identify individuals who have also appeared in footage captured by a camera.
The new technology, called XModal-ID, works essentially by using WiFi signals to measure a person as they pass by. Because these signals don't pass through people, they can be used to capture a model of an individual and, using an algorithm, can track how they move. The researchers believe that this information — in particular, a person's gait — can act as a sort of unique identifier. If the way that a person walks, as determined by the WiFi signals, matches the way that a person captured on camera moves, it is possible to conclude that they are the same person.
The development made by researchers at UC Santa Barbara builds upon existing research that uses WiFi to track or identify people. Researchers at the Massachusetts Institute of Technology's Computer Science and Artificial Intelligence Lab (CSAIL) have been working on ways to recognize human silhouettes through walls since 2013. Last year, the team at CSAIL developed a device that could use WiFi signals to track a person's movement. People have also used WiFi to accurately count the number of people in a room.
There are certainly applications to this technology. The team at CSAIL has used its research to develop a device called WiTrack, which can routinely and accurately detect if a person in a home falls down. Unlike installing cameras in a person's home, the technology that makes use of WiFi signals is less invasive. A person doesn't feel like they are being watched at all times and can maintain their privacy. It also doesn't require a device in every room, because WiFi signals travel through walls and still provide reliable information.
XModal-ID, the latest breakthrough from UC Santa Barbara, feels a little less altruistic. Researchers pose that the technology can be used to help identify a person suspected of committing a robbery. In such a scenario, the person is already caught on camera but can't clearly be identified. The WiFi technology could potentially be used to match their movements, using it as a sort of biometric determiner that could help to identify the alleged criminal.
There is science to back up the idea that a person's stride is unique to them and can be used to determine identity. Researchers at the Massachusetts Institute of Technology and Carnegie Mellon University have worked to develop technology that can measure a person's gait and identify them based on how they walk. The technique successfully matches a person with their gait between 90 and 95 percent of the time. Last year, China started to implement "gait recognition" technology to identify people on camera, in part because it is hard to fool and quite accurate.
The challenge, of course, is giving the technology information about a person's gait in order to match them. With XModal-ID, that data comes from a camera that has already captured a person in action. It's a novel solution to the problem, albeit an unsettling one. Most people who walk in front of a camera's lens aren't criminals, yet they are being captured all the same. The idea that their gait, the very way that they walk, could be tracked, analyzed and used to determine their identity — all without their knowledge or consent — is a terrifying proposition. The technology is impressive, and there are sure to be applications in which it could be useful. But the prospects of it being misused, or simply used to create another data point about your existence that could later identify you, make its usage feel a bit more dubious.
It's bad enough to have to worry about being tracked on camera at any given moment. This technology means anywhere there is WiFi, it's possible that your likeness and movements could be tracked.