Wi-Fi signals can be used to unobtrusively identify different people at a location, such as home. It promises to replace other forms of identification in those domestic environments, Chinese scientists say.
The system works by identifying body shapes along with the unique way that individuals move in a room. Those characteristics influence Wi-Fi propagation, researchers from Northwestern Polytechnical University in Xi’an claim. The Wi-Fi is affected by the people in the room, and that impact on the wireless access point can be detected and interpreted, they say.
“Each person has specific influence patterns to the surrounding Wi-Fi signal while moving indoors, regarding their body shape characteristics and motion patterns,” the team writes in an abstract to their paper, published in August.
FreeSense, as the waveform analysis system is called, could be further developed to ultimately replace forms of biometric authentication and even passwords. Face, gait and fingerprint recognition—methods that will become increasingly common in homes as the Internet of Things (IoT) takes off—could be made redundant with this system.
Indeed, personalization of IoT in the home is becoming ever more important, but having to laboriously enter user IDs and so on could well turn out to be a hindrance to adoption.
Despite technical hiccups overall, that identification is worth pursuing: A smart television, for example, becomes more useful the more it recognizes family members. The screen can produce listings customized for the individual, for example. It just needs to know who the individual is—the hard part.
Another use case example for FreeSense could be seamless, no-touch thermostat adjustment:
The temperature is set at one compromised level when a couple is at home and at another, preferred temperature, when an individual is there alone, for example. That kind of thing could sell residential IoT big time. Streamlining user authentication and identification, a hassle ordinarily, could tip the balance in favor of mass IoT adoption.
Security is another benefit, the scientists say. User privacy could conceivably be maintained at the router level, and passwords, including loss-of, and poor password habits, such as infrequent changing of them, become redundant.
The scientists say that sensing coverage range is increased over other human identification methods. With a fingerprint reader, one has to approach the device, for example.
“Fine-grained information regarding Wi-Fi communication” is what’s behind the system. The main method used, Channel State Information (CSI), “describes how the signal propagates from the transmitter to the receiver and reflects the combined effects of the surrounding objects,” the scientists describe in their paper (PDF).
“Scattering, fading and power decay with distance” are among the elements collected by their “CSI-enriched devices.”
Notably, the Chinese Wi-Fi system isn’t using received signal strength indication (RSSI). RSSI is another way indoor localization can be performed. It’s coarser, though, and although can be used for localization and gesture recognition, it isn’t as rich in its sensitivity to moving objects, the scientists say.
CSI isn’t without its challenges, too, though: Arbitrary walking about the house is difficult to track, for example. CSI needs a kind of collection zone.
The team members explain that their Wi-Fi indoor human identification is 88.9 percent to 94.5 percent accurate. They say they have tested it with nine users in a six meter by five meter room—those successful sample numbers, they say, will make it ideal for families.
This article is published as part of the IDG Contributor Network. Want to Join?