A research team at Carnegie Mellon University found a way to use the WiFi signal from a router to detect human movement, and display it in three-dimensional images.
The team is working with the AI research team at Facebook which is working on DensePose, which is an algorithm that can map the human body from two-dimensional photos or videos.
This DensePose is combined with a neural network worked on by the team from Carnegie Mellon, which can map WiFi waves from routers that hit the human body, the data is then sent to DensePose to be imaged into an image.
The result is an image or video that is similar to the result of a monitoring camera such as CCTV, but without light limitations or being covered by other objects. For example, a sensor in one room can detect a human body even when it is dark or when the body is covered by another object. The way it works is similar to a radar, which sends a wave and then hits an object and the wave is recaptured.
With this way of working, of course the results will not be perfect because there are often parts of the human body that are not read by the sensor. But considering that this technology only uses WiFi signals, which are common everywhere, it's quite impressive, isn't it?
But there is one other problem, namely the matter of privacy. The research team also mentions this privacy issue many times in their report. According to them, the sensor used is sufficient to protect the user because it cannot detect the body in detail, which can be used to identify a person.
"In fact, most households already have WiFi in their homes, and this technology might be able to improve its function to monitor the condition of parents who are at home or simply identify suspicious activity that is happening at home," wrote the research team in a paper published at Cornell University. .