With a WiFi signal and artificial intelligence it is possible to visualise the movement of people behind a wall. Researchers from the Massachusetts Institute of Technology (MIT) would like to use this technique for, among other applications, the monitoring of patients. The American MIT has been working on observing through walls for some time already. The research group, led by Dina Katabi, uses the (indoors nearly ubiquitous) WiFi radio signal for this. That part of this signal goes through walls is known by anyone who uses a WiFi signal at home that does not have a line-of-sight view of the router.

Minimal signal required

When a person is behind a wall, then part of this WiFi signal is again reflected back through the wall and can subsequently be detected by a receiver. This technique is imminently suitable for the detection of moving people, because these generate ever changing reflection patterns, which are easily distinguishable from stationary objects. For this detection to work only a very slight signal is sufficient – according to the MIT researchers a ten-thousand's of the signal strength that a mobile phone transmits.

Stylised figure

The group from Katabi has now added artificial intelligence to this motion detection, which can reconstruct the received radio signals into the movement of the person. For this, a large number of radio scans and pictures of the persons behind the wall where placed side-by-side, so that the program could learn that a certain radio signal corresponds to a particular position of the person. Radio signals received subsequently then show how a person's movement.

For this reconstruction of the moving person, Katabi uses a greatly simplified stick figure. This is because she wants to use the technique she developed for the monitoring of certain groups of patients or elderly in need of assistance, in situations where installing ordinary cameras would be inappropriate because of privacy concerns. Thanks to the stylised stick figures, patients only reveal their motion and not any detail. An advantage of the radio-graphical method is that there is no need for the patients to carry a sensor.

Separating people from each other

A further exceptional feature is that the software can distinguish between different people, because each person generates their own reflection profile. The software was able to distinguish 83 people from a group of 100, using only the radio signals and without photo or camera images. The software doesn't yet know who's who, but in any case is able to follow a certain individual within the group. The identification also goes wrong when two people come close together or are behind one another from the perspective of the WiFi signal.