A team of scientists from the Massachusetts Institute of Technology (MIT) invented a wireless device that tracks the movement of people through walls. It was called RF-Pose.
First, engineers collected data from two sensors: a wireless radio and a camera-both tracked people's actions, such as walking, talking, waiting for the elevator and opening the doors. Then, using camera images, they built a thin skeleton model for each person. This model was demonstrated by AI together with the radio signal. So the algorithm learned to recognize the connection between the signals of the radio transmitter and the constructed skeletons.
Specialists reported that their technology identifies an individual among a hundred people in 83 percent of cases. This feature is useful in search and rescue operations.
The modern version is two-dimensional, but now engineers are working on a three-dimensional one, which could display smaller movements, for example, hand tremors.
In May, researchers from the Georgia Institute of Technology created a robot that learned how to dress people.