2 min read
Computer scientists at Artificial Laboratory at MIT have successfully devised a mechanism that can see people through walls.
The system recreates your poses when you move by using wireless radio signals to sense where you are and then recreates your body like a plain stick figure called RF-Pose.
This technology is seen as a less intrusive way to monitor people for health and safety purposes for example, if an elderly person or a sick person has a risk of falling , you may want to make sure they don’t by using this technology for their surveillance.
The team believes that this system can be used to monitor people suffering from Parkinson’s disease or multiple sclerosis (MS). This will help doctors in better understanding of the progression of disease so, the doctors can prescribe medications.
This will further help elderly people live without stress as they have the knowledge of being monitored.
The main focus of the team lies in using this technology for healthcare, enabling passive monitoring of an individual. The researchers wrote that the data of the subjects has been collected after consent of the subject and has also been encrypted to ensure user privacy.
For applications in the real world, the team has a plan to integrate a consent mechanism, where a person who has installed the device will activate it for monitoring by doing a specific set of movements.
In order to train the neural network, the researchers showed machine, video of a person who is walking next to the RF interface that they made as they moved. By overlaying stick figures on the movement and training the network to do the same automatically, the researchers finalized this system. The researchers did not train the system to see through the walls. But the system generalized this knowledge so, it can handle through wall movement.
