MIT has developed a new artificial intelligence system that can detect human's pose and movement behind wall using radio signals. This can have positive impact in healthcare to track elder patients but at the same time it can have many privacy related issues.
Researchers at Computer science and Artificial intelligence laboratory at MIT has been working since long on radio signals and wi-fi to recognize human form and now they developed artificial intelligence system named 'RF-Pose' that uses neural network and wireless radio signals to see what people are doing behind the obstacles. The human pose and movements can be seen in the form of 2D stick figure. The new system harnesses the fact that human body reflects radio frequency signals.
MIT researchers collected data examples of people walking with both wireless signal pings and camera and trained the AI algorithm and neural network to generate stick figures for human pose and movements behind wall.
Let's see how this AI system will work:
The new system can recognise movement in poor lighting and detect multiple human as well. The system is still in its development phase but researchers believe that it could offer exponential benefits in military in surveillance and in search and rescue operation as well.
Also, it will have wide application in medical or healthcare sector. Experts says that this system can track and monitor patient's movement who are suffering from multiple sclerosis, Alzheimer and Parkinson's disease. The RF-Pose system is also helpful for elder people to send alerts if they fall or have in some trouble. Researchers believe that it can offer better care for elder people in living facilities.
This system has significant privacy concern where people may misuse this system to follow people to their destination who may be hiding themselves behind walls. To overcome this, MIT researcher is developing a solution that is 'consent mechanism' that would require performance specific movements before tracking.
This technology is the latest one in the project of using radio signals to mimic X-ray vision which is going on since long. MIT researchers says that the development in system may made possible for AI to generate 3D image of people rather 2D stick figures.