Comment by pavelstoev

In a dynamically changing environment (from the perspective of Wi-Fi signal), this will be difficult but not impossible with modern applications of ML algorithms. We worked on this technique back in 2016-18 at the University of Toronto WIRLab; take a look at the results video from back then. I think the person is somewhat identifiable. https://www.youtube.com/watch?v=lTOUBUhC0Cg