They've been experimenting with this since at least 1989.
https://muppet.fandom.com/wiki/Waldo_C._GraphicPeople have done simpler realtime rigs with off-the-shelf hardware and software in the past decade: see "The Dog Of Wisdom", https://www.youtube.com/watch?v=D-UmfqFjpl0 which was made with Blender and a Leap Motion: https://www.youtube.com/watch?v=0a_M9VsZ6Lk
And now we are completely drowning in VTubers, who use software like Live2d that analyzes a webcam image and uses it to control the motions of a pre-made 2D character. I've only ever seen it done to spice up the video of people streaming video games but I'm sure there's someone doing no-budget cartoons with it. There's also Adobe Character Animator, which has been used for various TV stuff like a live performance of Homer Simpson or a few low-budget shows.
And then there's VRChat; a few thousand dollars of head-mounted display/facial capture/body trackers and you can get realtime full-body tracking. There's probably someone fucking around with making movies this way too.
At this point I'm pretty sure that you could get most of the functionality of that hand-tooled puppetry gizmo by just taking a sock and gluing a couple of ping-pong balls onto it and tweaking some tracking software.