Also, why the hell are they feeding the data back into Unreal Engine? Just so they can have the projection on the animatronic update without actually having to redo the animation? So instead of just having a video play on a projector on a pre-canned animatronic animation, you're having an Unreal engine scene render this animation in real time?
Again, tail wagging the dog here on the complexity of a technological solution that wasn't needed, doesn't improve the scene, and in many cases actively harms it.