The first problem you identified is a real one.
The second one I anticipate might be resolved by the difference between a controlled voluntary movement and a sudden stop caused by hitting the floor. I surmise it is physically impossible for the human body in motion to effect immediate stoppage (physical dynamics), whereas hitting the floor (an immovable object) would cause sudden stoppage.
But your suggestion on heart rate will be investigated as the first possibility, since it could cover the eventuality of the "not-falling-out-of-the-chair" scenario; it's much more elegant and probably requires less power to provide a functional monitor.