I know what you mean, but I haven't figure out how to implement it yet. Maybe use video and pose recognition instead of physical movement.
