Markov-optimal sensing policy for user state estimation in mobile devices