Some new directions for adaptive control theory in robotics