A backstepping-like approach to adaptive control in discrete-time