Adaptive iterative learning consensus control for second-order multi-agent systems with unknown control gains.