Optimal iterative learning control design for continuous-time systems with nonidentical trial lengths using alternating projections between multiple sets.