On Dα-type iterative learning control