High order B-splme networks and its applications to learning feedforward control