Iterative trajectory tracking for flexible arms with approximate models

Learning control is applied to end-point trajectory tracking of a flexible robot link, performing repetitive tasks. The model of an experimental prototype is given and the main results from the frequency domain analysis of iterative learning controllers are recalled. Several possibilities of application of this methodology to flexible structures are presented and extensively tested by simulation. They differ for the level of detail of the link model which ranges from the exact model to a very rough approximation, which does not retain its nonminimum phase properties.<
>