Discrete predictive optimal ILC implemented on a non-minimum phase experimental test-bed

Predictive optimal ILC has been implemented on a nonminimum phase experimental test facility. Predictive optimal ILC has first been derived for discrete-time systems, having previously only been formulated for the continuous case. Practical implementation issues have been considered and methods have been proposed to limit the necessary memory and calculation time required. The plant has been described and experimental results then presented for two cases of predictive horizon. The effect of variation of parameters used in the cost function has been described. The convergence rate observed in practice has been compared against a theoretical bound.