EASY TUNING OF SMITH PREDICTOR IN PRESENCE OF DELAY UNCERTAINTY

This note presents a simple criterion for the tuning of a Smith predictor when the plant time delay is not precisely known. It is shown that a good choice of the dosed-loop bandwidth can be related to the amount of delay uncertainty in a straightforward way. The tuning guidelines here proposed can also be extended to the more general case of model mismatch in the rational part of the plant transfer function.