Self-tuning control of systems with infrequent and delayed output sampling

In the paper a stochastic self-tuning control algorithm is presented for MIMO systems. It is assumed that the input signals are allowed to change more frequently, then the output measurements are available. The presence of a time delay in the transducer performances is also considered. The algorithm is based on the generalised minimum-variance approach. Robust zero error regulation is ensured for constant set points and loaddisturbances. For known and open-loop stable plants, closed-loop stability is guaranteed by a suitable choice of the design parameters. A global convergence result is given for the adaptive version of the method. The algorithm is finally applied for control of a binary distillation column.