ABSTRACT
This chapter demonstrates how the robustness of Model Predictive Control can be improved in the absence of constraints. It looks at the effects the T-filter has on sensitivity. The chapter illustrates the impact of the T-filter on sensitivity for some simple examples. It shows how the effect of the Youla parameter upon sensitivity can be made more transparent and hence d.o.f. can be made available for a systematic robust design. The use of a T-filter is the most popular approach to robust design in generalised predictive control. The Youla parameterisation has been used as a tool to improve the robustness of fixed linear control loops and is well known in the robust control literature. An effective and simple strategy would take the T-filter based controller as a start point and use the Youla parameterisation approach to further improve sensitivity if required.
