ABSTRACT
From the proportional, integral, and derivative definition, the tuning problem is a three-dimensional search problem. The overall objective is to achieve a balanced closed-loop, that is, a good balance between servo and regulatory performance, on the one hand, and between robustness and performance, on the other hand. This chapter is devoted to explaining how to evaluate robustness and performance in order to assist in the tuning of λ and γ. In particular, extreme frequency equivalent complementary sensitivities posses similar initial rise time and the same sensitivity to high-frequency noise and modelling errors. For plants with a more balanced lead/lag ratio, the servo/regulation trade-off quickly disappears, and both the servo and regulator designs exhibit large overshoots in the set-point response due to the fundamental constraint. The overall conclusion is that one should clearly go for regulatory control when dealing with unstable plants.
