Gain-scheduling through continuation of observer-based realizations -: Applications to H∞ and μ controllers

The dynamic behavior of gain-scheduling controllers is highly depending on the state-space representations adopted for the family of linear controllers designed on a set of operating conditions. In this paper, a technique for determining a set of consistent and physically motivated linear state-space transformations to be applied to the original set of linear controllers is proposed. After transformation, these controllers exhibit an observer-based structure and are therefore easily interpolated and implemented. This method is applicable to discrete- or continuous-time and full- or augmented-order compensators, particularly including H-infinity and mu controllers, which do not generally enjoy ease of implementation.