We overview recent progress in the field of robust adaptive control with special emphasis on methodologies that use multiple-model architectures. We argue that the selection of the number of models, estimators and compensators in such architectures must be based on a precise definition of the robust performance requirements. We illustrate some of the concepts and outstanding issues by presenting a new methodology that blends robust non-adaptive mixed m-synthesis designs and stochastic hypothesis-testing concepts leading to the so-called robust multiple model adaptive control (RMMAC) architecture. A numerical example is used to illustrate the RMMAC design methodology, as well as its strengths and potential shortcomings. The later motivated us to develop a variant architecture, denoted as RMMAC/XI, that can be effectively used in highly uncertain exogenous plant disturbance environments.

We evaluate the performance of the robust multiple model adaptive control (RMMAC) methodology by considering a mass-spring-dashpot (MSD) system subject to high-frequency disturbances that strongly excite all its lightly damped oscillatory modes. The results demonstrate the superior performance of the RMMAC and its variant RMMAC/XI architecture for a much more difficult adaptive control problem than that designed and analysed in Reference (Int. J. Adaptive Control Signal Processing, in press).

We discuss recent progress in the field of robust adaptive control with special emphasis on methodologies that use multiple-model architectures. We emphasize that the selection of the number of models, estimators and compensators in such architectures must be based on precise definition of the robust performance requirements. We illustrate some of the concepts and outstanding issues for a new methodology that blends robust nonadaptive mixed µ-synthesis and stochastic hypothesis-testing concepts leading to the so-called RMMAC architecture. A numerical example is used to illustrate the RMMAC design methodology.