Adaptive Robust Control: A New Approach

In this paper we present a new approach to adaptive robust control. The essential idea involved is to fragment modeling uncertainty in a physical plant into "small" families of plant models for which robust LTI controllers can be readily designed. We then implement a convex combination of these controllers, the parameters of which are tuned based on error measurements. Using this approach we are able to provide absolute bounds on the ability to adaptively stabilize against unmodeled dynamics.