A high/low gain bundle of observers: application to the input estimation of a bioreactor model

This paper proposes a new interval observer design based on a high/low gain concept, developed for the estimation of the input of a bioreactor. In a first step, a high gain bounded error observer is proposed, which is able to estimate the unknown input of the considered system. This observer is useful when an accurate model and noise–free measurements are available. We show that the error of the high gain observer can be dynamically bounded, and then we generate guaranteed bounds on this error through an interval observer. The estimation scheme is then extended to an uncertain framework. Taking advantage of the nature of the interval estimates, we run in parallel various observers using high or low gains values and then we take the best estimates. The method is applied to the input estimation of a simple bioreactor model.