A bounded error observer with adjustable rate for a class of bioreactor models

In this paper, we build a bounded error observer for a common class of partially known bioreactor models. A bounded error observer reconstructs the state variables with a "reasonable" error depending on the knowledge of the model and the rate of convergence can be adjusted in some way. It is "between" the high gain observer, which has an adjustable rate of convergence but requires a perfect knowledge of the model, and an asymptotic observer which is very robust towards uncertainty but with a fixed rate of convergence. We study two different cases corresponding to two different outputs. The robustness and efficiency are illustrated by simulation studies.