An integrated nonlinear observer with sliding mode estimation for a class of nonlinear uncertain systems

Nonlinear observer is an effective approach for the state estimation of nonlinear systems. However, when modeling uncertainties (e.g. uncertain parameters) are not sufficiently random, the nonlinear observer design based on the nominal nonlinear dynamics may give certain estimation errors. To improve the estimation accuracy, in this paper we incorporate a sliding mode observer (SMO) into the nonlinear observer for a class of uncertain nonlinear systems . The measurable output estimation error is the sliding surface and we have proven the estimation convergence in the sliding mode. The proposed robust nonlinear estimator is applied to the on-line monitoring of a bioreactor. The simulation results demonstrate the effectiveness of the proposed method.