In this paper, a piecewise observer strategy is proposed for a non linear process with unknown inputs operating robustly over a wide range of varying conditions.  The proposed methodology is based on two stages: First, a new algorithm is developed to design multiple linear models able to describe the behaviour of a disturbed non linear process with unknown inputs. Second, a piecewise reduced order observer is scheduled on local models to estimate the unmeasured states of the non linear systems. Existence conditions of the estimation procedure are provided. The conceived observer ensures local stability and minimizes the effect of disturbances. Design and performances of the proposed approach are finally illustrated by means of simulation results on a biological process of water treatment.

Keywords: Robust observers, unknown inputs, piecewise observer, multiple linear models, nonlinear system, biological process.

BibTex:

@ARTICLE{P1120609001,

AUTHOR = {B. Sfaïhi and O. Boubaker},

TITLE = {Piecewise Robust Observer for a Nonlinear Biological process with Unknown Inputs},

}

(Full Paper, 816KB)