Design of a class of reduced order unknown inputs nonlinear observer for fault diagnosis

Damien Koenig (1) , Said Mammar (2)

1 LAG - Laboratoire d'automatique de Grenoble
2 LIVIC - Laboratoire sur les Interactions Véhicules-Infrastructure-Conducteurs

In this paper the problem of fault detection in non-linear systems is considered. Our goal is to extend the unknown inputs observer approach to nonlinear systems. The approach consists of two stages. At first stage, a change of coordinates is found that decoupages the nonlinear system into two subsystems: one which is independent of the unknown inputs and another with the states that can be expressed as linear combinations of the outputs and the states of the first subsystem. Secondly a nonlinear observer is designed for the former. The use of such observer in robust fault detection purposes is then discussed, and finally illustrated by a numerical example.

Domaines: Automatique

Damien Koenig, Said Mammar. Design of a class of reduced order unknown inputs nonlinear observer for fault diagnosis. American Control Conference, 2001, Arlington, United States. pp.CDROM. ⟨hal-00232860⟩