On the unknown input observer design : a decoupling class approach with application to sensor fault diagnosis

Souad Bezzaoucha (1) , Benoît Marx (1) , Didier Maquin (1) , José Ragot (1)

1 CRAN - Centre de Recherche en Automatique de Nancy

This paper addresses fault diagnosis for observerbased residual generators for linear discrete-time systems subject to unknown input. The proposed approach is a new method allowing to characterize a class of unknown inputs from which the estimation error is decoupled. This contribution is divided into two parts. The first one concerns the design of the UIO satisfying either an exact decoupling or an L2-attenuation of the unknown input to the state estimation error. The second part is dedicated to the implementation of a bank of such observers for sensor fault detection and isolation.

Domaines: Automatique; Automatique / Robotique

Souad Bezzaoucha, Benoît Marx, Didier Maquin, José Ragot. On the unknown input observer design : a decoupling class approach with application to sensor fault diagnosis. 1st International Conference on Automation and Mechatronics, CIAM'2011, Nov 2011, Oran, Algeria. pp.CDROM. ⟨hal-00630390⟩