Robust Residuals Generation for Faults Detection in Electric Powered Wheelchair

Abstract

A diagnostic approach for generation robust residuals to the dynamic model of an electric wheel chair is proposed in this paper. The method formulated here uses the observer presented in this work in order to design a residual generator that allows the detection of faults. The unknown input observer with perfect decoupling allows assessing these faults, which leads to their localization and detection. The present study focuses on faults arising from actuators for linear systems. Finally, the application of the observer algorithm is presented to study the performances of this observer.