Robust diagnosis to measurement uncertainties using bond graph approach: Application to intelligent autonomous vehicle

This paper deals with robust bond graph model-based fault detection and isolation to improve the robustness of the diagnosis system in presence of measurements and parameters uncertainties. We develop a procedure of measurement uncertainties modeling directly on the graph. By using the structural and causal properties of the bond graph, the robust diagnosis is performed. The interest of the developed methodology consists in using the graphical tool not only for measurement uncertainties modeling, but also for designing robust fault detection and isolation algorithms. Moreover, this method can be easily automated. The developed approach is validated by an application to an electromechanical traction system of intelligent autonomous vehicle.