Model-based fault diagnosis for a vehicle chassis system

Based on a linearized 3-DOF bicycle model, this paper compares two different fault detection and isolation (FDI) methods for the FDI of a command input fault and an output sensor fault within a vehicle chassis system. The first FDI approach uses parity equations to decouple the two types of faults, while at the same time, minimizes the influence of the disturbance on the residuals. The second FDI method uses sliding mode observer design, and by monitoring the equivalent control of the observers, in which the fault information is contained, achieves the detection and isolation of the faults. Lane change maneuvers are simulated, and the results showed that, under the given simulation condition, both diagnostic approaches can successfully detect and isolate the two types of faults despite the existence of a lateral wind gust disturbance.