Abstract
Keywords - Fault detection, fault diagnosis, steering system, vehicle system, x-by-wire systems.
Abstract - In recent years, increasing requirements for improved vehicle performance, reliability and safety has focused attention on the use of Fault Detection and Isolation (FDI) when designing vehicle control systems. In order to meet the increasing demand for better performance and reliability during operation, model-based FDI schemes have been developed to detect faults in sensors and actuators, and to apply appropriate corrective action without adding new hardware to the vehicle.
A real challenge in many applications is the design of a FDI scheme that is able to distinguish between model uncertainties and occurrence of faults. In principle, the problem can be tackled designing an intrinsically robust model-based algorithm. However, the advantages of this approach are undermined by the fact that the design of such algorithms is a difficult and complex task. A simpler alternative approach consists of a design carried out in two steps: the first step is the design of a FDI algorithm which neglects model uncertainties, and the second step consists in introducing opportune modifications to the FDI algorithm in order to make it robust to model uncertainties. Even if the second approach provides limited detection capabilities, its simplicity of implementation makes it more appealing for real implementation.
In this paper, we adopt the second approach and present a passive robust design technique for fault detection and isolation of sensor faults in a steer-by-wire system. The proposed method is based on classic Luenberger observers endowed with adaptive thresholds that are dynamically varying and follow the behavior of the residuals in absence of faults. In the presence of a fault, the residual will cross the adaptive threshold and a flag is activated.
Simulation results on a validated vehicle model are reported to demonstrate the effectiveness of the proposed approach.