RISF 기반의 Robust Fault Diagnosis 알고리즘

Abstract

With the growth of Advanced Driver Assistance System(ADAS), the automobile has been changed from traditional mechanical systems to electric/electronic control systems. For accurate control, it is essential to estimate of the vehicle states correctly. If the sensed value is incorrect, it can be a great safety hazard. The most important part to consider in Fault Detection and Isolation(FDI) Algorithm design is the false alarm rate. To reduce the false alarm rate, it could be a good way to design an algorithm more robust. Especially vehicle has high non-linearity due to tire slip or lock, body slip or cornering stiffness. These factors make it hard to estimate correct longitudinal and lateral velocity. In this paper, Reliability Indexed Sensor Fusion is applied to make a more robust fault diagnosis algorithm. In addition, Generalized Observer Scheme is also applied to detect and isolate the vehicle sensor fault.