Real-time implementation of IFDIA scheme in automotive system

Abstract

The paper describes the implementation of an Instrument Fault Detection, Isolation and Accommodation (IFDIA) scheme developed for real-time automotive applications. The IFDIA architecture, based on artificial neural networks (ANNs) and inference rules, was employed to locate and accommodate faults that could occur on the main sensors involved in the management of the engine operating. Many on-line tests carried out in several engine working and vehicle drive conditions, prove the goodness of the diagnostic procedure and its on-board applicability.