A Robust Calibration Methodology for an On-Board Diagnostic Car System

New car models are now by law equipped with on-board diagnostic (OBD) systems aimed at monitoring the state of health of strategic components that ensure low levels of polluting exhaust emissions. During development phases, for each new car model, the OBD system must be finely calibrated. This article presents a robust calibration methodology taking into account sources of variability mainly due to production process, operating, and environmental conditions. The methodology enables us to evaluate the false alarm and failure to detect risks intrinsically related to the adopted calibration. An application concerning an upstream oxygen sensor monitored by the OBD is presented.

Keywords:

• Applications in engineering and industry
• On-board diagnostics
• Robust parameter design
• Hypothesis testing
• Fault detection

ACKNOWLEDGMENTS

This work is the result of a joint research project of the University of Naples Federico II and ELASIS (FIAT Research Centre, Southern Italy). The authors want to thank the managers and engineers of ELASIS, Control Systems Department, for their energy and commitment to this project.

The authors are grateful to Professor Jeff Wu (Georgia Institute of Technology) for his prompt and helpful comments on a previous version of the article, and Professor Harriet Black Nembhard (Pennsylvania State University), whose review lead to a substantial improvement of the article.