Predicting NOx emissions during sensor light-off by leveraging sensor layout diversity in connected fleets

This paper reflects on a concept that leverages diverse sensor configurations across a fleet of connected vehicles to enhance their emissions monitoring and diagnostics. In this vision, the vehicles of a same family are equipped with different sensor layouts and grades, and share data to support the monitoring of the entire feet. Multiple applications within this framework are outlined, and a specific use case consisting in predicting the emissions during the light-off of the tailpipe NOx sensor with artificial neural networks is discussed, demonstrating the benefits of the proposed architecture.