Sensor modeling and data fusion for a safety warning system in hydrogen-based vehicles

Abstract

This document describes the process of analysis, modeling and compensation through data fusion algorithms of two hydrogen sensors for automotive applications: one is used to detect small leaks (e.g. in pipes ranging from the tank to the engine compartment), while the other has the task of detecting hydrogen concentrations near the explosivity level. Fast and accurate models of the sensors are developed allowing an accurate analysis of the main issues of a practical implementation of the warning system. Such analysis drives the development of effective data fusion techniques to reduce the measure uncertainty (reducing the risk of false alarms or missed detections). The implementation of such techniques on microcontroller-based embedded systems with limited resources in terms of memory and computational power is also addressed.