Health Index Modeling for Trustable Electronic Sensor Systems in an Autonomous Application

Abstract

This article presents a model to evaluate a relevant health index of an autonomous lane keeping robot in order to make the associated electronic sensing system more trustable. The system's functional operation relies on the quality of the image data from a mounted camera sensor. In order to emulate the relevant hardware and environmental faults of the sensor, faults with different intensities are injected to the image data. Based on the faults and the associated quality of the lane keeping performance, the presented model evaluates and assigns a suitable risk index for a given fault. The risk value is then used to determine the necessary control action to ensure the safe operation of the system. The presented concept can be easily adapted to different autonomous applications.