A Generic Risk Assessment Methodology and its Implementation as a Run-time Monitoring Device for Automated Vehicles

Abstract

In this paper, a generic run-time risk evaluation methodology utilizing sensor status and data quality metrics is proposed. The suggested risk quantification method is then utilized as a basis for a run-time monitoring device (MonDev) concept for automated vehicles. The MonDev concept utilizes an aggregation function of a set of risk factors associated with each sensor. A data-driven SVM method is used to generate weighting factors in the aggregation function. The implementation of the MonDev concept and the corresponding results are demonstrated using two example use cases in a simulation framework.