Evaluation and Sign-Off Methodology for Automated Vehicle Systems Based on Relevant Driving Situations

Abstract

Automated driving is currently under research in various projects and research activities. These activities are clustered in different research areas. The evaluation and sign-off methodology for automated vehicles is currently a challenge to be solved. An innovative approach for such a sign-off methodology is proposed. A state-of-the-art overview of existing evaluation methods for advanced driver assistance systems, which are already available on the market today, is given in the form of a systems engineering process model. The existing evaluation methods already require great effort for real-world testing. Applying these methods to automated vehicles will exceed reasonable budget and time. The proposed method for effective evaluation of automated vehicles considers a database of relevant driving situations, which are collected and simulated on the basis of accident data and field operational and dedicated studies under controlled conditions. The database of relevant driving situations is the basis for an effective evaluation to be applied in simulation, driving simulators, and in test track scenarios. The methodology is described, and necessary building blocks are provided to implement the proposed concept.