Paradigms in Scenario-Based Testing for Automated Driving

Abstract

Scenario-based testing is at present regarded as the enabling tool for the homologation of automated driving (AD) systems (SAE Level ≥ 3). This is mainly due to the aspiration of scenario-based testing for gaining understanding about the tested AD system, which provides support for arguments that shall convince ourselves about its capabilities. Since this way of testing involves system-dependent intertwinings, inconsistencies can thus impair the correctness of the homologation. In order to avoid this, we propose encapsulating these intertwinings by means of paradigms, which are individual mindsets towards homologation. For their definition, we introduce behaviors as a respective language, in combination with behavioral shifts, for the systematic derivation of what needs to be defined for a paradigm. The applicability is shown by mapping the derived paradigms against proposed approaches in automated driving. Furthermore, this systematics helps to identify concerns about using paradigms in scenario-based testing.