Identifying Difficult Environmental Conditions with Scenario-Based Hazard and Fault Analysis

Automated driving systems (ADS) are envisioned to operate in a complex open world. To ensure that difficult environmental conditions do not trigger inadequate system behavior, it is crucial to identify difficult conditions and test whether they expose specific functional insufficiencies. The standard ISO 21448 introduces the concept of Triggering Conditions but does not provide designated methods for finding such conditions. Thus, we present an analytical method to systematically identify triggering conditions. The method supports domain experts to comprehensively model given driving scenarios and exhaustively deduce potential hazardous behaviors of the ADS. We show the method's applicability by conducting a workshop series with 16 domain experts. There, 122 triggering conditions covering the sense-plan-act chain of ADS are elicited, based on which eight clusters are further identified.