Relevance and predictive role of human values in pedestrian and cyclist interactions with automated vehicles

As vehicle automation increases, vulnerable road users (VRUs) will need to negotiate right-of-way with level 4 and level 5 automated vehicles (AVs). Designing safe and comfortable interactions, therefore, requires a clear understanding of which human values are relevant for VRUs in such encounters. However, existing research lacks a comprehensive overview of relevant values, clear methods for quantifying them in concrete interaction scenarios, and evidence of their predictive role. To address these gaps, the present video-based laboratory study investigated VRU–vehicle interactions in a bottleneck scenario. Participants took the role of a pedestrian ( nₚ = 60) or a cyclist ( n c = 65; between-subjects factor) while interacting with an automated or a manual vehicle (within-subjects factor) and with or without another VRU (within-subjects factor). Value relevance was assessed quantitatively through ratings of value definitions and qualitatively by mapping values to participants' explanations of their behavioural intentions. The results show that a range of values, including comprehensibility, integrity, self-efficacy, and relaxedness, are relevant to VRU–AV interactions and should be considered in AV design. However, the predictive power of values for VRUs' behavioural intention was limited. Instead, participants' situation-independent decision preferences to wait or not wait were the strongest predictor of behavioural intention. These findings highlight the relevance of values for understanding VRU–AV interactions while indicating that their predictive power requires further investigation.