To pass or not to pass: a field experiment on human values in interactions with automated vehicles

Vulnerable road users (VRUs), such as pedestrians and cyclists, are at high risk in road traffic, accounting for more than half of all global traffic fatalities. Ensuring safe interactions with highly automated vehicles (AVs) requires understanding and predicting VRUs' behaviour. This study investigated the relevance and predictive role of human values alongside environmental factors in real-world VRU-AV interactions. In a field experiment using a Wizard-of-Oz paradigm, 28 pedestrians and 29 cyclists interacted with an oncoming vehicle in a space-sharing scenario. Human values were assessed both qualitatively and quantitatively, while distance to the vehicle and driving mode (AV vs. manually driven) were manipulated. Results show that numerous human values (e.g. comprehensibility, legal compliance, self-efficacy, relaxedness) were rated as highly relevant, but only values related to relaxed interaction significantly predicted pedestrians' behaviour. Distance predicted interaction behaviour for VRU groups, whereas driving mode had no effect. Overall, the findings highlight the importance of considering both environmental factors and human values. The study demonstrates that values provide a broader perspective for understanding VRU behaviour and informing the design of safe, trustworthy, and acceptable VRU-AV interactions.