Effect of System Capability Verification on Conflict, Trust, and Behavior in Automated Vehicles

Abstract

With automated driving, vehicles are no longer just tools but become teammates, which enable an increasing space of new interaction possibilities. By changing the relationship between the drivers and the automated vehicles (AV), conflicts regarding maneuver selection can occur. Conflicts can lead to safety-critical takeovers by the drivers. Current research mainly focuses on information requirements for takeovers, only a few works explored the factors necessary for automation engagement. Therefore, a fix-based driving simulator study with N=28 participants was conducted to investigate how verifiable information influences automation engagement, gaze behavior, trust, conflict, criticality, stress, and interaction perception. The results indicate, if drivers can verify the information given by the system, they perceive less conflict and more trust in the system, leading to a lower rejection frequency of an overtaking maneuver performed by an AV. The results indicate that systems that aim to prevent drivers initiated interventions should provide verifiable information.