Driver Adaptation to Partially Automated Driving in Urban Environments: Effects of Repeated Exposure and System Capabilities on Drivers’ Trust, Monitoring, and Response

Objective This paper aims to quantify how the repeated use of partially automated driving (PAD) systems in urban environments influences drivers’ trust, monitoring, and intervention, and how exposure to explicit system limits affects emerging adaptations. Background PAD is becoming increasingly available and expanding into urban environments. Meanwhile, safety depends on adequate driver behavior. Previous studies have reported decreases in vigilance and changes in trust with the repeated use of driver assistance systems. However, there is a general lack of studies examining driver adaptation to PAD systematically over time and in urban environments. Method We conducted two driving simulator studies with a total of 45 participants driving repeatedly on an urban commuting route for five drives with intermissions of one or two days. We investigated the influence of mere exposure and exposure to explicit system limits. Both studies employed a multimodal measurement approach, combining self-reported and observational data. Results Trust and single-glance aversions are moderated by system limits, alongside mere exposure. The number of critical responses to system failures reveals alarming driver behavior despite the use of a state-of-the-art driver monitoring system. Conclusion Driver adaptation is highly event-driven as system limits calibrate trust, sustain monitoring, and improve intervention behavior. Moreover, state-of-the-art driver monitoring systems may not be sufficient to ensure the long-term safety of PADs in urban environments. Application Urban PAD should adopt system limit warnings paired with gaze history-dependent prompts. Future studies and evaluations should extrapolate the revealed adaptation effects.