Drivers and automation: a systematic review on road suitability for assisted and partially/conditionally automated vehicles

Vehicle automation technology is advancing faster than our road system can adapt. It is vital to integrate automated vehicles effectively while ensuring drivers can safely and efficiently perform driving tasks when required. This systematic review examines vehicle automation’s impact on driver behaviour and performance with respect to the suitability of existing roads. Focusing on guidance tasks influenced by road geometry and traffic, the review screens 1569 studies; forty met eligibility criteria. The results reveal a significant deterioration in driver behaviour and performance during guidance tasks, characterized by protracted decision-making and reaction times, longer durations for guidance tasks, and increased variability in gaze movement, lateral displacement, and longitudinal and traversal kinematic parameters. While some technological solutions have been suggested to alleviate these declines, the need for more research into how road design affects driver behaviour in automated settings remains evident. A key finding is the limited evidence on both road geometric variables and traffic-related factors; the latter appears more often but remains limited and heterogeneously reported. Significant gaps in knowledge persist regarding automation’s impact on driving guidance in merging, diverging, weaving, and overtaking scenarios, especially on two-lane roads, requiring further study. Additionally, there are gaps related to environmental factors in the context of automation’s impact on guidance tasks and coverage of older drivers and interactions with pedestrians/cyclists and heavy vehicles (including professional drivers) were limited or absent. These points identify the most critical gaps and outline concrete directions for future work on road suitability under assisted and partially/conditionally automated driving.