Driverless vehicles' potential influence on bicyclist facility preferences

ABSTRACT Research in the field of autonomous vehicle technology focuses on its enhanced safety and convenience for vehicle occupants. The present paper seeks to establish a line of inquiry that addresses the implications of autonomous vehicle technology for nonmotorized road users, in the present case, bicyclists. Studies show that motorized traffic volume and speed affect nonmotorized agents' behavior and facility preference, but the degree to which this will apply to a driverless environment requires further study. We developed a stated-preference survey that had respondents select their preferred facility in a variety of hypothetical scenarios with and without the presence of driverless vehicles and on street types of varying motorized traffic volumes and speeds. A Random Parameters Logit Model is estimated to analyze the links between facility preferences, sociodemographics, street types, and the existence of driverless vehicles. The model results suggest that increases in motorized traffic volumes and speeds correlate with a greater preference for separated facilities. The presence of driverless vehicles amplifies this preference. Controlling for other factors, under driverless vehicle conditions, the odds of selecting protected facilities, such as buffered bicycle lanes and cycle tracks, were more than double the odds found under current conditions. We conclude with recommendations for infrastructure and policy and suggestions for future research in this nascent field of study.