Investigating the influence of connected information on driver behaviour: An analysis of pedestrian-vehicle conflicts in the middle section of urban road

Due to the vision obstruction caused by visually blind obstacles on urban roads, pedestrians suffer a high crash risk in pedestrian-vehicle conflicts. At the same time, the connected information can potentially improve driver behaviour with an earlier warning and driving aids. To ensure safer interactions between pedestrians and motor vehicles in the middle section of urban roads, this simulator-based study aims to investigate drivers' behaviour under the influence of connected information and predict crash risk during their interaction with pedestrians on urban roads, involving six conflict scenarios based on real-world traffic situations. The test employed a mixed experimental design, with connected information as the between-subject variable. A total of 70 participants were divided into a control group and an experimental group to complete the test. Results from linear mixed-effects models indicated that the presence of connected information and crosswalks positively influenced driver braking behaviour, resulting in a shorter reaction time, longer braking duration and distance, smaller maximum deceleration, and a reduced standard deviation of deceleration. Conversely, visual obstacles led to longer reaction times, while parked cars and buses negatively affected driver behaviour. Further, aggressive drivers exhibited poorer braking behaviour compared to neutral drivers. An explainable machine learning model was developed to predict pedestrian-vehicle crash risks during interactions, demonstrating satisfactory predictive accuracy. The presence of connected information and crosswalks was found to have a positive effect on reducing crash risks and improving safety margins. These findings provide valuable insights for implementing connected driving technology and developing measures to enhance pedestrian safety.