Modeling the Effect of Psychological Fear in Mediating the Critical Gap Behavior at Roundabouts for Heterogeneous, Disorderly Traffic Conditions: A Conditional Process Analysis Approach
摘要
Driver behavior at roundabouts, especially under heterogeneous, disorderly (HD) traffic conditions, involves complex decision-making that extends beyond geometric and traffic flow parameters. This study introduces psychological fear, a behaviorally inferred parameter representing the tendency of drivers to reject otherwise acceptable gaps, particularly when large-size vehicles are present in the circulating flow. A novel Conditional Process Analysis (CPA) model was developed using PROCESS macros to quantify how psychological fear mediates the relationship between large-size vehicles and critical gap values, while central diameter acts as a geometric moderator. The model was validated with field data from eleven roundabouts, yielding a mean square error of 8.5%. Furthermore, a Monte Carlo sensitivity analysis was carried out and confirmed psychological fear as the most influential factor. Compared to classical models like Raff’s method or logit-based approaches, CPA offers a more behaviorally grounded and flexible framework, capable of capturing both direct and indirect influences on driver decision-making. These insights contribute to more human-centered approaches in traffic flow modeling and intersection design, particularly in developing urban contexts where traffic compositions are highly diverse. The findings offer new insights into the psychological dynamics of gap acceptance, enhancing the understanding of driver decision-making at roundabouts.