Risk Assessment for Mixed Traffic in Freeway Weaving Areas: A Transformer-Based Method

This paper focuses on the traffic operation scenario of freeway weaving areas, taking the macro traffic flow of mixed manual and autonomous driving as the research object, and conducts research on traffic operation risk assessment methods. Considering the characteristics of traffic flow parameters such as existing road environment, traffic flow state, traffic composition, and traffic conflicts, a multi-dimensional index system for the operational characteristics of mixed traffic flow in weaving areas is established. Parameters significantly related to risks are screened out through correlation analysis, and a comprehensive risk assessment model is constructed to output the traffic operation risk status.