Research on Modeling and Simulation of Interaction Behavior in an Unsignalized Intersection

Commercial traffic simulation software is the main methodology to evaluate traffic efficiency. However, it is difficult to reproduce the microscopic driving behaviors at complex scenarios due to lane-based structure and simplified traffic models. This paper proposes an interactive behavior model integrated decision-making and motion-planning process at unsignalized intersection. The decision-making model is established based on cooperative game theory and the motion-planning model is established by improved hybrid A* algorithm. To test the proposed model, a discrete simulation model is built based on a refined grid structure where vehicles have two-dimensional freedom of movement compared with lane-based structures. The proposed simulation model can evaluate traffic safety and efficiency simultaneously. Results show that simulation data and field data have no significant difference in terms of interaction behaviors distribution, travel time distribution, post-encroachment time distribution, and trajectory distribution.