Abstract: To adapt to the new development trend of intelligent transportation, two traffic capacity models of signalized intersections are improved to make them suitable for an environment of connected autonomous vehicles (CAVs). The first is the minimum delay model, which obtains the green light time by minimizing the total delay at the intersection and obtains the cycle length according to the flow rate. This model introduces the ratio of the green light delay of each lane or lane group to the specific queuing service time as a parameter, which depends on the probability distribution of vehicle arrival at the intersection. The second is the hybrid model, which combines the static queuing model that estimates the queuing service time and a dynamic model that estimates the green light extension time from the unit delay, flow rate, lane speed limit, and detector length. A numerical example shows that the delay of the new model is significantly reduced in the environment of CAVs, which validates the two capacity models and provides the basis for the traffic management department to conduct intersection reconstruction.

Key words: signalized intersections, capacity, model, connected autonomous vehicles, traffic simulation