Travel time information serves as a basis for travel planning, route guidance and congestion avoidance. The advances of connected and autonomous vehicle (CAV) technologies offer vehicles the potential of reduced travel time compared to conventional human-driven vehicles (HDVs). In the future, the road will be shared by both HDVs and CAVs, leading to mixed traffic flows that can significantly differ from the single-class HDV traffic flow. Thus, to explore the characteristics of travel time with mixed traffic will be an essential issue for better traffic operation and management. In this study, travel time distribution was investigated along the arterial with three continuous intersections established by VISSIM microsimulation. A series of simulation experiments were conducted, and finite mixture of regression models were used to characterize the mean, variance and mixing weight of different components. The impact factors such as traffic volume, cycle length, CAV penetration rate and driving behavior variables were all scrutinized.

CICTP