Research Of Commuter Travel Behaviour And Network Effects In Autonomous Driving Environments With Vehicle Road Collaboration

With the rapid development of autonomous driving technology,the vehicle-infrastructure cooperation(VIC)environment has become a research hotspot in the transportation field.In the VIC environment,drivers can hand over driving tasks to the machine for engagement in in-vehicle activities,reducing travel time costs and increasing travel utility.This change can lead to changes in commuters’ travel behavior preferences and thus cause a series of impacts on urban traffic.The objective of this study is to explore the impact of autonomous driving technology on commuters’ travel behavior preferences after entering the VIC section,including departure time and travel route choices.It also analyzes the impact of these changes on urban traffic,such as the total travel time,total travel distance,and total delay.By analyzing the impact of autonomous driving technology on travel behavior,a better understanding of its application effectiveness in the VIC environment can be achieved,providing a scientific basis for future urban transportation planning and design.To better simulate the real VIC autonomous driving travel,this study constructed a commuters’ activity utility model based on the classic bottleneck model theory,fully revealing the utility changes of commuters in the traditional driving section and the VIC section.The activity utility of non-autonomous driving,home autonomous driving,work autonomous driving,and general autonomous driving was analyzed and studied separately.The feasibility of the commuters’ activity utility model in the VIC environment was proved through experimental simulations.The study adopted an agent-based model to simulate commuters’ travel during the morning peak period.In the agent-based model,commuters are not homogeneous,and each commuter has attributes such as income level,departure time and travel route choices,travel costs,and vehicle types.The information exchange between each agent is used to simulate vehicle-to-vehicle communication in the VIC environment.The agent-based model has strong applicability for analyzing the sensitivity of commuters to the activity utility model.The sumo-plexe autonomous driving simulation platform was used for experimental simulations,which established the formation process of the autonomous driving queue when entering the VIC section,and the splitting process of the autonomous driving queue when leaving the VIC section.The sumo-plexe platform is suitable for simulating the driving conditions of a large-scale autonomous driving fleet,enabling faster testing and optimization of the autonomous driving formation system,simulating various dangerous and emergency situations,and improving operational efficiency.The analysis of experimental results indicates that regardless of whether it is a single OD path simulation or a city road network simulation,when the activity utility of commuters tends towards an equilibrium state,the order of travel activity utility from small to large is:non-autonomous driving,family type,general type,and work type.The selection of departure time is in the order of general type,family type,work type,and non-autonomous driving.As the travel activity utility increases,the travel cost decreases,and the departure time becomes later.The departure time of the high-income population is generally later than that of the low-income population.In terms of travel path selection,autonomous driving cars tend to choose routes with a larger proportion of mileage in the vehicle-road coordination sections,while non-autonomous driving vehicles are more inclined to choose shorter routes.In addition,the high-income population is more inclined to choose routes with a larger proportion of mileage in the vehicle-road coordination sections compared to the low-income population.Compared to non-autonomous driving,autonomous driving cars will lead to longer travel time and vehicle mileage,as well as increased delays.In terms of population heterogeneity,the high-income population has shorter driving distances,longer driving times,and experiences more travel delays compared to the low-income population.