Research On Vehicle Lane-changing Behavior In Freeway Merging Areas In Heterogeneous Traffic Flow Environment

The inherent chaos of the ramp merging on highways is considered one of the main causes of traffic accidents and congestion.With the rapid development of autonomous driving and communication technologies,Connected and Autonomous Vehicles(CAV)will co-exist with Human-Driven Vehicles(HDV)for a long time until their full deployment is completed,and the microscopic traffic behaviors such as following and lane changing of heterogeneous traffic flows will be essential to improve traffic safety and traffic efficiency.The in-depth exploration of the behavioral characteristics and mechanisms of heterogeneous traffic flows will be crucial to improve the traffic safety and efficiency of mixed traffic systems.The paper combines natural driving data and focuses on the dynamic lanechanging behavior in highway merging area under heterogeneous traffic flow environment.First,based on the construction of vehicle field models of CAV and HDV respectively,the safety field model in heterogeneous traffic flow environment is formed by superimposing with the driving environment model.Specifically,based on the physical properties of the merging area environment and the dynamic characteristics of CAV vehicles,the environmental field model and CAV vehicle field model with physical significance are constructed,and the HDV vehicle field model is constructed in the psychological sense based on the driver’s personality and psychological characteristics.Then,combined with the safety field model,the potential lane-changing conflicts and congestion problems caused by ramp merging in the merging area are quantitatively analyzed and visually represented based on the state of the environment and vehicles.The early lane-changing model in the pre-merging area and the multiscenario forced lane-changing control strategy in the merging area are proposed,providing safety evaluation and risk reminders for vehicle interactions,driving safety,and dynamic lane-changing.Meanwhile,using the SUMO simulation tool,experiments in different scenarios of free and forced lane changing are constructed,and the improvement effect of the dynamic lane-changing model and control strategy on the safety conflict and traffic efficiency is simulated and analyzed.The main conclusions of this paper are summarized as follows:(1)To address the problem of modeling the heterogeneous traffic flow without differences in existing driving safety fields,we consider the differences in perception and forces between CAV vehicles and HDV vehicles,construct a vehicle field model and improve the action range constraints of the CAV vehicle field respectively,and optimize the model performance parameters using differential evolutionary algorithms.Considering the influence of HDV drivers’ personality differences and driving environment on driving safety,a comprehensive index of drivers’ environmental psychological visibility is constructed,and the psychological forces of HDV drivers in mixed driving are derived with the concept of psychological field,and the vehicle field model of HDV conventional vehicles is established.Furthermore,the vehicle field model determined was combined with the environmental field representing road boundaries,markings,and other features to form a safety field model that jointly describes the driving status and safety of heterogeneous traffic flow.The contextual analysis shows that the model can quantitatively analyze the driving risk of vehicles,and its contours can visualize the field intensity distribution of vehicles and characterize the driving safety space of vehicles.(2)To extend the safety field model of heterogeneous traffic flow to the merging area lane-changing scenario,for the early lane-changing active safety measure,considering the blind spot and lateral collision conflicts that may occur during lanechanging,the minimum safe distance for CAV vehicle lane-changing was modified based on the safety field model to enhance the safety margin for CAV and HDV vehicles,which was used as the lane-changing constraint.Then,an early lane-changing model was constructed with the goal of improving the CAV queue strength.Simulation experiments showed that the model,as one of the active safety and control measures to reduce weaving conflicts in the merging area,can effectively improve the vehicle speed during lane-changing and the traffic flow efficiency in the pre-merging area.(3)For the active safety measure of multi-vehicle combination lane-changing control strategy in the merging area,the risk of the target lane-changing position was measured by scalar measures such as safety potential and potential change rate,and a multi-vehicle combination lane-changing control strategy was constructed with the core of the driving safety index for lane-changing safety evaluation.The safety lanechanging timing was determined,and the driving speed of each vehicle after lanechanging was optimized.Simulation experiments showed that the strategy performed well in both traffic efficiency and safety effects,strengthened the interpretability of the safety field model of heterogeneous traffic flow,and provided decision-making support for vehicle lane-changing in the merging area.