Collaborative Control Of Intelligent Connected Vehicle Driving Trajectories And Intersection Signal Under Mixed Traffic Condition

With the rapid development of intelligent connected technology,intelligent connected vehicles(ICV)can share information with the intersection control system to provide a new optimization strategy for traffic control.However,in the next few years,the penetration rate of intelligent connected vehicles will continue to be low,and the mixed traffic condition of intelligent connected vehicles and traditional vehicles will exist for a long time.Studying the collaborative control of intelligent connected vehicle driving trajectories and intersection signal under mixed traffic condition has great significance to improve the efficiency of intersections.In this paper,the collaborative control of intelligent connected vehicle platoons driving trajectories and traffic signal timing under mixed traffic condition are studied based on the cellular automaton model,reinforcement learning technology and Q-learning algorithm.The main research works are as follows:(1)Construction of traffic flow model at signalized intersections under mixed traffic conditions.By analyzing the operating characteristics of vehicles under the condition of mixed traffic of intelligent connected vehicles and traditional vehicles,the concept of safe distance was introduced to improve the vehicle update rules for two different operating modes to build a new cellular automaton model of mixed traffic flow under intelligent connected environment.After that,the simulation platform was established.(2)Collaborative control optimization algorithm for vehicle trajectories and signal control based on reinforcement learning.The algorithm consists of two steps.Firstly,intelligent connected vehicles drive as platoons.Based on the reinforcement learning technology,the leader of the platoon was treated as an agent.The current position of the vehicle at the intersection,the state of the signal light and the speed of the vehicle were selected to describe the state characteristics of the vehicle,acceleration,cruise speed,and deceleration were used as the vehicle’s actions at the current state,and the delay time of the vehicle at the intersection was used as the reward basis for the vehicle control strategy to establish the vehicle trajectory and the traffic state of the intersection.The best mapping relationship between the intelligent connected vehicles to achieve the optimization of the trajectory of the intelligent connected vehicle.Secondly,the trajectory data of the intelligent connected vehicle platoon was collected to determine whether the vehicle platoon can pass completely within the same green light time.For the vehicle platoon which can not pass the intersection within the same green light time,the strategy of extending the green light time was adapted to optimize the signal control,and the convey speed recommendations was sent to the platoon to further optimize the platoon driving trajectory,realize the collaborative control optimization of the platoon driving trajectory and signal control and improve the traffic efficiency of the intersection.(3)Simulation and analysis.The simulation platform was used to analyze the driving trajectory and speed changes of intelligent connected vehicles,as well as the impact of intelligent connected vehicle penetration rates under different saturation levels on the traffic efficiency of intersections,thereby verifying the effectiveness of the algorithm.The research results showed that the average vehicle delay decreases with the increase in the penetration rate of intelligent connected vehicles.The effect was the most obvious in the case of medium saturation(0.7),followed by the effect in the case of low saturation(0.3),and the effect was the least obvious under condition of high saturation(0.9).