Research On CAV Cooperative Control Method And Effectiveness Evaluation At Unsignalized Intersections In Unreliable Communication Scene

Connected and automated vehicle(CAV)integrates vehicle-to-everything(V2X)and autonomous driving technologies.Its wide application can realize the cooperative and optimal control of CAVs at the unsignalized intersection,which provides a new solution to improve the effectiveness of traffic control at intersections and is a hot spot for research in the field of intelligent vehicle infrastructure cooperative control in recent years.However,the scenarios constructed by existing studies generally assume that V2 X communication is completely reliable,ignoring the impact of some inherent unreliable characteristics of V2X(such as packet loss or delay)on the effectiveness of cooperative control.In addition,there is still a lack of means to comprehensively evaluate the effectiveness of cooperative control of CAVs at intersections.Based on the above considerations,relying on National Key R&D Program of China “Automated driving simulation and digital twin test evaluation tool chain”(2021YFB2501200),this thesis focuses on the CAVs cooperative control and effectiveness evaluation at signal-free intersection under unreliable V2 X communication scenarios,and specifically carries out the following aspects of work:(1)A typical connected unsignalized intersection scenario is constructed,and the timevarying communication performance of V2 X was quantitatively analyzed.Firstly,based on the analysis of the traffic conflict patterns at the intersection,a typical application scenario of V2 X intersection with signal free is constructed,and the CAVs’ behavior,kinematics model and carfollowing model are constructed.Then,a V2 X communication performance test environment is built in the test field,and the real-vehicle closed-field test of V2 X time-varying communication performance is carried out under four static and dynamic test cases,and the time-varying communication performance is tested using delay and packet delivery rate.(2)A cooperative control method for CAV at unsignalized intersections considering unreliable V2 X communication is proposed.First,based on the systematic analysis of the influence of unreliable V2 X communication environment on CAV cooperative control,a communication packet loss model based on two-state Markov model is designed;Then,using the idea of centralized control,a layered framework for CAV cooperative control of unsignalized intersections based on the space-time resource reservation mechanism is proposed,and a timeout retransmission mechanism is introduced into the framework,and a V2 X communication based on a finite state machine is designed.Finally,a CAV trajectory planning method based on distributed Monte Carlo tree search is constructed.(3)A comprehensive evaluation method for the cooperative control efficiency of CAVs at intersections is designed.Firstly,the average vehicle speed at the intersection,the vehicle flow at the intersection,the vehicle density at the intersection,and the average queuing length of vehicles at the intersection are selected as the evaluation indicators for the cooperative control efficiency of CAVs at the intersection.Then the weights of the above evaluation indexes are determined based on the analytic hierarchy process using expert experience.Finally,based on the fuzzy comprehensive evaluation method,a comprehensive evaluation model for the cooperative control efficiency of CAVs at intersections is constructed,and the evaluation results can be divided into four grades: excellent,good,medium and poor.(4)A comparative experiment scheme is designed and the simulation test analysis is completed.Firstly,a simulation experiment environment is built based on the Veins platform,and six test cases of balanced traffic flow and unbalanced traffic flow are designed.Then,the fixed time control(FTC)strategy and the actuated light control(ALC)strategy are selected as the benchmark methods for comparison.The CAV cooperative control method of unsignalized intersections proposed in this thesis is tested and analyzed.Finally,the proposed evaluation method is used to comprehensively evaluate the control effectiveness.The experimental results show that the control effectiveness of the proposed CAV cooperative control method for signal-free intersections is significantly better than that of the FTC and ALC strategies,especially under the case of high traffic flow.The minimum average vehicle speed of the proposed method is still 2.5 times and 2.4 times the maximum values of FTC and ALC.The maximum intersection vehicle density is only 63.3% and 63.3% of the minimum values of FTC and ALC.The maximum intersection vehicle flow is three times that of FTC and ALC.The average queue length of vehicles at intersections always is 0,far lower than the results of FTC and ALC.In addition,based on the proposed comprehensive evaluation method,the evaluation results of this method are all excellent,but as the traffic flow increases,the evaluation results of FTC and ALC become worse and worse.