Optimal Control Of Connected And Automated Vehicle Driving Behavior For Expressway Traffic Oscillation

The main line of the expressway and the intersection area of the ramp,the construction area,and the sudden traffic accident area are the potential bottleneck areas of the expressway.Complex driving behaviors such as vehicle acceleration,deceleration,lane change,merging,and separation often occur frequently when vehicles travel from normal road sections to bottleneck road sections.Improper driving behavior can easily induce traffic flow oscillations,leading to serious traffic capacity reduction,hysteresis effects,rear-end collisions and other complex traffic problems and incidents.The connected and automated vehicle(CAV)technology realizes the spacetime holographic collection of vehicle microscopic data and real-time control of driving behavior,providing new solutions for the improvement of traffic flow oscillation and its derivative traffic problems.Based on the evolution characteristics of vehicle behavior under traffic flow shocks,formulating appropriate vehicle behavior optimization control methods are helpful to actively eliminate traffic flow shocks and vehicle conflict accidents.This paper analyzes the characteristics of vehicle trajectory data in traffic flow in the United States and China.Starting from the four aspects of mechanism,technology,effect,and simulation,this paper explores and studies the optimization control method of CAV driving behavior under oscillations,so as to improve the stability of expressway traffic flow,vehicle traffic efficiency and safety level.The paper first analyzes the evolution characteristics and conditions of the "generation-spread-dissipation" stage of the traffic flow oscillations,and then establishes a traffic flow simulation calculation model under the CAV environment.Subsequently,a series of optimization control strategies and algorithms for vehicle driving behavior oriented to traffic flow oscillation elimination and vehicle conflict avoidance were proposed.Then,the proposed optimization control methods are simulated and evaluated to eliminate traffic flow oscillations and improve traffic safety and traffic efficiency.And the robustness of the proposed optimization control methods under different control parameter levels and scenarios are tested and analyzed.The main research content of this paper can be divided into the following aspects:Firstly,this paper analyzes the stage evolution characteristics of expressway traffic flow shock wave,and constructs a traffic simulation model of typical traffic flow oscillation scene in the CAV environment.Based on the real vehicle trajectory data of domestic and foreign traffic flow oscillations,the asymmetric behavior theory and wavelet transform method are used to identify and analyze the trigger conditions and evolution characteristics of oscillations.Analyze vehicle driving behavior and changes in traffic flow operating status in oscillations.Analyze the traffic flow operation and vehicle control requirements in the expressway oscillation area to screen and calibrate the traffic simulation model oriented to the evolution of oscillations and vehicle behavior control.The effectiveness of the calibrated model in simulating different types of oscillations and vehicle behavior is tested.These contents lay a theoretical and simulation foundation for the proposed and tested CAV behavior optimization control strategies.Secondly,a series of CAV behavior optimization control strategies for improving traffic efficiency in oscillation areas are proposed.Apply traffic wave theory to determine the microcosmic behavior conditions of vehicles for the occurrence,spread and dissipation of oscillations.Based on the Jam Absorption Driving(JAD)theory,the synergistic condition of vehicle behavior that oscillations can be completely eliminated is evaluated.Combining the propagation and distribution characteristics of isolated oscillation,double oscillations in moving bottleneck and multiple oscillations in fixed bottleneck,the conditions for vehicle flow state fluctuations and cooperative vehicle driving behavior elimination conditions under different oscillation scenarios are analyzed.The complete JAD strategy,the continuous JAD strategy and the cooperative JAD strategy aiming at the optimal traffic efficiency are respectively proposed to completely eliminate isolated oscillation,double oscillations in moving bottleneck and multiple oscillations in fixed bottleneck.The effectiveness of the proposed vehicle behavior cooperative control strategy in eliminating shock waves is proved by simulation,and the robust control effect of the proposed strategy under different control parameters and different control scenarios is tested.Thirdly,a series of CAV behavior optimization control algorithms for traffic safety regulation in areas with traffic flow shocks are proposed.Considering the difference in the collision risk distribution of vector moving vehicles,the collision risk potential field representing the safety state of vehicle motion is constructed based on the theory of virtual potential field.Combined with Velocity Obstacle(VO)theory,the risk velocity obstacle algorithm,reciprocal velocity obstacle and generalized velocity obstacle algorithm are respectively constructed for individual vehicles,multi-vehicle interaction and CAV platoon conflict avoidance.In this way,the potential risk areas and conflict speeds of vehicles under different risk intensities can be identified.Based on the vehicle dynamics model to track the vehicle motion state,a dynamic window method is constructed to determine the vehicle collision risk avoidance velocity.A path planning model for vehicle collision risk avoidance is constructed based on the principle of model predictive control.The effectiveness of the proposed vehicle behavior optimization control algorithm in avoiding conflict risks is proved by simulation,and the robust control effect of the proposed algorithm under different control parameters and different control scenarios is tested.Finally,a multi-objective Pareto optimization vehicle behavior cooperative control strategy is proposed for the joint improvement of safety and efficiency in oscillation areas.Based on the multi-objective optimization theory,a dual-objective optimization function oriented towards the minimum cumulative vehicle transit time and vehicle collision risk is constructed.A predictive control rule for vehicle lateral lane change is constructed based on the theory of reciprocal velocity obstacle.A multi-lane shock wave complete elimination control strategy is constructed based on the JAD theory.Construct the constraint conditions for the safety guarantee and behavior control of vehicle following and changing lanes in a timely manner.Construct the non-dominated sorting genetic algorithm NSGA.II to solve the multi-objective optimization solution Pareto solution set,and determine the multi-vehicle behavior collaborative optimization control strategy set for the joint optimization of safety and efficiency in the multi-lane traffic flow oscillation problem.Finally,the simulation proves the effectiveness of the proposed multi-objective optimization vehicle behavior cooperative control strategy in improving traffic safety and traffic efficiency.And test the Pareto solution set distribution of the proposed strategy under different control parameters,which provides a more practical control scheme for the improvement of problems in different control demand scenarios.