Behavioral Decision-making Optimization And Simulation For Automated Vehicles

With the rapid development of Internet of Vehicles and autonomous driving technologies,autonomous driving is developing from single-vehicle intelligence to multi-vehicle cooperative driving,providing new solutions for improving traffic safety and traffic efficiency.The highway work zone is one of the major bottleneck areas in the highway system,where lane narrowing is prone to frequent crossing of vehicles.Due to the lack of effectively cooperative,vehicles may experience stop-and-go congestion in the area,which will greatly affect capacity and increase energy consumption.Therefore,it is important to study the behavioral decision-making optimization and simulation for automated vehicles in highway work zone scenario,which enables all vehicles to cooperative driving and pass the area safely and efficiently to improve traffic safety and reduce traffic congestion.The main contents of this thesis are as follows:（1）Aiming at the problem of vehicle cooperative optimization in highway work zone,the key to solve this problem is to optimize the traffic priority of vehicles in the conflict area.Therefore,this thesis constructs the behavioral decision-making optimization model for automated vehicles with the optimization objectives of minimizing the travel time and energy consumption while satisfying the safety constraints,as well as the vehicle kinematics,liveness,and right-of-way priority constraints.（2）Since the developed optimization model is a mixed integer nonlinear programming problem,it is difficult to solve it directly.However,the right-of-way priority problem in the constraints can be solved first,and then the vehicle trajectory planning can be completed by combining the right-of-way priority.Therefore,this thesis proposes a bi-level planning framework to reconfigure the optimization model,where the upper-level planning solves the vehicle’s priority order of passing seeking problem to achieve the optimal objective function;the lower-level planning solves the vehicle trajectory planning problem to ensure that the vehicle final passes through the conflict area according to the passing order given in the upper-level.In the upper-level planning,the passing priority search problem is converted into a tree search problem,and Monte Carlo tree search and heuristic rule algorithm are used to optimize the search;in the lower-level planning,the coordinated vehicle trajectory planning algorithm is designed to solve the vehicle trajectory planning problem,and all vehicle trajectories are planned by continuously iterating local trajectory planning,and the objective function values of the given passing priority order are feedback to the upper-level planning.Through continuous iteration of upper and lower planning,the optimal right-of-way priority order and vehicle trajectory scheme are finally returned.（3）The highway work zone traffic scenarios and the design of autonomous driving simulation experiments are built on the MATLAB software platform.Firstly,the case study was analysed and compared with the classical cooperative driving strategy,and the results showed that the proposed bi-level planning-based strategy can further save 4.15% of the total travel cost;secondly,the parameter search experiment was conducted to ensure the best performance of the proposed strategy.Finally,performance analysis is performed under different traffic flows.The results show that the strategy proposed in this thesis can deal with the complex situations under different traffic flows.And it has great performance in the indicators of traffic efficiency,energy consumption and safety.