| With the rapid development of China’s transportation industry,autonomous driving technology has provided new ideas and methods to ensure urban road transportation safety,improve traffic efficiency and build a smart transportation system.Urban road intersections are an important scene for intelligent vehicle traffic,where left-turn waiting area,as a traffic organization form that effectively improves the capacity and efficiency of left-turn lanes at road intersections,has been widely used in all major cities in China.How to plan a better trajectory for intelligent vehicles to turn left at such intersections to ensure vehicle safety and reduce energy consumption has become a challenging problem,and there is little research on intelligent vehicle left-turn trajectory planning at intersections with left-turn waiting area in the existing literature.Therefore,the systematic study of intelligent vehicle left-turn trajectory planning at left-turn waiting area intersections has important theoretical significance and application value for saving travel cost and improving utilization of road traffic resources and other practical transportation problems.Based on this background,this paper conducts research on the intelligent vehicle left-turn trajectory planning problem at left-turn waiting area intersections.The specific research work is as follows:First,based on the UAV video acquisition technology,overhead video is taken at four typical left-turn waiting area intersections,and video image processing technology is used to extract vehicle left-turn trajectory data and micro-motion parameters and pre-process them.According to the driving behavior,the vehicles are divided into stopping vehicles and nonstopping vehicles,and the spatial distribution,driving time distribution,lane distribution,speed distribution,acceleration distribution and other characteristics of the left-turn trajectories of the two types of vehicles are studied and analyzed,and the macroscopic distribution laws and microscopic motion parameters change laws of different types of vehicle trajectories are obtained.Based on the analysis of the above trajectory characteristics and vehicle driving motion characteristics,reference can be made for planning the left-turn trajectory which is in line with human driving characteristics.Secondly,on the basis of the quintic polynomial trajectory planning method,the intelligent vehicle left-turn trajectory planning method based on multi-objective optimization theory is proposed,and the trajectory planning problem is transformed into a trajectory optimization problem by combining with optimization theory.The multi-performance optimization objective function with comfort,efficiency and fuel consumption as indicators is constructed,and the driving state constraints are set by combining with the vehicle driving motion characteristics,and the collision constraints are established by considering the conflict with the opposite traffic flow.The adaptive weight update particle swarm intelligent optimization algorithm is selected as the solution method,and the penalty function is introduced to deal with the complex constraints,and the optimal solution of the model is sought within the limited feasible domain to improve the solution speed.Again,according to the left-turn waiting area intersection left-turn vehicle traffic rules and vehicle travel time-distance map to divide the intelligent vehicle left-turn traffic scenarios,put forward different scenarios of vehicle left-turn trajectory planning methods.Combined with the trajectory segmentation characteristics of vehicle left-turn process in specific scenes,the trajectory planning is carried out in stages,and the boundary conditions of the planning end are determined based on the vehicle driving characteristics and the boundary conditions of the starting point of planning,and the optimal trajectory of each stage is solved to complete the trajectory planning of the whole process of vehicle from driving in to driving out.Finally,based on the classification of intelligent vehicle left-turn passage scenarios at the intersection of left-turn waiting area,the simulation environment is built and six simulation conditions are designed based on the different vehicle driving states,and the trajectory planning simulation results are compared with human driving trajectories.The research results show that the vehicle stopping behavior significantly affects the trajectory space distribution,the vehicle left turn travel time,the instantaneous speed of driving in and out of the section are normally distributed,and the instantaneous acceleration size of vehicle movement is negatively correlated with the speed size.The method proposed in this paper can plan a smooth and collision-free left-turn reference trajectory under all working conditions.Compared with human driving,the driving efficiency of intelligent vehicles is improved by 11.7%on average,the fuel consumption is reduced by 23.28%on average,and the comfort is effectively improved,and the model still shows good adaptability under the condition of variable weights,which verifies the feasibility and effectiveness of the trajectory planning method proposed in this paper.The research results of the paper enriches the intelligent vehicle passing scenarios,provides technical and theoretical support for the safe and efficient passage of intelligent vehicles at left-turn waiting area intersections,and can provide reference for the research and analysis of vehicle left-turn trajectory planning methods. |
