The Construction Of Risk Field And Optimization Of Driving Behaviors For Signalized Intersections

With the development of intelligent automobile technology,autonomous vehicles have been gradually put into the automobile market.In the future of transportation system,the traffic flow will no longer be composed of manual driving vehicles,and the state of automatic driving vehicle and manual driving vehicle mixed driving will last for a long time.Therefore,it is of great importance that optimizing the driving behaviors of autonomous vehicles,which will improve the driving safety and efficiency,and benefits the prevention of traffic accidents and congestion.As the connection of urban road network,the urban intersection has complex driving environment,and various driving elements gather at intersection.Each direction,different types of traffic participants will gather,interweave and evacuate at the intersection.There are many factors that have impact on vehicles driving through an intersection.However,the factors considered in the existing traffic models are limited,it is difficult to accurately express the driving environment under the combined impact of various factors,and the existing traffic optimization models for autonomous vehicles usually focus on one of trajectory optimization or speed optimization.Therefore,this paper will analyze each kind of factors that affect vehicle driving at signalized intersections,establish a unified Driving Risk Field model based on the influencing factors,obtain the distribution of traffic risk in space and time dimensions,reconstruct the vehicle traffic trajectory based on the real-time distribution of traffic risk field,finally carry out automatic driving based on the benefit function considering multiple optimization objectives traffic behavior optimization of driving vehicles.First of all,this paper summarizes the main progress of the existing research on intersection crossing model.On this basis,this study analyzes the traffic behavior characteristics of vehicles at intersections,then discusses the factors that will affect the vehicle driving at signalized intersections,and analyzes the influence mechanism of various factors on vehicle driving.In this paper,we classify and summarize the factors that affect vehicle driving in signalized intersections into environmental factors and moving factors.Secondly,the environmental field is established to express the degree of influence of environmental factors on driving traffic behavior,and the moving field is established to express the degree of influence of moving factors on vehicle driving behavior.Then,superposing two kinds of field forms Driving Risk Field model,which can obtain the distribution of driving risk in space and time dimensions.Combined with the driving risk value of each space point at any time and the information of the vehicle itself,this paper proposes a reconstruction method to reconstruct the traffic behavior of the vehicle.On the basis of trajectory reconstruction,a multi-objective single vehicle benefit optimization function is established to optimize the traffic behavior of autonomous vehicles at intersections.Finally,the simulation system is established The trajectory of vehicles passing through the intersection in typical scenes is reconstructed,and the sample vehicles are selected for driving behavior optimization.After the completion of the simulation,the results have been output,calculate the reconstruction accuracy,verify the accuracy of the reconstruction method,calculate the single vehicle driving efficiency and intersection traffic efficiency,delay time,respectively verify the effectiveness of the optimization scheme for single vehicle traffic state and intersection operation state.The results show that,compared with the existing intersection trajectory reconstruction model,the proposed model improves the accuracy of displacement reconstruction,and the average displacement reconstruction accuracy is 95.40% for straight vehicle,91.44%for left turn,and 92.94% for right turn;In the aspect of automatic vehicle driving behavior optimization,it has significant effect on the traffic benefit of single vehicle and the overall passing benefit of the intersection.After the optimization of all vehicles in the observation period has been optimized,the crossing efficiency is improved by9.3% and the delay is reduced by 3.6%.Compared with the existing traffic model and optimization scheme,the proposed method has some innovation,and the feasibility of the method is proved in the simulation test.The results of this study can be applied to the intersection driving control of autonomous vehicles,and provide a model basis for the expression of driving environment in mixed flow environment and the safe driving control of autonomous vehicles.