Lane-Changing Safety Study Considering Next-Nearest Following Vehicle

Lane changing is a common driving behavior on the road,especially in the complex traffic conditions of urban roads.Drivers often have to change lanes in order to achieve their driving purposes.At the same time,the process of lane changing could produce conflict points,then reduce road safety,which in turn will cause road traffic accidents and result in economic losses.With the development of computer and information technology,the research on the safety of lane changing behavior is no longer limited to the traditional mathematical model.In this aper,with the help of the next generation of simulation data,based on the testing data,the safety of the next-nearest following vehicle is modeled and analyzed.In the actual lane changing behavior on the road,the next-nearest following vehicle is a conflict point that is easy to be ignored.As behind but not directly interact with lane changing vehicle,the next-nearest following vehicle is jointly affected by the vehicle in front of it,and the transmission of the influence is more complex.It is of great significance to establish the relationship between the driving condition of the vehicle in front and the next-nearest following vehicle for improving the safety.Firstly,the paper discusses the research background.Based on the research of traffic trajectory data,data-driven driving behavior and lane changing safety,the research status domestic and abroad is expounded,and the research route of the paper is designed on this basis.Secondly,the lane changing behavior characteristics of the nearest and next-nearest following vehicles are compared and analyzed.The lane changing process is divided into two stages,and the interaction between lane changing vehicle and nearest and next-nearest following vehicles is analyzed.Factors affecting the next-nearest following vehicles are obtained,and the parameters used are given.Thirdly,the LSTM neural network is used to establish the lane change speed prediction model for the next-nearest following vehicle.The next generation simulation data was selected to obtain the information of the next-nearest following vehicle corresponding to the lane changing time of the lane changing vehicle,and the statistical value of the relevant parameters was given.The lane change speed prediction model of the next-nearest following vehicle with multiple input parameters is established and the prediction results are analyzed.Finally,based on the assumption that the vehicle driving strategy changes in real time during lane changing,the concept of velocity variation intensity is introduced.Besides,combined with the initial lane change speed of lane change vehicle,the velocity change,space headway change and braking safety of the next-nearest following vehicle are studied.The relationship between the intensity of velocity variation,the initial speed of lane change,the speed difference between two adjacent vehicles and the safety of the next nearest vehicle is established.Suggestions and measures for the safety of lane changing are given.