Multi-vehicle Interactive Risk Assessment Method For Mainline Entrance Of Expressway

In recent years,with the constant improvement of the urban expressway system construction in our country,the number of vehicles is also increasing year by year,so traffic safety has been highlighted.Expressway mainline entrance is a important node for traffic flow conversion between expressway and urban road at all levels,and carries a lot of merging vehicles.Vehicles need to change lanes here and the interact with other vehicles frequently.Operation characteristics of vehicles are complex and may result in the occurrence of on-ramp accidents from time to time,which will further cause serious traffic congestion,property losses and casualties.Therefore,the study of vehicle operation risk assessment at mainline entrance of urban expressway plays an important role in improving traffic safety level of expressway and reducing traffic accidents.This paper first analyzes the type of mainline entrance,divides the study area into four sections according to research needs.It investigates the traffic volume,speed,acceleration,traffic conflicts and other parameters by using the method of Unmanned Aerial Vehicle photography.The target detection and tracking algorithm is used to preprocess the collected data.The relationship between vehicle lane-changing and traffic conflict at the mainline entrance is calculated.Secondly,the influence of driving behavior,vehicle and road factors on multivehicle interactive operation risk is analyzed.The changes of vehicle state at the ramp are determined.Based on SSAM's classification of traffic conflict types,vehicle operating risks at mainline entrances are classified into four categories.The mechanism of multi-vehicle interaction risk is studied through the spatial and temporal trajectory characteristics of vehicles.The relationship between traffic conflict and vehicle lanechanging behavior is determined,and the spatial distribution characteristics of multivehicle interaction risk at mainline entrances are displayed.Then,with vehicle spacing,velocity,acceleration and steering angle using as covariates,a bayesian hierarchical model based on vehicle physical state layer,multivehicle interaction layer and risk probability layer is constructed.The model parameters were demarcated by Gibbs sampling method.The severity index of vehicle interaction risk was defined.This paper determines the risk threshold of multi-vehicle operation and divides multi-vehicle interaction risk into four levels.Finally,ROC curve was used to verify the rationality of the risk threshold.The validity of model parameters was tested by using the posterior prediction of P value and variation of parameter quantile.The traffic micro-simulation software VISSIM and surrogate safety analysis model(SSAM)were used to collect conflict data,analyze the relationship between conflict indicator and covariates,which verifies the rationality of the selection of covariates.Risk assessment method in this paper can conduct risk diagnosis and assessment of multiple interactions of the whole vehicle safety at ramp entry.It can provide theoretical guidance for the pilot effectively avoid risk,and also offer a reference for managers to develop safety improvement measures at the same time.This study is of great significance for active safety prevention and control for vehicles and drivers.It can give early warning and improve the safety level of the whole expressway system.