Intersection rear-end collision accident frequency prediction method and system based on collision risk force

Through the frequency prediction method of rear-end collision accident at intersection based on collision risk force, a vehicle risk status and driver-vehicle combination feature model is constructed, which solves the problems of unbalanced data processing and parameter space dependence in the existing technology, and improves the rear-end collision accident at intersection The accuracy of frequency prediction promotes active prevention of traffic safety management.

CN119028123APending Publication Date: 2024-11-26GUILIN UNIV OF ELECTRONIC TECH +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410948418.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-16
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The existing traffic accident risk assessment model has deteriorated performance when processing unbalanced data, and cannot effectively predict the frequency of rear-end collisions at intersections, especially because the problem of parameter space dependence is difficult to solve.

Method used

The frequency prediction method of rear-end collision accidents based on collision risk force is adopted. By obtaining vehicle rear-end collision accident data, a vehicle risk status model is constructed, road trajectory data is collected, driver-vehicle combination feature model is established, and rear-end accidents are established using generalized Pareto distribution. Probability model, predicts the frequency of rear-end collision accidents at a specific time.

Benefits of technology

It has improved the accuracy of the frequency prediction of rear-end collision accidents at intersections, promoted urban traffic safety management, shifted from passive governance to active prevention, and enhanced the initiative and preventiveness of traffic safety management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119028123A_ABST
    Figure CN119028123A_ABST
Patent Text Reader

Abstract

The invention discloses an intersection rear-end collision accident frequency prediction method and system based on collision risk force. The method comprises the steps of obtaining intersection vehicle rear-end collision accident data; constructing a vehicle risk condition model of the intersection according to the vehicle condition data; and acquiring road track data of the intersection, and performing vehicle accident prediction on the intersection according to the vehicle risk condition model. The method has the advantages that urban traffic safety management can be effectively promoted, and accident-oriented passive treatment is converted into risk-oriented active prevention.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation technology, and in particular to a method and system for predicting the frequency of rear-end collision accidents at intersections based on collision risk force. Background Art

[0002] With the rapid development of economy and technology, there are more and more different types of vehicles, and the number of cars is increasing, such as cars, buses, trucks, electric vehicles, etc. However, traffic accidents have increased accordingly, leading to frequent traffic accidents and a sharp increase in casualties. These problems have become increasingly prominent. The increase in casualties has brought tremendous pressure to traffic safety and social stability management. Accurate traffic accident risk assessment can effectively reduce the probability of traffic accidents and avoid road congestion caused by accidents. At the same time, it can improve the efficiency of rescue resource dispatch and reduce secondary injuries caused by delayed rescue time. Therefore, studying traffic accident risk assessment helps to discover the potential laws of traffic accidents and provide a theoretical basis for avoiding accident risks.

[0003] Although there are already a large number of mature research methods in the field of traffic accident risk assessment, there is still the problem of low accuracy of accident risk assessment models. Specifically, on the one hand, deep feature learning helps to enrich data semantic information, enhance model generalization and adaptability, and improve model evaluation performance. However, the features contain rich semantic information, which lacks effective and sufficient mining and utilization. On the other hand, traffic accident data itself is unbalanced, and existing model methods have good learning effects on balanced data. However, when dealing with unbalanced data, the model performance will decline, resulting in inaccurate evaluation. In particular, the prediction of the frequency of rear-end collisions at intersections is crucial for active safety management, and existing statistical models and machine learning models generally cannot solve the problem of spatial dependency of parameters.

[0004] In view of the above situation, the present invention proposes a method and system for predicting the frequency of rear-end collision accidents at intersections based on collision risk force, which can effectively improve the existing technology and overcome its shortcomings. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a method and system for predicting the frequency of rear-end collision accidents at intersections based on collision risk force to solve the above problems existing in the prior art. The specific scheme is as follows:

[0006] In a first aspect, the present invention provides a method for predicting the frequency of rear-end collision accidents at intersections based on collision risk force, the method comprising:

[0007] Obtain intersection rear-end collision accident data;

[0008] Constructing a vehicle risk condition model for the intersection according to the vehicle condition data;

[0009] The intersection road trajectory data is collected, and vehicle accidents are predicted at the intersection according to the vehicle risk status model.

[0010] Preferably, the acquisition of the occurrence model of the vehicle rear-end collision accident at the intersection includes:

[0011]

[0012] Among them, T i is the collision time of vehicle i relative to the preceding vehicle at time t in a rear-end collision accident, V i (t) is the instantaneous speed of vehicle i at time t, V h (t) is the instantaneous speed of vehicle h at time t, l i is the length of the h car.

[0013] Preferably, the method for collecting the intersection road trajectory data comprises:

[0014] Determine the acquisition range of the video sensing device to cover each entrance and exit ramp of the intersection, so that the video sensing device can detect and identify the motor vehicles within the coverage range and the real-time speed of the motor vehicles in real time;

[0015] Selecting the pedestrian crossing line at the intersection as a benchmark, taking the lower left vertex of the pedestrian crossing line as a default coordinate origin O, and establishing a spatial rectangular coordinate system of the intersection;

[0016] Based on the spatial rectangular coordinate system, the position C of the motor vehicle within the coverage area is obtained. i (x, y), and based on the default coordinate origin O as the center point of the automatic tracking of the preceding vehicle, obtain the trajectory position of the preceding vehicle i+1 at time t:

[0017] C i+1 (x1,y1)→C i+1 (x2,y2)→...→C i+1 (x t ,y t )

[0018] Select the mass point position at the rear of the front vehicle as the coordinate origin, track the mass point at the front of the rear vehicle, and establish the relative distance and time change relationship between the front vehicle i+1 and the rear vehicle i:

[0019] Y(C i ,C i+1 )=kX(t i ,t i+1 )

[0020] Among them, Y is the relative distance between the two vehicles when the rear-end collision occurs, X is the time difference between the front and rear vehicles when the rear-end collision occurs, and k is the relative speed of the two vehicles. max The maximum time is T m , with the maximum relative speed k max Continue driving, the time when the two cars collide is T e , rear-end collision is Δt′=T e -T m .

[0021] Preferably, the method for constructing the vehicle risk status model at the intersection comprises:

[0022] A road user generates a repulsive collision risk force on another road user based on the vehicle condition data;

[0023] Based on the collision risk force, the road traffic risk of the intersection is evaluated, and a collision risk force model for representing rear-end collision is constructed according to the combined characteristics of the driver and the vehicle:

[0024]

[0025] Among them, F coll,21 is the collision risk force, F coll,21 The maximum value max(F coll,21 )=1,d 12 =0;Q i , i∈[1,2] is the combined driver-vehicle characteristics of the i-th road user; v r,12 is the relative speed of the two interacting vehicles, vehicle 1 and vehicle 2; Δt is the collision duration between the two interacting vehicles; subscript 1 represents the rear vehicle in a rear-end collision; subscript 2 represents the front vehicle in a rear-end collision; d 12 is the shortest distance between vehicle 1 and vehicle 2; Δd is d 12 →0 is the distance that the two vehicles continue to move forward after the collision; Δt′ is the rear-end collision, ε R is the road environment dielectric factor: ε R =λ6+λ7U dt +λ8U cu , U dt is the lighting condition; U cu is the road line shape; λ6, λ7, λ8 are the parameters to be estimated.

[0026] Preferably, the method for constructing the driver-vehicle combined feature includes:

[0027] The driver's reaction time and the vehicle's maximum available deceleration are used as the driver's characteristic parameters β r =λ0+λ1U sand vehicle characteristic parameter β v =λ2+λ3U c +λ4U hv +λ5U eb ;

[0028] The driver characteristic parameters and the vehicle characteristic parameters are combined to establish a driver-vehicle combination characteristic model:

[0029]

[0030] Among them, T r represents the reaction time, b max represents the maximum available deceleration, η is a constant; β r represents the driver characteristic parameter, U s Indicates whether the vehicle is speeding, λ0 and λ1 are parameters to be estimated; β v Represents vehicle characteristic parameters, U c Indicates a car; U hv Indicates heavy vehicles, U eb represents an electric bicycle, and λ3, λ4, λ5, and λ6 are parameters to be estimated.

[0031] Preferably, the method for predicting a vehicle accident at the intersection according to the vehicle risk status model comprises:

[0032] The collision risk force F coll,21 As a conflict indicator, F coll =1 is used as the dividing point between rear-end collision and rear-end collision accident. Based on the generalized Pareto distribution, the probability model of rear-end collision accident at intersection is established. The overall distribution function of collision risk force is:

[0033]

[0034] f=u+y

[0035] Z(f)=Z u (y)[1-Z(u)]+Z(u)f≥u

[0036] Where Z(f) is the collision risk force F coll,21 The value of the probability distribution function on f; Z u (y) is the conditional distribution function of the part of the collision risk force exceeding the threshold u; σ is the scale parameter; ξ is the shape parameter;

[0037] F coll = 1 as the dividing point between rear-end collision and rear-end collision accident, and the probability of rear-end collision accident is established as

[0038]

[0039] Among them, N u It represents the number of collision risk forces greater than the threshold u after the collision risk threshold u is determined; N represents the total number of samples;

[0040] According to the acquired trajectory data, the collision risk force x of the rear-end collision interaction vehicles at the intersection is extracted i , set the empirical mean excess function to

[0041]

[0042] Among them, u is the threshold, k is the collision risk force sample {x1,x2,...,x n}, the amount of collision risk force sample data greater than the threshold u;

[0043] The threshold value u is taken as the horizontal axis, and the average excess amount e is taken as the horizontal axis. k (u) is the vertical axis of the coordinate, and the collision risk data are arranged in descending order from large to small x1≥x2≥…≥x n , with k as the horizontal axis, (γ k,n ) -1 As the vertical axis, the statistic of the tail index is:

[0044]

[0045] Calculate the average excess e k (u) The point where the linear change is obvious after exceeding the threshold and satisfies the tail index γ k,n The horizontal coordinate corresponding to the point in the stable area is taken as the collision risk force threshold u;

[0046] According to the collision risk threshold u, the maximum likelihood estimation function L(y:σ) is solved to obtain the scale parameter σ and shape parameter ξ:

[0047]

[0048] According to the probability of rear-end collision Ω, the probability of rear-end collision at the intersection within the characteristic time T is predicted.

[0049]

[0050] Among them, A(T) is the expected number of rear-end collision accidents within time T; time is the actual observation time of the intersection.

[0051] Preferably, the method for obtaining intersection vehicle rear-end collision accident data comprises:

[0052] Select the accident data of signalized intersections within a preset time and space range to establish a historical accident database, and classify them according to the accident information master table, accident information personnel table, and accident information dictionary;

[0053] The classified accident information dictionary is divided into accident data information and specific accident content, and the accident data is cleaned according to the specific accident content to eliminate accident data occurring at non-signalized intersections.

[0054] Preferably, the method further comprises:

[0055] Based on the historical accident database, traffic accident collision types are divided into four types: frontal collision, side collision, rear-end collision, and other collisions;

[0056] Chi-square test was performed on each variable to determine the road environment that significantly affects the rear-end collision at the intersection. The road alignment, pavement structure, visibility, lighting conditions, and occurrence time were found to have significant differences with the accident type (p<0.05). The independent variables were tested for collinearity and the variance inflation factor was calculated. Variables with VIF values ​​greater than 10 were eliminated;

[0057] Select Y=1, i.e., the frontal collision type, as the reference item, and use the four influencing factors of road alignment, visibility, lighting conditions, and occurrence time after elimination as independent variables to establish a multivariate Logit regression model for rear-end collision accidents:

[0058]

[0059] Where h = 2, 3, 4; g 0h is the constant term of the h-th regression equation; G H is the Hth independent variable; g Hh is the Hth independent variable of the hth collision type.

[0060] Preferably, the method further comprises:

[0061] Calibrate the various parameters to be estimated for rear-end collision risk:

[0062]

[0063] The parameters to be estimated are solved according to the data of the historical accident database to obtain a solution set:

[0064] ξ i ={η,λ0,λ1,...,λ8}

[0065] Solving the solution set to obtain the optimal solution set when the collision risk force is close to 1 is used to calibrate the various parameters to be estimated for the rear-end collision risk force:

[0066] ξ best ={η,λ0,λ1,...,λ8},F coll,21 →1.

[0067] In a second aspect, the present invention provides a system for predicting the frequency of rear-end collision accidents at intersections based on collision risk force, the system comprising:

[0068] A modeling module, used for constructing a vehicle risk status model of the intersection according to the vehicle status data;

[0069] A prediction module, used to collect the intersection road trajectory data and predict vehicle accidents at the intersection according to the vehicle risk status model;

[0070] Wherein, the method for constructing the vehicle risk status model of the intersection includes: a road user generates a repulsive collision risk force on another road user based on the vehicle status data;

[0071] Based on the collision risk force, the road traffic risk of the intersection is evaluated, and a collision risk force model for representing rear-end collision is constructed according to the combined characteristics of the driver and the vehicle:

[0072]

[0073] Among them, F coll,21 is the collision risk force, F coll,21 The maximum value max(F coll,21 )=1,d 12 =0;Q i , i∈[1,2] is the combined driver-vehicle characteristics of the i-th road user; v r,12 is the relative speed of the two interacting vehicles, vehicle 1 and vehicle 2; Δt is the collision duration between the two interacting vehicles; subscript 1 represents the rear vehicle in a rear-end collision; subscript 2 represents the front vehicle in a rear-end collision; d 12 is the shortest distance between vehicle 1 and vehicle 2; Δd is d 12 →0 is the distance that the two vehicles continue to move forward after the collision; Δt′ is the rear-end collision, ε R is the road environment dielectric factor: ε R =λ6+λ7U dt +λ8U cu , U dt is the lighting condition; U cu is the road line shape; λ6, λ7, λ8 are the parameters to be estimated.

[0074] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method described in the first aspect.

[0075] Beneficial effects: The method and system for predicting the frequency of rear-end collision accidents at intersections based on collision risk force of the present invention obtains historical rear-end collision accident data at the intersection to be predicted and vehicle trajectory data collected by the sensing device, pre-processes the data, and constructs a collision risk force model under rear-end collision conflict; establishes a combined characteristic model of driver-vehicle, determines the road environment factors that significantly affect rear-end collision accidents at intersections, and calibrates various parameters to be estimated for collision risk force; establishes a probability model for rear-end collision accidents at intersections, and then predicts the frequency of rear-end collision accidents at intersections within a specific time. This method can effectively promote urban traffic safety management, and transform from accident-oriented passive governance to risk-oriented active prevention. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and the embodiments in the drawings do not constitute any limitation to the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0077] Figure 1 It is a flow chart of an embodiment of a method for predicting the frequency of rear-end collision accidents at intersections based on collision risk according to the present invention.

[0078] Figure 2 It is a schematic structural diagram of an embodiment of a system for predicting the frequency of rear-end collision accidents at intersections according to the present invention. DETAILED DESCRIPTION

[0079] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments, which are preferred embodiments of the present invention. It should be understood that the described embodiments are only part of the embodiments of the present invention, not all of the embodiments; it should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0080] The main idea of ​​the technical solution of the embodiment of the present invention is: obtaining vehicle rear-end collision accident data at the intersection; constructing a vehicle risk status model at the intersection based on the vehicle status data; collecting the road trajectory data of the intersection, and predicting vehicle accidents at the intersection based on the vehicle risk status model.

[0081] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0082] Embodiment 1

[0083] An embodiment of the present invention provides a method for predicting the frequency of rear-end collision accidents at intersections based on collision risk force, such as Figure 1 As shown, the rear-end collision accident frequency prediction method may specifically include the following steps:

[0084] Step S101, obtaining intersection vehicle rear-end collision accident data.

[0085] In this embodiment, obtaining the rear-end collision accident data at the intersection specifically involves obtaining the rear-end collision accident data within the time period of the intersection to be predicted and performing preprocessing. Specifically, firstly, the accident data of the signalized intersection within a certain time and space range is selected to establish a historical accident database, and then the accident information master table, accident information personnel table, and accident information dictionary are classified respectively;

[0086] Then the main table of accident information is divided into fields such as accident cause, location, collision type, accident time, road environment conditions at the accident point, etc. The table of accident information personnel is divided into fields such as basic information of accident personnel (name, personnel type, blood alcohol content, etc.), accident conditions (degree of injury, role in collision, etc.), driving vehicle status (vehicle turn signal status, vehicle lighting status, etc.), etc. The accident information dictionary is divided into accident data information and specific accident content;

[0087] Finally, the accident data was cleaned according to the content of the accident: accident data occurring at non-signalized intersections such as road sections were eliminated, accidents involving pedestrians were eliminated, traffic accident data not related to traffic conflicts were eliminated, traffic accidents with the causes of the accident identified as "mechanical failure", "tire blowout", "improper throttle control", "other improper operations", etc. were eliminated, accident data with incomplete records of accident collision types were eliminated, and secondary accident data were eliminated.

[0088] In this embodiment, the model for collecting the occurrence of the rear-end collision accident at the intersection may be:

[0089]

[0090] Among them, T i is the collision time of vehicle i relative to the preceding vehicle at time t in a rear-end collision accident, V i (t) is the instantaneous speed of vehicle i at time t, V h (t) is the instantaneous speed of vehicle h at time t, l i is the length of the h car.

[0091] In an optional embodiment, the method may further specifically include: based on the historical accident database, classifying traffic accident collision types into four types: frontal collision, side collision, rear-end collision, and other collisions;

[0092] Chi-square test was performed on each variable to determine the road environment that significantly affects the rear-end collision at the intersection. The road alignment, pavement structure, visibility, lighting conditions, and occurrence time were found to have significant differences with the accident type (p<0.05). The independent variables were tested for collinearity and the variance inflation factor was calculated. Variables with VIF values ​​greater than 10 were eliminated;

[0093] Select Y=1, i.e., the frontal collision type, as the reference item, and use the four influencing factors of road alignment, visibility, lighting conditions, and occurrence time after elimination as independent variables to establish a multivariate Logit regression model for rear-end collision accidents:

[0094]

[0095] Where h = 2, 3, 4; g 0h is the constant term of the h-th regression equation; G H is the Hth independent variable; g Hh is the Hth independent variable of the hth collision type.

[0096] In another optional embodiment, the method may further specifically include: calibrating various parameters to be estimated of the rear-end collision risk force:

[0097]

[0098] The parameters to be estimated are solved according to the data of the historical accident database to obtain a solution set:

[0099] ξ i ={η,λ0,λ1,...,λ8}

[0100] Solving the solution set to obtain the optimal solution set when the collision risk force is close to 1 is used to calibrate the various parameters to be estimated for the rear-end collision risk force:

[0101] ξ best ={η,λ0,λ1,...,λ8},F coll,21 →1.

[0102] Step S102: constructing a vehicle risk status model of the intersection according to the vehicle status data.

[0103] In this embodiment, in a traffic safety field, a road user may generate a repulsive collision risk force toward another road user. Based on the collision risk force, the road traffic risk assessment of the intersection is implemented, and a collision risk force model representing the rear-end collision is constructed:

[0104]

[0105] Among them, F coll,21is the collision risk force, F coll,21 The maximum value max(F coll,21 )=1,d 12 =0;Q i , i∈[1,2] is the combined driver-vehicle characteristics of the i-th road user; v r,12 is the relative speed of the two interacting vehicles, vehicle 1 and vehicle 2; Δt is the collision duration between the two interacting vehicles; subscript 1 represents the rear vehicle in a rear-end collision; subscript 2 represents the front vehicle in a rear-end collision; d 12 is the shortest distance between vehicle 1 and vehicle 2; Δd is d 12 →0 is the distance that the two vehicles continue to move forward after the collision; Δt′ is the rear-end collision, ε R is the road environment dielectric factor: ε R =λ6+λ7U dt +λ8U cu , U dt is the lighting condition; U cu is the road line shape; λ6, λ7, λ8 are the parameters to be estimated.

[0106] The method for constructing the combined characteristics of the driver and the vehicle in this embodiment may specifically be to use the driver's reaction time and the maximum available deceleration of the vehicle as the driver characteristic parameters β r =λ0+λ1U s and vehicle characteristic parameter β v =λ2+λ3U c +λ4U hv +λ5U eb ;

[0107] The driver characteristic parameters and the vehicle characteristic parameters are combined to establish a driver-vehicle combination characteristic model:

[0108]

[0109] Among them, T r represents the reaction time, b max represents the maximum available deceleration, η is a constant; β r represents the driver characteristic parameter, U s Indicates whether the vehicle is speeding, λ0 and λ1 are parameters to be estimated; β v Represents vehicle characteristic parameters, U c Indicates a car; U hv Indicates heavy vehicles, U eb represents an electric bicycle, and λ3, λ4, λ5, and λ6 are parameters to be estimated.

[0110] Step S103, collecting the intersection road trajectory data, and predicting vehicle accidents at the intersection according to the vehicle risk status model.

[0111] In this embodiment, the method for collecting the intersection road trajectory data can be specifically that the collection range of the video perception device is 120 meters vertically from the center of the intersection and the ground, and the coverage range includes each entrance and exit ramp of the intersection. The perception device can detect and identify the motor vehicles within the coverage range and the corresponding real-time speed in real time; select the pedestrian crossing line of the intersection as a benchmark, and use the lower left vertex of the pedestrian crossing line as the default coordinate origin O to establish a spatial rectangular coordinate system of the intersection; based on the spatial rectangular coordinate system, obtain the position C of the motor vehicle within the coverage range i (x, y), based on the default coordinate origin O, the automatic tracking of the preceding vehicle is realized, and the trajectory position C of the preceding vehicle i+1 at time t is obtained i+1 (x1,y1)→C i+1 (x2,y2)→...→C i+1 (x t ,y t ), select the mass point position of the front car tail as the coordinate origin, realize the mass point tracking of the rear car head, and establish the relative distance and time change relationship Y(C i ,C i+1 )=kX(t i ,t i+1 ), Y is the relative distance between the two vehicles when a rear-end collision occurs, k is the relative speed of the two vehicles, when the relative speed k max The maximum time is T m , with the maximum relative speed k max Continue driving, the time when the two cars collide is T e , rear-end collision is Δt′=T e -T m .

[0112] In this embodiment, the method for predicting vehicle accidents at the intersection according to the vehicle risk status model can be specifically to convert the collision risk force F coll,21 As a conflict indicator, F coll =1 is used as the dividing point between rear-end collision and rear-end collision accident. Based on the generalized Pareto distribution, the probability model of rear-end collision accident at intersection is established. The overall distribution function of collision risk force is:

[0113]

[0114] f=u+y

[0115] Z(f)=Z u (y)[1-Z(u)]+Z(u)f≥u

[0116] Where Z(f) is the collision risk force F coll,21 The value of the probability distribution function on f; Z u (y) is the conditional distribution function of the part of the collision risk force exceeding the threshold u; σ is the scale parameter; ξ is the shape parameter;

[0117] F coll = 1 as the dividing point between rear-end collision and rear-end collision accident, and the probability of rear-end collision accident is established as

[0118]

[0119] Among them, N u It represents the number of collision risk forces greater than the threshold u after the collision risk threshold u is determined; N represents the total number of samples;

[0120] According to the acquired trajectory data, the collision risk force x of the rear-end collision interaction vehicles at the intersection is extracted i , set the empirical mean excess function to

[0121]

[0122] Among them, u is the threshold, k is the collision risk force sample {x1,x2,...,x n}, the amount of collision risk force sample data greater than the threshold u;

[0123] The threshold value u is taken as the horizontal axis, and the average excess amount e is taken as the horizontal axis. k (u) is the vertical axis of the coordinate, and the collision risk data are arranged in descending order from large to small x1≥x2≥…≥x n , with k as the horizontal axis, (γ k,n ) -1 As the vertical axis, the statistic of the tail index is:

[0124]

[0125] Calculate the average excess e k (u) The point where the linear change is obvious after exceeding the threshold and satisfies the tail index γ k,n The horizontal coordinate corresponding to the point in the stable area is taken as the collision risk force threshold u;

[0126] According to the collision risk threshold u, the maximum likelihood estimation function L(y:σ) is solved to obtain the scale parameter σ and shape parameter ξ:

[0127]

[0128] According to the probability of rear-end collision Ω, the probability of rear-end collision at the intersection within the characteristic time T is predicted.

[0129]

[0130] Among them, A(T) is the expected number of rear-end collision accidents within time T; time is the actual observation time of the intersection.

[0131] Embodiment 2

[0132] An embodiment of the present invention provides a system for predicting the frequency of rear-end collision accidents at intersections based on collision risk force, such as Figure 2 As shown, the collaborative real-time scheduling may specifically include the following modules:

[0133] An acquisition module is used to acquire the data of rear-end collision accidents at intersections;

[0134] In this embodiment, the data of rear-end collision accidents at intersections are obtained and pre-processed. Specifically, the accident data of signalized intersections within a certain time and space range are first selected to establish a historical accident database, and then classified according to the accident information master table, accident information personnel table, and accident information dictionary;

[0135] Then the main table of accident information is divided into fields such as accident cause, location, collision type, accident time, road environment conditions at the accident point, etc. The table of accident information personnel is divided into fields such as basic information of accident personnel (name, personnel type, blood alcohol content, etc.), accident conditions (degree of injury, role in collision, etc.), driving vehicle status (vehicle turn signal status, vehicle lighting status, etc.), etc. The accident information dictionary is divided into accident data information and specific accident content;

[0136] Finally, the accident data was cleaned according to the content of the accident: accident data occurring at non-signalized intersections such as road sections were eliminated, accidents involving pedestrians were eliminated, traffic accident data not related to traffic conflicts were eliminated, traffic accidents with the causes of the accident identified as "mechanical failure", "tire blowout", "improper throttle control", "other improper operations", etc. were eliminated, accident data with incomplete records of accident collision types were eliminated, and secondary accident data were eliminated.

[0137] In this embodiment, the model for collecting the occurrence of the rear-end collision accident at the intersection may be:

[0138]

[0139] Among them, T i is the collision time of vehicle i relative to the preceding vehicle at time t in a rear-end collision accident, V i (t) is the instantaneous speed of vehicle i at time t, V h (t) is the instantaneous speed of vehicle h at time t, l i is the length of the h car.

[0140] In an optional embodiment, the method may further specifically include: based on the historical accident database, classifying traffic accident collision types into four types: frontal collision, side collision, rear-end collision, and other collisions;

[0141] Chi-square test was performed on each variable to determine the road environment that significantly affects the rear-end collision at the intersection. The road alignment, pavement structure, visibility, lighting conditions, and occurrence time were found to have significant differences with the accident type (p<0.05). The independent variables were tested for collinearity and the variance inflation factor was calculated. Variables with VIF values ​​greater than 10 were eliminated;

[0142] Select Y=1, i.e., the frontal collision type, as the reference item, and use the four influencing factors of road alignment, visibility, lighting conditions, and occurrence time after elimination as independent variables to establish a multivariate Logit regression model for rear-end collision accidents:

[0143]

[0144] Where h = 2, 3, 4; g 0h is the constant term of the h-th regression equation; G H is the Hth independent variable; g Hh is the Hth independent variable of the hth collision type.

[0145] In another optional embodiment, the method may further specifically include: calibrating various parameters to be estimated of the rear-end collision risk force:

[0146]

[0147] The parameters to be estimated are solved according to the data of the historical accident database to obtain a solution set:

[0148] ξ i ={η,λ0,λ1,...,λ8}

[0149] Solving the solution set to obtain the optimal solution set when the collision risk force is close to 1 is used to calibrate the various parameters to be estimated for the rear-end collision risk force:

[0150] ξ best ={η,λ0,λ1,...,λ8},F coll,21 →1.

[0151] A modeling module is used to construct a vehicle risk status model of the intersection based on the vehicle status data.

[0152] In this embodiment, in a traffic safety field, a road user may generate a repulsive collision risk force toward another road user. Based on the collision risk force, the road traffic risk assessment of the intersection is implemented, and a collision risk force model representing the rear-end collision is constructed:

[0153]

[0154] Among them, F coll,21 is the collision risk force, F coll,21 The maximum value max(F coll,21 )=1,d 12 =0;Q i , i∈[1,2] is the combined driver-vehicle characteristics of the i-th road user; v r,12 is the relative speed of the two interacting vehicles, vehicle 1 and vehicle 2; Δt is the collision duration between the two interacting vehicles; subscript 1 represents the rear vehicle in a rear-end collision; subscript 2 represents the front vehicle in a rear-end collision; d 12 is the shortest distance between vehicle 1 and vehicle 2; Δd is d 12 →0 is the distance that the two vehicles continue to move forward after the collision; Δt′ is the rear-end collision, ε R is the road environment dielectric factor: ε R =λ6+λ7U dt +λ8U cu , U dt is the lighting condition; U cu is the road line shape; λ6, λ7, λ8 are the parameters to be estimated.

[0155] The method for constructing the combined characteristics of the driver and the vehicle in this embodiment may specifically be to use the driver's reaction time and the maximum available deceleration of the vehicle as the driver characteristic parameters β r =λ0+λ1U s and vehicle characteristic parameter β v =λ2+λ3U c +λ4U hv +λ5U eb ;

[0156] The driver characteristic parameters and the vehicle characteristic parameters are combined to establish a driver-vehicle combination characteristic model:

[0157]

[0158] Among them, T r represents the reaction time, b max represents the maximum available deceleration, η is a constant; β r represents the driver characteristic parameter, U sIndicates whether the vehicle is speeding, λ0 and λ1 are parameters to be estimated; β v Represents vehicle characteristic parameters, U c Indicates a car; U hv Indicates heavy vehicles, U eb represents an electric bicycle, and λ3, λ4, λ5, and λ6 are parameters to be estimated.

[0159] The prediction module is used to collect the intersection road trajectory data and predict vehicle accidents at the intersection according to the vehicle risk status model.

[0160] In this embodiment, the method for collecting the intersection road trajectory data can be specifically that the collection range of the video perception device is 120 meters vertically from the center of the intersection and the ground, and the coverage range includes each entrance and exit ramp of the intersection. The perception device can detect and identify the motor vehicles within the coverage range and the corresponding real-time speed in real time; select the pedestrian crossing line of the intersection as a benchmark, and use the lower left vertex of the pedestrian crossing line as the default coordinate origin O to establish a spatial rectangular coordinate system of the intersection; based on the spatial rectangular coordinate system, obtain the position C of the motor vehicle within the coverage range i (x, y), based on the default coordinate origin O, the automatic tracking of the preceding vehicle is realized, and the trajectory position C of the preceding vehicle i+1 at time t is obtained i+1 (x1,y1)→C i+1 (x2,y2)→...→C i+1 (x t ,y t ), select the mass point position of the front car tail as the coordinate origin, realize the mass point tracking of the rear car head, and establish the relative distance and time change relationship Y(C i ,C i+1 )=kX(t i ,t i+1 ), Y is the relative distance between the two vehicles when a rear-end collision occurs, k is the relative speed of the two vehicles, when the relative speed k max The maximum time is T m , with the maximum relative speed k max Continue driving, the time when the two cars collide is T e , rear-end collision is Δt′=T e -T m .

[0161] In this embodiment, the method for predicting vehicle accidents at the intersection according to the vehicle risk status model can be specifically to convert the collision risk force F coll,21 As a conflict indicator, F coll =1 is used as the dividing point between rear-end collision and rear-end collision accident. Based on the generalized Pareto distribution, the probability model of rear-end collision accident at intersection is established. The overall distribution function of collision risk force is:

[0162]

[0163] f=u+y

[0164] Z(f)=Z u (y)[1-Z(u)]+Z(u)f≥u

[0165] Where Z(f) is the collision risk force F coll,21 The value of the probability distribution function on f; Z u (y) is the conditional distribution function of the part of the collision risk force exceeding the threshold u; σ is the scale parameter; ξ is the shape parameter;

[0166] F coll = 1 as the dividing point between rear-end collision and rear-end collision accident, and the probability of rear-end collision accident is established as

[0167]

[0168] Among them, N u It represents the number of collision risk forces greater than the threshold u after the collision risk threshold u is determined; N represents the total number of samples;

[0169] According to the acquired trajectory data, the collision risk force x of the rear-end collision interaction vehicles at the intersection is extracted i , set the empirical mean excess function to

[0170]

[0171] Among them, u is the threshold, k is the collision risk force sample {x1,x2,...,x n}, the amount of collision risk force sample data greater than the threshold u;

[0172] The threshold value u is taken as the horizontal axis, and the average excess amount e is taken as the horizontal axis. k (u) is the vertical axis of the coordinate, and the collision risk data are arranged in descending order from large to small x1≥x2≥…≥x n , with k as the horizontal axis, (γ k,n ) -1 As the vertical axis, the statistic of the tail index is:

[0173]

[0174] Calculate the average excess e k (u) The point where the linear change is obvious after exceeding the threshold and satisfies the tail index γ k,n The horizontal coordinate corresponding to the point in the stable area is taken as the collision risk force threshold u;

[0175] According to the collision risk threshold u, the maximum likelihood estimation function L(y:σ) is solved to obtain the scale parameter σ and shape parameter ξ:

[0176]

[0177] According to the probability of rear-end collision Ω, the probability of rear-end collision at the intersection within the characteristic time T is predicted.

[0178]

[0179] Among them, A(T) is the expected number of rear-end collision accidents within time T; time is the actual observation time of the intersection.

[0180] Embodiment 3

[0181] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for predicting the frequency of rear-end collision accidents at intersections based on collision risk force is implemented.

[0182] In summary, the embodiment of the present invention provides a method and system for predicting the frequency of rear-end collision accidents at intersections based on collision risk force. By acquiring historical rear-end collision accident data at the intersection to be predicted and vehicle trajectory data collected by the sensing device, the data is pre-processed to construct a collision risk force model under rear-end collision conflict; a combined characteristic model of the driver and vehicle is established to determine the road environment factors that significantly affect the rear-end collision accidents at the intersection, and various parameters to be estimated for the collision risk force are calibrated; a probability model for rear-end collision accidents at the intersection is established, and the frequency of rear-end collision accidents at the intersection within a specific time is predicted. This method can effectively promote urban traffic safety management, and transform from accident-oriented passive governance to risk-oriented active prevention.

[0183] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described order of actions, because according to the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0184] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded or executed on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, a computer, a server or a data center by wired (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or a data center containing one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0185] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0186] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0187] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0188] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0189] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0190] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0191] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0192] The present invention is described with reference to the flowcharts and / or block diagrams of the methods, apparatuses (devices) and computer program products of the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0193] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0194] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0195] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0196] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting the frequency of rear-end collision accidents at intersections based on collision risk force, characterized in that: The method comprises: Obtain intersection rear-end collision accident data; Constructing a vehicle risk condition model for the intersection according to the vehicle condition data; The intersection road trajectory data is collected, and vehicle accidents are predicted at the intersection according to the vehicle risk status model.

2. The method according to claim 1, characterized in that The acquisition of the occurrence model of the vehicle rear-end collision accident at the intersection includes: Among them, Ti is the collision time of vehicle i relative to the front vehicle at time t in the rear-end collision accident, V i (t) is the instantaneous speed of vehicle i at time t, V h (t) is the instantaneous speed of vehicle h at time t, l i is the length of the h car.

3. The method according to claim 1, characterized in that: The method for collecting the intersection road trajectory data comprises: Determine the acquisition range of the video sensing device to cover each entrance and exit ramp of the intersection, so that the video sensing device can detect and identify the motor vehicles within the coverage range and the real-time speed of the motor vehicles in real time; Selecting the pedestrian crossing line at the intersection as a benchmark, taking the lower left vertex of the pedestrian crossing line as a default coordinate origin O, and establishing a spatial rectangular coordinate system of the intersection; Based on the spatial rectangular coordinate system, the position C of the motor vehicle within the coverage area is obtained. i (x, y), and based on the default coordinate origin O as the center point of the automatic tracking of the preceding vehicle, obtain the trajectory position of the preceding vehicle i+1 at time t: C i+1 (x1,y1)→C i+1 (x2,y2)→...→C i+1 (x t ,y t ) Select the mass point position at the rear of the front vehicle as the coordinate origin, track the mass point at the front of the rear vehicle, and establish the relative distance and time change relationship between the front vehicle i+1 and the rear vehicle i: Y(C i ,C i+1 )=kX(t i ,t i+1 ) Among them, Y is the relative distance between the two vehicles when the rear-end collision occurs, X is the time difference between the front and rear vehicles when the rear-end collision occurs, and k is the relative speed of the two vehicles. max The maximum time is T m , with the maximum relative speed k max Continue driving, the time when the two cars collide is T e , rear-end collision is Δt′=T e -T m .

4. The method according to claim 1, characterized in that: The method for constructing the vehicle risk status model at the intersection comprises: A road user generates a repulsive collision risk force on another road user based on the vehicle condition data; Based on the collision risk force, the road traffic risk of the intersection is evaluated, and a collision risk force model for representing rear-end collision is constructed according to the combined characteristics of the driver and the vehicle: Among them, F coll,21 is the collision risk force, F coll,21 The maximum value max(F coll,21 )=1,d 12 =0;Q i , i∈[1,2] is the combined driver-vehicle characteristics of the i-th road user; v r,12 is the relative speed of the two interacting vehicles, vehicle 1 and vehicle 2; Δt is the collision duration between the two interacting vehicles; subscript 1 represents the rear vehicle in a rear-end collision; subscript 2 represents the front vehicle in a rear-end collision; d 12 is the shortest distance between vehicle 1 and vehicle 2; Δd is d 12 →0 is the distance that the two vehicles continue to move forward after the collision; Δt′ is the rear-end collision, ε R is the road environment dielectric factor: ε R =λ6+λ7U dt +λ8U cu , U dt is the lighting condition; U cu is the road line shape; λ6, λ7, λ8 are the parameters to be estimated.

5. The method according to claim 4, characterized in that The method for constructing the driver-vehicle combined feature includes: The driver's reaction time and the vehicle's maximum available deceleration are used as the driver's characteristic parameters β r =λ0+λ1U s and vehicle characteristic parameter β v =λ2+λ3U c +λ4U hv +λ5U eb ; The driver characteristic parameters and the vehicle characteristic parameters are combined to establish a driver-vehicle combination characteristic model: Among them, T r represents the reaction time, b max represents the maximum available deceleration, η is a constant; β r represents the driver characteristic parameter, U s Indicates whether the vehicle is speeding, λ0 and λ1 are parameters to be estimated; β v Represents vehicle characteristic parameters, U c Indicates a car; U hv Indicates heavy vehicles, U eb represents an electric bicycle, and λ3, λ4, λ5, and λ6 are parameters to be estimated.

6. The method according to claim 1, characterized in that The method of predicting a vehicle accident at the intersection according to the vehicle risk status model comprises: The collision risk force F coll,21 As a conflict indicator, F coll =1 is used as the dividing point between rear-end collision and rear-end collision accident. Based on the generalized Pareto distribution, the probability model of rear-end collision accident at intersection is established. The overall distribution function of collision risk force is: Z(f)=Z u (y)[1-Z(u)]+Z(u)f≥u Where Z(f) is the collision risk force F coll,21 The value of the probability distribution function on f; Z u (y) is the conditional distribution function of the part of the collision risk force exceeding the threshold u; σ is the scale parameter; ξ is the shape parameter; F coll = 1 as the dividing point between rear-end collision and rear-end collision accident, and the probability of rear-end collision accident is established as Among them, N u It represents the number of collision risk forces greater than the threshold u after the collision risk threshold u is determined; N represents the total number of samples; According to the acquired trajectory data, the collision risk force x of the rear-end collision interaction vehicles at the intersection is extracted i , set the empirical mean excess function to Among them, u is the threshold, k is the collision risk force sample {x1,x2,...,x n }, the amount of collision risk force sample data greater than the threshold u; The threshold value u is taken as the horizontal axis, and the average excess amount e is taken as the horizontal axis. k (u) is the vertical axis of the coordinate, and the collision risk data are arranged in descending order from large to small x1≥x2≥…≥x n , with k as the horizontal axis, (γ k,n ) -1 As the vertical axis, the statistic of the tail index is: Calculate the average excess e k (u) The point where the linear change is obvious after exceeding the threshold and satisfies the tail index γ k,n The horizontal coordinate corresponding to the point in the stable area is taken as the collision risk force threshold u; According to the collision risk threshold u, the maximum likelihood estimation function L(y:σ) is solved to obtain the scale parameter σ and shape parameter ξ: According to the probability of rear-end collision Ω, the probability of rear-end collision at the intersection within the characteristic time T is predicted. Among them, A(T) is the expected number of rear-end collision accidents within time T; time is the actual observation time of the intersection.

7. The method according to claim 1, characterized in that The method for obtaining intersection vehicle rear-end collision accident data comprises: Select the accident data of signalized intersections within a preset time and space range to establish a historical accident database, and classify them according to the accident information master table, accident information personnel table, and accident information dictionary; The classified accident information dictionary is divided into accident data information and specific accident content, and the accident data is cleaned according to the specific accident content to eliminate accident data occurring at non-signalized intersections.

8. The method according to claim 7, characterized in that The method further comprises: Based on the historical accident database, traffic accident collision types are divided into four types: frontal collision, side collision, rear-end collision, and other collisions; Chi-square test was performed on each variable to determine the road environment that significantly affects the rear-end collision at the intersection. The road alignment, pavement structure, visibility, lighting conditions, and occurrence time were found to have significant differences with the accident type (p<0.05). The independent variables were tested for collinearity and the variance inflation factor was calculated. Variables with VIF values ​​greater than 10 were eliminated; Select Y=1, i.e., the frontal collision type, as the reference item, and use the four influencing factors of road alignment, visibility, lighting conditions, and occurrence time after elimination as independent variables to establish a multivariate Logit regression model for rear-end collision accidents: Where h = 2, 3, 4; g 0h is the constant term of the h-th regression equation; G H is the Hth independent variable; g Hh is the Hth independent variable of the hth collision type.

9. The method according to claim 8, characterized in that The method further comprises: Calibrate the various parameters to be estimated for rear-end collision risk: The parameters to be estimated are solved according to the data of the historical accident database to obtain a solution set: x i ={η,λ0,λ1,...,λ8} Solving the solution set to obtain the optimal solution set when the collision risk force is close to 1 is used to calibrate the various parameters to be estimated for the rear-end collision risk force: x best ={η,λ0,λ1,...,λ8},F coll,21 →1.

10. A system for predicting the frequency of rear-end collision accidents at intersections based on collision risk force, characterized in that: The system comprises: An acquisition module, used to acquire vehicle status data at the intersection; A modeling module, used for constructing a vehicle risk status model of the intersection according to the vehicle status data; A prediction module, used to collect the intersection road trajectory data and predict vehicle accidents at the intersection according to the vehicle risk status model; Wherein, the method for constructing the vehicle risk status model of the intersection includes: a road user generates a repulsive collision risk force on another road user based on the vehicle status data; Based on the collision risk force, the road traffic risk of the intersection is evaluated, and a collision risk force model for representing rear-end collision is constructed according to the combined characteristics of the driver and the vehicle: Among them, F coll,21 is the collision risk force, F coll,21 The maximum value max(F coll,21 )=1,d 12 =0;Q i , i∈[1,2] is the combined driver-vehicle characteristics of the i-th road user; v r,12 is the relative speed of the two interacting vehicles, vehicle 1 and vehicle 2; Δt is the collision duration between the two interacting vehicles; subscript 1 represents the rear vehicle in a rear-end collision; subscript 2 represents the front vehicle in a rear-end collision; d 12 is the shortest distance between vehicle 1 and vehicle 2; Δd is d 12 →0 is the distance that the two vehicles continue to move forward after the collision; Δt′ is the rear-end collision, ε R is the road environment dielectric factor: ε R =λ6+λ7U dt +λ8U cu , U dt is the lighting condition; U cu is the road line shape; λ6, λ7, λ8 are the parameters to be estimated.