A safety evaluation method and system for automatic driving of a highway vehicle

Abstract

The application discloses a kind of expressway vehicle automatic driving safety evaluation method and system, belong to highway safety technical field, and the system includes: automatic driving behavior model layer, data layer, evaluation layer, display layer;Wherein simulation model layer includes the road three-dimensional model based on BIM technology, vehicle simulation model based on dynamics and automatic driving local tracking algorithm model, database layer includes road parameter and vehicle dynamics response data, evaluation layer includes the accident probability evaluation index based on fault tree model and the accident severity evaluation index based on analytic hierarchy process, display layer includes road three-dimensional model display, vehicle driving animation, dangerous section identification module.Based on automatic driving simulation model, the vehicle driving data under multiple conditions is obtained, the risk index of each section is calculated by safety evaluation model, the expressway safety is evaluated in design stage, to improve expressway linear optimization pertinently.

Description

TECHNICAL FIELD

[0001] The present application relates to a kind of expressway vehicle automatic driving safety evaluation method, further relates to the system for realizing expressway vehicle automatic driving safety evaluation method, belong to highway safety technical field. BACKGROUND

[0002] With the progress and popularization and application of automatic driving technology and vehicle networking etc., the driving behavior of vehicle on expressway has been obviously different compared with traditional manned driving. The current situation of automatic driving technology also puts forward new requirements for expressway design and construction. Expressway construction is primarily for safety, to ensure that vehicles equipped with automatic driving technology can drive safely on expressway, while ensuring engineering economy and ride comfort, it is necessary to give an expressway safety evaluation system that adapts to automatic driving technology, to better quantify the evaluation of the risk along the line and give guidance for the design and construction of expressway.

[0003] In the prior art, based on the scenario of driver driving the vehicle, a pre-safety evaluation system for expressway in design stage is designed, considering the factors such as vehicle operating speed, accident distribution and severity on adjacent sections of expressway, a database is constructed based on linear design, traffic characteristics and speed characteristics, and the safety of expressway is evaluated based on data information. The existing expressway portrait evaluation and safety risk research and judgment system based on multi-source data considers the multi-source fusion processing of traffic flow data, road indicators and accident data, and the output display of expressway risk hidden danger is obtained through the risk research and judgment processing system. And the expressway traffic operation safety evaluation method and system, by comprehensively analyzing virtual driving data, real vehicle test data, road traffic safety influencing factors and accident causation data, the safety hidden danger and risk section of expressway are obtained, and the traffic safety level of expressway is evaluated and improvement suggestions are put forward.

[0004] In summary, some existing expressway traffic safety evaluation methods and systems focus on traditional manned driving scenarios, consider the operating speed difference caused by driver physiological factors, do not consider the influence of laser radar and line control system equipped on automatic driving vehicle on vehicle driving behavior, and cannot reflect the vehicle driving behavior after applying automatic driving technology; The analysis of safety is mostly based on past traffic accident data, which does not have predictability and portability; The dynamics response under the condition of multi-degree of freedom of vehicle body is not considered. SUMMARY

[0005] The present application aims to solve the problems of strong experience dependence and not considering automatic driving characteristics in existing safety evaluation methods, and proposes an expressway vehicle automatic driving safety evaluation method and system;

[0006] To achieve the above functions, the application designs a safety evaluation method for automatic driving of expressway vehicles, for the expressway to be evaluated and each vehicle on the expressway to be evaluated, the vehicle is an unmanned vehicle for automatic driving, and the following steps S1-S4 are executed to complete the safety evaluation of automatic driving of the vehicle on the expressway to be evaluated, and improve the driving safety of the expressway in the operation stage after completion:

[0007] Step S1: collecting design data of the expressway to be evaluated, including traffic flow parameters, road geometric design parameters, and three-dimensional terrain data; wherein the traffic flow parameters include traffic volume, design speed, and density of the expressway to be evaluated; the road geometric design parameters include radius, super-elevation, widening, longitudinal slope speed, and synthetic slope in the linear design of the expressway to be evaluated; and the three-dimensional terrain data includes three-dimensional coordinate data of point clouds within a preset range of the linear design of the expressway to be evaluated;

[0008] Step S2: constructing an automatic driving vehicle-road coupling simulation model and an expressway three-dimensional BIM model, inputting the design data of the expressway to be evaluated into the automatic driving vehicle-road coupling simulation model and the expressway three-dimensional BIM model, and the automatic driving vehicle-road coupling simulation model includes a vehicle multi-degree-of-freedom dynamics simulation model, a road model, and an automatic driving tracking algorithm;

[0009] The vehicle multi-degree-of-freedom dynamics simulation model is used to output the dynamic response of the vehicle when driving on the expressway to be evaluated according to the vehicle characteristics, including tire force, vehicle body acceleration, and angular velocity;

[0010] The road model is used to reflect the geometric shape parameters, auxiliary facility conditions, and road surface friction coefficient of the expressway to be evaluated according to the design data of the expressway to be evaluated;

[0011] The automatic driving vehicle-road coupling simulation model is used to make the vehicle drive on the expressway to be evaluated along a preset route according to the vehicle characteristics;

[0012] The expressway three-dimensional BIM model contains three-dimensional position information of the ground within a preset range around the expressway to be evaluated, and reflects the driving sight distance condition;

[0013] Step S3: defining each vehicle accident type and calculating the occurrence probability of each vehicle accident type, establishing an expressway vehicle driving safety evaluation model according to the occurrence probability of each vehicle accident type, inputting each parameter output by each model in step S2 into the expressway vehicle driving safety evaluation model, calculating the along-line driving risk index of the expressway to be evaluated under the automatic driving environment, and completing the safety evaluation of automatic driving of the vehicle on the expressway to be evaluated;

[0014] Step S4: According to the along-line driving risk index, give the design improvement suggestions of the expressway in the design stage, and improve the driving safety of the expressway in the operation stage after being built.

[0015] As a preferred technical solution of the present application: step S31: divide the expressway to be evaluated into n segments according to a preset interval, and respectively for each segment of the divided expressway, perform the safety evaluation of the subsequent steps:

[0016]

[0017] In the formula, L is the full length of the expressway to be evaluated, the unit is meter, l i is the i-th segment of the divided expressway;

[0018] Define each vehicle accident type, and calculate the occurrence probability of each vehicle accident type, and further calculate the comprehensive accident occurrence probability;

[0019] Step S32: According to the three-dimensional BIM model of the expressway, quantize the environmental factors of the expressway to be evaluated, and obtain the accident severity index of the expressway to be evaluated;

[0020] Step S33: According to the comprehensive accident occurrence probability obtained in step S31 and the accident severity index obtained in step S32, the along-line driving risk index is calculated as follows:

[0021] r i =max(P i )·C i

[0022] In the formula, r i is the driving risk index of the i-th segment of the expressway, P i is the comprehensive accident occurrence probability of the i-th segment of the expressway, and C i is the accident severity index of the i-th segment of the expressway.

[0023] As a preferred technical solution of the present application: step S31 includes steps S311-S313 as follows:

[0024] Step S311: According to the vehicle characteristics, based on the vehicle multi-degree-of-freedom dynamics simulation model, obtain the dynamic response of the vehicle driving corresponding to each index of the single-vehicle accident occurrence for each segment of the expressway, wherein the dynamic response of the vehicle driving includes the vertical force of the tire, the lateral acceleration, and the driving speed. The indexes for evaluating single-vehicle accident occurrence include lateral load transfer rate LTR, lateral acceleration a y , stopping sight distance S t ; define single-vehicle accident types, including rollover accident type, sideslip accident type, and insufficient sight distance type;

[0025] The judgment method of the rollover accident type is as follows:

[0026] The lateral load transfer ratio LTR is calculated as follows:

[0027]

[0028] In the formula, F l1 , F l2 , F r1 , and F r2 are the vertical forces of the left front tire, the left rear tire, the right front tire, and the right rear tire when the vehicle is running, respectively, and the unit is N;

[0029] When 0≤LTR≤0.6, it is judged that the vehicle is not in danger of rollover accident; when 0.6<LTR≤1, it is judged that the vehicle is in danger of rollover accident; and when LTR=1, it is judged that the vehicle is in a critical state of rollover accident;

[0030] The judgment method of the side slip accident type is as follows:

[0031] According to the automatic driving vehicle road coupling simulation model, the lateral acceleration a y is obtained, and the unit is m / s 2 ; when 0≤a y ≤0.4g, it is judged that the vehicle is not in danger of side slip accident; when 0.4g<a y ≤0.8g, it is judged that the vehicle is in danger of side slip accident; and when a y >0.8g, it is judged that the vehicle is in a critical state of side slip accident, wherein g is the gravity acceleration, and the value is 9.81 m / s 2 ;

[0032] The judgment method of the insufficient sight distance type is as follows:

[0033]

[0034] In the formula, S t is the sight distance required for the vehicle to stop, v is the current driving speed of the vehicle, t is the time required for the vehicle to send a control-by-wire system command from perception, and f is the friction coefficient of the road; if the sight distance required for the vehicle to stop is greater than the sight distance that can be provided by the expressway to be evaluated, it is judged that the vehicle is in danger of insufficient sight distance, otherwise it is judged that the vehicle is not in danger of insufficient sight distance;

[0035] Step S312: According to the dynamic response of the vehicle running, the occurrence probability of each single vehicle accident type is calculated as follows:

[0036] The probability p1 of rollover accident is as follows:

[0037]

[0038] The probability p2 of the occurrence of the side slip accident is as follows:

[0039]

[0040] The probability p3 of the occurrence of the longitudinal sight distance deficiency is as follows:

[0041]

[0042] S is the sight distance provided by the expressway to be evaluated determined by the three-dimensional BIM model of the expressway;

[0043] Step S313: According to the occurrence probability of each single vehicle accident type, the comprehensive accident occurrence probability of the single vehicle accident is calculated based on the fault tree model. The fault tree model divides each single vehicle accident type into a lateral single vehicle accident and a longitudinal single vehicle accident. The lateral single vehicle accident includes a side slip accident type and a side slip accident type. The longitudinal single vehicle accident is a sight distance deficiency type. The comprehensive accident occurrence probability of the single vehicle accident is calculated as follows:

[0044]

[0045] In the formula, P is the comprehensive accident occurrence probability of the single vehicle accident, p j is the occurrence probability of the single vehicle accident type j.

[0046] As a preferred technical solution of the present application: step S32 includes the following steps S321-S325:

[0047] Step S321: Enumerate the environmental factors of the expressway to be evaluated, including three one-level indexes of roadside safety, road surface condition and design speed. The roadside safety includes two two-level indexes of slope gradient and guardrail setting.

[0048] Step S322: According to the three-dimensional BIM model of the expressway, the corresponding value of each index in step S321 is determined. According to the corresponding value of each index, the influence mode and degree of the single vehicle accident are classified. The classified levels include weakening, neutral and aggravation. Weakening means that the index weakens the severity of the single vehicle accident at this value, otherwise it is aggravation. Neutral means that the index has no effect on the severity of the single vehicle accident.

[0049] Step S323: According to the level of each index, the accident severity index of the expressway to be evaluated is calculated as follows:

[0050]

[0051] In the formula, c is the severity index, and l is the level of each index.

[0052] Step S324: according to the actual situation, the environmental factors of the to-be-evaluated expressway are selected, the analytic hierarchy process is used to determine the weight of each index, and the weight of each index is listed from the horizontal single-vehicle accident and the longitudinal single-vehicle accident;

[0053] Step S325: according to the accident severity index and the weight of each index, the accident severity index along the line is calculated as follows:

[0054]

[0055] In the formula, C j is the accident severity index in the j direction, m is the number of accident severity indexes participating in calculation, c n is the accident severity index, w n is the weight corresponding to the accident severity index c n .

[0056] The application also designs a safety evaluation system for expressway vehicle automatic driving, which comprises a model layer, a data layer, an evaluation layer and a display layer, so as to realize the safety evaluation method for expressway vehicle automatic driving.

[0057] The model layer is used for storing an automatic driving vehicle-road coupling simulation model and an expressway three-dimensional BIM model, wherein the automatic driving vehicle-road coupling simulation model comprises a vehicle multi-degree-of-freedom dynamics simulation model, a road model and an automatic driving tracking algorithm.

[0058] The data layer is used for storing road model data information, highway traffic design parameters and vehicle dynamics response time sequence data, wherein the road model data information comprises linear parameters and environmental parameters; the traffic parameters comprise traffic volume, design speed and traffic composition parameters; and the vehicle dynamics response time sequence data is time sequence data of each index obtained by automatic driving vehicle-road coupling simulation.

[0059] The evaluation layer is used for processing and calculating the data obtained from the data layer, and finally obtaining the along-line driving risk index of the to-be-evaluated expressway, which comprises two parts of accident occurrence probability and accident severity.

[0060] The display layer is used for displaying the to-be-evaluated expressway three-dimensional model, vehicle driving simulation animation and dangerous section identification for highway designers and managers, wherein the dangerous section identification needs to obtain the along-line driving risk index from the evaluation layer and display in the form of a key color block or a cloud chart in the road three-dimensional model in a visualized manner.

[0061] Advantages: compared with the prior art, the application has the following advantages:

[0062] This invention constructs an autonomous driving vehicle-road coupling simulation model based on vehicle dynamics simulation software and autonomous driving local path tracking algorithm, which can more accurately simulate vehicle behavior in autonomous driving environment and build an accurate simulation environment for highway safety evaluation in autonomous driving environment.

[0063] This invention considers various factors, including highway alignment parameters, road environment, and vehicle multi-degree-of-freedom dynamics, from the perspectives of lateral and longitudinal stability. It proposes evaluation indicators that can truly reflect the driving safety of autonomous vehicles. Based on the fault tree model and the analytic hierarchy process, a corresponding comprehensive evaluation model is established, making the evaluation system more comprehensive and objective, and avoiding the limitations of subjectivity and poor quantification in traditional highway safety evaluation.

[0064] The safety evaluation system of the highway vehicle driving safety evaluation method proposed in this invention, which is suitable for autonomous driving environments, consists of four independent and interactive parts: model layer, data layer, evaluation layer, and display layer. It can realize functions such as road model storage, traffic data integration, and risk section prediction after highway design is completed and before construction, which can meet the requirements of my country's highway informatization construction. Attached Figure Description

[0065] Figure 1 This is a flowchart of a safety evaluation method for autonomous driving of highway vehicles provided according to an embodiment of the present invention;

[0066] Figure 2 This is a schematic diagram of an accident tree model for calculating the overall accident probability of a single-vehicle accident, provided according to an embodiment of the present invention.

[0067] Figure 3 This is a schematic diagram of a safety evaluation method for autonomous driving of highway vehicles provided in an embodiment of the present invention. Detailed Implementation

[0068] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0069] Reference Figure 1 This invention provides a safety evaluation method for autonomous driving of vehicles on highways. The method targets the highway to be evaluated and the vehicles on it, wherein the vehicles are autonomous, unmanned vehicles. Steps S1-S4 are executed to complete the safety evaluation of autonomous driving on the highway to be evaluated, thereby improving the driving safety of the highway during its operational phase after construction.

[0070] Step S1: collect the design data of the expressway to be evaluated, including traffic flow parameters, road geometric design parameters, and three-dimensional terrain data; wherein the traffic flow parameters include traffic volume, design speed, and density of the expressway to be evaluated; the road geometric design parameters include radius, superelevation, widening, longitudinal slope speed, and synthetic slope in the linear design of the expressway to be evaluated; and the three-dimensional terrain data includes three-dimensional coordinate data of point clouds in a preset range of the linear design of the expressway to be evaluated;

[0071] Autodesk can provide complete road design modeling software, which can be used to establish a three-dimensional terrain map around a highway in the design stage, to conduct parameterized design of highway linear parameters in combination with local specifications, and to jointly use InfraWorks, Dynamo and other software to visually display three-dimensional models, to store and transmit various design data of the expressway described in step S1, and to provide a basis for the display layer in the safety evaluation system for automatic driving of vehicles on the expressway proposed in the present application.

[0072] According to the linear characteristics of the actual expressway, the following road model is designed in this embodiment, the design speed is 120 km / h, the road section is a cutting foundation, the radius is 600 m, and the linear parameters are as shown in Table 1:

[0073] Table 1

[0074]

[0075] The road superelevation calculation table of this section is as shown in Table 2:

[0076] Table 2

[0077]

[0078] Step S2: build an automatic driving vehicle-road coupling simulation model and an expressway three-dimensional BIM model, input the design data of the expressway to be evaluated into the automatic driving vehicle-road coupling simulation model and the expressway three-dimensional BIM model, and the automatic driving vehicle-road coupling simulation model includes a vehicle multi-degree-of-freedom dynamics simulation model, a road model, and an automatic driving tracking algorithm.

[0079] The vehicle multi-degree-of-freedom dynamics simulation model is used to output the dynamic response of the vehicle when driving on the expressway to be evaluated according to the vehicle characteristics, including tire force, vehicle body acceleration, and angular velocity.

[0080] The road model is used to reflect the geometric shape parameters, auxiliary facility conditions, and road surface friction coefficient of the expressway to be evaluated according to the design data of the expressway to be evaluated, and to provide road working conditions for the simulation of automatic driving vehicle-road coupling.

[0081] The automatic driving car road coupling simulation model is used to make the vehicle travel on the preset route on the expressway to be evaluated according to the vehicle characteristics, and can show better accuracy compared with the traditional human driver;

[0082] The expressway three-dimensional BIM model contains the three-dimensional position information of the ground within the preset range of the expressway to be evaluated and its periphery, and reflects the driving sight distance condition;

[0083] Carsim is a simulation software for vehicle dynamics, which can simulate the driving stability, braking performance and power performance of the vehicle under different driving controls and road conditions. Different linear roads can be established in the software, and the driving behavior of Carsim can be controlled by writing MPC automatic driving vehicle local tracking algorithm in Matlab / Simulink software to simulate the driving behavior of the automatic driving vehicle, so as to output the dynamic response of the automatic driving vehicle when driving on different road line shapes.

[0084] The expressway three-dimensional BIM model is used to establish an expressway model conforming to the actual engineering design. The model needs a script line shape parameter export function to maintain data consistency with the road module in Carsim. At the same time, the expressway three-dimensional BIM model needs to provide the sight distance S of each point along the expressway to be evaluated.

[0085] Step S3: defining each vehicle accident type and calculating the occurrence probability of each vehicle accident type, establishing an expressway vehicle driving safety evaluation model according to the occurrence probability of each vehicle accident type, inputting each parameter output by each model in step S2 into the expressway vehicle driving safety evaluation model, and calculating to obtain the along-line driving risk index of the expressway to be evaluated in the automatic driving environment, thereby completing the safety evaluation of the vehicle in the automatic driving on the expressway to be evaluated;

[0086] The specific method of step S3 is as follows:

[0087] Step S31: dividing the expressway to be evaluated into 31 segments according to the 30m interval and displaying the route special elements (the starting point of the flat curve, etc.) at the same time, and performing the safety evaluation of the subsequent steps for each divided expressway segment:

[0088]

[0089] In the formula, L is the total length of the expressway to be evaluated, the unit is meter, l i is the i-th divided expressway segment, and n is the total number of divided segments;

[0090] Defining each vehicle accident type and calculating the occurrence probability of each vehicle accident type, and further calculating the comprehensive accident occurrence probability;

[0091] Step S311: for each section of highway, respectively, according to the vehicle characteristics, based on the vehicle multi-degree-of-freedom dynamics simulation model, the dynamic response of the vehicle driving corresponding to each index evaluating single vehicle accident occurrence is obtained, wherein the dynamic response of the vehicle driving includes the vertical force of the tire, the lateral acceleration, the driving speed, and each index evaluating single vehicle accident occurrence includes the lateral load transfer ratio LTR, the lateral acceleration a y , the stopping sight distance S t ; define single vehicle accident types, including rollover accident type, sideslip accident type, and sight distance deficiency type;

[0092] The judgment method of the rollover accident type is as follows:

[0093] The lateral load transfer ratio LTR is calculated as follows:

[0094]

[0095] In the formula, F l1 , F l2 , F r1 , and F r2 are the vertical forces of the left front tire, the left rear tire, the right front tire, and the right rear tire respectively when the vehicle is driving, and the unit is N;

[0096] When 0≤LTR≤0.6, it is judged that the vehicle is not in danger of rollover accident, when 0.6<LTR≤1, it is judged that the vehicle is in danger of rollover accident, and when LTR=1, it is judged that the vehicle is in a critical state of rollover accident;

[0097] The judgment method of the sideslip accident type is as follows:

[0098] According to the automatic driving vehicle-road coupling simulation model, the lateral acceleration a y is obtained, and the unit is m / s 2 ; when 0≤a y ≤0.4g, it is judged that the vehicle is not in danger of sideslip accident, when 0.4g<a y ≤0.8g, it is judged that the vehicle is in danger of sideslip accident, and when a y >0.8g, it is judged that the vehicle is in a critical state of sideslip accident, wherein g is the gravitational acceleration, and the value is 9.81 m / s 2 ;

[0099] The sight distance deficiency type is determined by the maximum deceleration of the vehicle and the time required for the vehicle to start identifying the driving condition to the need for deceleration until the control chassis executes the deceleration operation, and considering the vehicle driving characteristics under normal conditions, the judgment method of this accident type is as follows:

[0100]

[0101] In the formula, S t is the required sight distance for vehicle stopping, v is the current driving speed of the vehicle, considering that the vehicle operating speed under the control of the autonomous driving algorithm is relatively small compared to the design speed, the design speed can be used to replace it in actual engineering; t is the time required for the vehicle to perceive and control the command line system to issue instructions, with a unit of s, in the embodiment, t takes 0.5s to meet the requirements of the scene; f is the friction coefficient of the road; if the required sight distance for vehicle stopping is greater than the sight distance that the expressway to be evaluated can provide, it is judged that the vehicle has a risk of insufficient sight distance, otherwise it is judged that the vehicle does not have a risk of insufficient sight distance;

[0102] In the embodiment, the expressway design data in step S1 is imported into the automatic driving vehicle-road coupling simulation model in step S2, and after simulation, the time sequence data of the vertical force (F) of the tire and the lateral acceleration a y of the vehicle can be obtained as shown in Table 3:

[0103] Table 3

[0104]

[0105]

[0106] Step S312: According to the dynamic response of the vehicle driving, the occurrence probability of each single vehicle accident type is calculated as follows:

[0107] The probability p1 of rolling accident is as follows:

[0108]

[0109] The probability p2 of side slip accident is as follows:

[0110]

[0111] The probability p3 of longitudinal sight distance deficiency is as follows:

[0112]

[0113] Wherein, S is the sight distance that the expressway to be evaluated can provide determined by the three-dimensional BIM model of the expressway;

[0114] According to the calculation formula, the probability p1 of rolling accident and the probability p2 of side slip accident are as follows:

[0115] Table 4

[0116]

[0117] The sight distance that the planned road can provide obtained from the sight distance checking function of the three-dimensional BIM model of the expressway is as follows:

[0118] The required stopping sight distance calculation formula is as follows:

[0119]

[0120] The minimum stopping sight distance is set to 83.29m in Civil 3D, and the sight distance check table is as shown in Table 5:

[0121] Table 5

[0122]

[0123]

[0124] It can be seen that the road sight distance can ensure the safety requirement, and the probability p3 of longitudinal accident caused by insufficient sight distance is 0 according to the calculation formula.

[0125] Step S313: According to the occurrence probability of each single vehicle accident type, the comprehensive accident occurrence probability of single vehicle accident is calculated based on the fault tree model; refer to Figure 2 , the fault tree model divides each single vehicle accident type into lateral single vehicle accident and longitudinal single vehicle accident, the lateral single vehicle accident includes rollover accident type and sideslip accident type, and the longitudinal single vehicle accident is insufficient sight distance type; the comprehensive accident occurrence probability of single vehicle accident is calculated as follows:

[0126]

[0127] In the formula, P is the comprehensive accident occurrence probability of single vehicle accident, p j is the occurrence probability of single vehicle accident type j.

[0128] According to the calculation formula of S313, the comprehensive accident occurrence probability P along the line is as shown in Table 6:

[0129] Table 6

[0130]

[0131] Step S32: According to the three-dimensional BIM model of the expressway, the environmental factors of the expressway to be evaluated are quantified to obtain the accident severity index of the expressway to be evaluated;

[0132] Step S32 includes steps S321-S325 as follows:

[0133] Step S321: Enumerate the environmental factors of the expressway to be evaluated, including one-layer indexes of roadside safety, road surface condition and design speed, wherein the roadside safety includes two-layer indexes of slope gradient and guardrail setting;

[0134] Step S322: According to the expressway three-dimensional BIM model, the value corresponding to each index in step S321 is determined, and the influence mode and degree of the single-vehicle accident are divided into grades according to the value corresponding to each index. The divided grades include weakening, neutral, and aggravating. Weakening means that the index weakens the severity of the single-vehicle accident at the value, otherwise it is aggravating. Neutral means that the index has no effect on the severity of the single-vehicle accident at the value.

[0135] The value range of each index is shown in Table 7 below:

[0136] Table 7

[0137]

[0138] According to the expressway three-dimensional BIM model in S1 and the road surface design material, the registration of several indexes along the highway is as follows:

[0139] No guardrail is set throughout the journey, the coefficient of friction is μ = 0.85, and the design speed is v0 = 120 km / h.

[0140] The slope of the side slope is shown in Table 8 below:

[0141] Table 8

[0142]

[0143] Step S323: According to the grade of each index, the severity index of the accident on the expressway to be evaluated is calculated as follows:

[0144]

[0145] In the formula, c is the severity index, and l is the grade of each index.

[0146] The meaning of the formula is that when a facility can reduce the severity of an accident, for example, setting a guardrail can reduce the severity of a lateral accident, the grade of the index is "weakening", and the severity index coefficient is 0.9 to reduce the final calculated severity index.

[0147] Step S324: According to the actual situation, select the environmental factors of the expressway to be evaluated, and use the analytic hierarchy process to determine the weight of each index. Since the vehicle is affected by different factors when it occurs in a lateral accident and a longitudinal accident, the weights of each index are listed from the aspects of lateral single-vehicle accidents and longitudinal single-vehicle accidents.

[0148] The severity index weight of the lateral safety accident obtained by the analytic hierarchy process is shown in Table 9 below:

[0149] Table 9

[0150]

[0151] The weight of the severity index of the longitudinal safety accident after the accident is obtained by the analytic hierarchy process as shown in Table 10.

[0152] Table 10

[0153]

[0154] Step S325: According to the accident severity index and the weight of each index, the severity index of the accident along the line is calculated as follows:

[0155]

[0156] In the formula, C j is the severity index of the accident in the j direction, m is the number of accident severity indexes participating in the calculation, c n is the severity index of the accident, w n is the weight corresponding to the severity index c n of the accident.

[0157] The calculation results are shown in Table 11.

[0158] Table 11

[0159]

[0160]

[0161] Step S33: According to the comprehensive accident probability obtained in step S31 and the accident severity index obtained in step S32, the accident probability and the accident severity index should be two numbers that change along the line in actual engineering. In actual engineering, the expressway can be divided into sections according to certain length and route elements, and the accident probability and the accident severity index in the section are calculated. The product of the two is the driving risk index of the section, and the driving risk index along the line is obtained by calculating different sections as follows:

[0162] r i = max(P i )·C i

[0163] In the formula, r i is the driving risk index of the i-th section of the expressway, P i is the comprehensive accident probability of the i-th section of the expressway, and C i is the accident severity index of the i-th section of the expressway.

[0164] The accident severity index is shown in Table 12.

[0165] Table 12

[0166]

[0167] According to the driving risk index obtained by S3, it can be obtained that driving on the curve section is relatively dangerous, the driving risk index is near 0.4, and a guardrail can be considered to be arranged on the section to reduce the severity after an accident occurs or to reduce the super height to reduce the accident probability, so as to reduce the driving risk index along the line as a whole.

[0168] Step S4: According to the driving risk index along the line, design improvement suggestions of the expressway are given in the design stage, and the driving safety of the expressway in the operation stage after being built is improved.

[0169] The embodiment of the application also provides a safety evaluation system for automatic driving of vehicles on an expressway, referring to Figure 3 , comprising a model layer, a data layer, an evaluation layer and a display layer, to realize the safety evaluation method for automatic driving of vehicles on an expressway.

[0170] The model layer is used to store an automatic driving vehicle-road coupling simulation model and an expressway three-dimensional BIM model, wherein the automatic driving vehicle-road coupling simulation model comprises a vehicle multi-degree-of-freedom dynamics simulation model and an automatic driving tracking algorithm.

[0171] The data layer is used to store road model data information, highway traffic design parameters and vehicle dynamics response time sequence data, wherein the road model data information comprises linear parameters and environmental parameters; the traffic parameters comprise traffic volume, design speed and traffic composition parameters; and the vehicle dynamics response time sequence data is time sequence data of each index obtained by automatic driving vehicle-road coupling simulation.

[0172] The evaluation layer is used to process and calculate the data obtained from the data layer, and finally obtain the driving risk index along the line of the expressway to be evaluated, including two parts of accident probability and accident severity.

[0173] The display layer is used to display the three-dimensional model of the expressway to be evaluated, vehicle driving simulation animation and dangerous section identification for highway designers and managers, wherein the dangerous section identification needs to obtain the driving risk index along the line from the evaluation layer, and is displayed in the form of a key color block or a cloud chart in the three-dimensional model of the road in a visual manner.

[0174] The embodiments of the application are described in detail above in combination with the drawings, but the application is not limited to the above embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the purpose of the application.

Claims

1. A safety evaluation method for automated driving of vehicles on highways, characterized in that, For the highway to be evaluated and the vehicles on it (the vehicles being autonomous, driverless vehicles), the following steps S1-S4 are performed to complete the safety evaluation of the autonomous driving of the vehicles on the highway to be evaluated, thereby improving the driving safety of the highway during its operation after construction: Step S1: Collect the design data of the highway to be evaluated, including traffic flow parameters, road geometric design parameters, and three-dimensional terrain data; The traffic flow parameters include the traffic volume, design speed, and density of the highway to be evaluated; the road geometry design parameters include the radius, superelevation, widening, longitudinal slope speed, and composite gradient in the alignment design of the highway to be evaluated; and the three-dimensional terrain data includes the three-dimensional coordinate data of the point cloud within the preset range of the alignment of the highway to be evaluated. Step S2: Construct an autonomous driving vehicle-road coupling simulation model and a highway 3D BIM model. Input the highway design data to be evaluated into the autonomous driving vehicle-road coupling simulation model and the highway 3D BIM model respectively. The autonomous driving vehicle-road coupling simulation model includes a vehicle multi-degree-of-freedom dynamics simulation model, a road model, and an autonomous driving tracking algorithm. Among them, the vehicle multi-degree-of-freedom dynamics simulation model is used to output the dynamic response of the vehicle when it is driving on the highway to be evaluated, including tire force, vehicle acceleration, and angular velocity, based on the vehicle characteristics. The road model is used to reflect the geometric parameters, ancillary facilities conditions, and pavement friction coefficient of the highway to be evaluated, based on the design data of the highway to be evaluated. The autonomous driving vehicle-road coupling simulation model is used to enable vehicles to travel along a preset route on the highway to be evaluated based on vehicle characteristics. The 3D BIM model of the highway contains the ground 3D location information of the highway to be evaluated and its surrounding pre-defined range, reflecting the driving sight distance conditions; Step S3: Define each type of vehicle accident and calculate the probability of occurrence of each type of vehicle accident. Based on the probability of occurrence of each type of vehicle accident, establish a highway vehicle driving safety evaluation model. Input the parameters output by each model in step S2 into the highway vehicle driving safety evaluation model, calculate the driving risk index along the highway to be evaluated under the autonomous driving environment, and complete the safety evaluation of autonomous driving of vehicles on the highway to be evaluated. Step S4: Based on the driving risk index along the route, provide design improvement suggestions for the expressway during the design phase to enhance driving safety during the operation phase after the expressway is completed.

2. The safety evaluation method for automated driving of highway vehicles according to claim 1, characterized in that, The specific method for step S3 is as follows: Step S31: Divide the highway to be evaluated into n segments according to a preset interval, and conduct a safety evaluation for each segment in subsequent steps. In the formula, L is the total length of the highway to be evaluated, in meters. i This represents the i-th segment of the highway. Define the accident types for each vehicle and calculate the probability of occurrence for each type of accident, and further calculate the overall accident probability; Step S32: Based on the 3D BIM model of the expressway, quantify the environmental factors of the expressway to be evaluated and obtain the accident severity index of the expressway to be evaluated. Step S33: Based on the comprehensive accident occurrence probability obtained in Step S31 and the accident severity index obtained in Step S32, calculate the driving risk index along the route as follows: r i =max(P i )·C i In the formula, r i Let P be the driving risk index of the i-th highway segment. i Let C be the overall accident probability of the i-th segment of the highway. i Let be the accident severity index for the i-th segment of the highway.

3. The safety evaluation method for automated driving of highway vehicles according to claim 2, characterized in that, Step S31 includes the following steps S311 - S313: Step S311: For each section of highway, based on vehicle characteristics and a multi-degree-of-freedom dynamics simulation model, obtain the vehicle's dynamic response corresponding to each indicator for evaluating a single-vehicle accident. The vehicle's dynamic response includes the vertical force of the tires, lateral acceleration, and speed. The indicators for evaluating a single-vehicle accident include the lateral load transfer rate (LTR) and lateral acceleration a. y Parking sight distance S t Define single-vehicle accident types, including rollover accidents, skid accidents, and insufficient visibility accidents. The method for judging the rollover accident type is as follows: Calculate the lateral load transfer ratio LTR as follows: In the formula, F l1 F l2 F r1 F r2 These are the vertical forces on the four tires (left front, left rear, right front, and right rear) when the vehicle is in motion, expressed in N. When 0 ≤ LTR ≤ 0.6, it is judged that the vehicle has no risk of rollover accident; when 0.6 < LTR ≤ 1, it is judged that the vehicle has a risk of rollover accident; when LTR = 1, it is judged that the vehicle is in the critical state of rollover accident; The method for judging the sideslip accident type is as follows: The lateral acceleration a is obtained from the autonomous driving vehicle-road coupling simulation model. y The unit is m / s 2 When 0≤a y When the concentration is ≤0.4g, it is determined that the vehicle is not at risk of skidding. y When the concentration is ≤0.8g, the vehicle is considered to be at risk of skidding. y When the force is greater than 0.8g, the vehicle is considered to be in a critical state where a skidding accident is likely to occur, where g is the acceleration due to gravity, with a value of 9.81 m / s². 2 ;​ The method for judging the insufficient sight distance type is as follows: In the formula, S t v is the sight distance required for the vehicle to stop, t is the current speed of the vehicle, f is the time required for the vehicle to receive the control command from the drive-by-wire system, and f is the friction coefficient of the road. If the sight distance required for the vehicle to stop is greater than the sight distance that the highway being evaluated can provide, then it is judged that the vehicle is at risk of insufficient sight distance; otherwise, it is judged that the vehicle is not at risk of insufficient sight distance. Step S312: According to the dynamic response of the vehicle running, calculate the occurrence probability of each single-vehicle accident type as follows: The probability p1 of rollover accident occurs as follows: The probability p2 of sideslip accident occurs as follows: The probability p3 of insufficient longitudinal sight distance is as follows: Where, S is the sight distance provided by the to-be-evaluated highway determined by the highway three-dimensional BIM model; Step S313: According to the occurrence probability of each single-vehicle accident type, based on the fault tree model, calculate the comprehensive accident occurrence probability of single-vehicle accidents; the fault tree model classifies each single-vehicle accident type into horizontal single-vehicle accidents and vertical single-vehicle accidents. Horizontal single-vehicle accidents include rollover accident type and sideslip accident type, and vertical single-vehicle accident is the insufficient sight distance type; calculate the comprehensive accident occurrence probability of single-vehicle accidents as follows: In the formula, P is the overall probability of a single-vehicle accident, p j Let be the probability of occurrence of single-vehicle accident type j.

4. The safety evaluation method for automated driving of highway vehicles according to claim 2, characterized in that, Step S32 includes the following steps S321 - S325: Step S321: List the environmental factors of the to-be-evaluated highway, including three first-level indicators: roadside safety situation, pavement condition, and design speed. Among them, the roadside safety situation includes two second-level indicators: slope gradient and guardrail setting; Step S322: According to the highway three-dimensional BIM model, determine the corresponding values of each index described in Step S321, and divide the levels of the influence mode and degree of each index on single-vehicle accidents. The divided levels include weakening, neutral, and aggravating; weakening means that this index weakens the severity after a single-vehicle accident occurs at this value, and vice versa for aggravating, and neutral means that this index has no influence on the severity of single-vehicle accidents at this value; Step S323: According to the levels of each index, calculate the accident severity index for evaluating the to-be-evaluated highway as follows: In the formula, c is the severity index, and l is the level of each index; Step S324: Select the environmental factors of the to-be-evaluated highway according to the actual situation, and use the analytic hierarchy process to determine the weights of each index, and list the weights of each index from two aspects of horizontal single-vehicle accidents and vertical single-vehicle accidents; Step S325: According to the accident severity index and the weights of each index, calculate the accident severity index along the line as follows: In the formula, C j Let m be the accident severity index in the j-direction, m be the number of accident severity indices used in the calculation, and c be the... n w is an index representing the severity of the accident. n The severity index of the accident, c n The corresponding weights.

5. A safety evaluation system for automated driving of highway vehicles, characterized in that, It includes a model layer, a data layer, an evaluation layer, and a display layer to implement a safety evaluation method for highway vehicle autonomous driving as described in any one of claims 1 - 4; The model layer is used to store the autonomous driving vehicle-road coupling simulation model and the highway three-dimensional BIM model. The autonomous driving vehicle-road coupling simulation model includes a vehicle multi-degree-of-freedom dynamics simulation model, a road model, and an autonomous driving tracking algorithm; The data layer is used to store road model data information, highway traffic design parameters, and vehicle dynamics response time-series data. The road model data information includes alignment parameters and environmental parameters; the traffic parameters include traffic volume, design speed, and traffic composition parameters; and the vehicle dynamics response time-series data is the time-series data of various indicators obtained from autonomous driving vehicle-road coupling simulation. The evaluation layer is used to process and calculate the data obtained from the data layer to finally obtain the driving risk index along the highway to be evaluated, which includes two parts: the probability of accident occurrence and the severity of accident. The display layer is used to show highway designers and managers the 3D model of the highway to be evaluated, vehicle driving simulation animation, and dangerous road section signs. The dangerous road section signs need to obtain the driving risk index along the route from the evaluation layer and be displayed in a visual way in the form of key color blocks or cloud maps in the 3D road model.

Images