A method, device, equipment and storage medium for automatic verification of accident scenes

Abstract

The present application discloses an automatic verification method, device, equipment and storage medium for accident scenarios. The method comprises: S1, obtaining accident road test data; S2, automatically classifying and labeling accidents according to the accident road test data, and obtaining accident types and labeling results; S3, automatically generating accident simulation scenario configuration according to the labeling results; S4, verifying the accident simulation scenario configuration. The present application automatically classifies and labels accidents according to accident road test data, and obtains accident types and labeling results; then automatically generates accident simulation scenario configuration and verifies scenario configuration according to the labeling results. Compared with the existing autonomous driving accident simulation process, it is not only conducive to reducing labor costs, but also the quality of the simulation scenario does not need to be limited by the experience of the labeler.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to an accident scene automatic verification method, device, equipment and storage medium. Background Art

[0002] The existing autonomous driving accident simulation process generally includes steps such as safety officers reporting accident road test data, manually reviewing and classifying accident road test data, manually configuring simulation time, and manually evaluating results. It can be seen that in this process, not only a large amount of manpower costs are required, but the quality of the simulation scene will depend on the experience of the annotator. With the rapid expansion of the road test scale and the difficulty of unified training of annotators, the number and quality of scenes are restricted by the annotators. Summary of the invention

[0003] To this end, the technical problem solved by the embodiments of the present application is to provide an automatic verification method, device, equipment and storage medium for accident scenarios, which can automatically classify accident types, configure them into simulation scenarios, and verify simulation scenarios. This not only helps to reduce labor costs, but also the quality of the simulation scenarios does not need to be limited by the experience of the labelers.

[0004] In order to solve the above technical problems, the technical solutions adopted in this application are as follows:

[0005] On the one hand, an embodiment of the present application provides an accident scenario automatic verification method, the method comprising:

[0006] S1: Obtain accident road test data;

[0007] S2: Automatically classify and label the accident according to the accident road test data to obtain the accident type and labeling result;

[0008] S3: Automatically generate accident simulation scenario configuration according to the annotation results;

[0009] S4: Verify the accident simulation scenario configuration.

[0010] Further, the S2 includes:

[0011] Set up decision tree conditions corresponding to different accident types;

[0012] The decision tree conditions are used to perform conditional verification on the accident road test data to determine the type of accident.

[0013] Preferably, the S2 further includes:

[0014] Mark the time period when the accident occurred for the type of accident described;

[0015] Label the expected driving behavior of the host vehicle.

[0016] Further, the S3 includes:

[0017] Configure the start and end time of the simulation scenario according to the accident occurrence period;

[0018] Automatically configure the main vehicle behavior instructions in the simulation scenario;

[0019] An evaluation criterion is automatically configured according to the expected driving behavior.

[0020] Further, the S4 includes:

[0021] Run the simulation scenario configuration on the simulation platform to verify the scenario reproduction capability;

[0022] Verify whether the evaluation criteria accurately evaluate the behavior of the main vehicle;

[0023] The report is automatically generated, uploaded to the scenario dataset, and submitted for manual review.

[0024] Preferably, S1' is further included between S1 and S2, and S1' includes:

[0025] The accident road test data is compressed.

[0026] On the other hand, an embodiment of the present application provides an automatic verification device for an accident scenario, the device comprising:

[0027] A data acquisition module, used to acquire accident road test data;

[0028] A classification and labeling module, used to automatically classify and label accidents according to the accident road test data to obtain accident types and labeling results;

[0029] A scenario configuration generation module is used to automatically generate accident simulation scenario configurations based on the annotation results;

[0030] The scenario configuration verification module is used to verify the accident simulation scenario configuration.

[0031] Furthermore, the device also includes:

[0032] The data compression module is used to compress the accident road test data.

[0033] On the other hand, an embodiment of the present application provides a device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor implements the steps of any one of the above-mentioned accident scenario automatic verification methods when executing the computer program.

[0034] On the other hand, an embodiment of the present application provides a storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of any one of the above-mentioned accident scenario automatic verification methods are performed.

[0035] In summary, compared with the prior art, the technical solution provided in the embodiment of the present application has at least the following beneficial effects:

[0036] 1. The embodiment of the present application automatically classifies and labels accidents according to the accident road test data to obtain the accident type and labeling results; then automatically generates the accident simulation scene configuration and verifies the scene configuration according to the labeling results. Compared with the existing autonomous driving accident simulation process, it is not only conducive to reducing labor costs, but also the quality of the simulation scene does not need to be limited by the experience of the labeler.

[0037] 2. The embodiments of the present application set decision tree conditions corresponding to different accident types and match them with the accident road test data, so that the accident types can be accurately classified.

[0038] 3. The embodiment of the present application reduces the amount of accident drive test data by setting a step of compressing the accident drive test data, thereby greatly improving the analysis efficiency of the accident drive test data. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flowchart of the accident scenario automatic verification method provided by the first exemplary embodiment of the present application.

[0040] Figure 2 It is a flowchart of an automatic verification method for accident scenarios provided by the second exemplary embodiment of the present application.

[0041] Figure 3 It is a schematic diagram of the structure of an automatic verification device for accident scenarios provided by the third exemplary embodiment of the present application.

[0042] Figure 4 It is a schematic diagram of the structure of the device provided by the fourth exemplary embodiment of the present application. DETAILED DESCRIPTION

[0043] This specific embodiment is merely an explanation of the present application and is not a limitation of the present application. After reading this specification, those skilled in the art may make modifications to the present embodiment without any creative contribution as needed, but such modifications are protected by the patent law as long as they are within the scope of the claims of the present application.

[0044] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0045] The term "comprise" and any variations thereof in the specification and claims of the present application are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products or apparatuses.

[0046] In the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0047] The embodiments of the present application are further described in detail below in conjunction with the drawings in the specification.

[0048] Figure 1 The first exemplary embodiment of the accident scene automatic verification method of the present application shown in the figure comprises:

[0049] S1: Obtain accident road test data;

[0050] S2: Automatically classify and label the accident according to the accident road test data to obtain the accident type and labeling result;

[0051] S3: Automatically generate accident simulation scenario configuration according to the annotation results;

[0052] S4: Verify the accident simulation scenario configuration.

[0053] The embodiment of the present application automatically classifies and labels accidents according to the accident road test data to obtain the accident type and labeling results; then automatically generates the accident simulation scenario configuration and verifies the scenario configuration according to the labeling results. Compared with the existing autonomous driving accident simulation process, it is not only beneficial to reduce labor costs, but also the quality of the simulation scenario does not need to be limited by the experience of the labeler.

[0054] In order to reduce the amount of accident drive test data and improve the analysis efficiency of accident drive test data, a second exemplary embodiment of the accident scene automatic verification method of the present application is as follows: Figure 2 As shown, in Figure 1 Further improvements are made on the first exemplary embodiment shown, and the specific contents are as follows:

[0055] S1' is also included between S1 and S2, and S1' includes:

[0056] The accident road test data is compressed.

[0057] In order to accurately classify the accident type, the third exemplary embodiment of the accident scene automatic verification method of the present application is Figure 1 Further improvements are made on the first exemplary embodiment shown, and the specific contents are as follows:

[0058] The S2 includes:

[0059] Set up decision tree conditions corresponding to different accident types;

[0060] The decision tree conditions are used to perform conditional verification on the accident road test data to determine the type of accident.

[0061] The embodiments of the present application set decision tree conditions corresponding to different accident types and match them with the accident road test data, so that the accident types can be accurately classified.

[0062] Specifically, the S2 further includes:

[0063] Mark the time period when the accident occurred for the type of accident described;

[0064] Label the expected driving behavior of the host vehicle.

[0065] Specifically, S3 includes:

[0066] Configure the start and end time of the simulation scenario according to the accident occurrence period;

[0067] Automatically configure the main vehicle behavior instructions in the simulation scenario;

[0068] An evaluation criterion is automatically configured according to the expected driving behavior.

[0069] Specifically, the S4 includes:

[0070] Run the simulation scenario configuration on the simulation platform to verify the scenario reproduction capability;

[0071] Verify whether the evaluation criteria accurately evaluate the behavior of the main vehicle;

[0072] The report is automatically generated, uploaded to the scenario dataset, and submitted for manual review.

[0073] The following is a detailed description of the content of the automatic verification method of the accident scene of this application, taking the accident type of the main vehicle being stuck by an obstacle as an example, as follows:

[0074] S1: Acquire accident road test data, the S1 step includes the following:

[0075] Record the map information near the location of the main vehicle at each moment; the map information mainly includes semantic map elements near the location of the main vehicle. The semantic map elements include traffic cones, road blocking warning signs and other elements related to road repair areas.

[0076] Record obstacle information at each moment; the obstacles include static roadblocks and dynamic roadblocks. Static roadblocks include traffic cones, road blocking warning signs, and stationary vehicles in road repair areas; dynamic roadblocks include pedestrians, very slow-moving vehicles, etc. The obstacle information includes obstacle ID, obstacle location coordinates at each moment, and obstacle speed at each moment.

[0077] Record the status information of the vehicle planning and control module of the main vehicle at each moment. The status information includes whether the main vehicle is taken over, the main vehicle position coordinates, the main vehicle heading angle, and the main vehicle speed.

[0078] Among them, the map information, obstacle information, and status information of the vehicle planning and control module all belong to the category of the accident road test data.

[0079] S1': compressing accident road test data. In this example, S1' is mainly implemented by compressing map information, which mainly includes the following steps:

[0080] Record the semantic map elements near the location of the main vehicle;

[0081] Extracting obstacles related to the road repair area from the semantic map elements;

[0082] Adjacent obstacles related to the road repair area are clustered and arranged by using a union-find algorithm and then drawn into the road repair area.

[0083] Specifically, the method of clustering and arranging the adjacent obstacles related to the road repair area by using a union-find algorithm and then drawing the road repair area includes the following steps:

[0084] Extract all obstacles related to the road repair area to form an obstacle set Ω(o1, o2, ..., o i ,…,o N );

[0085] Traverse the obstacle set Ω(o1,o2,…,o i ,…,o N ) in the obstacle element oi , determine whether the road construction area set C (c1, c2, …, c j , …, c M ) is an empty set:

[0086] If so, create a road construction area c1 related to the obstacle element o i , and add the road construction area c1 to the road construction area set C (c1, c2, …, c j , …, c M );

[0087] If not, traverse the road construction areas in the road construction area set C (c1, c2, …, c j , …, c M ), and find the obstacle element o j in the road construction area c i that is closest to the obstacle element o k :

[0088] If the distance d i between the obstacle element o k and the obstacle element o ik is less than Max_Interval, then add the obstacle element o i to the road construction area c j . Thus, the purpose of compressing map information is achieved by merging road construction areas.

[0089] Furthermore, if the same obstacle element belongs to more than two road construction areas at the same time, then merge the more than two road construction areas into one, so as to further reduce the road construction areas in the road construction area set C (c1, c2, …, c j , …, c M ).

[0090] After traversing all the obstacle elements in this way, the remaining road construction areas in the road construction area set C (c1, c2, …, c j , …, c M ) are the final results.

[0091] In order to visually display the road construction area, the road construction area can be represented by a convex hull.

[0092] It should be noted that the Max_Interval is a constant. N, M, i, j, k are all positive integers.

[0093] S2: Automatically classify and label the accident according to the accident road test data to obtain the accident type and the labeling result. The S2 specifically includes:

[0094] S21: Automatically classify the accident according to the accident road test data to obtain the accident type, specifically:

[0095] Set decision tree conditions corresponding to different accident types;

[0096] Use the decision tree conditions to perform condition verification on the accident road test data to determine the accident type.

[0097] For the accident type where the host vehicle is stuck, there are 4 decision tree conditions, specifically:

[0098] The first condition: The host vehicle must be taken over by the safety officer;

[0099] The second condition: The host vehicle must be in a stationary or very slow state before being taken over;

[0100] The third condition: There must be a stationary or very slow moving vehicle, or a road construction area blocking the front of the host vehicle during the period when the host vehicle is stuck;

[0101] The fourth condition: The host vehicle must successfully bypass the obstacle blocking its front after being taken over.

[0102] Use the above decision tree conditions to verify the accident road test data. If the above 4 decision tree conditions are met simultaneously, it can be determined that the accident type is successfully matched as the host vehicle being stuck.

[0103] The specific verification method includes the following:

[0104] From the step of recording the status information of the vehicle planning and control module of the host vehicle at each moment, the takeover information of the host vehicle can be obtained from the status information of the vehicle planning and control module. Use the takeover information to match the first condition. If the host vehicle is determined to be taken over, the first condition is successfully matched. At the same time, record the takeover time t of the host vehicle t .

[0105] From the step of recording the status information of the vehicle planning and control module of the host vehicle at each moment, the speed information of the host vehicle can be obtained from the status information of the vehicle planning and control module. If the average speed of the host vehicle within C1 seconds before the takeover time t t is less than the preset minimum speed (C1 is a constant. In this example, after multiple repeated verifications by the inventor, the value of C1 is 3), that is, it meets the requirement of the formula speed[t t -3,t t .average()<Min_Speed, it is considered that the host vehicle is in a stationary or very slow state before being taken over. At this time, the second condition is successfully matched.

[0106] From the step of recording the obstacle information at each moment, the speed of the obstacle during the period when the host vehicle is stuck can be obtained from the obstacle information, and the p90 speed of the obstacle during the stuck period is statistically calculated. If the p90 speed is less than the preset minimum speed, that is, it meets the requirement of the formula speed[t s ,t t .p90()<Min_Speed, then it is determined that the obstacle is a stationary or very slow-moving vehicle. At this time, the third condition is successfully matched.

[0107] It should be noted that the p90 is to arrange the speeds of all time points of the obstacle during the period when the host vehicle is stuck in ascending order, and the speed at the 90% position is the speed that beats 90% of the speeds during the entire stuck period.

[0108] Alternatively, from the step of recording the status information of the vehicle planning and control module of the host vehicle at each moment, the position coordinates (x e ,y e ) and the heading angle θ of the host vehicle can be obtained from the status information of the vehicle planning and control module; from the step of recording the obstacle information at each moment, the position coordinates (x o ,y o ) of the obstacle can be obtained from the obstacle information. At this time, the direction vector of the host vehicle is V dir (cosθ,sinθ), then the position vector of the obstacle relative to the host vehicle is V pos (x o -x e ,y o -y e ). If then it is considered that the obstacle is in front of the host vehicle. On the premise that the obstacle is in front of the host vehicle, the calculation formula for the deviation distance between the obstacle and the host vehicle is offset(e,w)=(x e -x w ,y e -y w )×(cosθ,sinθ). When there are two obstacles in the road construction area c j , namely obstacle o r and obstacle o l , if the cross product of the deviation distance between obstacle o r and the host vehicle and the deviation distance between obstacle o l and the host vehicle is less than 0, that is, it meets the requirement of offset(o r .position,ego.position)·offset(o l .position,ego.position)<0, then it can be considered that obstacle or With obstacles l Located on the left and right sides of the main vehicle, namely the road repair area c j It is blocking the front of the main vehicle. At this time, the third condition is also matched successfully.

[0109] From the step of recording the status information of the vehicle planning and control module of the main vehicle at each moment, the position coordinates and the orientation angle of the main vehicle can be obtained from the status information of the vehicle planning and control module. If the position coordinates and the orientation angle of the main vehicle can enable the main vehicle to successfully bypass the obstacle, the fourth condition is successfully matched. At the same time, the time t when the main vehicle bypasses the obstacle is recorded. e .

[0110] The S2 further includes:

[0111] S22: Automatically mark the accident according to the accident road test data to obtain a marking result, which is specifically:

[0112] Mark the time period when the accident occurred for the type of accident described;

[0113] Label the expected driving behavior of the host vehicle.

[0114] The method for obtaining the accident occurrence time period is as follows:

[0115] After the first condition is matched successfully, the sliding window method window(t,i)=(t-3-0.5·i,t-0.5·i) is used to continuously search forward (where i is a positive integer) until the average speed of the main vehicle is found to be greater than the preset minimum speed, that is, it meets the requirements of the formula speed[window(t,i)].average()>Min_Speed, thereby obtaining the time period (t s ,t t ), where t s =t-3-0.5·i. Record the time t when the main vehicle starts to decelerate s .

[0116] The method for marking the expected driving behavior of the host vehicle is specifically: taking the driving behavior after the safety officer takes over the host vehicle as the correct driving method, and marking it as the expected driving behavior of the host vehicle.

[0117] The S3 includes:

[0118] S31: configuring the start and end time of the simulation scenario according to the accident occurrence period;

[0119] S32: Automatically configure the main vehicle behavior indication in the simulation scenario;

[0120] S33: Automatically configure evaluation criteria according to the expected driving behavior.

[0121] Wherein, the S31 specifically includes the following steps:

[0122] S311: Based on the recorded time t when the host vehicle starts to decelerate s , set the start time of the simulation scenario to T s Configured as t s -C2, where C2 is a constant. In this example, the value of C2 is 1 second, that is, T s =t s -1sec;

[0123] S312: Based on the recorded time t for the main vehicle to bypass the obstacle e , set the end time of the simulation scenario to T e Configured as t e +C2, where C2 is a constant. In this example, the value of C2 is 1 second, that is, T e =t e +1sec.

[0124] The S32 specifically includes the following steps:

[0125] S321: Configuring the behavior instructions of the host vehicle before the accident occurs;

[0126] S322: According to the expected driving behavior, the position coordinates and the orientation angle of the main vehicle when it goes around the obstacle are obtained, and the position coordinates and the orientation angle are added to the simulation scene as behavior instructions.

[0127] By configuring the behavior instructions of the main vehicle before the accident, the behavior consistency of the vehicle before the accident can be maintained in the simulation scene, providing the prerequisite for accident reproduction. And adding the main vehicle position coordinates and orientation angle when bypassing obstacles as behavior instructions in the simulation scene according to the expected driving behavior can greatly improve the main vehicle's ability to bypass obstacles, so that the main vehicle can accurately bypass obstacles according to the behavior instructions.

[0128] The S4 includes:

[0129] S41: Run the simulation scenario configuration on the simulation platform to verify the scenario reproduction capability;

[0130] S42: Verify whether the evaluation criteria accurately evaluate the behavior of the main vehicle;

[0131] S43: Automatically generate a report, upload it to the scene dataset, and submit it to manual review.

[0132] Wherein, the S41 specifically includes:

[0133] Utilize a large-scale simulation platform to regularly run newly generated simulation scenarios in batches, use the same version of the autonomous driving algorithm as when the accident occurred, and use algorithms such as trajectory similarity to detect whether the problems in the original accident will be reproduced in the simulation; if the scenario is reproduced, enter S42.

[0134] Figure 3 The third exemplary embodiment of the present application provides an automatic verification device for an accident scenario, which corresponds one-to-one to the verification method in the above embodiment, and the device includes:

[0135] A data acquisition module, used to acquire accident road test data;

[0136] A classification and labeling module, used to automatically classify and label accidents according to the accident road test data to obtain accident types and labeling results;

[0137] A scenario configuration generation module is used to automatically generate accident simulation scenario configurations based on the annotation results;

[0138] The scenario configuration verification module is used to verify the accident simulation scenario configuration.

[0139] Each module of the above-mentioned verification device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each module above.

[0140] Figure 4 It is a device provided by the fourth exemplary embodiment of the present application, which may be a server. The device includes a processor, a memory and a communication interface connected via a system bus. Among them, the processor of the device is used to provide computing and control capabilities. The memory of the device can be implemented by any type of volatile or non-volatile storage device or a combination thereof, and the volatile or non-volatile storage device includes but is not limited to: a disk, an optical disk, an EEPROM, an EPROM, a SRAM, a ROM, a magnetic storage, a flash memory, and a PROM. The memory of the device provides an environment for the operation of the operating system and computer programs stored therein. The communication interface of the device is a network interface, and the network interface is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the verification method steps described in the above embodiment are implemented.

[0141] In another embodiment of the present application, a storage medium is provided, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the loading method described in the above embodiment are implemented. The storage medium includes but is not limited to: ROM, RAM, CD-ROM, magnetic disk, and floppy disk.

[0142] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device described in the present application is divided into different functional units or modules to complete all or part of the functions described above.

[0143] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

Claims

1. An automatic verification method for accident scenarios, characterized in that: The method comprises: S1: Acquire accident road test data; the accident road test data includes map information, obstacle information, and status information of vehicle planning and control modules; compressing the accident road test data; The compression of the accident road test data is achieved by compressing map information, including: recording semantic map elements near the location of the main vehicle; extracting obstacles related to the road repair area from the semantic map elements; clustering and arranging adjacent obstacles related to the road repair area by using a union-find algorithm and then drawing the road repair area; S2: Automatically classify and label the accident according to the accident road test data to obtain the accident type and labeling result; S2 also includes: Mark the time period when the accident occurred for the type of accident described; Mark the expected driving behavior of the host vehicle; S3: Automatically generate accident simulation scenario configuration according to the annotation results; S3 includes: Configure the start and end time of the simulation scenario according to the accident occurrence period; Automatically configure the main vehicle behavior instructions in the simulation scenario; automatically configuring evaluation criteria according to the desired driving behavior; S4: Verify the accident simulation scenario configuration; S4 includes: Run the simulation scenario configuration on the simulation platform to verify the scenario reproduction capability; Verify whether the evaluation criteria accurately evaluate the behavior of the main vehicle; The report is automatically generated, uploaded to the scenario dataset, and submitted for manual review.

2. The automatic verification method for accident scenarios according to claim 1, characterized in that: The S2 includes: Set up decision tree conditions corresponding to different accident types; The decision tree conditions are used to perform conditional verification on the accident road test data to determine the type of accident.

3. An automatic verification device for accident scenarios, characterized in that: The device comprises: A data acquisition module is used to acquire accident road test data; the accident road test data includes map information, obstacle information, and status information of a vehicle planning and control module; and is specifically used to compress the accident road test data; the compression of the accident road test data is achieved by compressing map information, including: recording semantic map elements near the location of the main vehicle; extracting obstacles related to the road repair area from the semantic map elements; and using a union-find algorithm to cluster and sort the adjacent obstacles related to the road repair area and then draw the road repair area; A classification and labeling module is used to automatically classify and label accidents according to the accident road test data to obtain accident types and labeling results; specifically, it is used to label the accident occurrence time period of the accident type; and label the expected driving behavior of the main vehicle; A scenario configuration generation module is used to automatically generate an accident simulation scenario configuration according to the annotation results; specifically, to configure the start and end time of the simulation scenario according to the accident occurrence period; automatically configure the main vehicle behavior indication in the simulation scenario; and automatically configure the evaluation criteria according to the expected driving behavior; The scenario configuration verification module is used to verify the accident simulation scenario configuration; specifically, it is used to run the simulation scenario configuration on the simulation platform to verify the scenario reproduction capability; verify whether the evaluation criteria accurately evaluate the main vehicle behavior; automatically generate a report, upload it to the scenario data set, and submit it to manual review.

4. A device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor implements the steps of the accident scene automatic verification method as described in any one of claims 1 or 2 when executing the computer program.

5. A storage medium, characterized in that: The storage medium stores a computer program, which, when executed by a processor, implements the steps of the accident scenario automatic verification method as described in any one of claims 1 or 2.

Images