Fault test scenario generation method, device, equipment and storage medium

Abstract

The present disclosure relates to a fault test scene generation method, device, equipment and storage medium, and belongs to the technical field of automatic driving. The generation method comprises: obtaining fault information generated by testing a to-be-tested automatic driving scheme in a plurality of sample fault test scenes, the sample fault test scene comprising scene feature information, the scene feature information comprising at least one of a driving environment parameter, a preset fault parameter and a traffic participant behavior parameter; performing clustering processing on the sample fault test scene of the generated fault information to obtain a plurality of test scene sets, each test scene set comprising at least one sample fault test scene; selecting at least one sample fault test scene corresponding to the same test scene set as a candidate scene; and generating a target fault test scene according to the scene feature information corresponding to the candidate scene. The fault test scene generation method provided by the present disclosure can improve the generation efficiency of the fault test scene and reduce the development cost of the fault test scene.

Description

Technical Field

[0001] The present disclosure relates to the field of autonomous driving technology, and in particular to a method, apparatus, device, and storage medium for generating a fault test scenario. Background Art

[0002] With the rise of deep learning and computer vision technologies, autonomous driving offers a new solution for improving traffic safety and efficiency. Vehicles can use various perception systems to obtain information about the road, vehicles, pedestrians, and road conditions within the driving scenario. Based on this information, they determine the appropriate autonomous driving solution, enabling the vehicle to respond autonomously to various driving scenarios without the need for a driver. Therefore, to ensure the safety of vehicles equipped with autonomous driving solutions, these solutions must be tested.

[0003] Autonomous vehicle testing is a crucial step in the development of autonomous driving solutions. Related technologies involve testers reconstructing driving scenarios by writing various fault test scenarios based on big data from human driver traffic accidents and natural driving information. These scenarios are then used to test the autonomous driving solution. However, manual fault test scenario writing by testers is inefficient and labor-intensive, hindering the generation of a large number of fault test scenarios. Furthermore, to cope with the ever-changing driving scenarios, a large number of fault test scenarios are required to test the autonomous driving solution and cover as many scenarios as possible. However, the number of fault test scenarios written by testers based on experience is limited, making it difficult to fully test the autonomous driving solution. Summary of the Invention

[0004] To overcome the problems existing in the related art, the present disclosure provides a method, apparatus, device and storage medium for generating a fault test scenario.

[0005] According to a first aspect of an embodiment of the present disclosure, a method for generating a fault test scenario is provided, comprising:

[0006] Obtaining fault information generated by testing the tested autonomous driving solution in a plurality of sample fault test scenarios, where the sample fault test scenarios include scenario feature information, and the scenario feature information includes at least one of a driving environment parameter, a preset fault parameter, and a traffic participant behavior parameter;

[0007] Clustering the sample fault test scenarios that generate fault information to obtain multiple test scenario sets, each test scenario set including at least one sample fault test scenario;

[0008] selecting at least one sample fault test scenario corresponding to the same test scenario set as a candidate scenario;

[0009] Generate a target fault test scenario based on the scenario feature information corresponding to the selected scenario.

[0010] In one embodiment, the sample fault test scenario also includes initial scenario parameters; the sample fault test scenario that generates fault information is clustered to obtain multiple test scenario sets, including: clustering the sample fault test scenario that generates fault information according to scenario feature information to obtain multiple fault scenario clusters, and the fault scenario clusters correspond to multiple sample fault test scenarios; clustering the multiple fault scenario clusters according to the initial scenario parameters to obtain multiple test scenario sets.

[0011] In one embodiment, multiple fault scenario clusters are clustered according to initial scenario parameters to obtain multiple test scenario sets, including: fuzzy random processing is performed on the initial scenario parameters corresponding to the first fault scenario cluster to obtain an initial scenario parameter range, the first fault scenario cluster is any one fault scenario cluster among the multiple fault scenario clusters; a second fault scenario cluster that meets the initial scenario parameter range is determined among the multiple fault scenario clusters to obtain a test scenario set, the multiple fault scenario clusters include the second fault scenario cluster, and the test scenario set includes at least one second fault scenario cluster.

[0012] In one embodiment, the sample fault test scenario also includes test classification information; based on the scenario feature information, the sample fault test scenario that generates fault information is clustered to obtain multiple fault scenario clusters, including: clustering the first sample fault test scenario based on the driving environment parameters to obtain multiple fault scenario clusters, the first sample fault test scenario is a sample fault test scenario that generates fault information and the corresponding test classification information is an environmental parameter test; wherein, the test classification information corresponding to the sample fault test scenario obtained by adding, deleting, and modifying the driving environment parameters in the basic scenario is the environmental parameter test.

[0013] In one embodiment, the sample fault test scenario also includes test classification information, and the scenario characteristic information includes multiple preset fault parameters, and the multiple preset fault parameters include at least one of a perception fault parameter, a decision fault parameter, and a control fault parameter; based on the scenario characteristic information, the sample fault test scenario that generates fault information is clustered to obtain multiple fault scenario clusters, including: clustering the second sample fault test scenario based on the preset fault parameters to obtain multiple fault scenario clusters, the second sample fault test scenario is a sample fault test scenario that generates fault information and the corresponding test classification information is a trigger fault test; wherein, the test classification information corresponding to the sample fault test scenario obtained by adding, deleting, and modifying the preset fault parameters in the basic scenario is the trigger fault test.

[0014] In one embodiment, the sample fault test scenario also includes test classification information, and the scenario feature information includes initial scenario parameters and traffic participant behavior parameters; based on the scenario feature information, the sample fault test scenario that generates fault information is clustered to obtain multiple fault scenario clusters, including: clustering the third sample fault test scenario based on the initial scenario parameters to obtain multiple fault scenario clusters, the third sample fault test scenario is a sample fault test scenario that generates fault information and the corresponding test classification information is a participant behavior test; wherein, the test classification information corresponding to the sample fault test scenario obtained by adding, deleting, and modifying the traffic participant behavior parameters in the basic scenario is the participant behavior test.

[0015] In one embodiment, the sample fault test scenario also includes a basic scenario, which includes road environment information, driving condition information, and traffic signal information; the method for generating a fault test scenario also includes: receiving a test instruction; determining the road environment information, driving condition information, and traffic signal information according to the test instruction, and generating a basic scenario based on the selected road environment information, driving condition information, and traffic signal information; adding, modifying, or deleting the scenario feature information on the basis of the basic scenario to generate a plurality of the sample fault test scenarios.

[0016] According to a second aspect of an embodiment of the present disclosure, there is provided a device for generating a fault test scenario, comprising:

[0017] The first module is configured to obtain fault information generated by testing the autonomous driving solution under test in a plurality of sample fault test scenarios, where the sample fault test scenarios include scenario feature information, and the scenario feature information includes at least one of a driving environment parameter, a preset fault parameter, and a traffic participant behavior parameter;

[0018] The second module is used to cluster the sample fault test scenarios that generate fault information to obtain multiple test scenario sets, each test scenario set including at least one sample fault test scenario;

[0019] The third module is used to select at least one sample fault test scenario corresponding to the same test scenario set as a candidate scenario;

[0020] The fourth module is used to generate a target fault test scenario based on the scenario feature information corresponding to the selected scenario.

[0021] In one embodiment, the sample fault test scenario also includes initial scenario parameters; the second module is also used to: cluster the sample fault test scenarios that generate fault information based on scenario feature information to obtain multiple fault scenario clusters, and the fault scenario clusters correspond to multiple sample fault test scenarios; cluster the multiple fault scenario clusters based on the initial scenario parameters to obtain multiple test scenario sets.

[0022] In one embodiment, the second module is also used to: perform fuzzy random processing on the initial scenario parameters corresponding to the first fault scenario cluster to obtain the initial scenario parameter range, the first fault scenario cluster is any one fault scenario cluster among multiple fault scenario clusters; determine the second fault scenario cluster that meets the initial scenario parameter range among the multiple fault scenario clusters to obtain a test scenario set, the multiple fault scenario clusters include the second fault scenario cluster, and the test scenario set includes at least one second fault scenario cluster.

[0023] In one embodiment, the sample fault test scenario also includes test classification information; the second module is also used to: cluster the first sample fault test scenario according to the driving environment parameters to obtain multiple fault scenario clusters, the first sample fault test scenario is a sample fault test scenario that generates fault information and the corresponding test classification information is an environmental parameter test; wherein, the test classification information corresponding to the sample fault test scenario obtained by adding, deleting, and modifying the driving environment parameters in the basic scenario is the environmental parameter test.

[0024] In one embodiment, the sample fault test scenario also includes test classification information, and the scenario characteristic information includes multiple preset fault parameters, and the multiple preset fault parameters include at least one of perception fault parameters, decision fault parameters and control fault parameters; the second module is also used to: cluster the second sample fault test scenario according to the preset fault parameters to obtain multiple fault scenario clusters, and the second sample fault test scenario is a sample fault test scenario that generates fault information and the corresponding test classification information is a trigger fault test; wherein, the test classification information corresponding to the sample fault test scenario obtained by adding, deleting and modifying the preset fault parameters in the basic scenario is the trigger fault test.

[0025] In one embodiment, the sample fault test scenario also includes test classification information, and the scenario feature information includes initial scenario parameters and traffic participant behavior parameters; the second module is also used to: cluster the third sample fault test scenario according to the initial scenario parameters to obtain multiple fault scenario clusters, and the third sample fault test scenario is a sample fault test scenario that generates fault information and the corresponding test classification information is a participant behavior test; wherein, the test classification information corresponding to the sample fault test scenario obtained by adding, deleting, and modifying the traffic participant behavior parameters in the basic scenario is the participant behavior test.

[0026] In one embodiment, the sample fault test scenario also includes a basic scenario, which includes road environment information, driving condition information, and traffic signal information; the device also includes a fifth module, which is used to: receive test instructions; determine the road environment information, driving condition information, and traffic signal information according to the test instructions, and generate a basic scenario based on the selected road environment information, driving condition information, and traffic signal information; add, modify, or delete the scenario feature information on the basis of the basic scenario to generate multiple sample fault test scenarios.

[0027] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, comprising: a processor; a memory for storing processor-executable instructions; and a processor for reading executable instructions from the memory and executing the instructions to implement the method for generating a fault test scenario provided in the first aspect of the present disclosure.

[0028] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which computer program instructions are stored. When the program instructions are executed by a processor, the steps of the method for generating a fault test scenario provided in the first aspect of the present disclosure are implemented.

[0029] The technical solutions provided by the embodiments of the present disclosure include at least the following beneficial effects: The target fault test scenarios are generated by combining the scene feature information corresponding to the candidate scenarios, thereby automatically generating a large number of target fault test scenarios based on the original sample fault test scenarios, thereby enriching the types of fault test scenarios and achieving sufficient testing of the autonomous driving solution; improving the efficiency of fault test scenario generation and reducing the development cost of fault test scenarios. Through clustering processing, sample fault test scenarios with certain similarities are combined into a test scenario set, so that the subsequently selected candidate scenarios have a certain degree of similarity, avoiding or reducing the possibility that the generated target fault test scenarios cannot be used to test the autonomous driving solution under test due to large differences in the scene feature information corresponding to the candidate scenarios.

[0030] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0032] Figure 1 The figure is a flowchart of a method for generating a fault test scenario according to an exemplary embodiment.

[0033] Figure 2 This is one of the flowcharts of a method for generating a test scenario set according to an exemplary embodiment.

[0034] Figure 3 This is a second flowchart of a method for generating a test scenario set according to an exemplary embodiment.

[0035] Figure 4 The figure is a flowchart of a method for generating a sample fault test scenario according to an exemplary embodiment.

[0036] Figure 5 It is a block diagram of a device for generating a fault test scenario according to an exemplary embodiment.

[0037] Figure 6 It is a block diagram of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION

[0038] Exemplary embodiments will be described in detail below with reference to the accompanying drawings.

[0039] It should be pointed out that the relevant embodiments and drawings are only for describing exemplary embodiments provided by the present disclosure, rather than all embodiments of the present disclosure, and it should not be understood that the present disclosure is limited to the relevant exemplary embodiments.

[0040] It should be noted that the terms "first", "second", etc. used in this disclosure are only used to distinguish different steps, devices or modules, etc. The relevant terms neither represent any specific technical meaning nor indicate the order or interdependence between them.

[0041] It should be noted that the modification of the term "at least one" used in the present disclosure is illustrative rather than restrictive. Unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0042] It should be noted that the term "and / or" used in this disclosure to describe an association relationship between associated objects generally indicates the existence of at least three types of association relationships. For example, "A and / or B" can represent at least three types of association relationships: the existence of A alone, the existence of both A and B, and the existence of B alone.

[0043] It should be noted that the steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. Unless otherwise specified, the scope of the present disclosure is not limited by the order in which the steps are described in the relevant embodiments.

[0044] It should be noted that all actions of acquiring signals, information or data in the present disclosure are carried out in compliance with the corresponding data protection laws and policies of the country where they are located and with the authorization given by the owner of the corresponding device.

[0045] Exemplary Methods

[0046] Figure 1 is a flowchart of a method for generating a fault test scenario according to an example embodiment, as Figure 1 As shown in the figure, the method for generating a fault test scenario is used in a fault test scenario generation device, and includes the following steps.

[0047] S110, obtaining fault information generated by a to-be-tested automatic driving scheme in a plurality of sample fault test scenarios, the sample fault test scenarios including scene feature information, and the scene feature information including at least one of a driving environment parameter, a preset fault parameter, and a traffic participant behavior parameter;

[0048] S120, performing clustering processing on the sample fault test scenarios with generated fault information to obtain a plurality of test scenario sets, and each test scenario set including at least one sample fault test scenario;

[0049] S130, selecting at least one sample fault test scenario corresponding to a same test scenario set as a to-be-selected scenario;

[0050] S140, generating a target fault test scenario according to scene feature information corresponding to the to-be-selected scenario.

[0051] The automatic driving scheme can be a program developed for a vehicle with an automatic driving function, and the vehicle can perform automatic driving by executing the automatic driving scheme, such as vehicle speed adjustment, driving trajectory planning, vehicle distance keeping, etc. The fault test scenario generation device tests the to-be-tested automatic driving scheme in the sample fault test scenarios, and generates fault information corresponding to the sample fault test scenarios in the case that the to-be-tested automatic driving scheme has a fault. The tester can debug and optimize the to-be-tested automatic driving scheme according to the fault information. The fault information can include vehicle collision, vehicle wear, vehicle non-response, etc. The fault information can be generated by the sample fault test scenarios, or can be generated by the automatic driving scheme.

[0052] In step S110, the scene feature information includes a plurality of parameters, which together define each sample fault test scenario, so that different functions of the to-be-tested automatic driving scheme can be tested through each sample fault test scenario. Before step S110, a basic scenario can be pre-set, and a plurality of sample fault test scenarios can be formed by modifying the scene feature information in the basic scenario, adding scene feature information to the basic scenario, etc. For other descriptions of the basic scenario and the sample fault test scenario generation method, please refer to the following embodiments.

[0053] Scene feature information includes at least one of driving environment parameters, preset fault parameters, and traffic participant behavior parameters. Driving environment parameters may include a traffic participant's initial speed, initial position, initial posture, traffic participant behavior triggering time or location, initial state of traffic lights, and duty cycle. Preset fault parameters may include perception fault parameters, decision fault parameters, and control fault parameters. These parameters are set based on perception, decision, and control, respectively. For example, perception fault parameters may provide environmental images of varying clarity to the autonomous driving solution under test, allowing the tested autonomous driving solution to obtain environmental information of varying richness based on the environmental images. Perception fault parameters may also limit the provision of environmental detection information to the tested autonomous driving solution. Control fault parameters may include control signal interruption. Preset fault parameters may also include the fault onset time, fault duration, and fault intensity corresponding to each of the perception fault parameters, decision fault parameters, and control fault parameters. Traffic participants include, but are not limited to, pedestrians, animals, and vehicles other than those implementing the autonomous driving solution under test. Traffic participant behavior parameters may include speed parameters, position parameters, and posture parameters of traffic participants that define abnormal traffic participant behavior.

[0054] In step S120, if the test generates fault information in the sample fault test scenario, it indicates that the autonomous driving solution under test cannot successfully pass the test of the sample fault test scenario; if the test does not generate fault information in the sample fault test scenario, it indicates that the autonomous driving solution under test successfully passes the test of the sample fault test scenario. The present disclosure clusters the sample fault test scenarios that generate fault information, while the sample fault test scenarios that do not generate fault information may not be processed. In one embodiment, the fault information can be filtered according to preset rules such as whether there is a collision and whether damage is caused to traffic participants, and only the sample fault test scenarios corresponding to the fault information that meets the preset rules are clustered.

[0055] Clustering can be performed using K-Means clustering, mean shift clustering, DBSCAN clustering, hierarchical clustering, etc., and clustering can be performed based on one or more parameters in the scenario feature information. The present disclosure does not limit the clustering method. After clustering, sample fault test scenarios with certain similarities form a test scenario set.

[0056] In step S130, one or more sample fault test scenarios corresponding to the same test scenario set can be randomly selected as candidate scenarios; in step S140, the scenario feature information corresponding to the candidate scenarios is randomly combined with the basic scenarios to generate target fault test scenarios, so that multiple target fault test scenarios can be generated based on the original sample fault test scenarios to enrich the types of fault test scenarios.

[0057] The method for generating a fault test scenario provided by the embodiments of the present disclosure combines the scene feature information corresponding to the candidate scenarios to generate a target fault test scenario, so that a large number of target fault test scenarios can be automatically generated on the basis of original sample fault test scenarios, to enrich the types of fault test scenarios and realize sufficient testing of the automatic driving scheme; improve the generation efficiency of fault test scenarios and reduce the development cost of fault test scenarios. By clustering processing, sample fault test scenarios with certain similarity are grouped into a test scenario set, so that the subsequently selected candidate scenarios have certain similarity, avoiding or reducing the situation that the target fault test scenario generated due to the large difference in the scene feature information corresponding to the candidate scenarios cannot be used to test the measured automatic driving scheme. In addition, the fault test scenario constructed by the method can also be used for hazard analysis and risk assessment in the design and development stage of the automatic driving scheme, which is beneficial to improving the safety of the automatic driving scheme and reducing the probability of dangerous events in the application process of the automatic driving scheme.

[0058] After the target fault test scenario is generated, the automatic driving scheme can be tested through the target fault test scenario to observe whether the automatic driving scheme has defects and debug the automatic driving scheme. Of course, in the case that fault information meeting the preset rule is generated in the target fault test scenario, a single factor variable can also be set according to the target fault test scenario to confirm the main factor of the fault information such as collision and dangerous behavior.

[0059] Please refer to Figure 2 In an embodiment, the sample fault test scenario further includes initial scene parameters; and the step S120 further includes:

[0060] S210, performing clustering processing on the sample fault test scenarios generating fault information according to the scene feature information, to obtain a plurality of fault scene clusters, the fault scene clusters corresponding to a plurality of sample fault test scenarios;

[0061] S220, performing clustering processing on the plurality of fault scene clusters according to the initial scene parameters, to obtain a plurality of test scenario sets.

[0062] In S210, the sample fault test scenarios can be clustered by one or more parameters in the scene feature information to form a plurality of fault scene clusters. The sample fault test scenarios in the same fault scene cluster have certain similarity. The clustering processing in S210 can also use K-Means clustering, mean shift clustering, DBSCAN clustering, hierarchical clustering, etc., and is based on one or more parameters in the scene feature information, which is not limited by the present disclosure.

[0063] The initial scenario parameters may include the initial speed, initial position, and initial posture of traffic participants; they may also include the initial speed, initial position, and initial posture of the autonomous driving solution being tested. In S220 , multiple fault scenario clusters are clustered based on the initial scenario parameters, so that the fault scenario clusters corresponding to the same test scenario set have similar initial scenario parameters.

[0064] In this embodiment, the sample fault test scenarios are clustered using scenario feature information and initial scenario parameters, so that the sample fault test scenarios corresponding to the same test scenario set have similarities, thereby increasing the possibility that the target fault test scenarios formed by subsequent combinations can be used for testing.

[0065] See also Figure 3 In one embodiment, the aforementioned step S220 further includes:

[0066] S310, performing fuzzy random processing on initial scenario parameters corresponding to the first fault scenario cluster to obtain an initial scenario parameter range, wherein the first fault scenario cluster is any one of the multiple fault scenario clusters;

[0067] S320, determining a second fault scenario cluster that meets the initial scenario parameter range from multiple fault scenario clusters, and obtaining a test scenario set, wherein the multiple fault scenario clusters include the second fault scenario cluster, and the test scenario set includes at least one second fault scenario cluster.

[0068] In S310, any fault scenario cluster is selected as the first fault scenario cluster, and fuzzy random processing is performed on the initial scenario parameters, so that an initial scenario parameter range can be obtained based on the initial scenario parameters of the first fault scenario cluster. For example, if the initial speed of a traffic participant is 4 km / h, the initial scenario parameter range obtained after fuzzy processing is 3.5 to 4.5 km / h.

[0069] In S320, the fault scenario clusters whose initial scenario parameters are within the range of the initial scenario parameters are taken as the second fault scenario clusters, and all or part of the second fault scenario clusters that meet the range of the initial scenario parameters are combined into a test scenario set, or the first fault scenario clusters and all or part of the second fault scenario clusters that meet the range of the initial scenario parameters are combined into a test scenario set, so that the initial scenario parameters of multiple sample fault test scenarios corresponding to the same test scenario set have certain similarities.

[0070] In one embodiment, the sample fault test scenario further includes test classification information; S210 includes:

[0071] The first sample fault test scenario is clustered according to the driving environment parameters to obtain multiple fault scenario clusters. The first sample fault test scenario is a sample fault test scenario that generates fault information and the corresponding test classification information is an environmental parameter test.

[0072] The test classification information corresponding to the sample fault test scenario generated by adding, deleting, or modifying driving environment parameters in the basic scenario is called the environmental parameter test. Specifically, modifications can be made to the initial speed, initial position, initial posture, behavior triggering time or location of traffic participants in the basic scenario, as well as the initial state and duty cycle of traffic lights.

[0073] For sample fault test scenarios whose test classification information is environmental parameter testing, clustering can be performed based on the driving environment parameters, so that sample fault test scenarios with similar driving environment parameters form a test scenario set.

[0074] In one embodiment, the sample fault test scenario further includes test classification information, and the scenario characteristic information includes multiple preset fault parameters, where the multiple preset fault parameters include at least one of a perception fault parameter, a decision fault parameter, and a control fault parameter. In the aforementioned step S210, the following is further included:

[0075] The second sample fault test scenario is clustered according to preset fault parameters to obtain multiple fault scenario clusters. The second sample fault test scenario is a sample fault test scenario that generates fault information and the corresponding test classification information is a sample fault test scenario that triggers a fault test.

[0076] The test classification information corresponding to the sample fault test scenario generated by adding, deleting, and modifying preset fault parameters in the basic scenario is called a triggered fault test. Specifically, based on the perspective of internal faults generated by the autonomous driving solution, fault injection can be performed on modules such as perception, decision-making, and control to test other faults caused by faults in the autonomous driving solution itself.

[0077] For the sample fault test scenarios whose test classification information is to trigger a fault test, clustering processing can be performed according to preset fault parameters, so that the sample fault test scenarios with similar preset fault parameters are combined into a test scenario set.

[0078] In one embodiment, the sample fault test scenario further includes test classification information, and the scenario feature information includes initial scenario parameters and traffic participant behavior parameters; the aforementioned step S210 further includes:

[0079] The third sample fault test scenario is clustered according to the initial scenario parameters to obtain multiple fault scenario clusters. The third sample fault test scenario is a sample fault test scenario that generates fault information and the corresponding test classification information is a participant behavior test.

[0080] The test classification information corresponding to the sample fault test scenario generated by adding, deleting, and modifying traffic participant behavior parameters in the basic scenario is called the participant behavior test. Specifically, the traffic participant behavior parameters can be set based on the perspective of abnormal behavior of traffic participants outside the autonomous driving solution to test whether the autonomous driving solution generates fault information when traffic participants behave abnormally.

[0081] For the sample fault test scenarios whose test classification information is participant behavior test, clustering processing can be performed based on the initial scenario parameters, so that the sample fault test scenarios with similar initial scenario parameters are composed of a test scenario set.

[0082] In this example, step S210 clusters the third sample fault test scenario based on initial scenario parameters to obtain multiple fault scenario clusters. Step S220 also clusters the multiple fault scenario clusters based on initial scenario parameters to obtain multiple test scenario sets. Therefore, the initial scenario parameters used in these two clustering processes can be the same or different. For example, both clustering processes can be based on initial scenario parameters such as the initial speed, initial position, and initial posture of traffic participants. For another example, the clustering process in S210 can be based on initial scenario parameters such as the initial speed, initial position, and initial posture of traffic participants, while the clustering process in S220 can be based on initial scenario parameters such as the initial speed, initial position, and initial posture of the tested autonomous driving solution. For another example, the clustering process in S210 can be based on the initial scenario parameter of the initial speed of traffic participants, while the clustering process in S220 can be based on the initial scenario parameter of the initial position of traffic participants. This example does not limit the specific content of the initial scenario parameters used in the two clustering processes.

[0083] It can be seen from the above embodiments that the present disclosure can use different clustering criteria to perform clustering processing according to different test classification information of sample fault test scenarios.

[0084] See also Figure 4 In some embodiments, the sample fault test scenario further includes a basic scenario, which includes road environment information, driving condition information, and traffic signal information. The method for generating the fault test scenario further includes:

[0085] S410, receiving a test instruction;

[0086] S420: Determine road environment information, driving condition information, and traffic signal information according to the test instruction, and generate a basic scenario based on the selected road environment information, driving condition information, and traffic signal information.

[0087] S430, adding, modifying or deleting scene feature information on the basis of the basic scene to generate a plurality of sample fault test scenes.

[0088] In some embodiments, the target to be tested can be tested in advance according to the needs, and a basic scene is constructed, and scene feature information is added, modified or deleted on the basis of the basic scene to generate a plurality of different sample fault test scenes. The basic scene includes road environment information, driving condition information and traffic signal information. According to the road environment information, the driving condition information and the traffic signal information, the basic map data available for the tested automatic driving scheme to execute can be generated. Optionally, the road environment information includes a straight road, a curved road, a road intersection type, auxiliary road information, roundabout information, etc., and can also include road length, road width, number of lanes, road transverse slope, longitudinal slope, road surface material, road surface roughness and other road geometric parameters. Optionally, the traffic signal information includes signal lamp parameters, road sign parameters, traffic marking parameters, traffic prompt parameters, height limit parameters, etc. Optionally, the driving condition information can include automatic parking, curb parking, roadside starting, etc.

[0089] Those skilled in the art can understand that in the process of constructing the basic scene, not only the parameters in the scene feature information need to be considered, but also the basic scene needs to be ensured to be able to efficiently and completely test the automatic driving scheme, that is, the settings of the parameters in the basic scene are ensured to be reasonable. The reasonableness of the basic scene can be verified by testing the pre-stored automatic driving scheme which has been debugged in the basic scene, and of course the reasonableness of the settings of the parameters in the basic scene can also be verified by manual driving.

[0090] Exemplary devices

[0091] Figure 5 is a fault test scene generation device block diagram according to an exemplary embodiment. Referring to Figure 5 The device 500 includes a first module 510, a second module 520, a third module 530 and a fourth module 540.

[0092] The first module 510 is configured to obtain fault information generated by testing a to-be-tested automatic driving scheme in a plurality of sample fault test scenes, the sample fault test scene including scene feature information, and the scene feature information including at least one of driving environment parameters, preset fault parameters and traffic participant behavior parameters.

[0093] The second module 520 is configured to perform clustering processing on the sample fault test scene in which the fault information is generated to obtain a plurality of test scene sets, and each test scene set includes at least one sample fault test scene.

[0094] The third module 530 is configured to select at least one sample fault test scene corresponding to the same test scene set as a candidate scene.

[0095] The fourth module 540 is configured to generate a target fault test scene according to scene characteristic information corresponding to the candidate scene.

[0096] In an embodiment, the sample fault test scene further includes an initial scene parameter; and the second module 520 is further configured to: cluster the sample fault test scene generating the fault information according to the scene characteristic information to obtain a plurality of fault scene clusters, the fault scene cluster corresponding to a plurality of sample fault test scenes; and cluster the plurality of fault scene clusters according to the initial scene parameter to obtain a plurality of test scene sets.

[0097] In an embodiment, the second module 520 is further configured to: perform fuzzy random processing on the initial scene parameter corresponding to the first fault scene cluster to obtain an initial scene parameter range, the first fault scene cluster being any one of the plurality of fault scene clusters; and determine a second fault scene cluster meeting the initial scene parameter range from the plurality of fault scene clusters to obtain a test scene set, the plurality of fault scene clusters including the second fault scene cluster, and the test scene set including at least one second fault scene cluster.

[0098] In an embodiment, the sample fault test scene further includes test classification information; and the second module 520 is further configured to: cluster the first sample fault test scene according to the driving environment parameter to obtain a plurality of fault scene clusters, the first sample fault test scene being a sample fault test scene generating the fault information and corresponding to test classification information of environmental parameter test; and wherein the test classification information corresponding to the sample fault test scene obtained by adding, deleting or modifying the driving environment parameter in the basic scene is the environmental parameter test.

[0099] In an embodiment, the sample fault test scene further includes test classification information, the scene characteristic information includes a plurality of preset fault parameters, and the plurality of preset fault parameters include at least one of a perception fault parameter, a decision fault parameter and a control fault parameter; and the second module 520 is further configured to: cluster the second sample fault test scene according to the preset fault parameter to obtain a plurality of fault scene clusters, the second sample fault test scene being a sample fault test scene generating the fault information and corresponding to test classification information of trigger fault test; and wherein the test classification information corresponding to the sample fault test scene obtained by adding, deleting or modifying the preset fault parameter in the basic scene is the trigger fault test.

[0100] In one embodiment, the sample fault test scenario also includes test classification information, and the scenario feature information includes initial scenario parameters and traffic participant behavior parameters; the second module 520 is also used to: cluster the third sample fault test scenario according to the initial scenario parameters to obtain multiple fault scenario clusters, and the third sample fault test scenario is a sample fault test scenario that generates fault information and the corresponding test classification information is a participant behavior test; wherein, the test classification information corresponding to the sample fault test scenario obtained by adding, deleting, and modifying the traffic participant behavior parameters in the basic scenario is the participant behavior test.

[0101] In one embodiment, the sample fault test scenario also includes a basic scenario, which includes road environment information, driving condition information, and traffic signal information; the device 500 also includes a fifth module, which is used to: receive test instructions; determine the road environment information, driving condition information, and traffic signal information according to the test instructions, and generate a basic scenario based on the selected road environment information, driving condition information, and traffic signal information; add, modify, or delete the scenario feature information on the basis of the basic scenario to generate multiple sample fault test scenarios.

[0102] The embodiments of the present disclosure provide Fault test scenario generation device Able to achieve the above Generation of fault test scenarios To avoid repetition, the various processes implemented in the embodiment of the method will not be described again here.

[0103] Exemplary electronic devices

[0104] Figure 6 FIG. 6 is a block diagram of an electronic device 600 according to an exemplary embodiment. The electronic device 600 may be a computer device, a notebook computer, a server, a vehicle controller, an in-vehicle terminal, an in-vehicle computer, or other types of electronic devices.

[0105] Reference Figure 6 The electronic device 600 may include at least one processor 610 and a memory 620. The processor 610 may execute instructions stored in the memory 620. The processor 610 is communicatively connected to the memory 620 via a data bus. In addition to the memory 620, the processor 610 may also be communicatively connected to an input device 630, an output device 640, and a communication device 650 via the data bus.

[0106] The processor 610 may be any conventional processor, such as a central processing unit (CPU), a graphics processing unit (GPU), a field programmable gate array (FPGA), a system on chip (SOC), an application specific integrated circuit (ASIC), or a combination thereof.

[0107] The memory 620 may be implemented by any type of volatile or non-volatile memory device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0108] In the embodiment of the present disclosure, executable instructions are stored in the memory 620. The processor 610 can read the executable instructions from the memory 620 and execute the instructions to implement all or part of the steps of the method for generating a fault test scenario in the above exemplary embodiment.

[0109] Exemplary computer-readable storage media

[0110] In addition to the above methods and apparatus, exemplary embodiments of the present disclosure also include a computer program product or a computer-readable storage medium storing the computer program product. The computer product includes computer program instructions that can be executed by a processor to implement all or part of the steps described in the above exemplary embodiments.

[0111] The computer program product may be written in any combination of one or more programming languages ​​to implement the program code for performing the operations of the disclosed embodiments, including object-oriented programming languages ​​such as Java, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages ​​and scripting languages ​​(e.g., Python). The program code may be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0112] The computer-readable storage medium may be any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may include, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination thereof. More specific examples of readable storage media include: a static random access memory (SRAM) electrically connected with one or more wires, an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic memory, a flash memory, a magnetic disk or an optical disk, or any suitable combination thereof.

[0113] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the present disclosure. The present disclosure is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are to be considered merely as exemplary, and the present disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and variations may be made without departing from the scope thereof.

Claims

1. A method for generating a fault test scenario, characterized in that: include: Obtaining fault information generated by testing the tested autonomous driving solution in multiple sample fault test scenarios, wherein the sample fault test scenarios include scenario feature information, basic scenarios, and initial scenario parameters, wherein the scenario feature information includes at least one of a driving environment parameter, a preset fault parameter, and a traffic participant behavior parameter, and the basic scenario includes road environment information, driving condition information, and traffic signal information; wherein the step of generating multiple sample fault test scenarios includes: Receive test instructions; determining road environment information, driving condition information, and traffic signal information according to the test instruction, and generating a basic scenario based on the selected road environment information, driving condition information, and traffic signal information; Add, modify or delete scenario feature information on the basis of the basic scenario to generate multiple sample fault test scenarios, while sample fault test scenarios that do not generate fault information are not processed; Clustering the sample fault test scenarios that generate the fault information to obtain multiple test scenario sets, each of the test scenario sets including at least one sample fault test scenario; wherein clustering the sample fault test scenarios that generate the fault information to obtain multiple test scenario sets includes: performing clustering processing on the sample fault test scenarios that generate the fault information according to the scenario feature information to obtain a plurality of fault scenario clusters, wherein the fault scenario clusters correspond to the plurality of sample fault test scenarios; Clustering the multiple fault scenario clusters according to the initial scenario parameters to obtain the multiple test scenario sets; wherein clustering the multiple fault scenario clusters according to the initial scenario parameters to obtain the multiple test scenario sets includes: Performing fuzzy random processing on the initial scenario parameters corresponding to the first fault scenario cluster to obtain an initial scenario parameter range, wherein the first fault scenario cluster is any one of the multiple fault scenario clusters; Determining a second fault scenario cluster that meets the initial scenario parameter range from the multiple fault scenario clusters to obtain the test scenario set, wherein the multiple fault scenario clusters include the second fault scenario cluster, and the test scenario set includes at least one second fault scenario cluster; selecting at least one of the sample fault test scenarios corresponding to the same test scenario set as a candidate scenario; A target fault test scenario is generated according to the scenario feature information corresponding to the candidate scenario.

2. The method for generating a fault test scenario according to claim 1, characterized in that: The sample fault test scenario also includes test classification information; the sample fault test scenario generating the fault information is clustered based on the scenario feature information to obtain multiple fault scenario clusters, including: The first sample fault test scenario is clustered according to the driving environment parameters to obtain a plurality of fault scenario clusters, wherein the first sample fault test scenario is the sample fault test scenario that generates the fault information and the corresponding test classification information is the environmental parameter test; wherein, the test classification information corresponding to the sample fault test scenario obtained by adding, deleting, and modifying the driving environment parameters in the basic scenario is the environmental parameter test.

3. The method for generating a fault test scenario according to claim 1, wherein: The sample fault test scenario further includes test classification information, the scenario feature information includes a plurality of preset fault parameters, and the plurality of preset fault parameters include at least one of a perception fault parameter, a decision fault parameter, and a control fault parameter; The clustering process of the sample fault test scenarios generating the fault information according to the scenario feature information to obtain multiple fault scenario clusters includes: The second sample fault test scenario is clustered according to the preset fault parameters to obtain a plurality of fault scenario clusters, wherein the second sample fault test scenario is the sample fault test scenario that generates the fault information and the corresponding test classification information is the triggered fault test; wherein, the test classification information corresponding to the sample fault test scenario obtained by adding, deleting, and modifying the preset fault parameters in the basic scenario is the triggered fault test.

4. The method for generating a fault test scenario according to claim 1, wherein: The sample fault test scenario further includes test classification information, and the scenario feature information includes the initial scenario parameters and the traffic participant behavior parameters; clustering the sample fault test scenario that generates the fault information based on the scenario feature information to obtain multiple fault scenario clusters, including: The third sample fault test scenario is clustered according to the initial scenario parameters to obtain a plurality of fault scenario clusters, wherein the third sample fault test scenario is the sample fault test scenario that generates the fault information and the corresponding test classification information is the participant behavior test; wherein, the test classification information corresponding to the sample fault test scenario obtained by adding, deleting, and modifying the traffic participant behavior parameters in the basic scenario is the participant behavior test.

5. A device for generating a fault test scenario, characterized in that: include: The first module is configured to obtain fault information generated by testing the autonomous driving solution under test in multiple sample fault test scenarios, wherein the sample fault test scenarios include scenario feature information, basic scenarios, and initial scenario parameters. The scenario feature information includes at least one of a driving environment parameter, a preset fault parameter, and a traffic participant behavior parameter. The basic scenario includes road environment information, driving condition information, and traffic signal information. The step of generating multiple sample fault test scenarios includes: Receive test instructions; determining road environment information, driving condition information, and traffic signal information according to the test instruction, and generating a basic scenario based on the selected road environment information, driving condition information, and traffic signal information; Add, modify or delete scenario feature information on the basis of the basic scenario to generate multiple sample fault test scenarios, while sample fault test scenarios that do not generate fault information are not processed; The second module is configured to cluster the sample fault test scenarios that generate the fault information to obtain multiple test scenario sets, each of which includes at least one sample fault test scenario; wherein the clustering of the sample fault test scenarios that generate the fault information to obtain multiple test scenario sets includes: performing clustering processing on the sample fault test scenarios that generate the fault information according to the scenario feature information to obtain a plurality of fault scenario clusters, wherein the fault scenario clusters correspond to the plurality of sample fault test scenarios; Clustering the multiple fault scenario clusters according to the initial scenario parameters to obtain the multiple test scenario sets; wherein clustering the multiple fault scenario clusters according to the initial scenario parameters to obtain the multiple test scenario sets includes: Performing fuzzy random processing on the initial scenario parameters corresponding to the first fault scenario cluster to obtain an initial scenario parameter range, wherein the first fault scenario cluster is any one of the multiple fault scenario clusters; Determining a second fault scenario cluster that meets the initial scenario parameter range from the multiple fault scenario clusters to obtain the test scenario set, wherein the multiple fault scenario clusters include the second fault scenario cluster, and the test scenario set includes at least one second fault scenario cluster; A third module is configured to select at least one of the sample fault test scenarios corresponding to the same test scenario set as a candidate scenario; The fourth module is used to generate a target fault test scenario according to the scenario feature information corresponding to the to-be-selected scenario.

6. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to read the executable instructions from the memory and execute the instructions to implement the method for generating a fault test scenario as described in any one of claims 1-4.

7. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the program instructions are executed by a processor, the steps of the method for generating a fault test scenario described in any one of claims 1 to 4 are implemented.

Images