A method and device for constructing a scenario library and test case set in autonomous driving

Abstract

The present application provides a method and device for constructing a scene library and a test case set in an autonomous driving system, which relates to the field of autonomous driving technology, and defines a scene element set in an autonomous driving system and an index value of each scene element in the scene element set with respect to an evaluation index; the scene elements are combined according to the scene label corresponding to each scene element, and multiple test scenes are generated and a scene library is formed; according to the index value of each scene element included in each test scene, the scene index value of each test scene with respect to the evaluation index is determined; according to the test requirements of the autonomous driving system and the scene index value of each test scene, and / or according to the design and operation conditions of the autonomous driving system and the scene label corresponding to each scene element, the target test scene is screened out from the scene library and a test case set is formed. In this way, the speed and efficiency of scene construction can be improved, the effective management and screening of massive scenes can be achieved, the test case set can be accurately screened, and the test targeting can be improved.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and more particularly to a method and device for constructing a scenario library and a test case set in autonomous driving. Background Art

[0002] With the development of autonomous driving technology, ensuring its functional safety has become a focus of the industry. Since traditional vehicle testing methods are difficult to fully cover the various complex traffic conditions that autonomous driving systems may encounter, scenario-based testing methods have gradually become a key means of verifying and confirming the safety of autonomous driving systems. Among them, the construction of scenario libraries and test cases is the basis for achieving this goal.

[0003] However, most existing scenario library construction technologies rely on manual experience or uploaded real scenarios, which results in a long scenario library construction cycle, uncontrollable scenario library generation results, and insufficient scenario library coverage. In addition, the manual selection of test cases from the generated scenario library also limits the flexibility and repeatability of the test. Once the test strategy needs to be adjusted or new security challenges need to be addressed, the built test case set may need to be re-evaluated or even reconstructed, increasing maintenance costs and workload. Summary of the invention

[0004] In view of this, the purpose of this application is to provide a method and device for constructing a scenario library and a test case set in autonomous driving. By constructing a limited but wide-coverage structured scenario element set and generating a scenario library in an automated manner, a large amount of repetitive manual design work can be avoided, and the speed and efficiency of scenario construction can be improved. The scenario elements are evaluated by evaluation indicators, and then the index values ​​of the scenario elements that constitute the test scenario are comprehensively considered to generate the scenario index values ​​of each test scenario, which helps to achieve effective management and screening of massive scenarios. The dual screening mechanism combined with test requirements and / or design operating conditions can accurately screen out a subset of test cases and improve the pertinence of the test.

[0005] The present application embodiment provides a method for constructing a scenario library and a test case set in autonomous driving, and the construction method includes:

[0006] Define a scene element set in an autonomous driving system and an index value of each scene element in the scene element set with respect to an evaluation index; wherein the scene element set is a data set organized in a multi-layer structure; each element at the bottom layer of the scene element set represents a scene element, and each upper layer element related to each element at the bottom layer represents scene labels at various levels corresponding to the scene element;

[0007] The scene elements are combined according to the scene labels corresponding to each scene element, multiple test scenes are generated and a scene library is formed;

[0008] Determine the scene index value of each test scene with respect to the evaluation index according to the index value of each scene element included in each test scene;

[0009] According to the test requirements of the autonomous driving system and the scenario index value of each test scenario, and / or according to the design operating conditions of the autonomous driving system and the scenario label corresponding to each scenario element, the target test scenario is screened out from the scenario library and a test case set is formed.

[0010] Further, according to the test requirements of the autonomous driving system and the scenario index value of each test scenario, according to the design operating conditions of the autonomous driving system and the scenario label corresponding to each scenario element, the target test scenario is screened out from the scenario library and a test case set is formed, including:

[0011] Filtering a first test scenario from the scenario library according to the test requirements of the autonomous driving system and the scenario index value of each test scenario;

[0012] Filtering a second test scene from the scene library according to the design operating conditions of the autonomous driving system and the scene label corresponding to each scene element;

[0013] The intersection between the first test scenario and the second test scenario is determined as the target test scenario, and the set of the target test scenarios is used as the test case set.

[0014] Furthermore, according to the test requirements of the autonomous driving system and the scenario index value of each test scenario, a target test scenario is screened out from the scenario library and a test case set is formed, including:

[0015] Decomposing the test requirements to determine the requirement items and requirement standards included in the test requirements;

[0016] According to the predefined mapping relationship between the requirement items and the evaluation indicators, the requirement items and the requirement standards are converted into screening rules;

[0017] According to the screening rules, the scenario indicator value of each test scenario and the scenario elements included in each test scenario, the target test scenario is screened out from the scenario library to form the test case set.

[0018] Furthermore, according to the design operating conditions of the autonomous driving system and the labels corresponding to each scenario element, a target test scenario is screened out from the scenario library to form a test case set, including:

[0019] For each function to be tested in the autonomous driving system, determining a target label and / or a filter label according to a design operating condition corresponding to the function to be tested;

[0020] For each test scene in the scene library, if the scene elements included in the test scene have the target tag and do not have the filter tag, the test scene is determined as a target test scene;

[0021] The determined target test scenario is used as the test case set corresponding to the function to be tested.

[0022] Further, the evaluation index includes: a frequency index, a risk index and a complexity index; the determining of the scenario index value of each test scenario with respect to the evaluation index according to the index value of each scenario element included in each test scenario includes:

[0023] For each test scenario, determining the minimum index value of each scenario element included in the test scenario with respect to the frequency index as the scenario index value with respect to the frequency index of the test scenario;

[0024] Determine the maximum index value of each scenario element included in the test scenario with respect to the risk index as the scenario index value with respect to the risk index of the test scenario;

[0025] The average index value of the complexity index of each scenario element included in the test scenario is determined as the first scenario index value of the complexity index of the test scenario; and the number of scenario elements in the test scenario whose index values ​​of the complexity index are greater than the average index value is determined as the second scenario index value of the complexity index of the test scenario.

[0026] Furthermore, the construction method further includes: defining a probability correction function when a combination of multiple target scene elements in the scene element set appears according to actual scene conditions;

[0027] Then, for the frequency index, determining the scene index value of each test scene with respect to the evaluation index according to the index value of each scene element included in each test scene also includes:

[0028] For each test scenario, determine whether the test scenario includes any combination of target scenario elements;

[0029] If included, determining a correction index value based on the probability correction function corresponding to each target scene element combination and the index value of each scene element included in each target scene element combination with respect to the frequency index;

[0030] The minimum value between the minimum index value of each scenario element included in the test scenario with respect to the frequency index and the correction index value is determined as the scenario index value of the test scenario with respect to the frequency index.

[0031] Furthermore, the scene tags of each level in the scene element set include a primary tag and a secondary tag subdivided under the primary tag; each secondary tag is associated with a scene element subset; the scene elements are combined according to the scene tag corresponding to each scene element to generate multiple test scenes and form a scene library, including:

[0032] Cartesian product operations are performed between the scene element subsets under each secondary tag, and scene elements are selected from the scene element subsets under each secondary tag and combined to obtain the multiple test scenes and form the scene library.

[0033] Furthermore, the construction method also includes:

[0034] According to the index value of each scenario element included in each target test scenario with respect to the risk index, it is identified whether the target test scenario contains a danger triggering condition.

[0035] The embodiment of the present application also provides a device for constructing a scenario library and a test case set in autonomous driving, the device comprising:

[0036] A definition module, used to define a scene element set in an autonomous driving system and an index value of each scene element in the scene element set with respect to an evaluation index; wherein the scene element set is a data set organized in a multi-layer structure; each element at the bottom layer of the scene element set represents a scene element, and each upper layer element related to each element at the bottom layer represents a scene label at each level corresponding to the scene element;

[0037] A generation module, used to combine the scene elements according to the scene labels corresponding to each scene element, generate multiple test scenes and form a scene library;

[0038] A determination module, used to determine the scene index value of each test scene with respect to the evaluation index according to the index value of each scene element included in each test scene;

[0039] A screening module is used to screen out target test scenarios from the scenario library and form a test case set according to the test requirements of the autonomous driving system and the scenario indicator value of each test scenario, and / or according to the design operating conditions of the autonomous driving system and the scenario label corresponding to each scenario element.

[0040] An embodiment of the present application also provides an electronic device, including: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, the steps of the method for constructing a scenario library and a test case set in autonomous driving are performed as described above.

[0041] An embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for constructing a scenario library and a test case set in autonomous driving are executed as described above.

[0042] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 A flowchart of a method for constructing a scenario library and a test case set in autonomous driving provided by an embodiment of the present application is shown;

[0045] Figure 2 An architectural diagram of a scene element set provided in an embodiment of the present application is shown;

[0046] Figure 3 A schematic diagram of the structure of a device for constructing a scenario library and a test case set in an autonomous driving system provided in an embodiment of the present application is shown;

[0047] Figure 4 A schematic structural diagram of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0048] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme 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 only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work belongs to the scope of protection of the present application.

[0049] Research has found that with the development of autonomous driving technology, ensuring its functional safety has become a focus of the industry. Since traditional vehicle testing methods are difficult to fully cover the various complex traffic conditions that autonomous driving systems may encounter, scenario-based testing methods have gradually become a key means of verifying and confirming the safety of autonomous driving systems. Among them, the construction of scenario libraries and test cases is the basis for achieving this goal.

[0050] However, most existing scenario library construction technologies rely on manual experience or uploaded real scenarios, which results in a long scenario library construction cycle, uncontrollable scenario library generation results, and insufficient scenario library coverage. In addition, the manual selection of test cases from the generated scenario library also limits the flexibility and repeatability of the test. Once the test strategy needs to be adjusted or new security challenges need to be addressed, the built test case set may need to be re-evaluated or even reconstructed, increasing maintenance costs and workload.

[0051] Based on this, an embodiment of the present application provides a method for constructing a scenario library and a test case set in autonomous driving, so as to improve the speed and efficiency of scenario construction, realize effective management and screening of massive scenarios, accurately screen out a subset of test cases, and improve the targeted nature of the test.

[0052] See also Figure 1 , Figure 1 The following is a flow chart of a method for constructing a scenario library and a test case set in an autonomous driving embodiment provided in the present application. Figure 1 As shown in, the construction method provided in the embodiment of the present application includes:

[0053] S101. Define a scene element set in an autonomous driving system and an index value of each scene element in the scene element set with respect to an evaluation index.

[0054] Among them, the scene element set is a data set with multi-layer structured organization; each element in the bottom layer of the scene element set represents a scene element, and each upper-layer element related to each element in the bottom layer represents each level of scene labels corresponding to the scene element.

[0055] In specific implementation, a series of scenario elements can be defined according to the needs of the autonomous driving system and the design and operation conditions (ODC), forming a limited but wide-coverage element set covering multiple dimensions such as road environment, traffic participant behavior, weather conditions, etc. Among them, the design and operation conditions (Operational Design Domain, ODC) refers to the general term for various conditions applicable to the functional operation of the autonomous driving system determined when designing the system, including the design and operation scope, vehicle status, driver and passenger status, traffic conditions, operating conditions and other necessary conditions.

[0056] It should be noted that the choice of scene architecture is not fixed, and the scene structure and scene element set can be customized according to different formats or requirements.

[0057] Furthermore, each scene element in the scene element set is evaluated to determine the index value of each scene element with respect to the evaluation index. The evaluation index includes: a frequency index, a risk index, and a complexity index. The index value of each evaluation index can be stored in the scene element set as the attribute information of the scene element.

[0058] Among them, frequency represents the probability of occurrence of scenario elements; risk represents the possibility of scenario elements causing accidents; complexity represents the contribution of scenario elements to the complexity of the scenario. These evaluation results will provide inheritance items for subsequent scenario evaluation. Specifically, the purpose of frequency evaluation is to mark the probability and number of occurrences of scenario elements and their constituent scenarios in the real world. For example, the Frequency of elements can be divided into 5 levels, namely F(E)={1,2,3,4,5}, where E is the scenario element, and the values ​​represent from very low frequency to very high frequency. The Risk of scenario elements can be divided into 5 levels, namely R(E)={1,2,3,4,5}, and the values ​​represent from very low risk to very high risk. The Risk of scenario elements can be directly associated with the potential triggering conditions of the scenario. Finally, the complexity of scenario elements can be divided into 5 levels, namely C(E)={1,2,3,4,5}, and the values ​​represent from not complex to very complex, and the contribution of the element to the complexity of the scenario.

[0059] It should be noted that the embodiment of the present application adopts the above five-level evaluation system, and other grading methods may also be adopted, such as a continuous scoring system or a more detailed evaluation system. In specific implementation, the evaluation criteria can be adjusted according to actual conditions, such as by introducing an expert scoring mechanism or combining historical accident data analysis to determine the index value of the adjustment scenario element with respect to the evaluation index.

[0060] See also Figure 2 , Figure 2 This is an architecture diagram of a scene element set provided in an embodiment of the present application. Figure 2 As shown in , each level of scene tags in the scene element set includes a primary tag 21 and secondary tags 22 subdivided under the primary tag; each secondary tag is associated with a scene element subset 23.

[0061] The first-level label is the top-level layer, which needs to cover the top types of all scene elements as much as possible. The first-level labels selected in the embodiment of the present application include four levels: road layer, environment layer, vehicle layer, and interaction layer. Each first-level label must be subdivided into second-level labels 22. For example, the road layer can be subdivided into second-level labels such as road type and road layout. The link layer can be subdivided into time period, rain and visibility, etc.; the vehicle layer can be subdivided into vehicle speed, driver status, etc.; the interaction layer can be subdivided into traffic participant (TP) type, TP behavior, etc. Under the second-level label, the scene element subset is further subdivided, including n specific scene elements, such as road types can include highways, urban expressways, etc.; road layouts include straight roads, curved roads, etc.; time periods include daytime, night, dusk, etc.; rain and visibility include heavy rain and low visibility, etc.; the vehicle speed can be 60, 100kph, etc.; driver status includes awake, sleepy, etc.; TP types include motor vehicles, pedestrians, etc.; TP behaviors include cut-in, emergency stop, etc.

[0062] Among them, the secondary label determines whether different scene elements can be combined when generating the test scene, that is, scene elements with the same secondary label cannot be combined, but different scene elements can be freely combined.

[0063] S102. Combining scene elements according to scene labels corresponding to each scene element, generating multiple test scenes and forming a scene library.

[0064] In this step, a Cartesian product operation is performed between the scene element subsets under each secondary label, and scene elements are selected from the scene element subsets under each secondary label and combined to obtain the multiple test scenes and form the scene library.

[0065] In one example, there is a scene feature set of , then:

[0066]

[0067] in, for A subset of It is a collection of scene elements with the same first-level label (such as road level), for which:

[0068]

[0069] in, for A subset of It is a collection of scene elements with the same secondary label (such as road type at the road level) under the primary label (road level). can be further decomposed, namely,

[0070]

[0071] in, For specific scene elements ( Concrete examples include: highways, city expressways...intersections).

[0072] In this regard, a scene It can be expressed as:

[0073]

[0074] Furthermore, a scene library based on scene elements It can be expressed as:

[0075]

[0076] in," ” denotes the Cartesian product.

[0077] S103. Determine the scene index value of each test scene with respect to the evaluation index according to the index value of each scene element included in each test scene.

[0078] In this step, based on the preset element-scenario evaluation inheritance rules, the characteristics of all scenario elements constituting each test scenario can be comprehensively considered to assign a scenario index value to each generated test scenario.

[0079] In specific implementation, step S103 may include:

[0080] For the frequency index, for each test scenario, the minimum index value of each scenario element included in the test scenario with respect to the frequency index is determined as the scenario index value of the test scenario with respect to the frequency index. The formula is:

[0081]

[0082] For the risk index, the maximum index value of each scenario element included in the test scenario with respect to the risk index is determined as the scenario index value of the test scenario with respect to the risk index. The formula is:

[0083]

[0084] For the complexity index, the average index value of each scenario element included in the test scenario with respect to the complexity index is determined as the first scenario index value of the test scenario with respect to the complexity index, that is, ; and determining the number of scene elements in the test scene whose index value of the complexity index is greater than the average index value as the second scene index value of the test scene with respect to the complexity index, that is, The formula is:

[0085]

[0086] in, is an indicator function, whose value is 1 when the condition is met, otherwise it is 0; It is the threshold value of the preset complexity index, which can be set as the average index value of each scene element with respect to the complexity index.

[0087] Furthermore, the embodiments of the present application also take into account that the combination of scene elements may affect the frequency of the scene they ultimately form, so only evaluating the frequency of the scene elements in isolation may not be in line with the actual situation, so it is necessary to perform the same frequency evaluation on some of the combined scene elements. For example, the probability of some scene elements appearing in the same scene is extremely low, or it is impossible for them to appear at the same time; for another example, the probability of some scene elements appearing in the same scene is extremely high, and they even often appear together.

[0088] Specifically, the construction method further includes: defining a probability correction function when a combination of multiple target scene elements in the scene element set appears according to the actual scene situation. The formula is expressed as:

[0089]

[0090] in, is the correction function.

[0091] For the frequency index, step S103 may further include:

[0092] For each test scenario, determine whether the test scenario includes any combination of target scene elements; if included, determine a correction index value based on the probability correction function corresponding to each combination of target scene elements and the index value of each scene element included in each combination of target scene elements with respect to the frequency index; and determine the minimum value between the minimum index value of each scene element included in the test scenario with respect to the frequency index and the correction index value as the scene index value of the test scenario with respect to the frequency index.

[0093] Correspondingly, the formula is expressed as:

[0094]

[0095] In this way, after correction, the scenario index value of the frequency index can more accurately reflect the probability of occurrence of the test scenario and be more in line with actual traffic conditions.

[0096] S104. According to the test requirements of the autonomous driving system and the scenario index value of each test scenario, and / or according to the design operating conditions of the autonomous driving system and the scenario label corresponding to each scenario element, the target test scenario is screened out from the scenario library and a test case set is formed.

[0097] In this step, you can filter by using only the test requirements and the scenario indicator values ​​of each test scenario, or by using only the design operating conditions of the autonomous driving system and the scenario labels corresponding to each scenario element, or you can combine these two methods to filter.

[0098] In a possible implementation, step S104 may include:

[0099] According to the test requirements of the autonomous driving system and the scenario index value of each test scenario, a first test scenario is screened out from the scenario library; according to the design operating condition (ODC) of the autonomous driving system and the scenario label corresponding to each scenario element, a second test scenario is screened out from the scenario library; the intersection between the first test scenario and the second test scenario is determined as the target test scenario, and the set of the target test scenarios is used as the test case set.

[0100] In this way, by combining the two methods for screening, it can be ensured that the final output test case subset not only meets the test requirements but also accurately reflects the autonomous driving function design statement of the system under test, thereby improving the targeted nature of the test.

[0101] In a possible implementation, in step S104, according to the test requirements of the autonomous driving system and the scenario index value of each test scenario, a target test scenario is screened out from the scenario library and a test case set is formed, which specifically includes:

[0102] Step 1: Decompose the test requirements to determine the requirement items and requirement standards included in the test requirements.

[0103] Among them, the test requirements are information described in natural language, from which the requirements items and requirements standards can be extracted through the relevant technologies of natural language processing. The requirements items refer to the items that the requirements are aimed at, such as the magnitude and granularity requirements of the test requirements (such as the number of scenarios), risk requirements (including nominal scenarios, dangerous scenarios or edge scenarios) and scenario coverage requirements (requirements for uncommon scenario elements). The requirements standards refer to the conditions that each requirement item should meet.

[0104] Step 2: According to the predefined mapping relationship between the requirement items and the evaluation indicators, the requirement items and the requirement standards are converted into screening rules.

[0105] Among them, the scenario coverage requirement corresponds to the frequency index of scenario evaluation, the risk requirement corresponds to the risk index of scenario evaluation, and the magnitude and granularity correspond to the complexity index of scenario evaluation. Therefore, the requirement items and the corresponding requirement standards can be converted into quantitative screening rules for the corresponding indicator values.

[0106] Step 3: According to the screening rules, the scenario indicator value of each test scenario and the scenario elements included in each test scenario, the target test scenario is screened out from the scenario library to form the test case set.

[0107] In this step, the scenario index value of each test scenario and the scenario elements included in each test scenario are matched according to the screening rules, the test scenarios that meet all the screening rules are screened out from the scenario library as target test scenarios, and the set of target test scenarios is determined as a test case set.

[0108] In another possible implementation, in step S104, according to the design operating conditions of the autonomous driving system and the label corresponding to each scenario element, a target test scenario is screened out from the scenario library and a test case set is formed, which specifically includes:

[0109] Step a: for each function to be tested in the autonomous driving system, determine a target label and / or a filter label according to the design operating conditions corresponding to the function to be tested.

[0110] Among them, the design operating conditions (ODC) refer to the various conditions applicable to the function operation of the autonomous driving system when it is designed. By disassembling the design operating conditions, the target label and / or filter label corresponding to each function to be tested can be determined.

[0111] Step b: for each test scene in the scene library, if the scene elements included in the test scene have the target tag but do not have the filter tag, the test scene is determined as a target test scene.

[0112] In this step, the target tag and / or filter tag are matched one by one with the tags of each level of the scene elements included in the test scene. If the scene elements included in a test scene have a target tag but no filter tag, the test scene is determined as a target test scene.

[0113] Step c: Use the determined target test scenario as the test case set corresponding to the function to be tested.

[0114] Furthermore, the construction method also includes:

[0115] According to the index value of each scenario element included in each target test scenario with respect to the risk index, it is identified whether the target test scenario contains a danger triggering condition.

[0116] Here, according to the explanation of trigger conditions and SOTIF causal model in the current standards ISO 21448-2022 and GB / T 43267-2023, there may be multiple elements in a scenario that have the potential to serve as trigger conditions, and there is a certain possibility that they will trigger the specific functional deficiencies of the system under test, leading to harmful behaviors or failure to prevent, detect and mitigate reasonably foreseeable indirect misuse. In other words, dangerous trigger conditions refer to elements or human misuse in the scenario that can cause risks. Therefore, it is a requirement for the subsequent SOTIF analysis to properly analyze the dangerous trigger conditions in the scenario during the construction of the scenario.

[0117] In this step, a risk threshold can be set and compared with the index value of the risk index of each scenario element included in the target test scenario, and the scenario elements higher than the risk threshold can be determined as danger trigger conditions.

[0118] The present application provides a method for constructing a scenario library and a test case set in autonomous driving, including: (1) Scenario library construction based on a structured scenario element set: a scenario library is formed by performing Cartesian product operations on a subset of elements based on labels through a set of structured scenario elements covering multiple dimensions such as road environment, interactive behavior, environment, and self-vehicle. This standardized and systematic scenario library construction method is the basis of the entire technical solution. (2) Multi-dimensional element evaluation system and trigger condition identification: The scenario elements are evaluated through three dimensions: frequency, risk, and complexity, ensuring the coverage and rationality of the generated scenario library, and the scenarios in the scenario library are marked with sufficient evaluation information to provide a basis for subsequent screening and sampling. In addition, identifying potential trigger conditions in the test case construction stage can provide support for subsequent expected functional safety analysis. (3) Scenario assignment method based on element-scenario evaluation inheritance rules: All element features that constitute the scene are comprehensively considered, and a comprehensive evaluation score is assigned to each scene according to preset rules to achieve effective management and screening of massive scenes. (4) Dual-track screening mechanism based on test requirements and design operating conditions: This mechanism combines on-demand screening with screening based on design operating conditions to ensure that the final output test case subset not only meets the test requirements but also accurately reflects the functional declaration of the system under test, thereby improving the targeted nature of the test.

[0119] See also Figure 3 , Figure 3 This is a schematic diagram of a structure of a device for constructing a scenario library and a test case set in an autonomous driving embodiment provided by the present application. Figure 3 As shown in , the construction device 300 includes:

[0120] A definition module 310 is used to define a scene element set in the autonomous driving system and an index value of each scene element in the scene element set with respect to an evaluation index; wherein the scene element set is a data set organized in a multi-layer structure; each element at the bottom layer of the scene element set represents a scene element, and each upper layer element related to each element at the bottom layer represents a scene label at each level corresponding to the scene element;

[0121] A generating module 320, for combining the scene elements according to the scene labels corresponding to each scene element, generating multiple test scenes and forming a scene library;

[0122] A determination module 330, configured to determine a scenario index value of each test scenario with respect to the evaluation index according to the index value of each scenario element included in each test scenario;

[0123] The screening module 340 is used to screen out target test scenarios from the scenario library and form a test case set according to the test requirements of the autonomous driving system and the scenario indicator value of each test scenario, and / or according to the design operating conditions of the autonomous driving system and the scenario label corresponding to each scenario element.

[0124] Furthermore, when the screening module 340 is used to screen out target test scenarios from the scenario library and form a test case set according to the test requirements of the autonomous driving system and the scenario index value of each test scenario, the screening module 340 is used to:

[0125] Filtering a first test scenario from the scenario library according to the test requirements of the autonomous driving system and the scenario index value of each test scenario;

[0126] Filtering a second test scene from the scene library according to the design operating conditions of the autonomous driving system and the scene label corresponding to each scene element;

[0127] The intersection between the first test scenario and the second test scenario is determined as the target test scenario, and the set of the target test scenarios is used as the test case set.

[0128] Furthermore, when the screening module 340 is used to screen out target test scenarios from the scenario library and form a test case set according to the test requirements of the autonomous driving system and the scenario index value of each test scenario, the screening module 340 is used to:

[0129] Decomposing the test requirements to determine the requirement items and requirement standards included in the test requirements;

[0130] According to the predefined mapping relationship between the requirement items and the evaluation indicators, the requirement items and the requirement standards are converted into screening rules;

[0131] According to the screening rules, the scenario indicator value of each test scenario and the scenario elements included in each test scenario, the target test scenario is screened out from the scenario library to form the test case set.

[0132] Furthermore, when the screening module 340 is used to screen out target test scenarios from the scenario library and form a test case set according to the design operating conditions of the autonomous driving system and the label corresponding to each scenario element, the screening module 340 is used to:

[0133] For each function to be tested in the autonomous driving system, determining a target label and / or a filter label according to a design operating condition corresponding to the function to be tested;

[0134] For each test scene in the scene library, if the scene elements included in the test scene have the target tag and do not have the filter tag, the test scene is determined as a target test scene;

[0135] The determined target test scenario is used as the test case set corresponding to the function to be tested.

[0136] Further, the evaluation index includes: a frequency index, a risk index and a complexity index; when the determination module 330 is used to determine the scenario index value of each test scenario with respect to the evaluation index according to the index value of each scenario element included in each test scenario, the determination module 330 is used to:

[0137] For each test scenario, determining the minimum index value of each scenario element included in the test scenario with respect to the frequency index as the scenario index value with respect to the frequency index of the test scenario;

[0138] Determine the maximum index value of each scenario element included in the test scenario with respect to the risk index as the scenario index value with respect to the risk index of the test scenario;

[0139] The average index value of the complexity index of each scenario element included in the test scenario is determined as the first scenario index value of the complexity index of the test scenario; and the number of scenario elements in the test scenario whose index values ​​of the complexity index are greater than the average index value is determined as the second scenario index value of the complexity index of the test scenario.

[0140] Furthermore, the determination module 330 is further used to: define a probability correction function when a combination of multiple target scene elements in the scene element set appears according to the actual scene situation;

[0141] Then, for the frequency index, when the determination module 330 is used to determine the scene index value of each test scene with respect to the evaluation index according to the index value of each scene element included in each test scene, the determination module 330 is used to:

[0142] For each test scenario, determine whether the test scenario includes any combination of target scenario elements;

[0143] If included, determining a correction index value based on the probability correction function corresponding to each target scene element combination and the index value of each scene element included in each target scene element combination with respect to the frequency index;

[0144] The minimum value between the minimum index value of each scenario element included in the test scenario with respect to the frequency index and the correction index value is determined as the scenario index value of the test scenario with respect to the frequency index.

[0145] Furthermore, the scene tags of each level in the scene element set include a primary tag and a secondary tag subdivided under the primary tag; each secondary tag is associated with a scene element subset; when the generation module 320 is used to combine the scene elements according to the scene tag corresponding to each scene element, generate multiple test scenes and form a scene library, the generation module 320 is used to:

[0146] Cartesian product operations are performed between the scene element subsets under each secondary tag, and scene elements are selected from the scene element subsets under each secondary tag and combined to obtain the multiple test scenes and form the scene library.

[0147] Furthermore, the construction device 300 further includes: an identification module; the identification module is used to:

[0148] According to the index value of each scenario element included in each target test scenario with respect to the risk index, it is identified whether the target test scenario contains a danger triggering condition.

[0149] See also Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 4 As shown in , the electronic device 400 includes a processor 410 , a memory 420 and a bus 430 .

[0150] The memory 420 stores machine-readable instructions executable by the processor 410. When the electronic device 400 is running, the processor 410 communicates with the memory 420 via the bus 430. When the machine-readable instructions are executed by the processor 410, the above-mentioned Figure 1 The steps of the method for constructing a scenario library and a test case set in autonomous driving in the method embodiment shown in the figure, and the specific implementation method can be found in the method embodiment, which will not be repeated here.

[0151] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1 The steps of the method for constructing a scenario library and a test case set in autonomous driving in the method embodiment shown in the figure, and the specific implementation method can be found in the method embodiment, which will not be repeated here.

[0152] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0153] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0154] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0155] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0156] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0157] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for constructing a scenario library and a test case set in autonomous driving, characterized in that: The construction method comprises: Define a scene element set in an autonomous driving system and an index value of each scene element in the scene element set with respect to an evaluation index; wherein the scene element set is a data set organized in a multi-layer structure; each element at the bottom layer of the scene element set represents a scene element, and each upper layer element related to each element at the bottom layer represents scene labels at various levels corresponding to the scene element; The scene elements are combined according to the scene labels corresponding to each scene element, multiple test scenes are generated and a scene library is formed; Determine the scene index value of each test scene with respect to the evaluation index according to the index value of each scene element included in each test scene; According to the test requirements of the autonomous driving system and the scenario index value of each test scenario, and / or according to the design operating conditions of the autonomous driving system and the scenario label corresponding to each scenario element, the target test scenario is screened out from the scenario library and a test case set is formed; According to the test requirements of the autonomous driving system and the scenario index value of each test scenario, according to the design operating conditions of the autonomous driving system and the scenario label corresponding to each scenario element, a target test scenario is screened from the scenario library to form a test case set, including: Filtering a first test scenario from the scenario library according to the test requirements of the autonomous driving system and the scenario index value of each test scenario; Filtering a second test scene from the scene library according to the design operating conditions of the autonomous driving system and the scene label corresponding to each scene element; The intersection between the first test scenario and the second test scenario is determined as the target test scenario, and the set of the target test scenarios is used as the test case set.

2. The construction method according to claim 1, characterized in that: According to the test requirements of the autonomous driving system and the scenario index value of each test scenario, a target test scenario is screened out from the scenario library to form a test case set, including: Decomposing the test requirements to determine the requirement items and requirement standards included in the test requirements; According to the predefined mapping relationship between the requirement items and the evaluation indicators, the requirement items and the requirement standards are converted into screening rules; According to the screening rules, the scenario indicator value of each test scenario and the scenario elements included in each test scenario, the target test scenario is screened out from the scenario library to form the test case set.

3. The construction method according to claim 1, characterized in that: According to the design operating conditions of the autonomous driving system and the labels corresponding to each scenario element, a target test scenario is screened out from the scenario library to form a test case set, including: For each function to be tested in the autonomous driving system, determining a target label and / or a filter label according to a design operating condition corresponding to the function to be tested; For each test scene in the scene library, if the scene elements included in the test scene have the target tag and do not have the filter tag, the test scene is determined as a target test scene; The determined target test scenario is used as the test case set corresponding to the function to be tested.

4. The construction method according to claim 1, characterized in that: The evaluation index includes: frequency index, risk index and complexity index; the determining of the scenario index value of each test scenario with respect to the evaluation index according to the index value of each scenario element included in each test scenario includes: For each test scenario, determining the minimum index value of each scenario element included in the test scenario with respect to the frequency index as the scenario index value with respect to the frequency index of the test scenario; Determine the maximum index value of each scenario element included in the test scenario with respect to the risk index as the scenario index value with respect to the risk index of the test scenario; The average index value of the complexity index of each scenario element included in the test scenario is determined as the first scenario index value of the complexity index of the test scenario; and the number of scenario elements in the test scenario whose index values ​​of the complexity index are greater than the average index value is determined as the second scenario index value of the complexity index of the test scenario.

5. The construction method according to claim 4, characterized in that: The construction method further includes: defining a probability correction function when a combination of multiple target scene elements in the scene element set appears according to actual scene conditions; Then, for the frequency index, determining the scene index value of each test scene with respect to the evaluation index according to the index value of each scene element included in each test scene also includes: For each test scenario, determine whether the test scenario includes any combination of target scenario elements; If included, determining a correction index value based on the probability correction function corresponding to each target scene element combination and the index value of each scene element included in each target scene element combination with respect to the frequency index; The minimum value between the minimum index value of each scenario element included in the test scenario with respect to the frequency index and the correction index value is determined as the scenario index value of the test scenario with respect to the frequency index.

6. The construction method according to claim 1, characterized in that: The scene tags of each level in the scene element set include a primary tag and secondary tags subdivided under the primary tag; each secondary tag is associated with a scene element subset; The step of combining the scene elements according to the scene labels corresponding to each scene element, generating multiple test scenes and forming a scene library includes: Cartesian product operations are performed between the scene element subsets under each secondary tag, and scene elements are selected from the scene element subsets under each secondary tag and combined to obtain the multiple test scenes and form the scene library.

7. The construction method according to claim 4, characterized in that: The construction method also includes: According to the index value of each scenario element included in each target test scenario with respect to the risk index, it is identified whether the target test scenario contains a danger triggering condition.

8. A device for constructing a scenario library and a test case set in autonomous driving, characterized in that: The construction device comprises: A definition module, used to define a scene element set in an autonomous driving system and an index value of each scene element in the scene element set with respect to an evaluation index; wherein the scene element set is a data set organized in a multi-layer structure; each element at the bottom layer of the scene element set represents a scene element, and each upper layer element related to each element at the bottom layer represents a scene label at each level corresponding to the scene element; A generation module, used to combine the scene elements according to the scene labels corresponding to each scene element, generate multiple test scenes and form a scene library; A determination module, used to determine the scene index value of each test scene with respect to the evaluation index according to the index value of each scene element included in each test scene; A screening module, configured to screen target test scenarios from the scenario library and form a test case set according to the test requirements of the autonomous driving system and the scenario index value of each test scenario, and / or according to the design operating conditions of the autonomous driving system and the scenario label corresponding to each scenario element; When the screening module is used to screen out target test scenarios from the scenario library and form a test case set according to the test requirements of the autonomous driving system and the scenario index value of each test scenario, the screening module is used to: Filtering a first test scenario from the scenario library according to the test requirements of the autonomous driving system and the scenario index value of each test scenario; Filtering a second test scene from the scene library according to the design operating conditions of the autonomous driving system and the scene label corresponding to each scene element; The intersection between the first test scenario and the second test scenario is determined as the target test scenario, and the set of the target test scenarios is used as the test case set.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor when running to execute the steps of the method for constructing a scenario library and a test case set in autonomous driving as described in any one of claims 1 to 7.

Images