Scene priority determination method and device

Abstract

The invention discloses a scene priority determination method and device, and the method comprises the steps: obtaining the evaluation information of a driving scene, and enabling the evaluation information of the driving scene to be used for describing the priorities of a plurality of scene features in the driving scene; and determining the priorities of the plurality of driving scenes according to the preset priorities of the plurality of scene features and the plurality of pieces of evaluation information of the plurality of driving scenes. Therefore, the accuracy of the priority of the evaluated driving scene can be improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent driving, and in particular to a method and device for determining scene priority. Background Art

[0002] Scenario priority is crucial for driving scenario testing. For example, accurately evaluating scenario priority can improve driving scenario testing efficiency while reducing testing resources and testing costs. Currently, the priority of driving scenarios is simply determined by multiplying multiple quantifiable indicators. The accuracy of the priority of test scenarios evaluated in this way is not high. Summary of the invention

[0003] The present application discloses a method and device for determining a scene priority, which can improve the accuracy of the priority of an evaluated driving scene.

[0004] In a first aspect, the present application provides a method for determining scene priorities, the method comprising: obtaining evaluation information of a driving scene, wherein the evaluation information of the driving scene is used to describe the priority of multiple scene features in the driving scene; determining the priorities of the M driving scenes according to the preset priorities of the multiple scene features and M evaluation information of the M driving scenes, where M is an integer greater than 1.

[0005] Exemplarily, scene features may also be referred to as scene assessment criteria, scene evaluation criteria, or scene evaluation parameters.

[0006] Here, the priority of the scene feature in the driving scene is used to indicate the importance of the classification type to which the scene feature belongs in the driving scene. It can be understood that for the same scene feature, the priority of the scene feature in different driving scenes may be different. In addition, different scene features have different classification types. The classification type of the scene feature makes the description of the driving scene of the scene feature more refined.

[0007] Here, the preset priority of the scene feature is used to indicate the importance of the scene feature among the above multiple scene features. The preset priority of the scene feature is independent of the driving scene, that is, the preset priority of the scene feature remains unchanged in different driving scenes. The preset priority of the scene feature can be used to know the importance ranking among the above multiple scene features.

[0008] Here, the driving scene is understood as the environment in which the vehicle is located, which includes one or more of the information such as road type, weather conditions, traffic conditions, driving tasks, etc. Typical driving scenes may be, for example, a vehicle cut-in scene, a vehicle cut-out scene, a vehicle following scene, or a front vehicle braking scene.

[0009] Here, the M evaluation information of the M driving scenarios means that each driving scenario corresponds to one evaluation information, so there are M evaluation information in total for the M driving scenarios.

[0010] In the above method, the importance ranking among multiple scene features is indicated by the preset priorities of these multiple scene features. The priorities of these multiple scene features in different driving scenarios can be used to more finely characterize different driving scenarios. In this way, the priorities of multiple driving scenarios are comprehensively determined from two dimensions: the preset priorities of scene features and the priorities of scene features in driving scenarios, which can improve the evaluation accuracy of the priorities of driving scenarios.

[0011] Exemplarily, in the evaluation information of the driving scene, the priorities of the multiple scene features in the driving scene are arranged according to the size relationship between the preset priorities of the multiple scene features. In this way, the evaluation information of the driving scene not only describes the priorities of the multiple scene features in the driving scene, but also indicates the size relationship between the preset priorities of the multiple scene features by specifying the arrangement order of the priorities of the multiple scene features in the driving scene, which is conducive to improving the evaluation efficiency of the priority of the driving scene.

[0012] Optionally, the method further includes: selecting N driving scenarios from the M driving scenarios for testing using the priorities of the M driving scenarios, wherein N is a positive integer less than M.

[0013] By implementing the above implementation, some driving scenarios are selected from a large number of driving scenarios for testing according to the priority of the driving scenarios, which can improve the testing effect of the driving scenarios. Furthermore, the priority of the driving scenarios can affect the test order. The driving scenarios with higher priorities can be tested first, which is conducive to saving test resources and test costs.

[0014] Optionally, the method further includes: when the test results of the N driving scenarios do not meet the test objectives, adjusting the preset priorities of the multiple scenario features according to the test results of the N driving scenarios.

[0015] Exemplarily, adjusting the preset priorities of the plurality of scene features based on the test result may be: adjusting the preset priorities of some or all of the plurality of scene features based on the test result.

[0016] Exemplarily, the test target is related to the number of scenario features involved in evaluating the scenario priority. The test target includes sub-test targets corresponding to each scenario feature in the plurality of scenario features.

[0017] For example, the above-mentioned multiple scene features include scene feature 1, scene feature 2 and scene feature 3, wherein the preset priorities of these multiple scene features are arranged from high to low as follows: scene feature 1>scene feature 2>scene feature 3. Taking the test results of the above-mentioned N driving scenes as an example that do not meet the sub-test targets corresponding to scene feature 3 in the above-mentioned test targets, in this case, adjusting the preset priorities of these multiple scene features according to the test results can be: increasing the preset priority of scene feature 3 among these multiple scene features according to the test results.

[0018] By implementing the above implementation method, the preset priorities of multiple scene features participating in the scene priority evaluation are adjusted through test result feedback to change the importance ranking between at least two scene features, so that the test results of the driving scene can be more in line with the test objectives.

[0019] Optionally, the multiple scene features include the following: scene complexity, scene danger, scene exposure, scene rationality and scene accuracy.

[0020] Exemplarily, the scenario hazard describes the possible impact on the safety of traffic participants and passengers of the autonomous driving system; the scenario exposure is used to describe the frequency of occurrence or appearance of driving scenarios; the scenario complexity is used to describe the degree of difficulty caused by various scenario attributes that may affect the driving task of the autonomous driving system (such as the behavior of participating road users, defined road infrastructure and environmental conditions, etc.); the scenario rationality is used to describe the degree of conformity between the scenario parameter data of the driving scenario used in the simulation test and the real physical world; the scenario accuracy is used to describe the degree of fit or similarity between the scenario extracted based on driving data and the actual scenario.

[0021] Optionally, the preset priorities of the multiple scene features are set based on the confidence levels of the multiple scene features or historical testing experience.

[0022] Exemplarily, the higher the confidence of a scene feature, the higher the preset priority of the scene feature can be set. In addition, historical test experience can indicate the setting of the preset priority of each scene feature corresponding to a small number of test results. In another implementation, the preset priority of multiple scene features can also be set based on multiple factors including the confidence of the scene feature, historical test experience (or prior experience), and user wishes, which are not specifically limited here.

[0023] By implementing the above implementation method, the dimension of preset priority of scene features is taken into consideration when evaluating the priority of driving scenes, which is conducive to improving the evaluation accuracy of scene priority.

[0024] Optionally, the multiple scene features include a first feature and a second feature, wherein a preset priority of the first feature is higher than a preset priority of the second feature; the M driving scenes include a first driving scene and a second driving scene, and the priority of the first driving scene and the priority of the second driving scene satisfy any of the following conditions:

[0025] If the priority of the first feature in the first driving scenario is higher than the priority of the first feature in the second driving scenario, the priority of the first driving scenario is higher than the priority of the second driving scenario; or,

[0026] If the priority of the first feature in the first driving scenario is the same as the priority of the first feature in the second driving scenario and the priority of the second feature in the first driving scenario is higher than the priority of the second feature in the second driving scenario, the priority of the first driving scenario is higher than the priority of the second driving scenario; or,

[0027] If the priority of the first feature in the first driving scenario is the same as the priority of the first feature in the second driving scenario and the priority of the second feature in the first driving scenario is the same as the priority of the second feature in the second driving scenario, the priority of the first driving scenario is the same as the priority of the second driving scenario.

[0028] That is to say, the priority of each driving scene can be determined in descending order according to the preset priority of the scene features. Assuming that the preset priority of scene feature 1 is the highest, the preset priority of scene feature 2 is the second, and the preset priority of scene feature 3 is the lowest, and the total number of driving scenes is M, where M is an integer greater than 1, the priorities of scene feature 1 in the M driving scenes are first compared in sequence, and the higher the priority of scene feature 1 in the driving scene, the higher the priority of the driving scene; if the priorities of scene feature 1 in P driving scenes in the M driving scenes are the same, and P is a positive integer less than M, then the priorities of scene feature 2 in the P driving scenes are compared in sequence, and the higher the priority of scene feature 2 in the driving scene, the higher the priority of the driving scene; if the priorities of scene feature 2 in L driving scenes in the P driving scenes are the same, and L is a positive integer less than P, then the priorities of scene feature 3 in the L driving scenes are compared in sequence, and the higher the priority of scene feature 3 in the driving scene, the higher the priority of the driving scene.

[0029] By implementing the above implementation method, the priorities of the corresponding scene features in each driving scene are compared in order from high to low according to the preset priorities of the scene features to determine the priority of each driving scene, thereby improving the accuracy of the priority of the evaluated driving scene.

[0030] Optionally, the total number of the multiple scene features is S, the preset priority of the i-th scene feature is greater than the preset priority of the i+1-th scene feature, i is an integer greater than or equal to 1 and less than S, and the priorities of the M driving scenes are determined according to the preset priorities of the multiple scene features and the M evaluation information of the M driving scenes, including: performing statistics on the preset priorities of the multiple scene features and the M evaluation information through a sorting function to determine the priorities of the M driving scenes; wherein the sorting function is the sum of S parts, the i-th part is represented by the i-th coefficient and the priority of the i-th scene feature in the driving scene, and the value of each coefficient in the sorting function satisfies the following constraint: the minimum value of the i-th part is greater than the maximum value of the sum of the i+1-th part, the i+2-th part, ..., the S-th part among the S parts.

[0031] In one implementation, the i-th part is represented by the i-th coefficient and the priority of the i-th scene feature in the driving scenario, which means that the i-th part is the product of the i-th coefficient and the priority of the i-th scene feature in the driving scenario.

[0032] In another implementation, the i-th part is represented by the i-th coefficient and the priority of the i-th scene feature in the driving scenario, which means that the i-th part is represented as a function with the i-th coefficient as the base and the priority of the i-th scene feature in the driving scenario as the exponent.

[0033] It can be understood that the evaluation information of the driving scene is substituted into the above-mentioned ranking function to obtain the ranking function value corresponding to the driving scene. The larger the ranking function value corresponding to the driving scene is, the higher the priority of the driving scene is.

[0034] By implementing the above implementation method, the ranking function can be defined to quickly calculate the ranking function value corresponding to each driving scenario, and the size of the ranking function value corresponding to each driving scenario can be compared to determine the priority of each driving scenario. This not only improves the evaluation accuracy of the scenario priority, but also improves the evaluation efficiency of the scenario priority.

[0035] Optionally, the preset priorities of the multiple scene features are input by the user.

[0036] By implementing the above implementation, a manual interaction interface may also be provided to the user, so that the user may set preset priorities for the multiple scene features respectively through the manual interaction interface.

[0037] In a second aspect, the present application provides a device for determining scene priorities, the device comprising: an acquisition unit, used to acquire evaluation information of a driving scene, wherein the evaluation information of the driving scene is used to describe the priority of multiple scene features in the driving scene; a processing unit, used to determine the priorities of the M driving scenes based on the preset priorities of the multiple scene features and M evaluation information of the M driving scenes, where M is an integer greater than 1.

[0038] Optionally, in the evaluation information of the driving scene, the priorities of the multiple scene features in the driving scene are arranged according to the size relationship between the preset priorities of the multiple scene features.

[0039] Optionally, the device further includes a testing unit, which is used to select N driving scenarios from the M driving scenarios for testing using the priorities of the M driving scenarios, wherein N is a positive integer less than M.

[0040] Optionally, the testing unit is further used to: when the test results of the N driving scenarios do not meet the test objectives, adjust the preset priorities of the multiple scenario features according to the test results of the N driving scenarios.

[0041] Optionally, the multiple scene features include the following: scene complexity, scene danger, scene exposure, scene rationality and scene accuracy.

[0042] Optionally, the preset priorities of the multiple scene features are set based on the confidence levels of the multiple scene features or historical testing experience.

[0043] Optionally, the multiple scene features include a first feature and a second feature, wherein a preset priority of the first feature is higher than a preset priority of the second feature; the M driving scenes include a first driving scene and a second driving scene, and the priority of the first driving scene and the priority of the second driving scene satisfy any of the following conditions:

[0044] If the priority of the first feature in the first driving scenario is higher than the priority of the first feature in the second driving scenario, the priority of the first driving scenario is higher than the priority of the second driving scenario; or,

[0045] If the priority of the first feature in the first driving scenario is the same as the priority of the first feature in the second driving scenario and the priority of the second feature in the first driving scenario is higher than the priority of the second feature in the second driving scenario, the priority of the first driving scenario is higher than the priority of the second driving scenario; or,

[0046] If the priority of the first feature in the first driving scenario is the same as the priority of the first feature in the second driving scenario and the priority of the second feature in the first driving scenario is the same as the priority of the second feature in the second driving scenario, the priority of the first driving scenario is the same as the priority of the second driving scenario.

[0047] Optionally, the total number of the multiple scene features is S, the preset priority of the i-th scene feature is greater than the preset priority of the i+1-th scene feature, i is an integer greater than or equal to 1 and less than S, and the processing unit is specifically used to: perform statistics on the preset priorities of the multiple scene features and the M evaluation information through a sorting function to determine the priorities of the M driving scenes; wherein the sorting function is the sum of S parts, the i-th part is represented by the i-th coefficient and the priority of the i-th scene feature in the driving scene, and the value of each coefficient in the sorting function satisfies the following constraints: the minimum value of the i-th part is greater than the maximum value of the sum of the i+1-th part, the i+2-th part, ..., the S-th part among the S parts.

[0048] Optionally, the preset priorities of the multiple scene features are input by the user.

[0049] In a third aspect, the present application provides a device for determining scene priority, the device comprising a processor and a memory, wherein the memory is used to store program instructions; the processor calls the program instructions in the memory so that the device executes the method in the first aspect or any possible implementation of the first aspect.

[0050] In a fourth aspect, the present application provides a computer-readable storage medium, comprising computer instructions, which, when executed by a processor, implement the method in the above-mentioned first aspect or any possible implementation manner of the first aspect.

[0051] In a fifth aspect, the present application provides a computer program product, which, when executed by a processor, implements the method in the first aspect or any possible embodiment of the first aspect. The computer program product, for example, may be a software installation package, and when the method provided by any possible design of the first aspect is needed, the computer program product may be downloaded and executed on a processor to implement the method in the first aspect or any possible embodiment of the first aspect.

[0052] The technical effects of the second to fifth aspects mentioned above can be referred to the description of the first aspect mentioned above, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1is a schematic diagram of the architecture of a priority determination device provided in an embodiment of the present application;

[0054] Figure 2 is a flow chart of a method for determining scene priority provided by an embodiment of the present application;

[0055] Figure 3 is a schematic diagram of a human interaction interface provided in an embodiment of the present application;

[0056] Figure 4 is a schematic diagram of a process for determining priorities of multiple driving scenarios provided in an embodiment of the present application;

[0057] Figure 5 It is a schematic diagram of another process for determining the priorities of multiple driving scenarios provided in an embodiment of the present application;

[0058] Figure 6 is a flowchart of another scene priority determination method provided in an embodiment of the present application;

[0059] Figure 7 It is a structural diagram of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0060] It should be noted that the prefixes such as "first" and "second" used in this application are only used to distinguish different description objects, and have no limiting effect on the position, order, priority, quantity or content of the described objects. For example, if the described object is a "field", the ordinal number before the "field" in the "first field" and the "second field" does not limit the position or order between the "fields". "First" and "second" do not limit whether the "fields" they modify are in the same message, nor do they limit the order of the "first field" and the "second field". For another example, if the described object is a "level", the ordinal number before the "level" in the "first level" and the "second level" does not limit the priority between the "levels". For another example, the number of described objects is not limited by the prefix and can be one or more. Taking "first device" as an example, the number of "devices" can be one or more. In addition, the objects modified by different prefixes can be the same or different. For example, if the object being described is a "device", then the "first device" and the "second device" can be the same device, the same type of device, or different types of devices; for another example, if the object being described is "information", then the "first information" and the "second information" can be information of the same content or information of different contents. In short, the use of prefixes used to distinguish the described objects in the embodiments of the present application does not constitute a limitation on the described objects. For the statement of the described objects, refer to the description in the context of the claims or embodiments, and no unnecessary limitation should be constituted due to the use of such prefixes.

[0061] It should be noted that the description methods such as "at least one of a1, a2, ... and an" used in the embodiments of the present application include the situation where any one of a1, a2, ... and an exists alone, and also include any combination of any multiple of a1, a2, ... and an, and each situation can exist alone. For example, the description method of "at least one of a, b and c" includes the situation where a is alone, b is alone, c is alone, a combination of a and b, a combination of a and c, b and c, or a combination of abc.

[0062] The technical solution in this application will be described below in conjunction with the accompanying drawings.

[0063] See also Figure 1 , Figure 1 is a schematic diagram of a structure of a priority determination device provided in an embodiment of the present application. The priority determination device is used to determine the priority of a driving scene. Figure 1 As shown, the priority determination device includes an acquisition unit and an output unit, wherein the acquisition unit and the output unit can communicate with each other in a wireless or wired manner. The priority determination device can be implemented in hardware, software, or a combination of hardware and software.

[0064] Exemplarily, the acquisition unit is used to obtain evaluation information of a driving scene, where the evaluation information of the driving scene is used to describe the priority of multiple scene features in the driving scene; the processing unit is used to determine the priority of the M driving scenes based on the preset priorities of the multiple scene features and M evaluation information of the M driving scenes, where M is an integer greater than 1.

[0065] Here, the driving scene can be understood as the environment in which the vehicle is located, which includes one or more of the information such as road type, weather conditions, traffic conditions, driving tasks, etc. Typical driving scenes can be, for example, a vehicle cut-in scene, a vehicle cut-out scene, a vehicle following scene, or a front vehicle braking scene.

[0066] Here, the scene feature may also be referred to as a scene assessment criterion, a scene evaluation criterion or a scene evaluation parameter.

[0067] Here, the priority of the scene feature in the driving scene can also be called the sub-priority of the scene feature. For the convenience of description, the sub-priority of the scene feature described below represents the priority of the scene feature in the driving scene.

[0068] It can be seen that the sub-priority of the scene feature is related to the driving scene. For the same scene feature, the sub-priority of the scene feature may be different in different driving scenes. The preset priority of the scene feature is independent of the driving scene, and the preset priority of the same scene feature in different driving scenes is the same. The sub-priority of the scene feature and the preset priority of the scene feature can refer to the relevant description of the following embodiment, and for the sake of brevity of the specification, it will not be repeated here.

[0069] Exemplarily, the priority determination device may be a network side device. The network side device may be, for example, a server deployed on the network side (e.g., a scenario priority evaluation server, a simulation test server, etc.), or a component or chip in the server, or a system-level device composed of multiple servers. The network side device may be deployed in a cloud environment or an edge environment, which is not specifically limited herein.

[0070] Exemplarily, the scene priority determination method provided in the embodiments of the present application can be applied not only to autonomous driving simulation systems, but also to other simulation test systems that require scene priority, such as model-in-the-loop test systems, hardware-in-the-loop test systems, etc.

[0071] Here, "autonomous driving" is not limited to fully autonomous driving systems, highly autonomous driving systems, conditional autonomous driving systems, or partially autonomous driving systems. Those skilled in the art can understand that non-fully manual driving systems that provide intelligent driving can be included in this concept for simulation processing.

[0072] It should be noted that Figure 1 This is only an exemplary architecture diagram of the priority determination device, but does not limit the priority determination device to only Figure 1 The structural diagram of the priority determination device may also include Figure 1 In addition, the method provided in the embodiment of the present application can be applied to Figure 1 The priority determination device shown, of course, the method provided in the embodiment of the present application can also be applied to other priority determination devices, and the embodiment of the present application is not limited to this.

[0073] Before introducing the scene priority determination method provided in the embodiment of the present application, it is necessary to first determine a plurality of scene features for evaluating the scene priority.

[0074] In one implementation, the multiple scenario features involved in the scenario priority assessment may be obtained from the scenario features defined by industry standards and / or expert experience. For example, the multiple scenario features involved in the scenario priority assessment may be obtained by screening from the scenario features defined by industry standards and / or expert experience based on the confidence of the scenario features, or may be custom set by the tester. It is understood that the above multiple scenario features are pre-set.

[0075] For example, the scenario characteristics defined by the industry standard include scenario criticality, scenario exposure and scenario complexity. Among them, scenario criticality is used to describe the possible impact on the safety of traffic participants and passengers of the autonomous driving system, scenario exposure is used to describe the frequency of occurrence or appearance of driving scenarios, and scenario complexity is used to describe the difficulty caused by various scenario attributes (such as the behavior of participating road users, defined road infrastructure and environmental conditions, etc.) that may affect the driving task of the autonomous driving system.

[0076] Exemplarily, for simulation tests, the scenario features defined by expert experience also include scenario rationality and scenario accuracy, wherein scenario rationality is used to describe the degree of conformity between the scenario parameter data of the driving scenario used in the simulation test and the real physical world, and scenario accuracy is used to describe the degree of fit or similarity between the scenario extracted based on driving data and the actual scenario.

[0077] Here, the number of multiple scene features involved in scene priority evaluation is not limited, for example, it can be 2, 3, 4 or more.

[0078] The following introduces two priorities of scene features defined in the embodiments of the present application: a preset priority of scene features and a sub-priority of scene features.

[0079] (1) Preset priority of scene features

[0080] Exemplarily, the preset priority of a scene feature may also be referred to as the block priority of the scene feature.

[0081] The preset priority of a scene feature is used to indicate the importance of the scene feature among the above-mentioned multiple scene features, and the higher the importance of the scene feature among the above-mentioned multiple scene features, the higher the preset priority of the scene feature.

[0082] For example, multiple scene features include scene feature 1, scene feature 2, and scene feature 3. Assuming that the preset priority of scene feature 1 is higher than the preset priority of scene feature 2, and the preset priority of scene feature 2 is higher than the preset priority of scene feature 3, it means that the importance of these three scene features is in the order of high to low: scene feature 1> scene feature 2> scene feature 3. It can be seen that the importance ranking among multiple scene features involved in evaluating scene priority can be known through the preset priority, that is, the influence of each scene feature on scene evaluation is considered from the dimension of scene feature dimension through the preset priority, which is conducive to improving the accuracy of evaluating scene priority.

[0083] In one implementation, the preset priorities of the above-mentioned multiple scene features are set based on the confidence level of the scene features or historical test experience, and can also be defined by the user based on the test objectives or his own wishes, which is not specifically limited here.

[0084] Taking multiple scene features including scene complexity, scene danger and scene exposure as an example, illustratively, based on historical testing experience or prior experience, it can be determined that safety is given priority and test complexity is secondarily considered. The preset priority of scene danger can be set higher than the preset priority of scene complexity, and the preset priority of scene complexity can be set higher than the preset priority of scene exposure. Exemplarily, the preset priorities of multiple scene features can also be set based on user wishes in order from high to low: scene danger > scene exposure > scene complexity. Exemplarily, the preset priorities of the above-mentioned multiple scene features can also be set based on the principle that the greater the confidence of a scene feature, the higher the preset priority of the scene feature. In some possible embodiments, for example, the preset priorities of multiple scene features can also be determined in combination with multiple items of the confidence of the scene feature, user wishes, historical testing experience, etc.

[0085] (2) Sub-priorities of scene features

[0086] The sub-priority of a scene feature is the priority of the scene feature in the driving scene.

[0087] Here, the sub-priority of the scene feature is used to indicate the importance of the classification type to which the scene feature belongs in the driving scene. For any driving scene, the higher the importance of the classification type to which the scene feature belongs in the driving scene, the higher the sub-priority of the scene feature. It can be understood that if the same scene feature belongs to different classification types in different driving scenes, it also means that the sub-priority of the scene feature in different driving scenes is also different.

[0088] Exemplarily, the classification type of the scene feature may be indicated by a numerical value or by a character.

[0089] Taking the numerical representation of the classification type as an example, the scene feature is the scene danger degree. Assuming that the scene danger degree has three classification types, namely basically no danger, slight danger and very dangerous, assuming that the higher the importance of the classification type to which the scene danger degree belongs in the driving scene, the greater the sub-priority value of the scene danger degree. When the classification type of the scene feature is basically no danger, the sub-priority value of the scene danger degree is 1; when the classification type of the scene feature is slightly dangerous, the sub-priority value of the scene danger degree is 2; when the classification type of the scene feature is very dangerous, the sub-priority value of the scene danger degree is 3. It can be seen that the importance or influence of the same scene feature can be further divided through the sub-priority, which can more finely characterize the importance or influence of the scene feature on the evaluation of different driving scenarios.

[0090] It can be understood that the three classification types (i.e., not dangerous, slightly dangerous, and very dangerous) of the scene danger level described above are only an example of classification. In some possible embodiments, the scene danger level may also have four classification types, namely, not dangerous, slightly dangerous, moderately dangerous, and very dangerous. In addition, different scene features may also have different numbers of classification types, and the number of classification types of each scene feature is not limited to be equal or unequal.

[0091] Exemplarily, the classification of scene features may be based on expert experience or measured data (such as motion trajectory data, traffic accident data, etc.), or a quantitative method may be used to classify scene features, which is not specifically limited here.

[0092] See also Figure 2 , Figure 2 is a flow chart of a scene priority determination method provided by an embodiment of the present application. The method can be applied to Figure 1 The processing device shown. The method includes but is not limited to the following steps:

[0093] S201: Obtain preset priorities of multiple scene features.

[0094] Here, the plurality of scene features are preset priorities for evaluating the driving scene. The preset priority of the scene feature is used to indicate the importance of the scene feature among the plurality of scene features.

[0095] The preset priorities of the above-mentioned driving scenarios and scenario features can refer to the description of the corresponding contents in the above-mentioned embodiments, which will not be repeated here.

[0096] Exemplarily, the multiple scene features include but are not limited to multiple ones of scene complexity, scene danger, scene exposure, scene rationality, and scene accuracy, etc. Here, the description of each scene feature can refer to the description of the corresponding scene feature in the above embodiment, which will not be repeated here.

[0097] In one implementation, obtaining the preset priorities of the plurality of scene features includes: receiving the preset priorities of the plurality of scene features input by a user, where the user may be, for example, a scene tester, an autonomous driving system developer, or the like.

[0098] It can be understood that obtaining the preset priorities of multiple scene features can be obtaining the values ​​of the preset priorities of multiple scene features, or can be obtaining the size relationship between the preset priorities of multiple scene features, which is not specifically limited here. When obtaining the values ​​of the preset priorities of multiple scene features, the size relationship between the preset priorities of the multiple scene features can be determined by the values ​​of the preset priorities of the multiple scene features, for example, the larger the value of the preset priority of a scene feature, the higher the preset priority of the scene feature.

[0099] Exemplarily, the preset priorities of multiple scene features input by the user may be received through any method such as a touch screen, a keyboard, a button, or voice.

[0100] See also Figure 3 , Figure 3 is a schematic diagram of a human interaction interface provided in an embodiment of the present application. Figure 3 It can be seen that Figure 3 A setting interface for the preset priority of scene features is shown, which includes an input box corresponding to the number of scene features participating in the evaluation, an input box corresponding to the ranking of the preset priorities of scene features from high to low, etc., wherein the user can define the number of scene features through the input box corresponding to the number of scene features participating in the evaluation, and can define the size relationship between the preset priorities of multiple scene features participating in the evaluation through the input box corresponding to the ranking of the preset priorities of scene features from high to low.

[0101] As you can see, Figure 3 The setting interface of the preset priority of the scene feature also includes a "Clear" key and a "Confirm" key, wherein the "Clear" key indicates a request to clear the information entered by the user on the setting interface of the preset priority of the scene feature, and the "Confirm" key indicates saving the information entered by the user on the setting interface of the preset priority of the scene feature.

[0102] Understandably, Figure 3 The setting interface of the preset priority of the scene feature is only an example, and does not limit the setting interface of the preset priority of the scene feature to only Figure 3As shown, for example, it may contain more or less information than the current preset priority setting interface. Figure 3 The input box corresponding to the number of scene features participating in the evaluation can be deleted, or, in the setting interface of the preset priority of the scene features, an input box corresponding to the scene features and an input box corresponding to the preset priority of the scene features can be added. Figure 3 In the setting interface of the preset priority of the scene features shown, the input boxes corresponding to the sorting of the preset priorities of the scene features from high to low are replaced by the input boxes corresponding to the scene features and the input boxes corresponding to the preset priorities of the scene features.

[0103] Figure 3 The artificial interaction interface shown can be presented on a touch screen. Figure 3 The following is an example of a user inputting preset priorities of multiple scene features through a touch screen. Figure 3 In the example, it is assumed that the user enters "3" in the input box corresponding to the number of scene features involved in the evaluation, and enters "scene danger, scene exposure, scene complexity" in the input boxes corresponding to the preset priorities of the scene features from high to low, and then clicks the "confirm" key. In this case, the control device can obtain that the scene features involved in the evaluation include scene danger, scene exposure, and scene complexity, among which the preset priority of scene danger is higher than the preset priority of scene exposure, and the preset priority of scene exposure is higher than the preset priority of scene complexity.

[0104] S202: Acquire evaluation information of a driving scene, where the evaluation information of the driving scene is used to describe the priorities of the above-mentioned multiple scene features in the driving scene.

[0105] The above sub-priorities can refer to the description of the corresponding contents of the above embodiments, which will not be repeated here.

[0106] The following scheme is exemplified by taking multiple scene features including scene complexity, scene danger and scene exposure as examples, assuming that the preset priorities of these multiple scene features are excluded from high to low as follows: preset priority of scene danger > preset priority of scene exposure > preset priority of scene complexity. The definitions of the sub-priorities of these three scene features can be referred to as shown in the following Table 1.

[0107] See Table 1, which shows the corresponding relationship between scene features and sub-priority values. As can be seen from Table 1, the sub-priorities of scene danger, scene exposure and scene complexity are all defined with 4 values.

[0108] For example, for the scene feature of scene exposure shown in Table 1, when the sub-priority value of scene exposure is "4", it means that the classification type of scene exposure is very high probability of occurrence; when the sub-priority value of scene exposure is "3", it means that the classification type of scene exposure is medium probability of occurrence; when the sub-priority value of scene exposure is "2", it means that the classification type of scene exposure is low probability of occurrence; when the sub-priority value of scene exposure is "1", it means that the classification type of scene exposure is basically not occurring.

[0109] For example, for the feature of scene complexity shown in Table 1, when the sub-priority value of scene complexity is "4", it means that the classification type of scene complexity is basically not complex; when the sub-priority value of scene complexity is "3", it means that the classification type of scene complexity is slightly complex; when the sub-priority value of scene complexity is "2", it means that the classification type of scene complexity is moderately complex; when the sub-priority value of scene complexity is "1", it means that the classification type of scene complexity is very complex.

[0110] In Table 1, the larger the sub-priority value of the scene feature is, the higher the sub-priority of the scene feature is. For example, when the sub-priority value of the scene danger is "4", compared with other values ​​of the sub-priority of the scene danger (such as "1", "2" or "3"), the sub-priority of the scene danger is the highest; when the sub-priority value of the scene danger is "1", compared with other values ​​of the sub-priority of the scene danger (such as "4", "2" or "3"), the sub-priority of the scene danger is the lowest; the sub-priority when the scene danger is "moderately dangerous" (that is, the sub-priority value is "3") is higher than the sub-priority when the scene danger is "slightly dangerous" (that is, the sub-priority value is "2").

[0111] Table 1

[0112]

[0113] It can be understood that Table 1 is only to clearly and intuitively present the possible division forms of the sub-priorities of scene features, and does not limit the correspondence between scene features and sub-priority values ​​to only those shown in Table 1. In one implementation, different scene features may have different numbers of corresponding classification types. For example, the scene feature of scene danger has three classification types, so the sub-priority of scene danger has three values; the scene feature of scene exposure has four classification types, so the sub-priority of scene exposure has four values. In another implementation, it can also be defined that the smaller the sub-priority value of the scene feature is, the higher the sub-priority of the scene feature is. In this case, taking the scene danger as an example, when the classification type of the scene danger is "very dangerous", the sub-priority value of the scene danger is "1"; when the classification type of the scene danger is "moderately dangerous", the sub-priority value of the scene danger is "2"; when the classification type of the scene danger is "slightly dangerous", the sub-priority value of the scene danger is "3"; when the classification type of the scene danger is "basically no danger", the sub-priority value of the scene danger is "4".

[0114] In some possible embodiments, in the evaluation information of the driving scene, the sub-priorities of the multiple scene features are arranged according to the size relationship between the preset priorities of the multiple scene features. For example, the size relationship between the preset priorities of the multiple scene features is the ascending order of the preset priorities of the multiple scene features or the descending order of the preset priorities of the multiple scene features.

[0115] Exemplarily, assuming that the preset priorities of multiple scene features are arranged from high to low as follows: pre-priority of scene hazard > pre-priority of scene exposure > pre-priority of scene complexity. Based on the definition of sub-priority of each scene feature shown in Table 1, driving scene 1 belongs to the above-mentioned M driving scenes. Assuming that the evaluation information of driving scene 1 is represented as "431", if the sub-priorities of these multiple scene features in the evaluation information of driving scene 1 are arranged in descending order of the preset priorities of these multiple scene features, it means that in driving scene 1, the sub-priority value of scene hazard is "4", the sub-priority value of scene exposure is "3", and the sub-priority value of scene complexity is "1"; if the sub-priorities of these multiple scene features in the evaluation information of driving scene 1 are arranged in ascending order of the preset priorities of these multiple scene features, it means that in driving scene 1, the sub-priority value of scene complexity is "4", the sub-priority value of scene exposure is "3", and the sub-priority value of scene hazard is "1".

[0116] In one implementation, obtaining evaluation information of a driving scene includes: receiving evaluation information of a driving scene input by a user. It can be understood that the amount of evaluation information of driving scenes received is related to the number of driving scenes, each driving scene corresponds to one evaluation information, M driving scenes have M evaluation information, and M is an integer greater than 1.

[0117] Here, users can be scenario testers, autonomous driving system developers, etc.

[0118] Exemplarily, the evaluation information of the driving scene input by the user may be received through any method such as a touch screen, a keyboard, a button, or voice.

[0119] See also Figure 3 , Figure 3 is a schematic diagram of a human interaction interface provided in an embodiment of the present application. Figure 3 It can be seen that Figure 3 Not only is the setting interface of the preset priority of the scene feature shown (i.e., the solid line box above the dotted line box), but also the setting interface of the sub-priority of the scene feature (i.e., the priority of the scene feature in the driving scene) is shown (i.e., the dotted line box). The setting interface of the preset priority of the scene feature can refer to the description of the corresponding content in the above S201, which will not be repeated here.

[0120] exist Figure 3 In the setting interface of the sub-priority of the scene feature, a selection box corresponding to the number of the driving scene and an input box for the priority of the corresponding scene feature in the corresponding driving scene are included. The number of the driving scene includes the number of each driving scene to be evaluated, for example, the number of the driving scene is 1, 2, 3, ..., M. On the setting interface of the sub-priority of the scene feature, when the user selects the number "2" of the driving scene, it means that the driving scene 2 is selected. Figure 3 The input box on the right side of each scene feature listed in the dashed box is used to set the priority (i.e., sub-priority) of the corresponding scene feature in driving scene 2. Here, each scene feature displayed on the sub-priority setting interface of the scene feature is based on Figure 3 Users in Figure 3 The preset priorities of the scene features shown are generated by the settings on the setting interface.

[0121] It can be seen that Figure 3 The setting interface of the sub-priority of the scene feature also includes a "clear" key and a "confirm" key, wherein the "clear" key indicates a request to clear the priorities of the scene features set by the user in the currently selected driving scene, and the "confirm" key indicates saving the priorities of the scene features set by the user in the currently selected driving scene. It can be understood that when the user clicks the "confirm" key, the evaluation information of the currently selected driving scene is generated.

[0122] The following is based on Figure 3 The following is an example of a user inputting driving scene evaluation information through a touch screen. Figure 3 In the figure, the driving characteristics that the driving user has set to participate in the evaluation include "scenario hazard, scene exposure, and scene complexity". For example, the user selects the driving scene number "2", that is, the user selects driving scene 2. Accordingly, the right side presents the input boxes for the priorities of each scene feature (i.e., scene hazard, scene exposure, and scene complexity) under driving scene 2. Assuming that the user enters "2" in the input box corresponding to the scene hazard, enters "3" in the input box corresponding to the scene exposure, enters "3" in the input box corresponding to the scene complexity, and clicks the "Confirm" button, it means that the evaluation information of driving scene 2 is generated, and the evaluation information of driving scene 2 is used to describe that the priority of scene hazard under driving scene 2 is 2, the priority of scene exposure under driving scene is 3, and the priority of scene complexity under driving scene 2 is 2.

[0123] Understandably, Figure 3 The setting interface of the sub-priority of the scene feature shown is only an example, and does not limit the setting interface of the sub-priority of the scene feature to only Figure 3 As shown, when the number of driving scenarios is multiple, for example Figure 3 The setting interface of the sub-priority of the scene features shown may also not have a driving scene number selection box, but directly display an input box for each driving scene and the priority of each scene feature in the corresponding driving scene, or, for example, directly set a text information import box, through which evaluation information of at least one driving scene can be imported.

[0124] In some possible embodiments, Figure 3 The setting interface of the preset priority of the scene features shown and Figure 3 The sub-priority setting interfaces of the scene features shown may not be displayed on the same interface, but may be split into multiple interfaces and displayed separately.

[0125] For example, when the number of driving scenarios to be evaluated is M, based on Figure 3 The manual interaction interface shown can obtain M evaluation information of M driving scenarios.

[0126] Furthermore, in order to improve the subsequent evaluation efficiency of the priorities of the M driving scenarios, the evaluation information of different driving scenarios can uniformly agree on the arrangement of the priorities of the above-mentioned multiple scene features in the driving scenarios. Figure 3The input information shown determines that the preset priorities of the scene features are arranged from high to low as follows: scene hazard > scene exposure > scene complexity. For example, it is agreed that the priorities of multiple scene features in the evaluation information of each driving scene in the driving scene are arranged in descending order according to the preset priorities of the multiple scene features. If the evaluation information of each driving scene is represented by an array, the first bit from the left of the array is set to store the priority of the scene hazard in the driving scene, the second bit from the left of the array is used to store the priority of the scene exposure in the driving scene, and the third bit from the left of the array is used to store the priority of the scene complexity in the driving scene. In this way, based on Figure 3 The input information of the priority of each scene feature in driving scene 2 is shown, and the evaluation information of driving scene 2 can be represented by array "232". In this way, the evaluation information of the driving scene not only describes the priority of the above-mentioned multiple scene features in the driving scene, but also indicates the size relationship between the preset priorities of the multiple scene features by specifying the arrangement order of the priorities of the multiple scene features in the driving scene, which is conducive to improving the evaluation efficiency of the priority of the driving scene.

[0127] S203: Determine priorities of the M driving scenarios according to the preset priorities of the above-mentioned multiple scene features and the M evaluation information of the M driving scenarios, where M is an integer greater than 1.

[0128] Exemplarily, the M evaluation information of the M driving scenarios are input by the user, for example, the user is based on Figure 3 The manual interaction interface shown inputs M evaluation information of M driving scenarios, that is, each driving scenario corresponds to one evaluation information.

[0129] Here, the preset priorities of the multiple scene features may be historically set by the user or currently input by the user. That is, in different driving scenes, the pre-set priorities of the multiple scene features remain unchanged.

[0130] In an embodiment of the present application, the priorities of the M driving scenes are determined based on the preset priorities of the above-mentioned multiple scene features and the M evaluation information of the M driving scenes, including: comparing the sub-priorities of the corresponding scene features in the M evaluation information in order of the preset priorities of the scene features from high to low, obtaining the sorting results of the M evaluation information, and determining the priorities of the M driving scenes based on the sorting results of the M evaluation information.

[0131] Exemplarily, it is assumed that three scene features are used to evaluate the priority of the driving scene, and the preset priorities of the three scene features are arranged from high to low as follows: scene feature 1>scene feature 2>scene feature 3. The specific process of determining the priorities of the M driving scenarios based on the above-mentioned sorting method may be: for the M evaluation information of the M driving scenarios, first compare the sub-priority of the scene feature 1 in each evaluation information, that is, the higher the priority of the scene feature 1 in the evaluation information in the driving scenario, the higher the ranking of the evaluation information in the M evaluation information, and the higher the priority of the driving scenario corresponding to the evaluation information; when the sub-priority of the scene feature 1 is the same, then compare the sub-priority of the scene feature 2, that is, the higher the priority of the scene feature 2 in the evaluation information in the driving scenario, the higher the ranking of the evaluation information in the M evaluation information, and the higher the priority of the driving scenario corresponding to the evaluation information; when the sub-priority of the scene feature 1 is the same and the sub-priority of the scene feature 2 is the same, then compare the sub-priority of the scene feature 3, that is, the higher the priority of the scene feature 3 in the evaluation information in the driving scenario, the higher the ranking of the evaluation information in the M evaluation information, and the higher the priority of the driving scenario corresponding to the evaluation information. It can be understood that the embodiments of the present application do not limit the number of scene features. When the number of scene features used to evaluate the scene priority can also be 2, 4, 5 or other values, the priority of the M driving scenes can also be determined by the sorting idea of ​​the above method.

[0132] In one implementation, the multiple scene features include a first feature and a second feature, wherein a preset priority of the first feature is higher than a preset priority of the second feature, and after the priorities of the M driving scenes are determined based on the above method, for any two driving scenes in the M driving scenes, taking the first driving scene and the second driving scene as examples, the priority of the first driving scene and the priority of the second driving scene meet any of the following conditions:

[0133] Case 1: The priority of the first feature in the first driving scenario is higher than the priority of the first feature in the second driving scenario, and the priority of the first driving scenario is higher than the priority of the second driving scenario;

[0134] Case 2: if the priority of the first feature in the first driving scenario is the same as the priority of the first feature in the second driving scenario and the priority of the second feature in the first driving scenario is higher than the priority of the second feature in the second driving scenario, then the priority of the first driving scenario is higher than the priority of the second driving scenario; or,

[0135] Case 3: If the priority of the first feature in the first driving scenario is the same as the priority of the first feature in the second driving scenario and the priority of the second feature in the first driving scenario is the same as the priority of the second feature in the second driving scenario, then the priority of the first driving scenario is the same as the priority of the second driving scenario.

[0136] In addition to the above-mentioned cases 1 to 3, it can be understood that if the priority of the first feature in the first driving scenario is lower than the priority of the first feature in the second driving scenario, the priority of the first driving scenario is lower than the priority of the second driving scenario; if the priority of the first feature in the first driving scenario is the same as the priority of the first feature in the second driving scenario and the priority of the second feature in the first driving scenario is lower than the priority of the second feature in the second driving scenario, the priority of the first driving scenario is lower than the priority of the second feature in the second driving scenario.

[0137] The following is based on Figure 4 The application of the above implementation method is specifically described.

[0138] Figure 4 : This is a schematic diagram of a process for determining the priorities of multiple driving scenarios provided by an embodiment of the present application. Taking the multiple scenario features of scenario complexity, scenario danger, and scenario exposure as an example, the preset priorities of the three are arranged from high to low as follows: scenario danger > scenario exposure > scenario complexity. Figure 4 It can be seen that Figure 4 Evaluation information of 20 driving scenarios is shown. The definition of the sub-priority of scenario features can refer to Table 1 above, and the larger the value of the sub-priority of the scenario feature in Table 1 (i.e., the priority of the scenario feature in the driving scenario), the higher the sub-priority of the scenario feature. Assuming that in the evaluation information of each driving scenario, the priorities of the three scenario features in the driving scenario are arranged from left to right in descending order of the preset priorities of the scenario features, then the evaluation information "431" of the driving scenario numbered "1" indicates that the priority of the scene hazard in driving scenario 1 is "4", the priority of the scene exposure in driving scenario 1 is "3", and the priority of the scene complexity in driving scenario 1 is "1". Figure 4 The evaluation information of the 20 driving scenarios shown in the figure is sorted in descending order according to the preset priorities of the scenario features, based on the sub-priorities of the corresponding scenario features in each evaluation information, to obtain a sorting result, namely Figure 4 The “Ranking of Evaluation Information” shown (the first column in the three-column table to the right of the arrow) is determined by Figure 4 From the sorting results of the evaluation information shown, the evaluation information “432” of the driving scene numbered “19” is ranked first, so the driving scene numbered “19” has the highest priority; the evaluation information “121” of the driving scene numbered “14” is ranked last, so the driving scene numbered “14” has the lowest priority; the evaluation information of the three driving scenes numbered “9”, “16” and “20” is the same, so the three driving scenes numbered “9”, “16” and “20” have the same priority.

[0139] Understandably, Figure 4 This is only an example of using three scenario features to evaluate the priorities of 20 driving scenarios from two dimensions: the preset priorities of the scenario features and the priorities of the scenario features in the driving scenarios. The number of scenario features used in evaluating the priorities of the driving scenarios and the number of driving scenarios are not limited.

[0140] In some possible embodiments, the total number of multiple scene features is S, assuming that the preset priority of the i-th scene feature is greater than the preset priority of the i+1-th scene feature, i is an integer greater than or equal to 1 and less than S, and the priorities of the M driving scenes are determined according to the preset priorities of the above-mentioned multiple scene features and the M evaluation information of the M driving scenes, including: determining the priorities of the M driving scenes by statistically analyzing the preset priorities of the multiple scene features and the M evaluation information of the M driving scenes through a sorting function, wherein the sorting function is the sum of S parts, the i-th part is represented by the i-th coefficient and the priority of the i-th scene feature in the driving scene, and the value of each coefficient in the sorting function satisfies the following constraint: the minimum value of the i-th part is greater than the maximum value of the sum of the i+1-th part, the i+2-th part, ..., the S-th part among the above-mentioned S parts.

[0141] In one implementation, the i-th part is represented by the i-th coefficient and the priority of the i-th scene feature in the driving scene, which can be: the i-th part is represented as a function with the i-th coefficient as the base and the priority of the i-th scene feature in the driving scene as the exponent. Taking the use of three scene features for evaluating scene priority as an example, in this case, the construction of the ranking function can be expressed as the following formula (1):

[0142] F1=a1 lp1 +b1 lp2 +c1 lp3 Formula (1)

[0143] Among them, F1 is the value of the sorting function, lp1 represents the priority of scene feature 1 in the driving scene, lp2 represents the priority of scene feature 2 in the driving scene, lp3 represents the priority of scene feature 3 in the driving scene, a1, b1, c1 are the above coefficients, assuming that the preset priority of scene feature 1 is higher than the preset priority of scene feature 2, and the preset priority of scene feature 2 is higher than the preset priority of scene feature 3, then the values ​​of a1, b1, c1 satisfy the following constraint condition 1: It can be understood that formula (1) is only an example of expressing a ranking function, and the expression of the ranking function is not limited to that shown in formula (1), and can also be based on some variations of formula (1).

[0144] For example, assuming that scene feature 1 is scene danger, scene feature 2 is scene exposure, and scene feature 3 is scene complexity, the definition of the sub-priority of each scene feature can be referred to as shown in Table 1. It can be seen from Table 1 that the value set of the sub-priority of scene danger is {1, 2, 3, 4}, the value set of the sub-priority of scene exposure is {1, 2, 3, 4}, and the value set of the sub-priority of scene complexity is {1, 2, 3, 4}. Based on constraint condition 1, the process of determining the values ​​of coefficients a1, b1, and c1 in the above formula (1) can be: for (b1 lp2 ) min >(c1 lp3 ) max This constraint, lp2 = 1 and lp3 = 4, first determine the values ​​of coefficients b1 and c1, and then for (a1 lp1 ) min >(b1 lp2 +c1 lp3 ) max This constraint, lp1=1 and lp2=lp3=4, and substituting the values ​​of b1 and c1 mentioned above determine the value of coefficient a1.

[0145] In another implementation, the ith part is represented by the ith coefficient and the priority of the ith scene feature in the driving scene, which can be: the ith part is represented by the product of the ith coefficient and the priority of the ith scene feature in the driving scene. Taking the use of three scene features for evaluating scene priority as an example, in this case, the construction of the ranking function can be expressed as the following formula (2):

[0146] F2=a2*lp1+b2*lp2+c2*lp3 Formula (2)

[0147] Wherein, F2 is the value of the sorting function, lp1 represents the priority of scene feature 1 in the driving scene, lp2 represents the priority of scene feature 2 in the driving scene, lp3 represents the priority of scene feature 3 in the driving scene, a2, b2, c2 are the above coefficients, wherein, assuming that the preset priority of scene feature 1 is higher than the preset priority of scene feature 2, and the preset priority of scene feature 2 is higher than the preset priority of scene feature 3, then the values ​​of a2, b2, c2 satisfy the following constraint condition 2: It can be understood that formula (2) is only an example of expressing a ranking function, and the representation of the ranking function is not limited to that shown in formula (2), and can also be some variations based on formula (2), which will not be elaborated here.

[0148] For example, assuming that scene feature 1 is scene danger, scene feature 2 is scene exposure, and scene feature 3 is scene complexity, the definition of the sub-priority of each scene feature can be referred to as shown in Table 1. It can be seen from Table 1 that the value set of the sub-priority of scene danger is {1, 2, 3, 4}, the value set of the sub-priority of scene exposure is {1, 2, 3, 4}, and the value set of the sub-priority of scene complexity is {1, 2, 3, 4}. Based on constraint condition 2, the process of determining the values ​​of a2, b2, and c2 in the above formula (2) can be: for (b2*lp2) min >(c2*lp3) max This constraint, lp2 = 1 and lp3 = 4, first determine the values ​​of coefficients b2 and c2, and then for (a2*lp1) min >(b2*lp2+c2*lp3)max, lp1=1 and lp2=lp3=4, and substituting the values ​​of b2 and c2 determined above to determine the value of coefficient a2.

[0149] The following is based on Figure 5 The application of the above implementation method is specifically described.

[0150] Figure 5 This is another schematic diagram of a process for determining the priorities of multiple driving scenarios provided by an embodiment of the present application. Taking the multiple scenario features of scenario complexity, scenario danger, and scenario exposure as an example, the preset priorities of the three are ranked from high to low as follows: scenario danger > scenario exposure > scenario complexity. Figure 5 The 20 evaluation information of the 20 driving scenarios shown can be referred to Figure 4 The description of the corresponding content in will not be repeated here. Figure 5 The 20 evaluation information shown, for example, use the ranking function F2 shown in the above formula (2), where, in formula (2), lp1 corresponds to the scene danger, lp2 corresponds to the scene exposure, lp3 corresponds to the scene complexity, and the values ​​of the coefficients a2, b2, and c2 are taken as a set of values ​​a2=25 and b2= and c2=1 that satisfy the above constraint condition 2. For example, the ranking function F2 is used to perform statistics on the preset priorities of the three scene features and the 20 evaluation information to obtain the ranking function value corresponding to each driving scene; the ranking function values ​​corresponding to the driving scenes are sorted from large to small. The larger the ranking function value corresponding to the driving scene, the higher the priority of the driving scene. In this way, the priority of the 20 driving scenes can be determined. Figure 5 It can be seen from the sorting result of F2 that the driving scene numbered "19" has the highest priority, the driving scene numbered "14" has the lowest priority, and the three driving scenes numbered "9", "16" and "20" have the same priority.

[0151] Understandably, Figure 5 This is only an example of using three scenario features to evaluate the priorities of 20 driving scenarios from two dimensions: the preset priorities of the scenario features and the priorities of the scenario features in the driving scenarios. The number of scenario features used in evaluating the priorities of the driving scenarios and the number of driving scenarios are not limited.

[0152] It can be seen that, by implementing the embodiments of the present application, the importance ranking between the multiple scene features involved in the evaluation is indicated by defining the preset priorities of the multiple scene features. The priority information of the multiple scene features in different driving scenarios can be used to more finely characterize different driving scenarios. In this way, the priorities of multiple driving scenarios are comprehensively determined from two dimensions: the preset priorities of the scene features and the priorities of the scene features in the driving scenarios. This can improve the evaluation accuracy of the priorities of the driving scenarios.

[0153] In some possible embodiments, the priority of driving scenarios may be applied to the screening of test scenarios.

[0154] See also Figure 6 , Figure 6 is a flowchart of another method for determining scene priority provided by an embodiment of the present application. This method can be applied to Figure 1 The processing device shown. Figure 6 The details of S601-S603 in the method shown can be referred to above. Figure 2 The relevant descriptions of S201-S203 in the embodiment are not repeated here. In addition, the method further comprises the following steps:

[0155] S604: Select N driving scenarios from the M driving scenarios for testing using the priorities of the M driving scenarios, where N is a positive integer less than M.

[0156] In one implementation, based on the priorities of the M driving scenarios, N driving scenarios with higher priorities are selected from the M driving scenarios for testing. Here, the N driving scenarios with higher priorities can be determined based on the ranking results of the priorities of the M driving scenarios from high to low. In this way, when test resources and test costs are limited, some driving scenarios with higher importance can be selected for testing based on the priority ranking of multiple driving scenarios, which can improve the test efficiency of driving scenarios.

[0157] For example, the N driving scenarios include a first driving scenario and a second driving scenario, wherein the priority of the first driving scenario is higher than the priority of the second driving scenario, and the first driving scenario is tested before the second driving scenario. In this way, the priority of the driving scenario will affect the test order of the driving scenario, and the driving scenario with a higher priority can be tested first.

[0158] S605: When the test results of the N driving scenarios do not meet the test objectives, adjust the preset priorities of the multiple scenario features according to the test results.

[0159] Here, the test target is related to the scene feature, for example, the test target includes sub-test targets corresponding to the above multiple scene features.

[0160] Exemplarily, adjusting the preset priorities of the plurality of scene features based on the test result may be: adjusting the preset priorities of some or all of the plurality of scene features based on the test result.

[0161] Exemplarily, the above-mentioned multiple scene features include scene feature 1, scene feature 2 and scene feature 3, wherein the preset priorities of the multiple scene features are arranged from high to low as follows: scene feature 1>scene feature 2>scene feature 3. Taking the test results of the above-mentioned N driving scenes not meeting the sub-test target corresponding to scene feature 3 in the above-mentioned test target as an example, in this case, adjusting the preset priorities of the multiple scene features according to the test results may be: raising the preset priority of scene feature 3 among the multiple scene features according to the test results. For example, after the above adjustment, the preset priorities of the multiple scene features may be arranged from high to low as follows: scene feature 1>scene feature 3>scene feature 2, or scene feature 3>scene feature 1>scene feature 2, or scene feature 3>scene feature 2>scene feature 1, which is not specifically limited here.

[0162] For example, scenario feature 3 is scenario complexity, and the sub-test target corresponding to scenario feature 3 may be that the scenario coverage must reach a preset threshold.

[0163] Here, the test results of the above-mentioned N driving scenarios may be a statistical result obtained based on the test outputs of the N driving scenarios, or may refer to the test output of at least one driving scenario among the N driving scenarios.

[0164] It can be seen that by implementing the embodiment of the present application, some driving scenes with higher priorities are selected from multiple driving scenes for testing according to the priority of the driving scenes, which can save the testing resources and testing costs of the driving scenes and improve the testing efficiency. In addition, by adjusting the preset priorities of multiple scene features involved in the scene priority evaluation through the test result feedback to change the importance ranking between at least two scene features, the test results of the driving scenes can be made more in line with the test objectives.

[0165] In the embodiments of the present application, Figure 1 The priority determination device shown can be implemented by hardware, software or a combination of hardware and software.

[0166] For example, Figure 1 The priority determination device shown can be used to implement Figure 2 The method described in the embodiment. Figure 2 In an embodiment, the acquisition unit may be used to execute S201 and S202, and the processing unit may be used to execute S203.

[0167] In some possible embodiments, Figure 1 The priority determination device shown also includes a testing unit, in which case Figure 1 The priority determination device shown can also realize Figure 6 The method described in the embodiment. Figure 6 In the embodiment, the acquisition unit may be used to execute S601 and S602, the processing unit may be used to execute S603, and the testing unit may be used to execute S604 and S605.

[0168] It should be understood that the above Figure 1 The division of the units in the priority determination device shown is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. In addition, the units in the device can be implemented in the form of a processor calling software; for example, the device includes a processor, the processor is connected to a memory, the memory stores instructions, and the processor calls the instructions stored in the memory to implement any of the above methods or implement the functions of the units of the device, wherein the processor is, for example, a general-purpose processor, such as a central processing unit (CPU) or a microprocessor, and the memory is a memory inside the device or a memory outside the device. Alternatively, the units in the device can be implemented in the form of hardware circuits, and the functions of some or all of the units can be realized by designing the hardware circuits, and the hardware circuits can be understood as one or more processors; for example, in one implementation, the hardware circuit is an application-specific integrated circuit (ASIC), and the functions of some or all of the above units are realized by designing the logical relationship of the components in the circuit; for another example, in another implementation, the hardware circuit can be realized by a programmable logic device (PLD), taking a field programmable gate array (FPGA) as an example, which can include a large number of logic gate circuits, and the connection relationship between the logic gate circuits is configured by a configuration file, so as to realize the functions of some or all of the above units. All units of the above devices can be realized in the form of a processor calling software, or in the form of hardware circuits, or in part by a processor calling software, and the rest by hardware circuits.

[0169] In the embodiment of the present application, the processor is a circuit with signal processing capability. In one implementation, the processor can be a circuit with instruction reading and running capability, such as a central processing unit (CPU), a microprocessor, a graphics processing unit (GPU) (which can be understood as a microprocessor), or a digital signal processor (DSP), etc.; in another implementation, the processor can realize certain functions through the logical relationship of the hardware circuit, and the logical relationship of the hardware circuit is fixed or reconfigurable, such as a hardware circuit implemented by a processor as an application-specific integrated circuit (ASIC) or a programmable logic device (PLD), such as an FPGA. In a reconfigurable hardware circuit, the processor loads a configuration document to implement the process of hardware circuit configuration, which can be understood as the process of the processor loading instructions to implement the functions of some or all of the above units. In addition, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as a neural network processing unit (NPU), a tensor processing unit (TPU), a deep learning processing unit (DPU), etc.

[0170] It can be seen that each unit in the above device can be one or more processors (or processing circuits) configured to implement the above method, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor forms.

[0171] In addition, all or part of the units in the above device can be integrated together, or can be implemented independently. In one implementation, these units are integrated together and implemented in the form of a system-on-a-chip (SOC). The SOC may include at least one processor for implementing any of the above methods or implementing the functions of each unit of the device. The type of the at least one processor may be different, for example, including a CPU and an FPGA, a CPU and an artificial intelligence processor, a CPU and a GPU, etc.

[0172] See also Figure 7 , Figure 7 Schematic diagram of a computing device provided in an embodiment of the present application. Figure 7As shown, the computing device 40 includes: a processor 401, a communication interface 402, a memory 403 and a bus 404. The processor 401, the memory 403 and the communication interface 402 communicate with each other via the bus 404. It should be understood that the present application does not limit the number of processors and memories in the computing device 40.

[0173] In one implementation, the computing device 40 may be a network-side device. The network-side device may be, for example, a server deployed on the network side (e.g., a scenario priority evaluation server, a simulation test server, etc.), or a component or chip in the server, or a system-level device composed of multiple servers. The network-side device may be deployed in a cloud environment or an edge environment, which is not specifically limited herein.

[0174] The bus 404 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 The bus 404 is represented by only one line, but it does not mean that there is only one bus or one type of bus. The bus 404 may include a path for transmitting information between various components of the computing device 40 (eg, the memory 403, the processor 401, and the communication interface 402).

[0175] The processor 401 can refer to the related description of the processor in the above embodiment, which will not be repeated here.

[0176] The memory 403 is used to provide a storage space, and the storage space can store data such as an operating system and a computer program. The memory 403 can be a random access memory (RAM), an erasable programmable read only memory (EPROM), a read-only memory (ROM), or a portable read-only memory (CD-ROM), etc., or a combination of multiple types. The memory 403 can exist alone or be integrated into the processor 401.

[0177] The communication interface 402 may be used to provide information input or output for the processor 401. Alternatively, the communication interface 402 may be used to receive data sent externally and / or send data externally, and may be a wired link interface such as an Ethernet cable, or a wireless link interface (such as Wi-Fi, Bluetooth, general wireless transmission, etc.). Alternatively, the communication interface 402 may also include a transmitter (such as a radio frequency transmitter, an antenna, etc.) coupled to the interface, or a receiver, etc.

[0178] In some possible embodiments, when the computing device 40 further includes a display 405. The display 405 is connected or coupled to the processor 401 via a bus 404. The display 405 can be used to present an artificial interaction interface to the user, for example, the user can input the preset priorities of the above-mentioned multiple scene features and the M evaluation information of the above-mentioned M driving scenes through the artificial interaction interface, where M is an integer greater than 1. The display 405 can be a display screen, and the display screen can be a liquid crystal display (LCD), an organic or inorganic light-emitting diode (OLED), an active matrix organic light emitting diode panel (AMOLED), etc.

[0179] The processor 401 in the computing device 40 is used to read the computer program stored in the memory 403 to execute the aforementioned method, for example Figure 2 or Figure 6 The method described.

[0180] In one possible design, the computing device 40 may be a processor for executing Figure 2 The processor 401 may be used to read one or more computer programs stored in the memory to perform the following operations:

[0181] Acquiring evaluation information of a driving scene through an acquisition unit, wherein the evaluation information of the driving scene is used to describe the priority of multiple scene features in the driving scene;

[0182] The priorities of the M driving scenarios are determined according to the preset priorities of the multiple scenario features and the M evaluation information of the M driving scenarios, where M is an integer greater than 1.

[0183] In the above-mentioned embodiments of this invention, the description of each embodiment has its own emphasis. For the part that is not described in detail in a certain embodiment, please refer to the relevant description of other embodiments. In addition, in the various embodiments of this application, if there is no special explanation and logical conflict, the terms and / or descriptions between the various embodiments are consistent and can be referenced to each other. The technical features in different embodiments can be combined to form a new embodiment according to their internal logical relationship.

[0184] It should be noted that, a person skilled in the art can see that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electrically-erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0185] The essence of the technical solution of the present application or the part that makes the contribution or all or part of the technical solution can be embodied in the form of a software product. The computer program product is stored in a storage medium and includes a number of instructions for enabling a device (which can be a personal computer, a server, or a network device, a robot, a single-chip microcomputer, a chip, a robot, etc.) to execute all or part of the steps of the method described in each embodiment of the present application.

Claims

1. A method for determining scene priority, characterized in that: The method comprises: Acquiring evaluation information of a driving scenario, wherein the evaluation information of the driving scenario is used to describe the priority of multiple scenario features in the driving scenario; The priorities of the M driving scenarios are determined according to the preset priorities of the multiple scenario features and the M evaluation information of the M driving scenarios, where M is an integer greater than 1.

2. The method according to claim 1, characterized in that The method further comprises: N driving scenarios are selected from the M driving scenarios for testing using the priorities of the M driving scenarios, where N is a positive integer less than M.

3. The method according to claim 2, characterized in that The method further comprises: When the test results of the N driving scenarios do not meet the test objectives, the preset priorities of the multiple scenario features are adjusted according to the test results of the N driving scenarios.

4. The method according to any one of claims 1 to 3, characterized in that: The multiple scene features Including the following: Scenario complexity, scenario danger, scenario exposure, scenario rationality and scenario accuracy.

5. The method according to any one of claims 1 to 4, characterized in that: The preset priorities of the multiple scene features are set based on the confidence levels of the multiple scene features or historical testing experience.

6. The method according to any one of claims 1 to 5, characterized in that: The multiple scene features include a first feature and a second feature, wherein a preset priority of the first feature is higher than a preset priority of the second feature; the M driving scenes include a first driving scene and a second driving scene, and the priority of the first driving scene and the priority of the second driving scene satisfy any of the following conditions: If the priority of the first feature in the first driving scenario is higher than the priority of the first feature in the second driving scenario, the priority of the first driving scenario is higher than the priority of the second driving scenario; or, If the priority of the first feature in the first driving scenario is the same as the priority of the first feature in the second driving scenario and the priority of the second feature in the first driving scenario is higher than the priority of the second feature in the second driving scenario, the priority of the first driving scenario is higher than the priority of the second driving scenario; or, If the priority of the first feature in the first driving scenario is the same as the priority of the first feature in the second driving scenario and the priority of the second feature in the first driving scenario is the same as the priority of the second feature in the second driving scenario, the priority of the first driving scenario is the same as the priority of the second driving scenario.

7. The method according to any one of claims 1 to 5, characterized in that: The total number of the multiple scene features is S, the preset priority of the i-th scene feature is greater than the preset priority of the i+1-th scene feature, i is an integer greater than or equal to 1 and less than S, Determining priorities of the M driving scenarios according to the preset priorities of the multiple scene features and the M evaluation information of the M driving scenarios, comprising: performing statistics on the preset priorities of the multiple scene features and the M evaluation information through a ranking function to determine the priorities of the M driving scenarios; The ranking function is the sum of S parts, the i-th part is represented by the i-th coefficient and the priority of the i-th scene feature in the driving scene, and the value of each coefficient in the ranking function satisfies the following constraints: The minimum value of the i-th part is greater than the maximum value of the sum of the i+1-th part, the i+2-th part, ..., the S-th part among the S parts.

8. The method according to any one of claims 1 to 7, characterized in that: The preset priorities of the multiple scene features are input by the user.

9. A device for determining scene priority, characterized in that: The device comprises: An acquisition unit, configured to acquire evaluation information of a driving scene, wherein the evaluation information of the driving scene is used to describe the priority of multiple scene features in the driving scene; A processing unit is used to determine the priorities of the M driving scenarios according to the preset priorities of the multiple scene features and the M evaluation information of the M driving scenarios, where M is an integer greater than 1.

10. The device according to claim 9, characterized in that The device also includes a testing unit: The testing unit is used to select N driving scenarios from the M driving scenarios for testing using the priorities of the M driving scenarios, where N is a positive integer less than M.

11. The device according to claim 10, characterized in that The test unit is also used to: When the test results of the N driving scenarios do not meet the test objectives, the preset priorities of the multiple scenario features are adjusted according to the test results of the N driving scenarios.

12. The device according to any one of claims 9 to 11, characterized in that: The multiple scene features Including the following: Scenario complexity, scenario danger, scenario exposure, scenario rationality and scenario accuracy.

13. The device according to any one of claims 9 to 12, characterized in that: The preset priorities of the multiple scene features are set based on the confidence levels of the multiple scene features or historical testing experience.

14. The device according to any one of claims 9 to 13, characterized in that: The multiple scene features include a first feature and a second feature, wherein a preset priority of the first feature is higher than a preset priority of the second feature; the M driving scenes include a first driving scene and a second driving scene, and the priority of the first driving scene and the priority of the second driving scene satisfy any of the following conditions: If the priority of the first feature in the first driving scenario is higher than the priority of the first feature in the second driving scenario, the priority of the first driving scenario is higher than the priority of the second driving scenario; or, If the priority of the first feature in the first driving scenario is the same as the priority of the first feature in the second driving scenario and the priority of the second feature in the first driving scenario is higher than the priority of the second feature in the second driving scenario, the priority of the first driving scenario is higher than the priority of the second driving scenario; or, If the priority of the first feature in the first driving scenario is the same as the priority of the first feature in the second driving scenario and the priority of the second feature in the first driving scenario is the same as the priority of the second feature in the second driving scenario, the priority of the first driving scenario is the same as the priority of the second driving scenario.

15. The device according to any one of claims 9 to 13, characterized in that: The total number of the multiple scene features is S, the preset priority of the i-th scene feature is greater than the preset priority of the i+1-th scene feature, i is an integer greater than or equal to 1 and less than S, The processing unit is specifically used to: perform statistics on the preset priorities of the multiple scene features and the M evaluation information through a ranking function to determine the priorities of the M driving scenes; The ranking function is the sum of S parts, the i-th part is represented by the i-th coefficient and the priority of the i-th scene feature in the driving scene, and the value of each coefficient in the ranking function satisfies the following constraints: The minimum value of the i-th part is greater than the maximum value of the sum of the i+1-th part, the i+2-th part, ..., the S-th part among the S parts.

16. The device according to any one of claims 9 to 15, characterized in that: The preset priorities of the multiple scene features are input by the user.

17. A communication device, characterized in that: The device comprises at least one processor and an interface circuit, wherein the processor is used to execute instructions and / or data interaction through the interface circuit, so that the device executes the method according to any one of claims 1 to 8.

18. A computer-readable storage medium, characterized in that: The method comprises computer instructions, which, when executed by a processor, implement the method according to any one of claims 1 to 8.

Images