Method and device for determining scene complexity and function complexity of automatic driving system

Abstract

The invention discloses a scene complexity and function complexity determination method and device of an automatic driving system. Taking the scene complexity determination method as an example, the method comprises the steps of determining description information of each scene attribute in a plurality of scene attributes of a driving scene; determining the complexity of each scene attribute in the plurality of scene attributes based on first mapping information, wherein the first mapping information is used for indicating a corresponding relationship between the description information and the complexity; determining the weight of each scene attribute in the plurality of scene attributes; and performing weighted summation on the complexity of each scene attribute in the plurality of scene attributes based on the weight to obtain the complexity of the driving scene. Therefore, the complexity of the driving scene can be reasonably and accurately evaluated, and the accuracy of the test result of the automatic driving vehicle can be improved.

Description

Method and device for determining scene complexity and functional complexity of autonomous driving system Technical Field

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

[0002] Autonomous vehicle testing is influenced by many factors, including the complexity of the driving scenario and the functional complexity of the autonomous driving system. Driving scenario complexity guides the design and selection of driving scenarios for autonomous vehicle evaluations, while the functional complexity of the autonomous driving system can be used to test its performance in driving scenarios of varying complexity. Improving the accuracy of autonomous vehicle test results is an urgent issue.

[0003] Summary of the Invention

[0004] The present application discloses a scene complexity determination method and device, which can reasonably and effectively evaluate the complexity of driving scenes, which is conducive to improving the accuracy of driving scene complexity assessment, thereby helping to improve the accuracy of test results of autonomous driving vehicles.

[0005] In a first aspect, the present application provides a method for determining scene complexity, the method comprising: determining descriptive information of each scene attribute among multiple scene attributes of a driving scene; determining the complexity of each scene attribute among the multiple scene attributes based on first mapping information, wherein the first mapping information is used to indicate the correspondence between the descriptive information and the complexity; determining a weight for each scene attribute among the multiple scene attributes; and obtaining the complexity of the driving scene by weighted summing the complexity of each scene attribute among the multiple scene attributes based on the weight.

[0006] Exemplarily, the description information may be a description of a state, presentation form, or type, or may be a value or parameter range, which is not specifically limited here.

[0007] Exemplarily, determining the description information of each scene attribute among the multiple scene attributes of the driving scene may include: acquiring the description information of each scene attribute among the multiple scene attributes of the driving scene.

[0008] Here, the first mapping information is pre-configured. For example, the first mapping information may be configured by relevant standard setters or autonomous driving testers.

[0009] In the above method, the complexity of the driving scene is obtained by weighted summing up the complexities of multiple scene attributes in the driving scene, and the complexity of each scene attribute can be obtained by searching for known mapping information based on the description information of the scene attribute in the driving scene. The mapping information indicates the correspondence between the description information of the scene attribute and the complexity. This not only simplifies the calculation, but also makes the evaluation of the complexity of the driving scene more accurate and reasonable.

[0010] Optionally, determining the weight of each of the multiple scene attributes includes: determining the weight of each of the multiple scene attributes based on second mapping information, wherein the second mapping information is used to indicate the correspondence between the scene attributes and the weights.

[0011] Here, the second mapping information is also pre-set. The second mapping information may be configured by, for example, relevant standard setters or autonomous driving testers.

[0012] By implementing the above embodiment, the weight of the scene attribute can be obtained by searching for known mapping information based on the scene attribute. The mapping information indicates the correspondence between the scene attribute and the weight, which is conducive to improving the evaluation efficiency of the complexity of the driving scene.

[0013] Optionally, the sum of the weights of the multiple scene attributes is 1.

[0014] Optionally, the multiple scene attributes include multiple items of road, weather, environment, host vehicle and network connection.

[0015] By implementing the above-mentioned implementation method, the complexity of the driving scene is evaluated from multiple dimensions including road, weather, environment, host vehicle and network connection, which increases the evaluation dimensions and makes the evaluation of the complexity of the driving scene more accurate and reasonable.

[0016] Optionally, the scene attributes include first-level scene attributes and second-level scene attributes. When the first-level scene attribute is road, the second-level scene attribute is at least one of road structure, lane line and slope; when the first-level scene attribute is weather, the second-level scene attribute includes weather type; when the first-level scene attribute is environment, the second-level scene attribute is at least one of traffic signs, temporary traffic events, types of road users and the number of road users; when the first-level scene attribute is the main vehicle, the second-level scene attribute includes the main vehicle speed; when the first-level scene attribute is networking, the second-level scene attribute includes networking information type.

[0017] Here, the second-level scene attributes are a subdivision of the first-level scene attributes.

[0018] Exemplarily, when the first-level scene attribute is the host vehicle, the second-level scene attribute further includes at least one of the host vehicle's functional status and the host vehicle's driver and passenger fatigue status.

[0019] For example, in the first mapping information, the descriptive attributes of the road structure include: long straight roads, curved roads, routes, and special roads (such as roundabouts, tunnels, ramps, etc.).

[0020] Illustratively, the lane line description information is related to at least one of whether the lane line exists, whether the lane line is worn or obscured, whether the lane line is regular, and whether the lane line is covered. For example, in the first mapping information described above, the lane line description information includes clear, worn or obscured, unclear due to water or ice on the road, irregular, and no lane line.

[0021] For example, in the first mapping information, the description of the slope includes two descriptions: slope and no slope. Furthermore, slope can also be divided according to the slope range to which it belongs, which is not specifically limited here.

[0022] For example, in the first mapping information, the weather type description information includes good daytime visibility, poor daytime visibility, good nighttime visibility, and poor nighttime visibility. It can be seen that the configuration of the weather type description information is related to at least one of visibility, whether there is ambient light, whether it is nighttime, and the current weather (e.g., sunny, rainy, or foggy).

[0023] Here, traffic signs can also be understood as traffic facilities, facility signs, etc., without specific limitation. For example, the configuration of the traffic sign description information is related to at least one of the following: the presence of the traffic sign, whether the traffic sign is regular, whether the traffic sign is clear, whether the traffic sign is far away, whether the traffic sign is reflective or dirty, etc.

[0024] For example, the configuration of the descriptive information of temporary traffic conditions is related to at least one of the following: whether there is a temporary traffic event, whether the temporary traffic conditions are maintained by a dedicated person, whether the temporary traffic event has a significant impact on driving, whether it is highly sporadic and difficult to foresee, and whether it is a temporary traffic event indicated by a warning sign.

[0025] Exemplarily, the configuration of the descriptive information of the road user type is related to at least one of whether there is a road user in the driving scene, the road users included, whether the behavior of the road user is standard, whether the road user is common, etc.

[0026] In the implementation of the above embodiment, the above-mentioned road, weather, environment, main vehicle and network connection are used as first-level scene attributes. The multiple first-level scene attributes can be further subdivided to obtain multiple second-level scene attributes. The complexity of the driving scene can be evaluated through these multiple second-level scene attributes, making the evaluation more accurate.

[0027] Optionally, the weight of each scene attribute is the same, or the weight of the scene attribute is associated with the degree of influence of the scene attribute on the complexity of the driving scene.

[0028] Exemplarily, the weight of each scene attribute is the same, which may be the weight of each first-level scene attribute is the same, or the weight of each second-level scene attribute is the same.

[0029] By implementing the above implementation method, when evaluating the complexity of the driving scene, the influence of different scene attributes on the complexity of the driving scene is considered, thereby achieving a reasonable and effective evaluation of the complexity of the driving scene.

[0030] Exemplarily, the weight of the second-level scene attribute is related to the weight of the first-level scene attribute.

[0031] For example, when the first-level scene attribute is A, if the second-level scene attribute is only a1, then the weight of the second-level scene attribute a1 is the weight of the first-level scene attribute A. If the second-level scene attributes include a1, a2, and a3, then the weights of these three second-level scene attributes are obtained by distributing them based on the weight of the first-level scene attribute A. The distribution method can be, for example, uniform distribution or other distribution methods. It can be understood that the sum of the weights of the second-level scene attributes a1, a2, and a3 is the weight of the first-level scene attribute A.

[0032] By implementing the above implementation, the weights of the corresponding second-level scene attributes are determined by the weights of the first-level scene attributes, which fully considers the impact of the first-level scene attributes on the complexity of the driving scene.

[0033] The present application also discloses a method and device for determining the functional complexity of an autonomous driving system, which can solve the problem of inconsistent evaluations of different autonomous driving systems and achieve a reasonable and accurate evaluation of the functional complexity of the autonomous driving system, thereby helping to improve the accuracy of the test results of autonomous driving vehicles.

[0034] In second aspect, the present application provides a method for determining the functional complexity of an autonomous driving system, the method comprising: determining the descriptive information of each of multiple conditional constraints of the autonomous driving system under a driving scenario; determining the complexity of each of the multiple conditional constraints based on first mapping information, wherein the first mapping information is used to indicate the correspondence between the descriptive information and the complexity; determining the weight of each of the multiple conditional constraints; and obtaining the functional complexity of the autonomous driving system under the driving scenario by weighted summing the complexity of each of the multiple conditional constraints based on the weight.

[0035] Here, "autonomous driving system" is not limited to fully autonomous driving system, highly autonomous driving system, conditional autonomous driving system, or partially autonomous driving system. Those skilled in the art can understand that non-fully manual driving systems that provide intelligent driving can be covered under this concept.

[0036] Here, the first mapping information is pre-configured. For example, the first mapping information may be configured by relevant standard setters or autonomous driving testers.

[0037] In the above method, the functional complexity of the autonomous driving system in a driving scenario is obtained by weighted summing the complexities of multiple constraints in the driving scenario. These multiple constraints are common to different autonomous driving systems. In addition, the complexity of each constraint can be obtained by searching known mapping information based on the description of the constraint in the driving scenario. This mapping information indicates the correspondence between the description of the constraint and the complexity. This allows for a reasonable and accurate assessment of the functional complexity of the autonomous driving system. Furthermore, this evaluation method is applicable to different autonomous driving systems and provides a reference for performance evaluation of different autonomous driving systems, which is conducive to improving the accuracy of autonomous vehicle test results.

[0038] Optionally, determining the weight of each of the multiple conditional constraints includes: determining the weight of each of the multiple conditional constraints based on second mapping information, wherein the second mapping information is used to indicate the correspondence between the conditional constraints and the weights.

[0039] Here, the second mapping information is also pre-set. The second mapping information may be configured by, for example, relevant standard setters or autonomous driving testers.

[0040] By implementing the above-mentioned embodiment, the weight of the conditional constraint can be obtained by searching for known mapping information based on the conditional constraint. The mapping information indicates the corresponding relationship between the conditional constraint and the weight, which is conducive to improving the evaluation efficiency of the functional complexity of the autonomous driving system.

[0041] Optionally, the sum of the weights of the multiple conditional constraints is 1.

[0042] Optionally, the multiple conditional constraints include multiple items of design operating range ODD boundary constraints, function activation condition constraints, takeover condition constraints, ODD dependency constraints, self-vehicle constraints, emergency response constraints and perception function limitation constraints.

[0043] By implementing the above-mentioned implementation method, the functional complexity of the autonomous driving system in any driving scenario is evaluated from multiple dimensions including ODD boundary constraints, function activation condition constraints, takeover condition constraints, ODD dependency constraints, ego-vehicle constraints, emergency response constraints, and perception function limitation constraints. The evaluation dimensions are rich and comprehensive, which improves the accuracy of the evaluation.

[0044] Optionally, the conditional constraints include first-level conditional constraints and second-level conditional constraints. When the first-level conditional constraint is a designed operating range (ODD) boundary constraint, the second-level conditional constraint includes an ODD boundary range; when the first-level conditional constraint is a function activation conditional constraint, the second-level conditional constraint is at least one of the speed difference of the function and the stable following time; when the first-level conditional constraint is a takeover conditional constraint, the second-level conditional constraint includes a takeover time range; when the first-level conditional constraint is an ODD dependency constraint, the second-level conditional constraint is at least one of the speed difference of the ODD and the change of the guide vehicle; when the first-level conditional constraint is a self-vehicle constraint, the second-level conditional constraint is at least one of the main vehicle speed, the driver and passenger fatigue status, and the vehicle function status; when the first-level conditional constraint is an emergency response constraint, the second-level conditional constraint includes automatic braking; when the first-level conditional constraint is a perception function limited constraint, the second-level conditional constraint includes sensor detection.

[0045] Here, the second-level conditional constraints are a subdivision of the first-level conditional constraints.

[0046] Exemplarily, in the first mapping information, the description information of the ODD boundary range includes three descriptions: being within the ODD, being at the ODD boundary, and being outside the ODD.

[0047] For example, the configuration of the speed difference description information for the function is related to whether the relative speed difference between the two vehicles meets the requirements for activating the autonomous driving system function. For example, in the first mapping information, the speed difference description information for the function includes two descriptions: the relative speed difference between the two vehicles that meets the requirements for activating the autonomous driving system function and the relative speed difference between the two vehicles that does not meet the requirements for activating the autonomous driving system function.

[0048] For example, the configuration of the descriptive information of the stable following time is related to whether the stable following time required for activating the automatic driving system function is met.

[0049] Exemplarily, the configuration of the description information of the takeover time range is related to whether takeover occurs within the takeover time range. For example, in the first mapping information, the description information of the takeover time range includes two attributes: takeover occurs and takeover does not occur.

[0050] Exemplarily, in the first mapping information, the description information of the speed difference of the ODD includes two attributes: the relative speed difference between the two vehicles satisfies the ODD range and the relative speed difference between the two vehicles does not satisfy the ODD range; the description information of the change of the guide vehicle includes two descriptions: the guide vehicle has not undergone a change that causes the guide vehicle to exit the ODD and the guide vehicle has undergone a change that causes the guide vehicle to exit the ODD.

[0051] Exemplarily, in the first mapping information, the description information of the automatic braking includes two descriptions: no automatic braking occurs and automatic braking occurs.

[0052] For example, in the first mapping information, the description information of the sensor detection includes three descriptions: normal, limited detection range, and unable to work normally.

[0053] In implementing the above embodiment, the ODD boundary constraints, function activation condition constraints, takeover condition constraints, ODD dependency constraints, ego vehicle constraints, emergency response constraints, and perception function restriction constraints are used as first-level constraints. These first-level constraints can be further subdivided to obtain multiple second-level constraints. These second-level constraints are then used to assess the functional complexity of the autonomous driving system, resulting in a more accurate assessment. Furthermore, this method is applicable not only to the functional complexity assessment of autonomous driving systems based on ODD, but also to the assessment of functional complexity of autonomous driving systems based on ODC.

[0054] Optionally, the weight of each conditional constraint is the same, or the weight of the conditional constraint is associated with the degree of influence of the conditional constraint on the functional complexity of the autonomous driving system.

[0055] Exemplarily, the weight of each conditional constraint is the same, which may be the weight of each first-level conditional constraint is the same, or the weight of each second-level conditional constraint is the same.

[0056] By implementing the above implementation method, when evaluating the functional complexity of the autonomous driving system, the degree of influence of different condition constraints in the driving scenario on the functional complexity of the autonomous driving system is considered, thereby achieving a reasonable and effective evaluation of the functional complexity of the autonomous driving system.

[0057] Optionally, the weight of the second-level conditional constraint is related to the weight of the first-level conditional constraint.

[0058] For example, if the first-level constraint is B and the second-level constraint is only b1, the weight of the second-level constraint b1 is the weight of the first-level constraint B. If the second-level constraints include b1, b2, and b3, the weights of these three second-level constraints are distributed based on the weight of the first-level constraint B. The distribution method can be, for example, even or in other ways. In other words, the sum of the weights of the second-level constraints b1, b2, and b3 is the weight of the first-level constraint B.

[0059] By implementing the above implementation method, the weight of the corresponding second-level condition constraint is determined by the weight of the first-level condition constraint, which fully considers the impact of the first-level condition constraint on the functional complexity of the autonomous driving system.

[0060] In a third aspect, the present application provides a device for determining scene complexity, the device comprising: an acquisition unit for determining descriptive information of each scene attribute among multiple scene attributes of a driving scene; a processing unit for determining the complexity of each scene attribute among the multiple scene attributes based on first mapping information, wherein the first mapping information is used to indicate the correspondence between the descriptive information and the complexity; the processing unit is also used to determine the weight of each scene attribute among the multiple scene attributes; the processing unit is also used to obtain the complexity of the driving scene by performing a weighted summation of the complexity of each scene attribute among the multiple scene attributes based on the weight.

[0061] Optionally, the processing unit is specifically configured to determine a weight of each of the multiple scene attributes based on second mapping information, wherein the second mapping information is used to indicate a correspondence between the scene attributes and the weights.

[0062] Optionally, the sum of the weights of the multiple scene attributes is 1.

[0063] Optionally, the multiple scene attributes include multiple items of road, weather, environment, host vehicle and network connection.

[0064] Optionally, the scene attributes include first-level scene attributes and second-level scene attributes. When the first-level scene attribute is road, the second-level scene attribute is at least one of road structure, lane line and slope; when the first-level scene attribute is weather, the second-level scene attribute includes weather type; when the first-level scene attribute is environment, the second-level scene attribute is at least one of traffic signs, temporary traffic events, types of road users and the number of road users; when the first-level scene attribute is the main vehicle, the second-level scene attribute includes the main vehicle speed; when the first-level scene attribute is networking, the second-level scene attribute includes networking information type.

[0065] Optionally, the weight of each scene attribute is the same, or the weight of the scene attribute is associated with the degree of influence of the scene attribute on the complexity of the driving scene.

[0066] Optionally, the weight of the second-level scene attribute is related to the weight of the first-level scene attribute.

[0067] In a fourth aspect, the present application provides a device for determining the functional complexity of an autonomous driving system, the device comprising: an acquisition unit for determining the descriptive information of each of multiple conditional constraints of the autonomous driving system under a driving scenario; a processing unit for determining the complexity of each of the multiple conditional constraints based on first mapping information, wherein the first mapping information is used to indicate the correspondence between the descriptive information and the complexity; the processing unit is also used to determine the weight of each of the multiple conditional constraints; the processing unit is also used to obtain the functional complexity of the autonomous driving system under the driving scenario by performing a weighted summation of the complexity of each of the multiple conditional constraints based on the weight.

[0068] Optionally, the processing unit is specifically configured to determine a weight of each of the multiple conditional constraints based on second mapping information, wherein the second mapping information is used to indicate a corresponding relationship between the conditional constraints and the weights.

[0069] Optionally, the sum of the weights of the multiple conditional constraints is 1.

[0070] Optionally, the multiple conditional constraints include multiple items of design operating range ODD boundary constraints, function activation condition constraints, takeover condition constraints, ODD dependency constraints, self-vehicle constraints, emergency response constraints and perception function limitation constraints.

[0071] Optionally, the conditional constraints include first-level conditional constraints and second-level conditional constraints. When the first-level conditional constraint is a designed operating range (ODD) boundary constraint, the second-level conditional constraint includes an ODD boundary range; when the first-level conditional constraint is a function activation conditional constraint, the second-level conditional constraint is at least one of the speed difference of the function and the stable following time; when the first-level conditional constraint is a takeover conditional constraint, the second-level conditional constraint includes a takeover time range; when the first-level conditional constraint is an ODD dependency constraint, the second-level conditional constraint is at least one of the speed difference of the ODD and the change of the guide vehicle; when the first-level conditional constraint is a self-vehicle constraint, the second-level conditional constraint is at least one of the main vehicle speed, the driver and passenger fatigue status, and the vehicle function status; when the first-level conditional constraint is an emergency response constraint, the second-level conditional constraint includes automatic braking; when the first-level conditional constraint is a perception function limited constraint, the second-level conditional constraint includes sensor detection.

[0072] Optionally, the weight of each conditional constraint is the same, or the weight of the conditional constraint is associated with the degree of influence of the conditional constraint on the functional complexity of the autonomous driving system.

[0073] Optionally, the weight of the second-level conditional constraint is related to the weight of the first-level conditional constraint.

[0074] In a fifth aspect, the present application provides a device for determining scene complexity, 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.

[0075] In the sixth aspect, the present application provides a device for determining the functional complexity of an autonomous driving system, 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 second aspect or any possible implementation of the second aspect.

[0076] In the seventh aspect, the present application provides a vehicle, which includes a device as described in the third aspect or any possible implementation of the third aspect, or includes a device as described in the fourth aspect or any possible implementation of the fourth aspect, or includes a device as described in at least one of the fifth and sixth aspects.

[0077] In an eighth 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 of the first aspect.

[0078] In the ninth aspect, the present application provides a computer program product, which, when executed by a processor, implements the method in the above-mentioned first aspect or any possible embodiment of the first aspect, or implements the method in the above-mentioned second aspect or any possible embodiment of the second aspect.

[0079] Exemplarily, the computer program product may be a software installation package. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] FIG1 is a schematic diagram of the architecture of a communication system provided in an embodiment of the present application;

[0081] FIG2 is a schematic diagram of a scene complexity assessment framework provided in an embodiment of the present application;

[0082] FIG3 is a flow chart of a method for determining scene complexity provided by an embodiment of the present application;

[0083] FIG4A is a schematic diagram of a driving scenario provided in an embodiment of the present application;

[0084] FIG4B is a schematic diagram of another driving scenario provided in an embodiment of the present application;

[0085] FIG4C is a schematic diagram of another driving scenario provided in an embodiment of the present application;

[0086] FIG5 is a schematic diagram of a functional complexity assessment framework for an autonomous driving system provided in an embodiment of the present application;

[0087] FIG6 is a flowchart of a method for determining functional complexity of an autonomous driving system provided in an embodiment of the present application;

[0088] FIG7 is a schematic diagram of another driving scenario provided in an embodiment of the present application;

[0089] FIG8 is a schematic diagram of the structure of a computing device provided in an embodiment of the present application;

[0090] FIG9 is a schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0091] It should be noted that the prefixes such as "first" and "second" used in this application are only for distinguishing different description objects, and do not have any limiting effect on the position, order, priority, quantity or content of the described objects. For example, if the described object is a "field", then the ordinal number before the "field" in the "first field" and the "second field" does not limit the position or order between the "fields", and "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", then 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 described object 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 described object 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 to distinguish the described objects in the embodiments of this application does not constitute a limitation on the described objects. For the description of the described objects, please refer to the description in the context of the claims or embodiments, and the use of such prefixes should not constitute an unnecessary limitation.

[0092] It should be noted that the descriptions used in the embodiments of the present application, such as "at least one of a1, a2, ..., and an" and the like, include any one of a1, a2, ..., and an existing alone, and any combination of any multiple of a1, a2, ..., and an, each of which can exist alone. For example, the description "at least one of a, b, and c" includes a alone, b alone, c alone, a combination of a and b, a combination of a and c, a combination of b and c, or a combination of ab and c.

[0093] To facilitate understanding, the following first introduces relevant terms that may be involved in the embodiments of this application.

[0094] (1) Design operating range (ODD)

[0095] The operational domain design (ODD) refers to the external environmental conditions within which an autonomous driving system must operate, as determined during its design. For example, these conditions may include geographic location, road type, speed range, weather, time of day, and national and local traffic laws and regulations. For example, the Highway Pilot (HWP) system, once it recognizes that the vehicle is within the ODD (e.g., the vehicle is currently traveling on a highway, the weather is clear, the speed is appropriate, the lighting conditions are good, and the global positioning system (GPS) signal is stable), and the driver confirms activation of the system, the HWP system will continue to perform all dynamic driving tasks.

[0096] (2) Design operating conditions ODC

[0097] Operational design conditions (ODC) refer to the general term for various conditions applicable to the functional operation of an automated driving system, determined when the system is designed. These include the design operating range (ODD) and further internal conditions required for the system to start and operate safely, such as vehicle status, driver and passenger status, and other necessary conditions, such as traffic conditions and operational conditions.

[0098] Different autonomous driving systems have different ODDs. For example, in a highway autonomous driving system, System A can only be activated and operated during daytime, while System B can be activated and operated during daytime and clear nights.

[0099] Vehicle status includes vehicle speed and functional status, including software and hardware status. Vehicle speed includes an activation speed range, which is used to determine whether the autonomous driving system can be activated. Before the autonomous driving system can be safely started and operated, the vehicle status must meet the conditions for safe startup and operation. For example, autonomous driving systems at high speeds must have functional self-test capabilities and perform a functional self-test before startup. The sensing function status, positioning function status, and computing function status must meet system design requirements.

[0100] The driver and passengers include the driver / dynamic driving task backup user and passengers. The driver and passenger status includes the status attributes for monitoring the driver / dynamic driving task backup user and passengers. For example, monitoring whether the driver is fatigued, driving under the influence of alcohol, whether the driver is wearing a seat belt, etc., so as to determine whether the vehicle has driving risks. For another example, monitoring the dynamic driving task backup user to monitor whether he is fatigued, driving under the influence of alcohol, etc., can determine whether the dynamic driving task backup user can take over. For another example, monitoring whether the passenger has any behavior of snatching the autonomous driving device can determine whether the current autonomous driving system can operate safely.

[0101] The technical solutions in the embodiments of the present application will be described below with reference to the accompanying drawings.

[0102] See Figure 1, which is a schematic diagram of the architecture of a communication system provided in an embodiment of the present application. The system is used to determine the complexity of the driving scenario and to evaluate the functional complexity of the vehicle's autonomous driving system in any driving scenario. Here, "autonomous driving system" 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 will understand that non-fully manual driving systems that provide intelligent driving can be covered under this concept.

[0103] As shown in FIG1 , the communication system includes a data source device and a first device, wherein the data source device and the first device can be connected and communicated with each other wirelessly or by wire.

[0104] Here, the data source device can be an unmanned aerial vehicle (UAV) equipped with a data collection device, a data collection vehicle, roadside equipment, surrounding vehicles, etc. The roadside equipment can be, for example, a roadside unit (RSU), multi-access edge computing (MEC), or sensor, or a component or chip within these devices, or a system-level device consisting of an RSU and MEC, or a system-level device consisting of an RSU and sensors, or a system-level device consisting of an RSU, MEC, and sensors.

[0105] Here, the first device can be a network side device, which can be, for example, a server deployed on the network side (such as a scenario evaluation server, an autonomous driving system evaluation server, etc.), or a component in the server (the component can be, for example, a chip, an integrated circuit, etc.), or a system-level device composed of multiple servers. The network side device can be deployed in a cloud environment or an edge environment, which is not specifically limited here. When the first device is a network side device, the data collected by the data source device can also be transferred to the network side device through a removable storage medium such as a mobile hard disk.

[0106] In some possible embodiments, the first device may also be a vehicle or a component in a vehicle, and the component may be, for example, a chip, an integrated circuit, etc. Here, the vehicle refers to a vehicle equipped with an automatic driving system. In addition, depending on the power source of the vehicle, the vehicle may be, for example, a new energy vehicle or a traditional vehicle, wherein a traditional vehicle refers to a fuel vehicle, such as a gasoline vehicle, a diesel vehicle, etc., and a new energy vehicle may be, for example, an electric vehicle (EV), a hybrid electric vehicle (HEV), a range extended EV, a plug-in hybrid vehicle (Plug-in HEV), a fuel cell vehicle or other new energy vehicles, which are not specifically limited here.

[0107] For example, the data source device is used to collect data required to assess the complexity of the driving scenario, such as descriptions of scene attributes in the driving scenario and descriptions of the conditional constraints of the autonomous driving system. In the embodiments of the present application, the number of data source devices is not limited and can be one or more.

[0108] In the embodiment of the present application, the description information may be a description of a state, presentation form or type, or a value or parameter range, which is not specifically limited here.

[0109] In one implementation, a first device is used to evaluate the complexity of a driving scene. Exemplarily, the first device determines descriptive information for each of a plurality of scene attributes of the driving scene; determines the complexity of each of the plurality of scene attributes based on mapping information 1, wherein mapping information 1 is used to indicate a correspondence between descriptive information and complexity; determines a weight for each of the plurality of scene attributes; and obtains the complexity of the driving scene by weighted summing the complexity of each of the plurality of scene attributes based on the weight. The evaluation process of the complexity of the driving scene may be specifically described in the following embodiments and will not be repeated here.

[0110] Here, the driving scene may be, for example, a vehicle cutting-in scene, a vehicle cutting-out scene, a vehicle following scene, or a preceding vehicle braking scene.

[0111] For example, the description information of the scene attributes can be obtained by the first device from the data source device. The mapping information 1 can be pre-stored locally on the first device or obtained by the first device from other devices (such as a cloud server), which is not specifically limited here.

[0112] In another implementation, the first device may also be used to evaluate the functional complexity of the vehicle's autonomous driving system in the driving scenario. The evaluation process of the functional complexity of the autonomous driving system can be specifically described in the following embodiments and will not be repeated here.

[0113] The communication system shown in Figure 1 can be applied to a variety of network types, for example, one or more of the following network types: SparkLink or NearLink, long term evolution (LTE) network, fifth generation mobile communication technology (5G), wireless local area network (for example, Wi-Fi), Bluetooth (BT), Zigbee, or vehicle-mounted short-range wireless communication network, etc.

[0114] Here, the system architecture shown in Figure 1 is only an example and does not limit the number of network elements included in the system shown in Figure 1. In addition, the method provided in the embodiment of the present application can be applied not only to the system shown in Figure 1, but also to other systems. For example, the communication system also includes a cloud server, which is used to transmit the above-mentioned mapping information 1 to the first device, so as to assist the first device in realizing the evaluation of the complexity of the driving scene and / or the evaluation of the functional complexity of the autonomous driving system.

[0115] Before introducing the scene complexity determination method provided by the embodiment of the present application, a scene complexity evaluation framework provided by the embodiment of the present application is first introduced based on Figure 2. Referring to Figure 2, Figure 2 is a schematic diagram of a scene complexity evaluation framework provided by the embodiment of the present application.

[0116] As can be seen from Figure 2, the multiple scene attributes used to evaluate scene complexity include multiple items of road, weather, environment, main vehicle, and network connection. In order to accurately evaluate the complexity of the driving scene, the scene attributes include first-level scene attributes and second-level scene attributes, among which the second-level scene attributes are a subdivision of the first-level scene attributes.

[0117] Exemplarily, when the first-level scene attribute is road, the second-level scene attribute is at least one of road structure (Road Structure), lane lines (Lane Lines) and slope (Slope); when the first-level scene attribute is weather, the second-level scene attribute includes weather type (Weather Type); when the first-level scene attribute is environment, the second-level scene attribute is at least one of traffic sign (Traffic Sign), temporary traffic event (Temporary Traffic Event), type of road users (Types of Traffic Participants) and number of road users (Number of Traffic Participants).

[0118] In an embodiment of the present application, each scene attribute is pre-configured with multiple descriptions and the complexity corresponding to each description. Here, the correspondence between the descriptions and the complexity can be stored as mapping information. For example, when the description of scene attribute 1 is the first description, the complexity of scene attribute 1 is the complexity corresponding to the first description. It will be understood that for the same scene attribute, different descriptions correspond to different complexities.

[0119] In one implementation, the multiple scene attributes include the aforementioned road, weather, and environment.

[0120] Taking the first-level scene attribute of road and the second-level scene attributes of road structure, lane line, and slope as an example, the corresponding relationship between the description information and complexity of each second-level scene attribute is described respectively:

[0121] (1) Road structure

[0122] As an example, Table 1 shows the correspondence between preconfigured road structure description information and the complexity of the road structure. As can be seen from Table 1, "Road Structure" is configured with four types of description information: long straight roads, curves, intersections, and special roads. Special roads here include, but are not limited to, tunnels, ramps, and roundabouts. The complexity corresponding to the "long straight road" road structure description is 0.25, the complexity corresponding to the "curve" road structure description is 0.5, the complexity corresponding to the "intersection" road structure description is 0.75, and the complexity corresponding to the "special road" road structure description is 1. The configuration of road structure description information and the corresponding complexity of the description information is not limited to the configuration shown in Table 1.

[0123] Table 1

[0124] (2) Lane lines

[0125] As an example, Table 2 shows the correspondence between preconfigured lane line description information and lane line complexity. As can be seen from Table 2, "lane line" is configured with five description information: clear, worn or obscured, unclear due to water / ice, irregular, and no lane line. The complexity corresponding to the "clear" lane line description is 0, the complexity corresponding to the "worn or obscured" lane line description is 0.25, the complexity corresponding to the "unclear due to water / ice" lane line description is 0.5, the complexity corresponding to the "irregular" lane line description is 0.75, and the complexity corresponding to the "no lane line" lane line description is 1. The configuration of lane line description information and the corresponding complexity of the description information is not limited to the configuration shown in Table 2.

[0126] Table 2

[0127] (3) Slope

[0128] As an example, Table 3 shows the correspondence between the pre-configured slope description information and the complexity of the slope. It can be seen from Table 3 that the "slope" configuration has two description information, namely no slope and with slope. Among them, the complexity corresponding to the slope description information "no slope" is 0, and the complexity corresponding to the slope description information "with slope" is 1.

[0129] Table 3

[0130] Here, the configuration of the slope description information and the complexity corresponding to the description information is not limited to that shown in Table 2. For example, in some possible embodiments, the slope description information can also be set based on the degree of inclination of the road compared to the horizontal line, and different degrees of inclination correspond to different complexities.

[0131] It is understandable that the above-mentioned roads are not limited to being subdivided into the above-mentioned three second-level scene attributes of road type structure, lane line and slope.

[0132] The first-level scene attribute weather and the second-level scene attribute weather type are used as an example for illustrative explanation. Referring to Table 4, Table 4 shows the correspondence between the description information of the pre-configured weather type and the complexity of the weather type. As can be seen from Table 4, "Weather Type" is configured with four description information, namely, good daytime visibility, poor daytime visibility, good nighttime visibility, and poor nighttime visibility. Among them, the complexity corresponding to the description information of the weather type "good daytime visibility" is 0.25, the complexity corresponding to the description information of the weather type "poor daytime visibility" is 0.5, the complexity corresponding to the description information of the weather type "good nighttime visibility" is 0.75, and the complexity corresponding to the description information of the weather type "poor nighttime visibility" is 1.

[0133] Table 4

[0134] Here, the description information of the weather type and the configuration of the complexity corresponding to the description information are not limited to those shown in Table 4.

[0135] For example, in another implementation, the "weather type" is configured with five types of descriptive information, namely, high visibility on sunny days, moderate visibility on rainy days or in the evening, ambient light at night, no ambient light at night, and extremely low visibility on foggy days. Among them, the complexity corresponding to the weather type description information "high visibility on sunny days" is 0, the complexity corresponding to the weather type description information "moderate visibility on rainy days or in the evening" is 0.25, the complexity corresponding to the weather type description information "ambient light at night" is 0.5, the complexity corresponding to the weather type description information "no ambient light at night" is 0.75, and the complexity corresponding to the weather type description information "extremely low visibility on foggy days" is 1.

[0136] Taking the first-level scene attribute as the environment and the second-level scene attributes as traffic signs, temporary traffic events, road user types, and road user numbers as examples, the corresponding relationship between the descriptive information configured for each second-level scene attribute and the complexity is described respectively:

[0137] (1) Traffic signs

[0138] Here, traffic signs can also be understood as traffic facilities, facility signs, etc., without specific limitation.

[0139] As an example, Table 5 shows the correspondence between preconfigured traffic sign description information and traffic sign complexity. As can be seen from Table 4, "Traffic Sign" is configured with five description information: no traffic sign, clear, distant, reflective / dirty, and irregular. The complexity corresponding to the traffic sign description "no traffic sign" is 0, the complexity corresponding to the traffic sign description "clear" is 0.25, the complexity corresponding to the traffic sign description "distant" is 0.5, the complexity corresponding to the traffic sign description "reflective / dirty" is 0.75, and the complexity corresponding to the traffic sign description "irregular" is 1. The configuration of traffic sign description information and the complexity corresponding to the description information is not limited to the configuration shown in Table 5.

[0140] Table 5

[0141] (2) Temporary traffic incidents

[0142] As an example, Table 6 shows the correspondence between the pre-configured description information of temporary traffic events and the complexity of temporary traffic events. As can be seen from Table 6, "temporary traffic events" are configured with five types of description information, namely, no temporary traffic events, maintenance by dedicated personnel (such as traffic control), warning signs (such as road construction), significant impact on driving (such as traffic accidents), and sporadic and difficult to foresee (such as falling rocks or wheel detachment). Among them, the complexity corresponding to the description information of the temporary traffic event "no temporary traffic events" is 0, the complexity corresponding to the description information of the temporary traffic event "maintenance by dedicated personnel" is 0.25, the complexity corresponding to the description information of the temporary traffic event "warning signs" is 0.5, the complexity corresponding to the description information of the temporary traffic event "significant impact on driving" is 0.75, and the complexity corresponding to the description information of the temporary traffic event "strong sporadic and difficult to foresee" is 1. Here, the configuration of the description information of temporary traffic events and the complexity corresponding to the description information is not limited to that shown in Table 6.

[0143] Table 6

[0144] (3) Types of road users

[0145] Here, road users can also be referred to as traffic participants.

[0146] As an example, Table 7 shows the correspondence between the preconfigured descriptions of road user types and their complexity. As can be seen from Table 7, "Road User Type" is configured with four descriptions: the complexity of "only motor vehicles" is 0.25; the complexity of "motor vehicles, pedestrians, or non-motor vehicles, all located in legally required locations" is 0.5; the complexity of "motor vehicles, pedestrians, or non-motor vehicles, not located in legally required locations" is 0.75; and the complexity of "uncommon road users" is 1. For example, uncommon road users include pedestrians on horseback and animals crossing the road (e.g., horses, cattle, sheep, deer, etc.).

[0147] Table 7

[0148] Here, the description information of the road user type and the configuration of the complexity corresponding to the description information are not limited to those shown in Table 7.

[0149] For example, in another implementation, the type of road user further includes description information “no road user”, and the complexity corresponding to the description information “no road user” is 0.

[0150] (4) Number of road users

[0151] As an example, Table 8 shows the correspondence between the pre-configured description information of the number of road users and the complexity of the number of road users. As can be seen from Table 8, "Number of Road Users" is configured with three description information, namely less than 3, greater than or equal to 3 and less than or equal to 6, and greater than 6. Among them, the complexity corresponding to the description information "less than 3" is 0; the complexity corresponding to the description information "greater than or equal to 3 and less than or equal to 6" is 0.5; and the complexity corresponding to the description information "more than 6" is 1.

[0152] Table 8

[0153] Here, the description information of the number of road users and the configuration of the complexity corresponding to the description information are not limited to those shown in Table 8.

[0154] It is understandable that the above environment is not limited to being subdivided into the four second-level scene attributes of the above traffic signs, temporary traffic events, types of road users and number of road users.

[0155] In some possible embodiments, the multiple scene attributes further include at least one of a host vehicle and a network connection.

[0156] In one implementation, when the first-level scene attribute is the host vehicle, the second-level scene attribute includes the host vehicle's speed. As an example, the host vehicle's speed description can be normalized based on the legally mandated maximum speed to determine the complexity of the description. For example, if the host vehicle's speed description is 80 km / h, and the legally mandated maximum speed is 120 km / h, then the complexity of the host vehicle's speed description at 80 km / h is 0.67, which is 80 divided by 120.

[0157] In one implementation, when the first-level scene attribute is network connection, the second-level scene attribute includes the network connection information type. As an example, the pre-configured description information for the network connection information type and the complexity corresponding to the description information can be: the network connection information type can be divided into V2V (vehicle to vehicle), V2I (vehicle to infrastructure), V2P (vehicle to people), V2N (vehicle to network) and satellite positioning. When the description information of the network connection information type is "including one of V2V, V2I, V2P, V2N and satellite positioning", the complexity of the network information type is 0.2; when the description information of the network connection information type is "including two of V2V, V2I, V2P, V2N and satellite positioning", the complexity of the network information type is 0.2. items", the complexity of this network information type is 0.4; when the description information of the network information type is "including three items of V2V, V2I, V2P, V2N and satellite positioning", the complexity of this network information type is 0.6; when the description information of the network information type is "including four items of V2V, V2I, V2P, V2N and satellite positioning", the complexity of this network information type is 0.8; when the description information of the network information type is "including five items of V2V, V2I, V2P, V2N and satellite positioning", the complexity of this network information type is 1.

[0158] In the embodiment of the present application, each scene attribute used to evaluate scene complexity is also configured with a weight. The sum of the weights of multiple scene attributes is 1. When the scene attributes include first-level scene attributes and second-level scene attributes, each second-level scene attribute also corresponds to a weight, and the weight of the second-level scene attribute is related to the weight of the first-level scene attribute. For example, as can be seen from Figure 2, the weight of the road structure is w 11 , the weight of the lane line is w 12 , the slope weight is w 13 , the weight of weather type is w 21 ,…, the weights of the other second-level scene attributes can be found in Figure 2 and will not be described here in detail.

[0159] Exemplarily, the sum of the weights of multiple second-level scene attributes participating in the scene complexity evaluation should be 1.

[0160] In one implementation, the complexity of the driving scene is evaluated from three dimensions: road, weather, and environment. It is assumed that the first-level scene attribute "road" is further subdivided into the above-mentioned second-level scene attributes: road structure, lane line, and slope. The weight of the road structure is denoted as W RS , the weight of the lane line is recorded as W LL , the slope weight is recorded as W S The first-level scene attribute "weather" is further subdivided into the above-mentioned second-level scene attributes: weather type, and the weight of the weather type is recorded as W WTThe first-level scene attribute “environment” is further subdivided into the above-mentioned second-level scene attributes: traffic signs, temporary traffic events, types of road users and the number of road users, among which the weight of traffic signs is recorded as W TS The weight of temporary traffic events is recorded as W TTE The weight of the number of road users is recorded as W NTP , the weight of the road user type is recorded as W TTP , then the complexity of any driving scenario (e.g., driving scenario A) can be evaluated using the above eight second-level scenario attributes. The complexity of driving scenario A can be calculated using the following formula (1):

[0161] Among them, Soc A represents the complexity of driving scenario A, RS represents the description information of “road structure” in driving scenario A, LL represents the description information of “lane line” in driving scenario A, S represents the description information of “slope” in driving scenario A, WT represents the description information of “weather type” in driving scenario A, TS represents the description information of “traffic sign” in driving scenario A, TTE represents the description information of “temporary traffic event” in driving scenario A, NTP represents the description information of “number of road users” in driving scenario A, TTP represents the description information of “type of road user” in driving scenario A, and weight W RS 、W LL 、W S 、W WT 、W TS 、W TTE 、W NTP and W TTP Please refer to the above description and will not repeat them here.

[0162] In addition, in formula (1), the weight W RS 、W LL 、W S 、W WT 、W TS 、W TTE 、W NTP and W TTP The sum is 1.

[0163] Here, the weight of the second-level scene attribute is related to the weight of the first-level scene attribute. It can be understood that the weight of the second-level scene attribute is obtained based on the weight distribution of the first-level scene attribute.

[0164] For example, when the first-level scene attribute is A, if the second-level scene attribute is only a1, then the weight of the second-level scene attribute a1 is the weight of the first-level scene attribute A. If the second-level scene attributes include a1, a2, and a3, then the weights of these three second-level scene attributes are obtained by distributing them based on the weight of the first-level scene attribute A. The distribution method can be, for example, uniform distribution or other distribution methods. It can be understood that the sum of the weights of the second-level scene attributes a1, a2, and a3 is the weight of the first-level scene attribute A.

[0165] Based on the evaluation framework shown in FIG2 above, see FIG3, which is a flow chart of a method for determining scene complexity provided by an embodiment of the present application. This method can be applied to the first device shown in FIG1. ​​The method includes but is not limited to the following steps:

[0166] S301: Determine description information of each scene attribute among a plurality of scene attributes of a driving scene.

[0167] Exemplarily, the multiple scene attributes include multiple items of road, weather, environment, host vehicle, and network connection.

[0168] In one implementation, the scene attributes include first-level scene attributes and second-level scene attributes. Here, the correspondence between the first-level scene attributes and the second-level scene attributes can refer to the description of the corresponding content in Figure 2 above, which will not be repeated here.

[0169] In one implementation, the descriptive information for each scene attribute may be obtained by the first device from at least one of a roadside device, a vehicle in the driving scene, a drone, or other data collection device. It will be appreciated that when the scene attributes include first-level scene attributes and second-level scene attributes, determining the descriptive information for each scene attribute in the driving scene is equivalent to determining the descriptive information for each second-level scene attribute in the driving scene.

[0170] Below, when the first-level scene attributes include road, weather, environment, host vehicle and network connection, nine second-level scene attributes are further used to evaluate the complexity of the driving scenarios shown in Figures 4A, 4B and 4C below. These nine second-level scene attributes include road structure, lane lines, slope, weather type, traffic signs, temporary traffic events, type of road users, host vehicle speed and network connection information type.

[0171] Referring to Figure 4A, Figure 4A is a schematic diagram of a driving scenario provided in an embodiment of the present application. The driving scenario shown in Figure 4A is Driving Scenario 1. Assume that the first device determines that in Driving Scenario 1, the descriptive information of the road structure is intersection, the descriptive information of the lane line is clear, the descriptive information of the slope is no slope, the descriptive information of the weather type is good visibility during the day, the descriptive information of the traffic sign is no traffic sign, the descriptive information of the temporary traffic event is no temporary traffic event, the descriptive information of the road user type is only motor vehicles, the descriptive information of the main vehicle speed is 60 km / h, and the descriptive information of the network information type is only V2V.

[0172] Referring to Figure 4B, Figure 4B is a schematic diagram of another driving scenario provided in an embodiment of the present application. The driving scenario shown in Figure 4B is driving scenario 2. Assume that the scene description information 2 acquired by the first device indicates that in driving scenario 2, the description information of the road structure is a long straight road, the description information of the lane line is clear, the description information of the slope is no slope, the description information of the weather type is night without ambient light, the description information of the traffic sign is far distance, the description information of the temporary traffic event is road construction, the description information of the road user type is only motor vehicles, the description information of the main vehicle speed is 60 km / h, and the description information of the network information type is that none of the items are included.

[0173] Referring to FIG4C , FIG4C is a schematic diagram of another driving scenario provided in an embodiment of the present application. The driving scenario shown in FIG4C is driving scenario 3. Assume that the scene description information 3 acquired by the first device indicates that in driving scenario 3, the description information of the road structure is an intersection, the description information of the lane line is unclear due to ice on the road surface, the description information of the slope is no slope, the description information of the weather type is poor visibility at night, the description information of the traffic sign is far away, the description information of the temporary traffic event is road construction, the description information of the type of road user is including motor vehicles and pedestrians but pedestrians are crossing the road, the description information of the main vehicle speed is 60 km / h, and the description information of the type of network information is only V2V.

[0174] S302: Determine the complexity of each of the multiple scene attributes based on mapping information 1, where mapping information 1 is used to indicate a correspondence between the above description information and the complexity.

[0175] Exemplarily, the mapping information 1 includes the corresponding relationships shown in Tables 1 to 7. In some possible embodiments, the mapping information 1 also includes the corresponding relationships shown in Table 8.

[0176] In one implementation, since the scene attributes include first-level scene attributes and second-level scene attributes, the complexity of each of the multiple scene attributes is determined based on the mapping information 1, including: determining the complexity of each second-level scene attribute in the multiple second-level scene attributes based on the mapping information 1. In this case, the mapping information 1 stores the correspondence between the descriptive information of the second-level scene attributes and the complexity.

[0177] Taking the second-level scene attribute 1 as an example, determining the complexity of the second-level scene attribute 1 based on the mapping information 1 includes: determining the target description information of the second-level scene attribute 1 in the mapping information 1 that is closest to the description information of the second-level scene attribute 1 in the driving scenario, and using the complexity corresponding to the target description information of the second-level scene attribute 1 in the mapping information 1 as the complexity of the second-level scene attribute 1 in the driving scenario. In some possible embodiments, the description information of the second-level scene attribute 1 in the driving scenario belongs to the pre-configured description information of the second-level scene attribute 1 in the mapping information 1. In this case, the complexity corresponding to the description information of the second-level scene attribute 1 in the driving scenario found in the mapping information 1 is used as the complexity of the second-level scene attribute 1 in the driving scenario.

[0178] For example, in the driving scene 2 shown in FIG4B , it is assumed that the second-level scene attribute 1 is a temporary traffic event. Based on the description of FIG4B above, it can be known that the description information of the temporary traffic event is road construction. Based on the configuration information of the temporary traffic event shown in Table 6 above, it is determined that the description information of the temporary traffic event closest to the description information “road construction” in Table 6 is “there is a warning sign”. Therefore, in the driving scene shown in FIG4B , the complexity of the second-level scene attribute 1 “temporary traffic event” is the complexity corresponding to the description information “there is a warning sign”, that is, 0.5.

[0179] For another example, in the driving scene 2 shown in FIG4B , it is assumed that the second-level scene attribute 1 is the road structure. Based on the description of FIG4B , it can be known that the description information of the road structure is a long straight road. Based on the configuration information of the road structure shown in Table 1 , the complexity of the second-level scene attribute 1 "road structure" is determined to be the complexity corresponding to the description information "long straight road" in Table 1, that is, 0.25.

[0180] For the driving scenario 1 shown in FIG4A above, the above method can be used to determine that in driving scenario 1, the complexity of the second-level scenario attribute “road structure” is 0.75, the complexity of the second-level scenario attributes “lane line”, “slope”, “weather type”, “traffic sign” and “temporary traffic event” are all 0, the complexity of the second-level scenario attribute “type of road user” is 0.25, the complexity of the second-level scenario attribute “main vehicle speed” is 0.5, and the complexity of the second-level scenario attribute “connected information type” is 0.2.

[0181] For driving scenario 2 shown in FIG4B above, the above method can be used to determine that in driving scenario 2, the complexity of the second-level scenario attribute “road structure” is 0.25, the complexity of the second-level scenario attributes “lane line”, “slope” and “connected information type” are all 0, the complexity of the second-level scenario attribute “weather type” is 0.75, the complexity of the second-level scenario attribute “traffic sign” is 0.5, the complexity of the second-level scenario attribute “temporary traffic event” is 0.5, the complexity of the second-level scenario attribute “type of road user” is 0.25, and the complexity of the second-level scenario attribute “main vehicle speed” is 0.5.

[0182] For driving scenario 3 shown in FIG4C above, the above method can be used to determine that in driving scenario 3, the complexity of the second-level scenario attribute “road structure” is 0.75, the complexity of the second-level scenario attribute “lane line” is 0.5, the complexity of the second-level scenario attribute “slope” is 0, the complexity of the second-level scenario attribute “weather type” is 0.75, the complexity of the second-level scenario attribute “traffic sign” is 0.5, the complexity of the second-level scenario attribute “temporary traffic event” is 0.75, the complexity of the second-level scenario attribute “type of road user” is 0.75, the complexity of the second-level scenario attribute “main vehicle speed” is 0.5, and the complexity of the second-level scenario attribute “connected information type” is 0.2.

[0183] S303: Determine the weight of each of the multiple scene attributes, and perform weighted summation on the complexity of each of the multiple scene attributes using the weight to obtain the complexity of the driving scene.

[0184] The sum of the weights of multiple scene attributes is 1.

[0185] In one implementation, determining the weight of each of the multiple scene attributes includes: determining the weight of each of the multiple scene attributes based on mapping information 2, where mapping information 2 is used to indicate a correspondence between scene attributes and weights.

[0186] Furthermore, since scene attributes include first-level scene attributes and second-level scene attributes, determining the weight of each of these multiple scene attributes based on mapping information 2 includes: determining the weight of each first-level scene attribute based on mapping information 2, and then determining the weight of each corresponding second-level scene attribute based on the weight of each first-level scene attribute. In this case, mapping information 2 stores the correspondence between first-level scene attributes and weights. It will be understood that the sum of the weights of the multiple second-level scene attributes corresponding to these multiple first-level scene attributes is 1.

[0187] In some possible embodiments, the mapping information 2 may also store the correspondence between the second-level scene attributes and the weights. In this case, the weight of each second-level scene can be determined directly based on the mapping information 2. Alternatively, the mapping information 2 may store both the correspondence between the first-level scene attributes and the weights and the correspondence between the second-level scene attributes and the weights, which is not specifically limited here.

[0188] In one implementation, the weight of the second-level scene attribute is obtained based on the weight of the first-level scene attribute corresponding to the second-level scene attribute.

[0189] In one implementation, the weight of each second-level scene attribute is the same, or the weight of the second-level scene attribute is associated with the degree of influence of the second-level scene attribute on the complexity of the driving scene.

[0190] For example, the weights of the plurality of second-level scene attributes may be allocated in any of the following ways:

[0191] evenly distributed; or,

[0192] The influence of the above-mentioned multiple first-level scene attributes on the complexity of the driving scene is unevenly distributed.

[0193] Assuming that the aforementioned multiple first-level scene attributes include road, weather, environment, host vehicle, and network connection, the two weight allocation methods are explained:

[0194] Method 1: Use uniform distribution

[0195] That is, the weights of the five first-level scene attributes of road, weather, environment, main vehicle and network connection are Furthermore, in Figure 2, the weights of the three second-level scene attributes, road structure, lane line, and slope, are obtained by assigning the weights of the corresponding first-level scene attribute "road", for example The weight of the second-level scene attribute "weather type" is the weight of the first-level scene attribute "weather", so Assuming that the first-level scene attribute is environment, and the second-level scene attributes are traffic signs, temporary traffic events, and types of road users, the weights of the three second-level scene attributes, traffic signs, temporary traffic events, and types of road users, are obtained by distributing them based on the weight of the first-level scene attribute “environment”, for example: The weight of the second-level scene attribute "main vehicle speed" The weight of the second-level scenario attribute "connected information type"

[0196] In this case, based on the weights determined above, the complexity of each second-level scene attribute under the driving scenario 1 determined in S302 above is weighted and summed, and the complexity of the driving scenario 1 is obtained to be 0.21; the complexity of each second-level scene attribute under the driving scenario 2 determined in S302 above is weighted and summed, and the complexity of the driving scenario 2 is obtained to be 0.35; the complexity of each second-level scene attribute under the driving scenario 3 determined in S302 above is weighted and summed, and the complexity of the driving scenario 3 is obtained to be 0.49.

[0197] Method 2: Using non-uniform distribution

[0198] For example, based on the degree of influence of multiple first-level scene attributes on the complexity of the driving scene, the weights of these multiple first-level scene attributes should meet the following conditions: weight of environment > weight of road > weight of weather > weight of main vehicle > weight of network connection. A set of values ​​for the weights of these five first-level scene attributes is, for example, Furthermore, in Figure 2, the weights of the three second-level scene attributes, road structure, lane lines, and slope, are based on the weight of the first-level scene attribute "road". To allocate, for example The weight of the second-level scene attribute "weather type" is the weight of the first-level scene attribute "weather", so Assuming that the first-level scene attribute is environment, the second-level scene attributes are traffic signs, temporary traffic events, and types of road users. The weights of the three second-level scene attributes, traffic signs, temporary traffic events, and types of road users, are based on the weight of the first-level scene attribute "environment". To allocate, for example The weight of the second-level scene attribute "main vehicle speed" The weight of the second-level scenario attribute "connected information type"

[0199] In this case, based on the weights determined above, the complexity of each second-level scene attribute under the driving scenario 1 determined in S302 above is weighted and summed, and the complexity of the driving scenario 1 is obtained to be 0.17; the complexity of each second-level scene attribute under the driving scenario 2 determined in S302 above is weighted and summed, and the complexity of the driving scenario 2 is obtained to be 0.5; the complexity of each second-level scene attribute under the driving scenario 3 determined in S302 above is weighted and summed, and the complexity of the driving scenario 3 is obtained to be 0.54.

[0200] As can be seen, the implementation of the present application embodiment, which assesses scene complexity through a hierarchical and detailed evaluation of at least three dimensions—road, weather, and environment—can reasonably and effectively assess the complexity of driving scenarios, thereby improving the accuracy of driving scene complexity assessments. Furthermore, the driving scene complexity obtained in this manner can provide a reference for establishing a baseline for access control.

[0201] In some possible embodiments, the present application also provides a framework for evaluating the functional complexity of a vehicle's autonomous driving system. Referring to FIG5 , FIG5 is a schematic diagram of a framework for evaluating the functional complexity of an autonomous driving system provided by an embodiment of the present application.

[0202] As shown in Figure 5, the multiple constraints used to assess the functional complexity of an autonomous driving system include multiple levels of ODD boundary constraints, function activation constraints, takeover constraints, ODD dependency constraints, ego vehicle constraints, emergency response constraints, and perception function restriction constraints. To accurately assess the functional complexity of an autonomous driving system, the constraints include both first-level and second-level constraints, where the second-level constraints are a refinement of the first-level constraints.

[0203] Exemplarily, when the first-level condition constraint is an ODD boundary constraint, the second-level condition constraint includes an ODD boundary range; when the first-level condition constraint is a function activation condition constraint, the second-level condition constraint is at least one of the speed difference of the function and the stable following time; when the first-level condition constraint is a takeover condition constraint, the second-level condition constraint includes a takeover time range; when the first-level condition constraint is an ODD dependency constraint, the second-level condition constraint is at least one of the speed difference of the ODD and the change of the lead vehicle; when the first-level condition constraint is an ego-vehicle constraint, the second-level condition constraint is at least one of the main vehicle speed, the driver and passenger fatigue status, and the vehicle function status; when the first-level condition constraint is an emergency response constraint, the second-level condition constraint includes automatic braking; when the first-level condition constraint is a perception function limited constraint, the second-level condition constraint includes sensor detection.

[0204] In an embodiment of the present application, each conditional constraint is pre-configured with multiple descriptive information and the complexity corresponding to each descriptive information. Here, the correspondence between the descriptive information and the complexity of the conditional constraint can be stored as a kind of mapping information. Exemplarily, when the descriptive information of conditional constraint 1 is the second description, the complexity of conditional constraint 1 is the complexity corresponding to the second description. Here, conditional constraint 1 can be a first-level conditional constraint or a second-level conditional constraint, which is not specifically limited here. It can be understood that for the same conditional constraint, different descriptive information corresponds to different complexities.

[0205] In one implementation, the multiple conditional constraints include the above-mentioned ODD boundary constraint, function activation conditional constraint, and takeover conditional constraint.

[0206] In one implementation, the first-level conditional constraint is an ODD boundary constraint, and the second-level conditional constraint includes an ODD boundary range. As an example, see Table 9, which shows the correspondence between the description information of the pre-configured ODD boundary range and the complexity of the ODD boundary range. As can be seen from Table 9, the "ODD boundary range" is configured with three types of description information, namely, being within the ODD, being at the ODD boundary, and being outside the ODD. When the description information of the ODD boundary range is "being within the ODD", its corresponding complexity is 0; when the description information of the ODD boundary range is "being at the ODD boundary", its corresponding complexity is 0.5; when the description information of the ODD boundary range is "being outside the ODD", its corresponding complexity is 1.

[0207] Table 9

[0208] Here, the description information of the ODD boundary range and the configuration of the complexity corresponding to the description information are not limited to those shown in Table 9.

[0209] In one implementation, taking the first-level constraint as the function activation constraint and the second-level constraint as the speed difference and stable following time as an example, the corresponding relationship between the descriptive information and complexity of each second-level constraint configuration is described separately:

[0210] (1) Functional speed difference

[0211] As an example, Table 10 shows the correspondence between the description information of the preconfigured speed difference function and the complexity of the speed difference function. As can be seen from Table 10, the "speed difference function" is configured with two description information: the relative speed difference between the two vehicles that satisfies the activation of the autonomous driving system function, and the relative speed difference between the two vehicles that does not satisfy the activation of the autonomous driving system function. When the speed difference function description information is "the relative speed difference between the two vehicles that satisfies the activation of the autonomous driving system function," its corresponding complexity is 0; when the speed difference function description information is "the relative speed difference between the two vehicles that does not satisfy the activation of the autonomous driving system function," its corresponding complexity is 1.

[0212] Table 10

[0213] Here, the description information of the speed difference of the function and the configuration of the complexity corresponding to the description information are not limited to those shown in Table 10.

[0214] (2) Stable following time

[0215] As an example, Table 11 shows the correspondence between the preconfigured description information of the stable following time and the complexity of the stable following time. As can be seen from Table 11, the "stable following time" configuration has two description information: the stable following time that satisfies the activation of the autonomous driving system function and the stable following time that does not. When the description information of the stable following time is "stable following time that satisfies the activation of the autonomous driving system function," the corresponding complexity is 0; when the description information of the stable following time is "stable following time that does not satisfy the activation of the autonomous driving system function," the corresponding complexity is 1.

[0216] Table 11

[0217] Here, the description information of the stable following time and the configuration of the complexity corresponding to the description information are not limited to those shown in Table 11.

[0218] In one implementation, the first-level condition constraint is a takeover condition constraint, and the second-level condition constraint includes a takeover time range. As an example, see Table 12, which shows the correspondence between the descriptive information of the pre-configured takeover time range and the complexity of the takeover time range. As can be seen from Table 12, the "takeover time range" is configured with two descriptive information, namely, when takeover occurs and when takeover does not occur. When the descriptive information of the takeover time range is "takeover occurs", its corresponding complexity is 0; when the descriptive information of the takeover time range is "takeover does not occur", its corresponding complexity is 1.

[0219] Table 12

[0220] Here, the description information of the takeover time range and the configuration of the complexity corresponding to the description information are not limited to those shown in Table 12.

[0221] In some possible embodiments, the multiple conditional constraints further include at least one of an ODD dependency constraint, a self-vehicle constraint, an emergency response constraint, and a perception function limitation constraint.

[0222] In one implementation, taking the first-level constraint as an ODD dependency constraint and the second-level constraints including the ODD speed difference and the change of the leading vehicle as an example, the corresponding relationship between the descriptive information and complexity of each second-level constraint configuration is described separately:

[0223] (1) ODD speed difference

[0224] As an example, Table 13 shows the correspondence between the pre-configured ODD speed difference description information and the complexity of the ODD speed difference. As can be seen from Table 13, the "ODD speed difference" configuration has two description information, namely, the relative speed difference between the two vehicles that meets the ODD range and the relative speed difference between the two vehicles that does not meet the ODD range. When the ODD speed difference description information is "the relative speed difference between the two vehicles that meets the ODD range," its corresponding complexity is 0; when the ODD speed difference description information is "the relative speed difference between the two vehicles that does not meet the ODD range," its corresponding complexity is 1. Here, the configuration of the ODD speed difference description information and the complexity corresponding to the description information is not limited to only those shown in Table 13.

[0225] Table 13

[0226] (2) Changes in the guide vehicle

[0227] As an example, Table 14 shows the correspondence between the preconfigured description information for the guide vehicle change and the complexity of the guide vehicle change. As can be seen from Table 14, "Guide vehicle change" is configured with two description information: the relative speed difference between the two vehicles that satisfies the ODD range, and the relative speed difference between the two vehicles that does not. When the description information for the guide vehicle change is "the guide vehicle has not undergone a change that causes it to exit the ODD," the corresponding complexity is 0; when the description information for the guide vehicle change is "the guide vehicle has undergone a change that causes it to exit the ODD," the corresponding complexity is 1. The description information for the guide vehicle change and the corresponding complexity are not limited to the configuration shown in Table 14.

[0228] Table 14

[0229] In one implementation, the first-level conditional constraint is the ego vehicle constraint, and the second-level conditional constraint includes the host vehicle speed. As an example, Table 15 shows the correspondence between the pre-configured host vehicle speed description information and the complexity of the host vehicle speed. As can be seen from Table 15, the "host vehicle speed" configuration has two description information, namely, satisfying the ego vehicle speed restriction and not satisfying the ego vehicle speed restriction. When the host vehicle speed description information is "satisfying the ego vehicle speed restriction", the corresponding complexity is 0; when the host vehicle speed description information is "not satisfying the ego vehicle speed restriction", the corresponding complexity is 1. Here, the configuration of the host vehicle speed description information and the complexity corresponding to the description information is not limited to only those shown in Table 15.

[0230] Table 15

[0231] In some possible embodiments, when the first-level constraint is the ego-vehicle constraint, the second-level constraint also includes at least one of the two second-level constraints: the driver fatigue state and the vehicle functional state. In this case, the functional complexity of the autonomous driving system can be evaluated based on ODC. Taking the second-level constraint of driver fatigue state as an example, if the description of the driver fatigue state is "not fatigued", the corresponding complexity can be set to 0; if the description of the driver fatigue state is "fatigued", the corresponding complexity can be set to 1.

[0232] In one implementation, the first-level conditional constraint is an emergency response constraint, and the second-level conditional constraint includes automatic braking. As an example, Table 16 shows the correspondence between the pre-configured descriptive information of automatic braking and the complexity of automatic braking. As can be seen from Table 16, the "automatic braking" configuration has two descriptive information, namely, satisfying the conditions of no automatic braking and automatic braking. When the descriptive information of the automatic braking is "no automatic braking", the corresponding complexity is 0; when the descriptive information of the automatic braking is "automatic braking", the corresponding complexity is 1. Here, the configuration of the descriptive information of the automatic braking and the complexity corresponding to the descriptive information is not limited to that shown in Table 16.

[0233] Table 16

[0234] In one implementation, the first-level constraint is a perception function restriction constraint, and the second-level constraint includes sensor detection. As an example, Table 17 shows the correspondence between the pre-configured sensor detection descriptive information and the complexity of sensor detection. As can be seen from Table 17, the second-level constraint configuration of sensor detection has three types of descriptive information, namely normal, detection range restricted, and malfunctioning. When the sensor detection descriptive information is "normal," the corresponding complexity is 0; when the sensor detection descriptive information is "detection range restricted," the corresponding complexity is 0.5; and when the sensor detection descriptive information is "malfunctioning," the corresponding complexity is 1. Here, the configuration of the sensor detection descriptive information and the complexity corresponding to the descriptive information is not limited to that shown in Table 17.

[0235] Table 17

[0236] In the embodiment of the present application, each conditional constraint used to evaluate the functional complexity of the autonomous driving system is also configured with a weight. The sum of the weights of multiple conditional constraints is 1. When the conditional constraints include first-level conditional constraints and second-level conditional constraints, each second-level conditional constraint also corresponds to a weight, and the weight of the second-level conditional constraint is related to the weight of the first-level conditional constraint. For example, as can be seen from Figure 5, the weight of the ODD boundary range is w 11 , the weight of the speed difference of the function is w 21 , the weight of the stable following time is w 22 , the weight of the takeover time range is w 31 ,…, the weights of the remaining second-level condition constraints can be specifically shown in Figure 5 and will not be repeated here.

[0237] For example, the sum of the weights of multiple second-level condition constraints participating in the functional complexity evaluation of the autonomous driving system should be 1.

[0238] In one implementation, the weight of each second-level conditional constraint is related to the weight of the first-level conditional constraint corresponding to the second-level conditional constraint.

[0239] For example, if the first-level constraint is B and the second-level constraint is only b1, the weight of the second-level constraint b1 is the weight of the first-level constraint B. If the second-level constraints include b1, b2, and b3, the weights of these three second-level constraints are distributed based on the weight of the first-level constraint B. The distribution method can be, for example, even or in other ways. In other words, the sum of the weights of the second-level constraints b1, b2, and b3 is the weight of the first-level constraint B.

[0240] Refer to Figure 6, which is a flow chart of a method for determining the functional complexity of an autonomous driving system provided in an embodiment of the present application. The method can be applied to the first device shown in Figure 1, and the scheme is exemplified by taking the first device evaluating the functional complexity of the autonomous driving system of the first vehicle in the driving scenario as an example. The first vehicle is the main vehicle, and the first vehicle is equipped with an autonomous driving system. Accordingly, the first device can be the first vehicle or a component within the first vehicle, or it can be a network-side device independent of the first vehicle or a component within a network-side device. The component can be, for example, a chip or an integrated circuit, etc., which is not specifically limited here. The method includes but is not limited to the following steps:

[0241] S601: Determine descriptive information of each of multiple conditional constraints of an autonomous driving system in a driving scenario.

[0242] Here, the autonomous driving system is not limited to a fully autonomous driving system, a highly autonomous driving system, a conditional autonomous driving system, or a partially autonomous driving system. Those skilled in the art can understand that non-fully manual driving systems that provide intelligent driving can be covered under this concept.

[0243] Exemplarily, the multiple conditional constraints include multiple items of ODD boundary constraints, function activation condition constraints, takeover condition constraints, ODD dependency constraints, self-vehicle constraints, emergency response constraints, and perception function limitation constraints.

[0244] In one implementation, the conditional constraints include first-level conditional constraints and second-level conditional constraints. The corresponding relationship between the first-level conditional constraints and the second-level conditional constraints can be referred to the description of the corresponding content in Figure 5 above, which will not be repeated here.

[0245] In one implementation, when the first device is a first vehicle, the description information of each conditional constraint of the autonomous driving system in the driving scenario may be locally obtained by the first device. When the first device is a network-side device independent of the first vehicle, obtaining the system description information of the first vehicle means receiving the system description information from the first vehicle. In some possible embodiments, the description information of each conditional constraint may be obtained from the first vehicle, or partially from roadside equipment, vehicles surrounding the first vehicle, etc.

[0246] It can be understood that when the conditional constraints include first-level conditional constraints and second-level conditional constraints, determining the descriptive information of each conditional constraint of the autonomous driving system under the driving scenario is equivalent to determining the descriptive information of each second-level conditional constraint of the autonomous driving system under the driving scenario.

[0247] Refer to Figure 7, which is a schematic diagram of another driving scenario provided by an embodiment of the present application. The driving scenario shown in Figure 7 is called driving scenario 4. In Figure 7, it is assumed that the automatic driving system of the first vehicle only supports single-lane cruising (that is, there needs to be a guide vehicle in front of the vehicle. Once there is no guide vehicle, the driver must be reminded to take over. If the driver does not take over within the preset time, automatic braking is performed). Taking multiple second-level conditional constraints including the second-level conditional constraints shown in Figure 5 (including ODD boundary range, speed difference of function, stable following time, takeover time range, ODD speed difference, change of guide vehicle, main vehicle speed, automatic braking and sensor detection) as an example, it is assumed that the first device determines that the description information of "ODD boundary range" in driving scenario 4 is at the ODD boundary (because the guide vehicle in front of the vehicle Changing lanes), the descriptive information of "Functional speed difference" is the relative speed difference between the two vehicles that satisfies the activation of the automatic driving system function, the descriptive information of "Stable following time" is the stable following time that satisfies the activation of the automatic driving system function, the descriptive information of "Takeover time range" is that takeover occurs, the descriptive information of "ODD speed difference" is the relative speed difference between the two vehicles that satisfies the ODD range, the descriptive information of "Lead vehicle change" is that the lead vehicle has undergone a change that causes the exit from ODD, the descriptive information of "Main vehicle speed" is that the speed limit of the own vehicle is met, the descriptive information of "Automatic braking" is that automatic braking does not occur, and the descriptive information of "Sensor detection" is normal.

[0248] S602: Determine the complexity of each of the multiple conditional constraints based on mapping information 3, where mapping information 3 is used to indicate a corresponding relationship between description information and complexity.

[0249] Exemplarily, mapping information 3 includes the corresponding relationships shown in Tables 9 to 17 above.

[0250] In one implementation, since the conditional constraints include first-level conditional constraints and second-level conditional constraints, the complexity of each of the multiple conditional constraints is determined based on the mapping information 3, including: determining the complexity of each second-level conditional constraint in the multiple second-level conditional constraints based on the mapping information 3. In this case, the mapping information 3 stores the correspondence between the descriptive information of the second-level conditional constraints and the complexity.

[0251] Taking the second-level conditional constraint 1 as an example, the complexity of the second-level conditional constraint 1 is determined based on the mapping information 3, including: determining the target description information of the second-level conditional constraint 1 in the mapping information 3 that is closest to the description information of the second-level conditional constraint 1 of the automatic driving system under the driving scenario, and taking the complexity corresponding to the target description information of the second-level conditional constraint 1 in the mapping information 3 as the complexity of the second-level conditional constraint 1 under the driving scenario.

[0252] In some possible embodiments, the description information of the second-level condition constraint 1 of the automatic driving system in the driving scenario (for example, description information A) belongs to the pre-configured description information of the second-level condition constraint 1 in the mapping information 3. In this case, the complexity corresponding to the description information A of the second-level condition constraint 1 found in the mapping information 3 is used as the complexity of the second-level condition constraint 1 in the driving scenario.

[0253] For the first vehicle (as the main vehicle) in the driving scenario shown in Figure 7, based on the description of Figure 7 and mapping information 3, it can be determined that in the above driving scenario 4, the complexity of the second-level condition constraint "ODD boundary range" is 0.5, the complexity of the second-level condition constraint "change of guide vehicle" is 1, the complexity of the second-level condition constraint "takeover time range" is 1, and the complexity of the second-level condition constraints "speed difference of function", "stable following time", "speed difference of ODD", "main vehicle speed", "automatic braking" and "sensor detection" are all 0.

[0254] S603: Determine the weight of each of the multiple conditional constraints, and perform weighted summation of the complexity of each of the multiple conditional constraints using the weights to obtain the functional complexity of the autonomous driving system under the driving scenario.

[0255] The sum of the weights of multiple condition constraints is 1.

[0256] In one implementation, determining the weight of each of the multiple conditional constraints includes: determining the weight of each of the multiple conditional constraints based on mapping information 4, where mapping information 4 is used to indicate a correspondence between conditional constraints and weights.

[0257] Furthermore, since the conditional constraints include first-level conditional constraints and second-level conditional constraints, determining the weight of each of these multiple conditional constraints based on mapping information 4 includes: determining the weight of each first-level conditional constraint based on mapping information 4, and then determining the weight of each corresponding second-level conditional constraint based on the weight of each first-level conditional constraint. In this case, mapping information 4 stores the correspondence between first-level conditional constraints and weights. It will be understood that the sum of the weights of the multiple second-level conditional constraints corresponding to these multiple first-level conditional constraints is 1.

[0258] In some possible embodiments, the mapping information 4 may also store the correspondence between the second-level conditional constraints and weights. In this case, the weight of each second-level scenario can be determined directly based on the mapping information 4. Alternatively, the mapping information 4 may store both the correspondence between the first-level conditional constraints and weights and the correspondence between the second-level conditional constraints and weights, which is not specifically limited here.

[0259] In one implementation, the weight of the second-level conditional constraint is obtained based on the weight of the first-level conditional constraint corresponding to the second-level conditional constraint.

[0260] In one implementation, the weight of each second-level constraint is the same, or the weight of the second-level constraint is associated with the degree of influence of the second-level constraint on the complexity of the driving scenario.

[0261] For example, the weights of the above-mentioned multiple second-level condition constraints can be allocated in any of the following ways:

[0262] evenly distributed; or,

[0263] Based on the above multiple first-level condition constraints, the degree of influence on the functional complexity of the autonomous driving system is unevenly distributed.

[0264] Taking Figure 5 as an example, assuming that the multiple first-level constraints mentioned above include ODD boundary constraints, function activation constraints, takeover constraints, ODD dependency constraints, ego vehicle constraints, emergency response constraints, and perception function restriction constraints, these two weight distribution methods are explained:

[0265] Method 1: Use uniform distribution

[0266] That is, the weights of the seven first-level constraints, namely, ODD boundary constraints, function activation condition constraints, takeover condition constraints, ODD dependency constraints, vehicle constraints, emergency response constraints, and perception function restriction constraints, are all Based on the correspondence between the first-level constraints and the second-level constraints in Figure 5, the weights of the second-level constraints can be further determined. Among them, the weight of the second-level constraint "ODD boundary range" is The weights of the two second-level constraints, speed difference and stable following time, are obtained by distributing the weights of the first-level constraint, “function activation constraint”. For example, The second-level condition constrains the weight of the "takeover time range" The weights of the two second-level constraints, ODD speed difference and lead car change, are obtained by distributing the weights of the first-level constraint "ODD dependency constraint", for example The weight of the second-level condition constraint "main vehicle speed" The second-level condition constrains the weight of "automatic braking" The second-level condition constrains the weight of "sensor detection"

[0267] In this case, based on the weights determined above, the complexity of each of the multiple second-level condition constraints in the driving scenario determined in S602 above is weightedly summed to obtain the functional complexity of the automatic driving system in the driving scenario as 0.29.

[0268] Method 2: Using non-uniform distribution

[0269] For example, based on the degree of influence of multiple first-level constraints on the functional complexity of the autonomous driving system, the weights of these multiple first-level constraints should meet the following conditions: weight of ODD boundary constraint > weight of perception function restriction constraint > weight of ODD dependency constraint > weight of vehicle constraint > weight of function activation constraint > weight of takeover constraint > weight of emergency response constraint. A set of values ​​for the weights of these seven first-level constraints is, for example, Furthermore, in Figure 5, the second-level condition constrains the weight of the “ODD boundary range” The weights of the two second-level constraints, speed difference and stable following time, are obtained by distributing the weights of the first-level constraint, “function activation constraint”. For example, The second-level condition constrains the weight of the "takeover time range" The weights of the two second-level constraints, ODD speed difference and lead car change, are obtained by distributing the weights of the first-level constraint "ODD dependency constraint", for example The weight of the second-level condition constraint "main vehicle speed" The second-level condition constrains the weight of "automatic braking" The second-level condition constrains the weight of "sensor detection"

[0270] In this case, based on the weights determined above, the complexity of each second-level conditional constraint in the multiple second-level conditional constraints in the driving scenario determined in S602 above is weightedly summed to obtain the functional complexity of the automatic driving system in the driving scenario as 0.29.

[0271] It is understood that when any vehicle equipped with an autonomous driving system in a driving scenario serves as the primary vehicle, the functional complexity of the vehicle's autonomous driving system in that driving scenario can be evaluated using the above method. For example, in a driving scenario, there are vehicle 1 and vehicle 2. When vehicle 1 serves as the primary vehicle, the functional complexity of vehicle 1's autonomous driving system in that driving scenario is determined to be a first value using the above method. When vehicle 2 serves as the primary vehicle, the functional complexity of vehicle 2's autonomous driving system in that driving scenario is determined to be a second value using the above method. If the first value is greater than the second value, it indicates that the performance of vehicle 1's autonomous driving system is superior to that of vehicle 2's autonomous driving system.

[0272] As can be seen, the implementation of the present invention provides a hierarchical and detailed evaluation of the complexity of an autonomous driving system based on at least three dimensions: ODD boundary constraints, function activation condition constraints, and takeover condition constraints. This allows for a reasonable and effective assessment of the functional complexity of autonomous driving systems in driving scenarios. Furthermore, this evaluation method is applicable to different autonomous driving systems and provides a reference for performance evaluation of these systems.

[0273] 8 is a schematic diagram of the structure of a computing device provided in an embodiment of the present application, wherein the computing device 30 includes an acquisition unit 310 and a processing unit 312. The computing device 30 can be implemented by hardware, software, or a combination of hardware and software.

[0274] In one implementation, the acquisition unit 310 is used to determine the descriptive information of each scene attribute in multiple scene attributes of the driving scene; the processing unit 312 is used to: determine the complexity of each scene attribute in the multiple scene attributes based on first mapping information, wherein the first mapping information is used to indicate the correspondence between the descriptive information and the complexity of the scene attribute; determine the weight of each scene attribute in the multiple scene attributes; and obtain the complexity of the driving scene by weighted summing the complexity of each scene attribute in the multiple scene attributes based on the above weight.

[0275] In this case, the computing device 30 may be used to implement the method described in the embodiment of Figure 3. In the embodiment of Figure 3, the acquiring unit 310 may be used to execute S301, and the processing unit 312 may be used to execute S302 and S303.

[0276] In another implementation, the acquisition unit 310 is used to determine the descriptive information of each of the multiple conditional constraints of the autonomous driving system in a driving scenario; the processing unit 312 is used to: determine the complexity of each of the multiple conditional constraints based on the first mapping information, wherein the first mapping information is used to indicate the correspondence between the descriptive information and the complexity of the conditional constraint; determine the weight of each of the multiple conditional constraints; and obtain the functional complexity of the autonomous driving system in the driving scenario by weighted summing the complexity of each of the multiple conditional constraints based on the above-mentioned weight.

[0277] In this case, the computing device 30 may also be used to implement the method described in the embodiment of Figure 6. In the embodiment of Figure 6, the acquiring unit 310 may be used to execute S601, and the processing unit 312 may be used to execute S602 and S603.

[0278] It should be understood that the division of the various units in the computing device 30 described above is merely a division of logical functions. In actual implementation, they may be fully or partially integrated into a single physical entity, or they may be physically separated. Furthermore, the units in the device may be implemented in the form of a processor calling software; for example, the device may include a processor connected to a memory storing instructions, and the processor calling the instructions stored in the memory to implement any of the above methods or functions of the various units of the device, wherein the processor may be, for example, a general-purpose processor such as a central processing unit (CPU) or a microprocessor, and the memory may be a memory within the device or a memory external to 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. The hardware circuit can be understood as one or more processors. For example, in one implementation, the hardware circuit is an application-specific integrated circuit (ASIC), which realizes the functions of some or all of the above units by designing the logical relationship of the components in the circuit. For another example, in another implementation, the hardware circuit can be implemented by a programmable logic device (PLD). Taking a field programmable gate array (FPGA) as an example, it can include a large number of logic gate circuits, and the connection relationship between the logic gate circuits is configured by configuring the configuration file, thereby realizing the functions of some or all of the above units. All units of the above devices can be implemented in the form of software called by the processor, or in the form of hardware circuits, or in part by software called by the processor, and the rest by hardware circuits.

[0279] In an embodiment of the present application, a processor is a circuit with a signal processing capability. In one implementation, the processor can be a circuit with instruction reading and execution 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). In another implementation, the processor can implement certain functions through the logical relationship of a hardware circuit. 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.

[0280] 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.

[0281] In addition, the various units in the above devices can be fully or partially 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 the various units of the device. The type of the at least one processor can be different, for example, including a CPU and FPGA, a CPU and an artificial intelligence processor, a CPU and a GPU, etc.

[0282] See Figure 9, which is a schematic diagram of the structure of a computing device provided in an embodiment of the present application. As shown in Figure 9, computing device 40 includes: a processor 401, a communication interface 402, a memory 403, and a bus 404. Processor 401, memory 403, and communication interface 402 communicate with each other via bus 404. It should be understood that this application does not limit the number of processors and memories in computing device 40.

[0283] 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 evaluation server, an autonomous driving system evaluation server, etc.), or a component in the server (e.g., a chip, an integrated circuit, etc.), or a system-level device composed of multiple servers. The network-side device may be deployed in a cloud environment or an edge environment.

[0284] In another implementation, the computing device 40 is a vehicle or a component in a vehicle, and the component may be, for example, a chip, an integrated circuit, or the like.

[0285] Bus 404 may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, among others. Buses may be classified as address buses, data buses, control buses, and the like. For ease of illustration, FIG9 illustrates a single bus line, but this does not imply a single bus or type of bus. Bus 404 may include a path for transmitting information between various components of computing device 40 (e.g., memory 403, processor 401, and communication interface 402).

[0286] The processor 401 can refer to the relevant description of the processor in the above embodiment, which will not be repeated here.

[0287] Memory 403 is used to provide storage space for storing data such as the operating system and computer programs. Memory 403 can be one or a combination of random access memory (RAM), erasable programmable read-only memory (EPROM), read-only memory (ROM), or compact disc read-only memory (CD-ROM). Memory 403 can exist independently or be integrated into processor 401.

[0288] The communication interface 402 can be used to provide information input or output for the processor 401. Alternatively, the communication interface 402 can be used to receive data transmitted externally and / or transmit data externally. It can 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 can also include a transmitter (such as a radio frequency transmitter, antenna, etc.) or a receiver coupled to the interface.

[0289] In some possible embodiments, when the computing device 40 further includes a display (not shown), the display is connected or coupled to the processor 401 via a bus 404. The display can be used to display the complexity of the above-mentioned driving scene, or to display the functional complexity of the above-mentioned first automatic driving system in the driving scene. The display 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. In another implementation, the display can also be a vehicle-mounted tablet, a vehicle-mounted display, or a head-up display (HUD) system, etc.

[0290] 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, such as the method described in FIG. 3 or FIG. 6 .

[0291] In one possible design, the computing device 40 may be one or more modules in an execution entity that executes the method shown in FIG. 3 , and the processor 401 may be configured to read one or more computer programs stored in a memory to perform the following operations:

[0292] Determining, by the acquisition unit 310, descriptive information of each of the plurality of scene attributes of the driving scene;

[0293] The complexity of each of the multiple scene attributes is determined based on the first mapping information, where the first mapping information is used to indicate the correspondence between the descriptive information of the scene attribute and the complexity; a weight is determined for each of the multiple scene attributes; and the complexity of the driving scene is obtained by weighted summing the complexity of each of the multiple scene attributes based on the weight. In one possible design, the computing device 40 may be one or more modules in the execution body of the method shown in FIG6 , and the processor 401 may be configured to read one or more computer programs stored in the memory to perform the following operations:

[0294] Determining, by the acquisition unit 310, descriptive information of each of the plurality of conditional constraints of the autonomous driving system in the driving scenario;

[0295] Based on the first mapping information, the complexity of each of the multiple conditional constraints is determined, wherein the first mapping information is used to indicate the correspondence between the descriptive information of the conditional constraint and the complexity; the weight of each of the multiple conditional constraints is determined; and the functional complexity of the autonomous driving system in the driving scenario is obtained by weighted summing the complexity of each of the multiple conditional constraints based on the above weight.

[0296] In the embodiments described above, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a particular embodiment, please refer to the relevant descriptions of other embodiments. In addition, in the various embodiments of this application, unless otherwise specified or there is a 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 new embodiments based on their inherent logical relationships.

[0297] It should be noted that, those 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 a program to instruct relevant hardware. 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.

[0298] The technical solution of the present application may essentially or contribute to the part or all or part of the technical solution 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 may 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 complexity, characterized in that: The method comprises: Determining descriptive information of each of a plurality of scene attributes of a driving scene; Determining the complexity of each of the multiple scene attributes based on first mapping information, wherein the first mapping information is used to indicate a corresponding relationship between the description information and the complexity; determining a weight of each of the plurality of scene attributes; The complexity of the driving scene is obtained by performing weighted summation on the complexity of each of the multiple scene attributes based on the weight.

2. The method according to claim 1, characterized in that The determining the weight of each scene attribute in the plurality of scene attributes comprises: The weight of each of the plurality of scene attributes is determined based on second mapping information, wherein the second mapping information is used to indicate a corresponding relationship between the scene attribute and the weight.

3. The method according to claim 1 or 2, characterized in that: The sum of the weights of the multiple scene attributes is 1.

4. The method according to any one of claims 1 to 3, characterized in that: The multiple scene attributes include multiple items of road, weather, environment, host vehicle and network connection.

5. The method according to any one of claims 1 to 4, characterized in that: The scene attributes include first-level scene attributes and second-level scene attributes. When the first-level scene attribute is road, the second-level scene attribute is at least one of road structure, lane line and slope; when the first-level scene attribute is weather, the second-level scene attribute includes weather type; when the first-level scene attribute is environment, the second-level scene attribute is at least one of traffic signs, temporary traffic events, types of road users and the number of road users; when the first-level scene attribute is the main vehicle, the second-level scene attribute includes the main vehicle speed; when the first-level scene attribute is networking, the second-level scene attribute includes networking information type.

6. The method according to any one of claims 1 to 5, characterized in that: The weight of each scene attribute is the same, or the weight of the scene attribute is associated with the degree of influence of the scene attribute on the complexity of the driving scene.

7. A method for determining the functional complexity of an autonomous driving system, characterized in that: The method comprises: Determine description information of each of multiple condition constraints of the autonomous driving system in a driving scenario; Determining the complexity of each of the multiple conditional constraints based on first mapping information, wherein the first mapping information is used to indicate a corresponding relationship between the description information and the complexity; Determining a weight of each of the plurality of conditional constraints; The functional complexity of the automatic driving system in the driving scenario is obtained by performing weighted summation on the complexity of each of the multiple conditional constraints based on the weight.

8. The method according to claim 7, characterized in that Determining the weight of each of the plurality of conditional constraints comprises: The weight of each of the multiple conditional constraints is determined based on second mapping information, wherein the second mapping information is used to indicate a corresponding relationship between the conditional constraints and the weight.

9. The method according to claim 7 or 8, characterized in that: The sum of the weights of the multiple conditional constraints is 1.

10. The method according to any one of claims 7 to 9, characterized in that: The multiple conditional constraints include multiple items of designed operating range ODD boundary constraints, function activation condition constraints, takeover condition constraints, ODD dependency constraints, self-vehicle constraints, emergency response constraints and perception function limitation constraints.

11. The method according to any one of claims 7 to 10, characterized in that: The conditional constraints include first-level conditional constraints and second-level conditional constraints. When the first-level conditional constraint is a designed operating range ODD boundary constraint, the second-level conditional constraint includes the ODD boundary range; when the first-level conditional constraint is a function activation conditional constraint, the second-level conditional constraint is at least one of the speed difference of the function and the stable following time; when the first-level conditional constraint is a takeover conditional constraint, the second-level conditional constraint includes the takeover time range; when the first-level conditional constraint is an ODD dependency constraint, the second-level conditional constraint is at least one of the speed difference of the ODD and the change of the guide vehicle; when the first-level conditional constraint is an ego-vehicle constraint, the second-level conditional constraint is at least one of the main vehicle speed, the fatigue state of the driver and the vehicle function state; when the first-level conditional constraint is an emergency response constraint, the second-level conditional constraint includes automatic braking; when the first-level conditional constraint is a perception function limited constraint, the second-level conditional constraint includes sensor detection.

12. The method according to any one of claims 7 to 11, characterized in that: The weight of each conditional constraint is the same, or the weight of the conditional constraint is associated with the degree of influence of the conditional constraint on the functional complexity of the autonomous driving system.

13. A device for determining scene complexity, characterized in that: The device comprises: An acquisition unit, used to determine description information of each scene attribute among a plurality of scene attributes of a driving scene; a processing unit, configured to determine the complexity of each of the plurality of scene attributes based on first mapping information, wherein the first mapping information is used to indicate a corresponding relationship between the description information and the complexity; The processing unit is further configured to determine a weight of each of the plurality of scene attributes; The processing unit is further configured to obtain the complexity of the driving scene by performing weighted summation on the complexity of each of the multiple scene attributes based on the weight.

14. The device according to claim 13, characterized in that The processing unit is specifically used for: The weight of each of the plurality of scene attributes is determined based on second mapping information, wherein the second mapping information is used to indicate a corresponding relationship between the scene attribute and the weight.

15. The device according to claim 13 or 14, characterized in that The sum of the weights of the multiple scene attributes is 1.

16. The device according to any one of claims 13 to 15, characterized in that: The multiple scene attributes include multiple items of road, weather, environment, host vehicle and network connection.

17. The device according to any one of claims 13 to 16, characterized in that: The scene attributes include first-level scene attributes and second-level scene attributes. When the first-level scene attribute is road, the second-level scene attribute is at least one of road structure, lane line and slope; when the first-level scene attribute is weather, the second-level scene attribute includes weather type; when the first-level scene attribute is environment, the second-level scene attribute is at least one of traffic signs, temporary traffic events, types of road users and the number of road users; when the first-level scene attribute is the main vehicle, the second-level scene attribute includes the main vehicle speed; when the first-level scene attribute is networking, the second-level scene attribute includes networking information type.

18. The device according to any one of claims 13 to 17, characterized in that: The weight of each scene attribute is the same, or the weight of the scene attribute is associated with the degree of influence of the scene attribute on the complexity of the driving scene.

19. A device for determining the functional complexity of an autonomous driving system, characterized in that: The device comprises: An acquisition unit, used to determine description information of each of a plurality of condition constraints of the automatic driving system under a driving scenario; A processing unit, configured to determine the complexity of each of the plurality of conditional constraints based on first mapping information, wherein the first mapping information is used to indicate a corresponding relationship between the description information and the complexity; The processing unit is further configured to determine a weight of each of the plurality of conditional constraints; The processing unit is further used to obtain the functional complexity of the automatic driving system in the driving scenario by performing weighted summation on the complexity of each of the multiple conditional constraints based on the weight.

20. The device according to claim 19, characterized in that The processing unit is specifically used for: The weight of each of the multiple conditional constraints is determined based on second mapping information, wherein the second mapping information is used to indicate a corresponding relationship between the conditional constraints and the weight.

21. The device according to claim 19 or 20, characterized in that The sum of the weights of the multiple conditional constraints is 1.

22. The device according to any one of claims 19 to 21, characterized in that The multiple conditional constraints include multiple items of designed operating range ODD boundary constraints, function activation condition constraints, takeover condition constraints, ODD dependency constraints, self-vehicle constraints, emergency response constraints and perception function limitation constraints.

23. The device according to any one of claims 19 to 22, characterized in that The conditional constraints include first-level conditional constraints and second-level conditional constraints. When the first-level conditional constraint is a designed operating range ODD boundary constraint, the second-level conditional constraint includes the ODD boundary range; when the first-level conditional constraint is a function activation conditional constraint, the second-level conditional constraint is at least one of the speed difference of the function and the stable following time; when the first-level conditional constraint is a takeover conditional constraint, the second-level conditional constraint includes the takeover time range; when the first-level conditional constraint is an ODD dependency constraint, the second-level conditional constraint is at least one of the speed difference of the ODD and the change of the guide vehicle; when the first-level conditional constraint is an ego-vehicle constraint, the second-level conditional constraint is at least one of the main vehicle speed, the fatigue state of the driver and the vehicle function state; when the first-level conditional constraint is an emergency response constraint, the second-level conditional constraint includes automatic braking; when the first-level conditional constraint is a perception function limited constraint, the second-level conditional constraint includes sensor detection.

24. The device according to any one of claims 19 to 23, characterized in that The weight of each conditional constraint is the same, or the weight of the conditional constraint is associated with the degree of influence of the conditional constraint on the functional complexity of the autonomous driving system.

25. 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 12.

26. A vehicle, characterized in that: The vehicle comprises an apparatus as claimed in any one of claims 13-24.

27. 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 12.

Images