Scene parameter determination method and device
By determining the parameter set related to the driving scenario, satisfying the motion constraints and conducting driving simulation tests, the problem of determining parameters in the test scenario of autonomous driving vehicles is solved, the rationality and effectiveness of the test scenario is improved, and the rationality and safety of the parameters are ensured. .
Patent Information
- Application Number
- CN202310544377.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-05-15
- Publication Date
- 2024-12-03
AI Technical Summary
The prior art cannot effectively determine the test scenario parameters of autonomous driving vehicles, resulting in the inability to meet the safety testing needs of autonomous driving vehicles.
By determining the set of parameters related to the driving scenario, including a motion parameter group, it satisfies the motion constraints, and obtains the parameter group that meets the test purpose through driving simulation tests, for safety testing of autonomous vehicles.
It improves the rationality and effectiveness of autonomous driving safety testing scenarios, ensures that the parameter values are reasonable and foreseeable in the physical world, and can pass the driving scenarios safely, enhancing the accuracy of the test results.
Smart Images

Figure CN119064023A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of vehicle testing, and in particular to a method and device for determining scene parameters. Background Art
[0002] The parameter values of the test scenario are very important for the safety testing of autonomous vehicles. For example, the parameter values of the test scenario are related to whether the autonomous vehicle can pass safely in the test scenario.
[0003] The current method for determining parameter values for test scenarios cannot meet the safety testing requirements of autonomous driving vehicles. How to determine reasonable values for the parameters of test scenarios is an urgent problem that needs to be solved. Summary of the invention
[0004] The present application discloses a method and device for determining scene parameters, which can determine reasonable values for the parameters of a driving scene, and is conducive to improving the rationality and effectiveness of autonomous driving safety test scenes.
[0005] In a first aspect, the present application provides a scene parameter determination method, the method comprising: determining a first parameter set related to a first driving scene, the first parameter set comprising M motion parameter groups, each motion parameter group comprising a set of values of n motion parameters, the n motion parameters being used to describe a state of a subject participating in the first driving scene, each motion parameter group in the M motion parameter groups satisfying a motion constraint, the motion constraint being used to indicate a correlation between multiple motion parameters in the n motion parameters, M and n being integers greater than 1; performing a driving simulation test using the first parameter set to obtain a second parameter set, the second parameter set comprising K motion parameter groups in the M motion parameter groups, the K motion parameter groups meeting the test purpose of the first driving scene, and K being a positive integer less than or equal to M.
[0006] Exemplarily, the method can be applied to a network side device. The network side device can be, for example, a computing device (e.g., a server) deployed on the network side, or a component or chip in the server, 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.
[0007] Here, a set of values of n motion parameters can constitute a motion parameter group, that is, a motion parameter group includes n values, each of the n values corresponds to one motion parameter among the n motion parameters, so the above M motion parameter groups are actually a grouping form of multiple sets of values of n motion parameters, and multiple motion parameter groups can constitute a parameter set, for example, the above M motion parameter groups constitute a first parameter set. Exemplarily, motion parameters can also be called scene parameters.
[0008] Exemplarily, the process of determining the first parameter set is also equivalent to the process of determining multiple sets of values of n motion parameters.
[0009] Exemplary, the driving scenario may be, for example, a vehicle entry scenario, a vehicle exit scenario, a vehicle follow-up scenario, or a front vehicle brake scenario, or the like. It can be understood that the vehicle cutting scene can also be called the front car cutting scene, and the vehicle cutting scene can also be called the front car cutting scene.
[0010] Exemplary, based on the type of cutting into a vehicle, the vehicle cutting into a vehicle cutting into a vehicle cutting into a vehicle cutting into a vehicle cutting into a vehicle cutting into a vehicle cutting into a vehicle cutting into a vehicle understanding. It can be understood that the vehicle cutting into a vehicle and a vehicle cutting into a vehicle cutting into a vehicle are different driving scenarios.
[0011] Exemplarily, the motion parameters include, but are not limited to, at least one of speed, distance, relative speed, relative distance, displacement, acceleration, angular velocity, posture, lane change time, and the like.
[0012] Here, the motion constraint is used to indicate the correlation between multiple motion parameters among n motion parameters. The correlation between multiple motion parameters can be represented by a joint distribution between multiple parameters, for example. Taking two motion parameters among multiple motion parameters as an example, the correlation between the two motion parameters can be represented by a two-dimensional joint distribution between the two parameters. The two-dimensional joint distribution is used to measure the probability distribution relationship between two parameters. Here, the constraint corresponding to the two-dimensional joint distribution is an example of the above-mentioned motion constraint. Exemplarily, the correlation between multiple motion parameters can also be represented by a three-dimensional joint distribution of parameters, or other joint distributions, etc., which are not specifically limited here.
[0013] For example, a joint distribution, also known as a joint probability distribution, may be a probability distribution of a random vector composed of two or more (including two) random variables.
[0014] Exemplary, the driver model used in the driving simulation test is obtained through driver capability training. The trained driver model has the same driving ability as the human driver, so that it can accurately simulate whether the human driver can meet the test purpose under the motion parameters in the corresponding driving scenario in the driving simulation test.
[0015] Exemplarily, the test objectives of the first driving scenario include one or more of: no collision, no distance from the vehicle in front being less than a preset threshold during braking, no skidding when exiting a curve, etc.
[0016] In the above method, a driving simulation test is performed on multiple groups of values of motion parameters that satisfy the motion constraints to determine the values of the motion parameters of the driving scene. In this way, the values of the parameters in the driving scene can be reasonable and predictable in the physical world, and the human driver can pass safely (for example, without collision), which provides reasonable values for the parameters of the driving scene, which is not only conducive to improving the accuracy of the test results, but also conducive to improving the rationality and effectiveness of the autonomous driving safety test scenario.
[0017] Optionally, the motion constraint is obtained based on natural driving data statistics in the first driving scenario.
[0018] Exemplarily, the natural driving data in the first driving scenario at least includes: the possible values of the above n parameters when a natural person participates in driving in the first driving scenario. Furthermore, the natural driving data of the first driving scenario also includes the driving trajectory information and / or motion state information of the vehicle driven by a natural person and the surrounding vehicles in the first driving scenario. Here, the possible values of the above n parameters when a natural person participates in driving in the first driving scenario can be obtained based on the driving trajectory information and / or motion state information of the vehicle driven by a natural person and the surrounding vehicles in the first driving scenario. It can be understood that the natural driving data of the first driving scenario can be driving data collected in the physical world, or data obtained based on statistical analysis of driving data collected in the physical world, or driving data provided by human drivers based on driving experience, which is not specifically limited here.
[0019] Implementing the above implementation method, since the natural driving data is actually data that will appear in the physical world, the motion constraints obtained based on the statistical analysis of the natural driving data are also reasonable and foreseeable. Therefore, the M motion parameter groups in the first parameter set are the values that will appear in the first driving scenario.
[0020] Optionally, the second parameter set is used for safety testing of the autonomous driving vehicle in the first driving scenario.
[0021] Optionally, the second parameter set is used to test whether the vehicle meets the test purpose.
[0022] Exemplary, the second parameter set can be provided to the developer of the vehicle or the autonomous driving system, or to the test mechanism of the autonomous driving vehicle, etc., which is conducive to improving the safety of the autonomous driving system and improving the rationality and effectiveness of the autonomous driving safety test scenario.
[0023] Optionally, the first motion parameter group is any one of the M motion parameter groups, and the each motion parameter group satisfies a motion constraint, including: the value of the i-th motion parameter in the first motion parameter group satisfies the constraints corresponding to the two-dimensional joint distribution of the i-th motion parameter and each motion parameter in the remaining (n-1) motion parameters, and i is a positive integer less than or equal to n.
[0024] Here, the two-dimensional joint distribution can be used to measure the probability distribution relationship between two parameters. For example, the jth motion parameter is any one of the remaining (n-1) motion parameters, and the constraints corresponding to the two-dimensional joint distribution of the i-th motion parameter and the j-th motion parameter include: the constraints corresponding to the first joint distribution when the i-th motion parameter is the independent variable and the j-th motion parameter is the dependent variable; and, the constraints corresponding to the second joint distribution when the i-th motion parameter is the dependent variable and the j-th motion parameter is the independent variable; wherein, the constraints corresponding to the first joint distribution include the upper limit of the variance corresponding to the first joint distribution and the lower limit of the variance corresponding to the first joint distribution; the constraints corresponding to the second joint distribution include the upper limit of the variance corresponding to the second joint distribution and the lower limit of the variance corresponding to the second joint distribution.
[0025] Here, the constraints corresponding to the two-dimensional joint distribution of motion parameters are not limited to only include the upper limit of the variance corresponding to the two-dimensional joint distribution of motion parameters and the lower limit of the variance corresponding to the two-dimensional joint distribution of motion parameters. Exemplarily, the constraints corresponding to the two-dimensional joint distribution of motion parameters may also include the upper limit of the variance corresponding to the two-dimensional joint distribution of motion parameters and the mean line corresponding to the two-dimensional joint distribution of motion parameters, or include the mean line corresponding to the two-dimensional joint distribution of motion parameters and the lower limit of the variance corresponding to the two-dimensional joint distribution of motion parameters, or other constraints, which are not limited here.
[0026] Implementing the above implementation method, the constraints corresponding to the two-dimensional joint distribution of motion parameters are analyzed based on the statistics of natural driving data. Compared with the spatial fitting of more dimensional motion parameters, the correlation between motion parameters is simpler and more efficient, which is conducive to improving data processing efficiency.
[0027] Optionally, determining the first parameter set related to the first driving scenario includes: determining the first parameter set related to the first driving scenario based on the motion constraint and a third parameter set, wherein the third parameter set includes the Cartesian product of the generalized value sets of each of the n motion parameters, and the value range corresponding to each of the n motion parameters in the third parameter set at least covers a preset value range of the motion parameter.
[0028] For example, the preset value range of each motion parameter is the preset value range of the motion parameter.
[0029] For another example, the value range corresponding to each parameter covers the preset value range of the parameter, that is, it can be a user-defined setting, such as setting a larger range.
[0030] Taking the i-th parameter among n parameters as an example, i is a positive integer less than or equal to n, and the preset value range of the i-th parameter can be associated with the natural driving data of the driving scene, or can be set by the user based on historical experience.
[0031] Exemplarily, the preset value range of the ith parameter can be associated with the natural driving data of the driving scene and can be: the preset value range of the ith parameter is determined by the constraint corresponding to the two-dimensional joint distribution of the ith motion parameter and each of the remaining (n-1) motion parameters, where the constraint corresponding to the two-dimensional joint distribution of the ith motion parameter and each of the remaining (n-1) motion parameters is determined based on the natural driving data statistics of the driving scene.
[0032] By implementing the above implementation method, the first parameter set can be obtained from the third parameter set (i.e., the initial parameter set) based on motion constraints, which is conducive to improving the rationality of the autonomous driving safety test scenario.
[0033] Optionally, the generalized value set of the i-th motion parameter among the n motion parameters is obtained according to a value range corresponding to the i-th motion parameter and a value step size of the i-th motion parameter.
[0034] Exemplarily, the parameter value step size may be a default setting or a user-defined setting.
[0035] In one implementation method, a level can be set for the driving scene, and the value step of the parameter is determined based on the level of the driving scene. This level can be used for example to indicate the safety level of the driving scenario, that is, the lower the level, the more dangerous the driving scenario, and the smaller the value step of the parameter.
[0036] Taking the small car cutting-in scenario and the large car cutting-in scenario as examples, the level of the small car cutting-in scenario is level 1, and the level of the large car cutting-in scenario is level 2. Among them, level 1 is greater than level 2, which means that the danger level of the large car cutting-in scenario is higher than that of the small car cutting-in scenario. Taking parameter a as an example, the value step length of parameter a in the large car cutting-in scenario is smaller than the value step length of parameter a in the small car cutting-in scenario. In this way, the large car cutting-in scenario with a higher degree of danger can be more finely characterized, and subsequent tests can be more refined, which is conducive to improving the effectiveness of autonomous driving safety test scenarios.
[0037] Optionally, the value range corresponding to the i-th motion parameter includes a first interval and a second interval, the value step of the i-th motion parameter includes a first value corresponding to the first interval and a second value corresponding to the second interval, and the first interval and the second interval have no overlap; wherein, when the i-th motion parameter takes a value within the first interval, the danger level of the driving scene is higher than the danger level of the driving scene when the i-th motion parameter takes a value within the second interval, and the first value is smaller than the second value.
[0038] Taking the parameter "the distance between the cutting vehicle and the main vehicle when cutting in" in the vehicle cutting-in scenario as an example, assuming that the value range corresponding to the "distance between the cutting vehicle and the main vehicle when cutting in" is set to [0, 150]. Since the smaller the distance between the cutting vehicle and the main vehicle when cutting in is when other parameters remain unchanged, the higher the danger level of the driving scene, in this case, the smaller the value step of the motion parameter can be set. For example, the value range [0, 150] is divided into multiple intervals, such as the intervals [0, 60], (60, 100] and (100, 150, then when the "distance between the cutting vehicle and the main vehicle when cutting in" takes values in the interval [0, 60], the value step is set to 5m, for example; when it takes values in the interval (60, 100], the value step is set to 10m, for example; when it takes values in the interval (100, 150], the value step is set to 25m, for example.
[0039] By implementing the above implementation, for the same driving scene, the step length of the motion parameter can be adaptively determined based on the value of the motion parameter to the degree of danger of the driving scene, that is, for the same driving scene, the same motion parameter can have multiple different step lengths. In this way, the step length of the parameter in the driving scene with a lower degree of danger is greater than the step length of the parameter in the driving scene with a higher degree of danger, which can not only more accurately describe the boundary scene (i.e., the driving scene with a higher degree of danger), but also help improve the test efficiency without affecting the accuracy.
[0040] Optionally, it is characterized in that the first driving scene is a vehicle cutting-in scene, the first vehicle is the vehicle to be tested, the second vehicle is the cutting-in vehicle that will cut in front of the first vehicle, and the n motion parameters include the speed of the first vehicle, the distance between the second vehicle and the first vehicle when cutting in, the speed of the second vehicle relative to the first vehicle and the lane changing time of the second vehicle.
[0041] Here, the first vehicle and the second vehicle are both entities participating in the vehicle cut-in scenario. The above implementation method is implemented to provide an example of a driving simulation test in a vehicle cut-in scenario.
[0042] Optionally, the first driving scene is a vehicle cutting-out scene, the third vehicle is a vehicle to be tested, the fourth vehicle is a vehicle cutting out in front of the third vehicle, the fifth vehicle is a vehicle in front of the fourth vehicle before cutting out, and the n motion parameters include the speed of the fourth vehicle, the speed of the fifth vehicle relative to the fourth vehicle, the distance between the fourth vehicle and the fifth vehicle when cutting out, and the lane changing time of the fourth vehicle.
[0043] Here, the first vehicle, the third vehicle and the fourth vehicle are all entities participating in the vehicle cut-out scenario. The above implementation provides an example of a driving simulation test in a vehicle cut-out scenario.
[0044] Optionally, the test purpose of the first driving scenario is to avoid collision.
[0045] Exemplarily, no collision occurs means that no collision occurs for the human driver in the first driving scenario, including no collision occurs for at least one of the average driving level, the lowest driving level and the highest driving level of ordinary human drivers in the first driving scenario, or no collision occurs for at least one of the average driving level, the lowest driving level and the highest driving level of cautious and skilled human drivers in the first driving scenario, or no collision occurs for at least one of the average driving level, the lowest driving level and the highest driving level of mature human drivers in the first driving scenario.
[0046] It can be understood that the driving ability of the driver model used in the driving simulation test is the same as or matches the driving level of the human driver required by the test purpose of the first driving scenario. For example, if the test purpose of the first driving scenario is that an ordinary human driver with an average driving level does not collide in the first driving scenario, then the driving ability of the driver model used in the driving simulation test is the average driving level of ordinary human drivers. In this case, it also means that the ability requirement for the automatic driving system is to reach the average driving level of ordinary human drivers.
[0047] Here, ordinary human drivers, cautious and skilled human drivers and mature human drivers are just some examples of human drivers, and examples of human drivers are not limited to only ordinary human drivers, cautious and skilled human drivers and mature human drivers.
[0048] Implementing the above implementation method, for each of the K motion parameter groups in the second parameter set, the values of n motion parameters are realized not only that they will actually appear in the physical world, but also that when the values of n motion parameters are applied to the first driving scenario, the subjects (such as multiple vehicles) participating in the first driving scenario will not collide, which is conducive to improving the effectiveness of the autonomous driving safety test scenario.
[0049] In a second aspect, the present application provides a device for determining scene parameters, the device comprising: a processing unit, for determining a first parameter set related to a first driving scene, the first parameter set comprising M motion parameter groups, each motion parameter group comprising a set of values of n motion parameters, the n motion parameters being used to describe the state of a subject participating in the first driving scene, each motion parameter group in the M motion parameter groups satisfying a motion constraint, the motion constraint being used to indicate a correlation between multiple motion parameters in the n motion parameters, M and n being integers greater than 1; a simulation unit, for performing a driving simulation test using the first parameter set to obtain a second parameter set, the second parameter set comprising K motion parameter groups in the M motion parameter groups, the K motion parameter groups meeting the test purpose of the first driving scene, and K being a positive integer less than or equal to M.
[0050] Optionally, the motion constraint is obtained based on natural driving data statistics in the first driving scenario.
[0051] Optionally, the second parameter set is used for safety testing of the autonomous driving vehicle in the first driving scenario.
[0052] Optionally, the second parameter set is used to test whether the vehicle meets the test purpose.
[0053] Optionally, the first motion parameter group is any one of the M motion parameter groups, and the each motion parameter group satisfies a motion constraint, including: the value of the i-th motion parameter in the first motion parameter group satisfies the constraints corresponding to the two-dimensional joint distribution of the i-th motion parameter and each motion parameter in the remaining (n-1) motion parameters, and i is a positive integer less than or equal to n.
[0054] For example, the jth motion parameter is any one of the remaining (n-1) motion parameters, and the constraints corresponding to the two-dimensional joint distribution of the ith motion parameter and the jth motion parameter include: the constraints corresponding to the first joint distribution when the ith motion parameter is the independent variable and the jth motion parameter is the dependent variable; and, the constraints corresponding to the second joint distribution when the ith motion parameter is the dependent variable and the jth motion parameter is the independent variable; wherein, the constraints corresponding to the first joint distribution include the upper limit of the variance corresponding to the first joint distribution and the lower limit of the variance corresponding to the first joint distribution; the constraints corresponding to the second joint distribution include the upper limit of the variance corresponding to the second joint distribution and the lower limit of the variance corresponding to the second joint distribution.
[0055] Optionally, the processing unit is specifically used to: determine the first parameter set related to the first driving scene based on the motion constraint and a third parameter set, wherein the third parameter set includes the Cartesian product of the generalized value sets of each of the n motion parameters, and the value range corresponding to each of the n motion parameters in the third parameter set at least covers the preset value range of the motion parameter.
[0056] Optionally, the generalized value set of the i-th motion parameter among the n motion parameters is obtained according to a value range corresponding to the i-th motion parameter and a value step size of the i-th motion parameter.
[0057] Optionally, the value range corresponding to the i-th motion parameter includes a first interval and a second interval, the value step of the i-th motion parameter includes a first value corresponding to the first interval and a second value corresponding to the second interval, and the first interval and the second interval have no overlap; wherein, when the i-th motion parameter takes a value within the first interval, the danger level of the driving scene is higher than the danger level of the driving scene when the i-th motion parameter takes a value within the second interval, and the first value is smaller than the second value.
[0058] Optionally, the first driving scenario is a vehicle entry scenario, the first vehicle is a vehicle to be tested, and the second vehicle is a cutting vehicle to be cut into front of the first vehicle. The n motion parameters include the speed of the first vehicle, the distance from the first vehicle when the second vehicle is cut into, the speed of the second vehicle relative to the first vehicle, and the lane change time of the second vehicle.
[0059] Optionally, the first driving scene is a vehicle cutting-out scene, the third vehicle is a vehicle to be tested, the fourth vehicle is a vehicle cutting out in front of the third vehicle, the fifth vehicle is a vehicle in front of the fourth vehicle before cutting out, and the n motion parameters include the speed of the fourth vehicle, the speed of the fifth vehicle relative to the fourth vehicle, the distance between the fourth vehicle and the fifth vehicle when cutting out, and the lane changing time of the fourth vehicle.
[0060] Optionally, the test purpose of the first driving scenario is to avoid collision.
[0061] In a third aspect, the present application provides a communication device, the device comprising at least one processor and a communication interface, the communication interface being used to provide information input and / or output to the at least one processor. The device may be a chip or an integrated circuit, or may be the device described in the second aspect above, the device being used to implement the method in the first aspect or any possible embodiment of the first aspect.
[0062] In a fourth aspect, the present application provides a computer readable storage medium, including computer instructions, which implements a method in any possible implementation of the first aspect or the first aspect as described above when the computer instructions are run by a processor.
[0063] In a fifth aspect, the present application provides a computer program product, which, when executed by a processor, implements the method in the first aspect or any possible embodiment of the first aspect. The computer program product, for example, may be a software installation package, and when the method provided by any possible design of the first aspect is needed, the computer program product may be downloaded and executed on a processor to implement the method in the first aspect or any possible embodiment of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS
[0064] Figure 1 It is a schematic diagram of the architecture of a communication system provided by an embodiment of the present application;
[0065] Figure 2 is a flow chart of a method for determining scene parameters provided in an embodiment of the present application;
[0066] Figure 3 is a flow chart of a method for constructing an initial parameter set provided in an embodiment of the present application;
[0067] Figure 4 It is a schematic diagram of constraints corresponding to two-dimensional joint distribution of parameters in some vehicle cut-in scenarios provided in embodiments of the present application;
[0068] Figure 5 is a schematic diagram of a driving simulation test interface in a vehicle cut-in scenario provided in an embodiment of the present application;
[0069] Figure 6 is a schematic diagram of a driving simulation test interface in a vehicle cut-out scenario provided in an embodiment of the present application;
[0070] Figure 7 It is a schematic diagram of constraints corresponding to two-dimensional joint distribution of parameters in some vehicle cut-out scenarios provided in an embodiment of the present application;
[0071] Figure 8 is a flow chart of another method for determining scene parameters provided in an embodiment of the present application;
[0072] Fig. 9 is a structural schematic diagram of a data processing device provided in an embodiment of the present application;
[0073] Fig.10 It is a structural diagram of a data processing device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0074] It should be noted that the prefixes such as "first" and "second" used in this application are only used to distinguish different description objects, and have no limiting effect on the position, order, priority, quantity or content of the described objects. For example, if the described object is a "field", the ordinal number before the "field" in the "first field" and the "second field" does not limit the position or order between the "fields". "First" and "second" do not limit whether the "fields" they modify are in the same message, nor do they limit the order of the "first field" and the "second field". For another example, if the described object is a "level", the ordinal number before the "level" in the "first level" and the "second level" does not limit the priority between the "levels". For another example, the number of described objects is not limited by the prefix and can be one or more. Taking "first device" as an example, the number of "devices" can be one or more. In addition, the objects modified by different prefixes can be the same or different. For example, if the object being described is a "device", then the "first device" and the "second device" can be the same device, the same type of device, or different types of devices; for another example, if the object being described is "information", then the "first information" and the "second information" can be information of the same content or information of different contents. In short, the use of prefixes used to distinguish the described objects in the embodiments of the present application does not constitute a limitation on the described objects. For the statement of the described objects, refer to the description in the context of the claims or embodiments, and no unnecessary limitation should be constituted due to the use of such prefixes.
[0075] It should be noted that the description methods such as "at least one of a1, a2, ... and an" used in the embodiments of the present application include the situation where any one of a1, a2, ... and an exists alone, and also include any combination of any multiple of a1, a2, ... and an, and each situation can exist alone. For example, the description method of "at least one of a, b and c" includes the situation where a is alone, b is alone, c is alone, a combination of a and b, a combination of a and c, b and c, or a combination of abc.
[0076] See also Figure 1 , Figure 1 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 value of the motion parameter of the driving scene. Figure 1 As shown, the system includes a data source device and a network side device, wherein the data source device and the network side device can be connected and communicated through wireless or wired mode. In one possible implementation, 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.
[0077] The data source device is used to provide the network side device with natural driving data of the driving scene, and the natural driving data at least includes the value of the motion parameter corresponding to the driving scene when a natural person participates in driving in the driving scene, and the motion parameter is used to describe the state of the subject participating in the driving scene. In some possible embodiments, the natural driving data of the driving scene also includes the driving trajectory information and / or motion state information of the vehicle driven by a natural person and the surrounding vehicles in the driving scene. Here, the value of the motion parameter corresponding to the driving scene when a natural person participates in driving in the driving scene can be obtained by the data source device based on the collected driving trajectory information and / or motion state information of the vehicle driven by a natural person and the surrounding vehicles in the driving scene.
[0078] Here, the driving scene may be, for example, a vehicle cut-in scene, a vehicle cut-out scene, a vehicle following scene, or a front vehicle braking scene, etc. Taking the vehicle cut-in scene as an example, assuming that vehicle 1 is the main vehicle and vehicle 2 is the cut-in vehicle that will cut in front of vehicle 1, the motion parameters corresponding to the driving scene include: the speed of vehicle 1, the distance between vehicle 2 and vehicle 1 when cutting in, the speed of vehicle 2 relative to vehicle 1, and the lane change time of vehicle 2.
[0079] Exemplarily, the data source device may be an unmanned aerial vehicle device, a data collection vehicle, etc. equipped with a collection device.
[0080] Here, the network side device may be, for example, a computing device (eg, a server) deployed on the network side, or a system-level device composed of components or chips in the server, or a system-level device composed of multiple servers. Network-side devices can be deployed in a cloud environment or an edge environment, and there are no specific limitations here.
[0081] Exemplarily, the network-side device deploys a data screening module and a simulation test module, wherein the data screening module is used to determine M groups of values of n motion parameters related to the driving scene, the n motion parameters are used to describe the state of the subject participating in the driving scene, each group of values of the n parameters can be called a motion parameter group, that is, a motion parameter group includes n values, each value of the n values corresponds to a motion parameter of the n motion parameters, and the M groups of values of the n parameters can constitute a parameter set, for example, the first parameter set, that is, the first parameter set includes M motion parameter groups, M motion parameter groups, Each motion parameter group in the motion parameter group satisfies a motion constraint, and the motion constraint is used to indicate the correlation between multiple motion parameters in the n motion parameters, and M and n are integers greater than 1; the simulation test module is used to perform a driving simulation test through each group of values in the M groups of values of the n motion parameters, and determine K groups of values in the M groups of values of the n motion parameters, and the K groups of values meet the test purpose of the driving scenario. The K groups of values of the n motion parameters can constitute a parameter set, such as a second parameter set, that is, the second parameter set includes K motion parameter groups, and K is a positive integer less than or equal to M.
[0082] Exemplarily, the test objectives of the driving scenario include, but are not limited to, one or more of: no collision occurs, the distance to the vehicle in front is not less than a preset threshold during braking, and no skidding occurs when exiting a curve.
[0083] here, Figure 1 The system architecture shown is only an example and is not intended to limit Figure 1 In addition, the method provided in the embodiment of the present application can be applied not only to Figure 1 The system shown can also be applied to other systems. For example, Figure 1 The data filtering module and the simulation test module in the data may not be integrated together, i.e., it is deployed separately. For example, the data filtering module is integrated on the network side device, the simulation test module is integrated on the simulation device, and the network side device and the simulation device can communicate in a wireless or wired manner. In this case, the above-mentioned system includes a data source device, a network side device and a simulation device.
[0084] Exemplarily, after the network side device obtains the above-mentioned first parameter set, the network side device sends the first parameter set to the simulation device; accordingly, the simulation device receives the first parameter set from the network side device, and uses each group of values of the n motion parameters in the first parameter set to perform a driving simulation test, outputs the above-mentioned second parameter set, and finally the simulation device sends the above-mentioned second parameter set to the network side device.
[0085] Exemplarily, the method provided in the embodiment of the present application is not only applicable to the autonomous driving simulation system, and realizes the determination of the range of scene parameters of the autonomous driving simulation, but can also be applied to other test systems that need to set scene parameters, such as real vehicle test systems, vehicle-in-the-loop test systems, model-in-the-loop test systems, hardware-in-the-loop test systems, etc. It should be noted here that "autonomous driving" is not limited to fully autonomous driving systems, highly autonomous driving systems, conditional autonomous driving systems, or partially autonomous driving systems, etc. Those skilled in the art can understand that non-fully manual driving systems that provide intelligent driving can be included in this concept for simulation processing.
[0086] See also Figure 2 , Figure 2 is a flow chart of a method for determining scene parameters provided by an embodiment of the present application. The method can be applied to the above Figure 1 The network side device shown. The method includes but is not limited to the following steps:
[0087] S201: Determine a first parameter set related to the driving scene. The first parameter set includes M motion parameter groups, each motion parameter group includes a set of values of n motion parameters, and each motion parameter group satisfies motion constraints. n motion parameters are used to describe the state of the subject participating in the driving scene. M and n are integers greater than 1.
[0088] Exemplarily, the process of determining the first parameter set is also equivalent to the process of determining multiple sets of values of n motion parameters.
[0089] Here, a set of values of n motion parameters can form a group of motion parameters, that is, a group of motion parameters includes n values, and each of the n values corresponds to one of the n motion parameters. Therefore, the above M motion parameter groups are actually a grouping form of multiple sets of values of n motion parameters, and multiple groups of motion parameter groups can form a parameter set.
[0090] Here, motion constraints are used to indicate the correlation between multiple motion parameters of n motion parameters. The correlation between multiple motion parameters can be represented, for example, by a joint distribution between multiple parameters. For example, a joint distribution, also known as a joint probability distribution, can be a probability distribution of a random vector composed of more than two (including two) random variables.
[0091] Taking two motion parameters among multiple motion parameters as an example, the correlation between the two motion parameters can be represented by a two-dimensional joint distribution between the two parameters. The two-dimensional joint distribution can be used to measure the probability distribution relationship between the two parameters. Here, the constraint corresponding to the two-dimensional joint distribution is an example of the above-mentioned motion constraint. In some possible embodiments, the correlation between multiple motion parameters can also be represented by a three-dimensional joint distribution of parameters, or other joint distributions, etc., which are not specifically limited here.
[0092] It can be understood that the driving scenario involves n motion parameters, that is, the driving scenario corresponds to n motion parameters. These n motion parameters are used to describe the state of the subject (eg, multiple vehicles) participating in the driving scenario. Exemplary, motion parameters may also be referred to as scene parameters. For convenience of narrative, the motion parameters in the following description are referred to as parameters simply. Accordingly, in the following description, the motion parameter group is also referred to as the parameter group simply.
[0093] Exemplary, driving scenarios include, but are not limited to, vehicle entry scenarios, vehicle cut-out scenarios, vehicle follow-up scenarios, front-vehicle braking scenarios, and the like. It can be understood that the vehicle cutting scene can also be called the front car cutting scene, and the vehicle cutting scene can also be called the front car cutting scene.
[0094] Exemplary, the types of vehicle entry scenarios can also be subdivided, such as classification as car entry scenarios and car entry scenarios. It can be understood that car entry scenarios and car entry scenarios are different driving scenarios.
[0095] For example, the parameters include but are not limited to at least one of speed, distance, relative speed, relative distance, displacement, acceleration, angular velocity, posture, lane change time, and the like.
[0096] For example, the driving scenario is a vehicle entry scenario. Assuming that the vehicle 1 is a vehicle to be tested and the vehicle 2 is a cutting vehicle to be cut in front of the vehicle 1, the above n parameters include: the speed of the vehicle 1, the distance from the vehicle 2 when the vehicle 2 is cut in, the speed of the vehicle 2 with respect to the vehicle 1 and the lane change time of the vehicle 2.
[0097] For example, the driving scene is a vehicle cutout scene. Assuming that vehicle 1 is a vehicle to be tested, vehicle 3 is a cutout vehicle to be cut out in front of vehicle 1, and vehicle 4 is a front vehicle before vehicle 3 is cut out, the above n parameters include: the speed of vehicle 3, the speed of vehicle 4 with respect to vehicle 3, the distance from vehicle 4 when vehicle 3 is cut out, and the lane change time of vehicle 3.
[0098] In the embodiment of the present application, the first parameter set may be directly obtained by the network side device from other network side devices, or may be locally generated by the network side device.
[0099] In one implementation, when the first parameter set is generated locally by the network-side device, before executing S201, natural driving data of the driving scene and an initial parameter set may be acquired first.
[0100] In one implementation, obtaining natural driving data of a driving scene includes: obtaining natural driving data of a driving scene from a data source device. The data source device can refer to Figure 1 The description of the corresponding content will not be repeated here.
[0101] Exemplarily, the natural driving data of the driving scene at least includes the values of the parameters corresponding to the driving scene when a natural person participates in driving in the driving scene. In some possible embodiments, the natural driving data of the driving scene also includes the driving trajectory information and / or motion state information of the vehicle driven by a natural person and the surrounding vehicles in the driving scene. For example, when a natural person participates in driving in the driving scene, the values of the parameters corresponding to the driving scene can be obtained based on the driving trajectory information and / or motion state information of the vehicle driven by a natural person and the surrounding vehicles in the driving scene. It can be understood that the natural driving data of the driving scene can be driving data collected in the physical world, or data obtained based on statistical analysis of driving data collected in the physical world, or driving data provided by human drivers based on driving experience, which is not specifically limited here.
[0102] In one implementation, obtaining the initial parameter set may be: generating the initial parameter set locally. In some possible embodiments, the initial parameter set may also be obtained by the network side device from other devices, which is not specifically limited here. Exemplarily, the initial parameter set may also be referred to as a complete set of scene parameters.
[0103] The initial parameter set includes the Cartesian product of the generalized value sets of n parameters, and the value range corresponding to each parameter of the n parameters in the initial parameter set at least covers the preset value range of the parameter.
[0104] See also Figure 3 , Figure 3 is a flow chart of a method for constructing an initial parameter set provided by an embodiment of the present application. For example, the method can be applied to Figure 1 The network side device shown. The method includes but is not limited to the following steps:
[0105] S301: Determine the value ranges corresponding to n parameters respectively, where the value range corresponding to each parameter at least covers the preset value range of the parameter.
[0106] For example, the value range corresponding to each parameter is the preset value range of the parameter.
[0107] For another example, the value range corresponding to each parameter covers the preset value range of the parameter, that is, it can be a user-defined setting, such as setting a larger range.
[0108] Exemplarily, the preset value range of each parameter may be associated with natural driving data of the driving scene.
[0109] Taking the i-th parameter among n parameters as an example, i is a positive integer less than or equal to n, and the association between the preset value range of the i-th parameter and the natural driving data of the driving scene means: the preset value range of the i-th parameter is determined by the above-mentioned motion constraint, and the motion constraint is obtained based on the statistics of the natural driving data of the driving scene.
[0110] For example, assuming that the correlation between multiple parameters indicated by the motion constraint is represented by a two-dimensional joint distribution of the parameters, the preset value range of the i-th parameter can be determined by the constraint corresponding to the two-dimensional joint distribution of the i-th parameter and each of the remaining (n-1) parameters, wherein the constraints corresponding to the two-dimensional joint distribution of the i-th parameter and each of the remaining (n-1) parameters are all obtained based on natural driving data statistics.
[0111] Further, assuming that the jth parameter is any one of the remaining (n-1) parameters mentioned above, the constraints corresponding to the two-dimensional joint distribution of the ith parameter and the jth parameter include: the constraints corresponding to the first joint distribution when the ith parameter is the independent variable and the jth parameter is the dependent variable; and, the constraints corresponding to the second joint distribution when the ith parameter is the dependent variable and the jth parameter is the independent variable. Exemplarily, the constraints corresponding to the first joint distribution include the upper bound of the variance corresponding to the first joint distribution and the lower bound of the variance corresponding to the first joint distribution; the constraints corresponding to the second joint distribution include the upper bound of the variance corresponding to the second joint distribution and the lower bound of the variance corresponding to the second joint distribution. In this case, the upper limit of the preset value range of the ith parameter is the maximum value of the ith parameter under the constraints of the upper and lower bounds of the variance corresponding to the two-dimensional joint distribution of the ith parameter and each of the remaining (n-1) parameters, and the lower limit of the preset value range of the ith parameter is the minimum value of the ith parameter under the constraints of the upper and lower bounds of the variance corresponding to the two-dimensional joint distribution of the ith parameter and each of the remaining (n-1) parameters. Here, the constraints corresponding to the two-dimensional joint distribution of the parameters are not limited to only the upper bound of the variance corresponding to the two-dimensional joint distribution of the parameters and the lower bound of the variance corresponding to the two-dimensional joint distribution of the parameters. In some possible embodiments, the constraints corresponding to the two-dimensional joint distribution of the parameters may also include the upper bound of the variance corresponding to the two-dimensional joint distribution of the parameters and the mean line corresponding to the two-dimensional joint distribution of the parameters, or include the mean line corresponding to the two-dimensional joint distribution of the parameters and the lower bound of the variance corresponding to the two-dimensional joint distribution of the parameters, or other constraints, which are not limited here.
[0112] Take the driving scene as an example, see Figure 4 , Figure 4 Schematic diagram of constraints corresponding to the two-dimensional joint distribution of parameters in some vehicle cut-in scenarios provided by the embodiments of the present application. Figure 4 It can be seen that in the vehicle cut-in scenario, the parameters include "main vehicle speed", "speed of the cut-in vehicle relative to the main vehicle ( Figure 4 relative speed in Chinese), "the lane changing time of the cutting-in vehicle" and "the distance between the cutting-in vehicle and the main vehicle when cutting in ( Figure 4 These four parameters are combined in pairs to construct a two-dimensional joint distribution. There are 12 combinations in total. For details, see Figure 4 There are 12 pictures listed in it.
[0113] Understandably, Figure 4 The dense data points in each picture are only an example of the distribution form of the corresponding data in the vehicle cut-in scenario, and do not limit the distribution form of the data to only Figure 4 As shown. Similarly, Figure 4 The schematic diagram of the constraints corresponding to the two-dimensional joint distribution of parameters in the vehicle cut-in scenario is only given as an example, and does not limit the schematic diagram of the constraints corresponding to the two-dimensional joint distribution of parameters in the vehicle cut-in scenario to only Figure 4 shown.
[0114] The following takes the parameter "main vehicle speed" as an example. Figure 4 The process of determining the preset value range of the parameter is described in detail. Figure 4 It can be seen that Figure 4 (1), (2), (3), (4), (7) and (10) are related to the parameter “host vehicle speed”. Figure 4 (1), Figure 4 (1) shows a schematic diagram of the constraints corresponding to the two-dimensional joint distribution when "main vehicle speed" is the independent variable and "relative speed" is the dependent variable, wherein the dense data points represent the data distribution of the two parameters "main vehicle speed" and "speed of the cutting vehicle relative to the main vehicle" in the natural driving data of the vehicle cutting-in scene. For example, the joint probability distribution between the two parameters can be statistically analyzed first to determine the mean line ( Figure 4 (1) is not shown), the upper and lower limits of the variance corresponding to the two-dimensional joint distribution are determined by taking the mean as the center and the variance of the upper and lower distances as preset multiples (for example, 3 times), where the upper limit is Figure 4 The upper line segment on the two line segments shown in (1) has a lower boundary of Figure 4 The line segment located above the two line segments shown in (1).
[0115] It can be seen that in Figure 4Under the constraints of the upper and lower limits shown in (1), the maximum speed of the main vehicle is 60 (denoted as v 1max ) and the minimum value of the main vehicle speed is 3 (denoted as v 1min ), and determine Figure 4 The maximum value of the main vehicle speed and the minimum value of the main vehicle speed under the constraints of the upper and lower limits shown in (2), (3), (4), (7) and (10) are denoted as {v 2max , v 2min}、{v 3max , v 3min}、{v 4max , v 4min}、{v 7max , v 7min} and {v 10max , v 10min}, finally, the upper limit of the preset value range of the main vehicle speed is the set {v 1max , v 2max , v 3max , v 4max , v 7max , v 10max}, the lower limit of the preset value range of the main vehicle speed is the set {v 1min , v 2min , v 3min , v 4min , v 7min , v 10min}, for example, the lower limit of the preset value range of the main vehicle speed is determined to be 0 and the upper limit of the preset value range of the main vehicle speed is 80, that is, the preset value range of the main vehicle speed is [0, 80]. It can be understood that the preset value range of the main vehicle speed obtained in this way is reasonable and real and predictable in the physical world, and is meaningful for subsequent simulation tests in driving scenarios. Similarly, Figure 4 You can also refer to the other figures in Figure 4 The description of (1) is based on Figure 4 Based on the same analysis method, the preset value range of the parameter "relative speed", the preset value range of the parameter "lane change time for cutting in the vehicle", and the preset value range of the parameter "relative distance" can also be determined. For the sake of brevity of the specification, they will not be repeated here.
[0116] In some possible embodiments, the preset value range of the parameter is not limited to being determined only by the constraints corresponding to the two-dimensional joint distribution of the parameters among the above n parameters. For example, the preset value range of the parameter can also be determined by the constraints corresponding to the three-dimensional joint distribution of the parameters among the above n parameters, and there is no specific limitation on the number of parameters participating in the joint distribution.
[0117] In some possible embodiments, the preset value range of each parameter can also be set by the user. For example, for the parameter "main vehicle speed", if the user's setting information is received to test the performance of the automatic driving system at 0-60 km / h (kilometers per hour), the maximum value of the "main vehicle speed" is determined to be 60 and the minimum value of the "main vehicle speed" is 0, that is, the preset value range of the "main vehicle speed" is [0, 60].
[0118] S302: performing generalized value selection within the value range corresponding to each parameter based on the value step of the parameter to obtain a generalized value set for each parameter.
[0119] In the embodiment of the present application, the parameter value step size can be a default setting or a user-defined setting.
[0120] In one implementation, a level can be set for the driving scenario, and the parameter value step size is determined based on the level of the driving scenario. The level can be used to indicate the safety level of the driving scenario, that is, the lower the level, the more dangerous the driving scenario, and the smaller the parameter value step size.
[0121] Taking the small car cutting in scene and the large car cutting in scene as examples, the level of the small car cutting in scene is level 1, and the level of the large car cutting in scene is level 2. Among them, level 1 is greater than level 2, which means that the danger level of the large car cutting in scene is higher than that of the small car cutting in scene. Taking parameter a as an example, the value step length of parameter a in the large car cutting in scene is smaller than the value step length of parameter a in the small car cutting in scene. In this way, the large car cutting in scene with a higher degree of danger can be more finely characterized, and subsequent tests can also be made more refined.
[0122] In another implementation, the parameter value step size can be adaptively determined based on the value of the parameter to the degree of danger of the driving scene. In this case, for the same driving scene, the same parameter can have multiple different value step sizes.
[0123] Take the i-th parameter as an example to illustrate the method of determining the value step length: assuming that the value range of the i-th parameter includes the first interval and the second interval, the first interval and the second interval do not overlap, the value step length of the i-th parameter in the first interval is the first value, and the value step length of the i-th parameter in the second interval is the second value. If the degree of danger of the driving scene when the i-th parameter is in the first interval is greater than the degree of danger of the driving scene when the i-th parameter is in the second interval, then the first value is less than the second value. In this way, the higher the degree of danger of the driving scene, the smaller the value step length of the parameter. In other words, the value step length of the parameter can be adaptively adjusted to the degree of danger of the driving scene based on the parameter, so that subsequent tests are more refined.
[0124] For example, if the driving scenario is a vehicle cut-in scenario, we can know that, when other parameters remain unchanged, the smaller the distance between the cut-in vehicle and the main vehicle when cutting in, the more dangerous the driving scenario is; when other parameters remain unchanged, the shorter the lane change time of the cut-in vehicle, the more dangerous the driving scenario is; when other parameters remain unchanged, the greater the speed of the cut-in vehicle relative to the main vehicle, the more dangerous the driving scenario is; when other parameters remain unchanged, the greater the speed of the main vehicle, the more dangerous the driving scenario is. Assume that the value range corresponding to the parameter "the distance between the cut-in vehicle and the main vehicle when cutting in" is set to [0, 150]. Since, when other parameters remain unchanged, the smaller the distance between the cut-in vehicle and the main vehicle when cutting in, the more dangerous the driving scenario is, in this case, the value range [0, 150] can be divided into multiple intervals, such as the intervals [0, 60], (60, 100] and (100, 150). Then, "the distance between the cut-in vehicle and the main vehicle when cutting in" is in the interval [0, 60]. When the value is taken within the interval [60, 100], the value step length is, for example, 5m; when the value is taken within the interval [100, 150], the value step length is, for example, 25m, so that the generalized value set of the parameter “the distance between the cutting vehicle and the main vehicle when cutting in” is {5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 70, 80, 90, 100, 125, 150}.
[0125] In some possible embodiments, only the step size of the parameters that are strongly associated with the degree of danger of the driving scene may be adaptively determined. For example, in a vehicle cut-in scene, the parameters that are strongly associated with the degree of danger of the driving scene include "the distance between the cut-in vehicle and the host vehicle when cutting in" and "the lane change time of the cut-in vehicle". Referring to Table 1, Table 1 exemplarily provides a corresponding relationship table between a driving scene, a parameter, a value range of the parameter, and a generalized value of the parameter, wherein a plurality of generalized values of the parameter constitute a generalized value set of the parameter. In Table 1, the corresponding relationship of "vehicle cut-in-main vehicle speed-value range (0-60)-parameter generalized value set {5,10,15,20,25,30,35,40,45,50,55,60}" is taken as an example. This corresponding relationship indicates that in the vehicle cut-in scenario, the value range of the main vehicle speed parameter is, for example, 0-60 km / h, and the generalized value set of this parameter can be {5,10,15,20,25,30,35,40,45,50,55,60}. It can be seen from Table 1 that among the four parameters, only the value step length of the two parameters "lane change time of the cut-in vehicle" and "distance between the cut-in vehicle and the main vehicle when cutting in" is determined by adaptive method, while the two parameters "main vehicle speed" and "speed of the cut-in vehicle relative to the main vehicle" are generalized by fixed value step length.
[0126] Table 1
[0127]
[0128] Here, the above Table 1 is only used as an example to reflect the correspondence between driving scenarios, parameters, value ranges and generalized values of parameters. In actual applications, the text content and storage method of the correspondence may also be in other forms, which are not specifically limited here.
[0129] Here, the method for determining the parameter values provided above can ensure that the parameter values are complete and without omissions, and the parameter value step size is determined by an adaptive adjustment method, that is, the parameter value step size used in driving scenarios with a lower degree of danger is larger than the parameter value step size used in driving scenarios with a higher degree of danger. This can not only more accurately characterize boundary scenarios (i.e., driving scenarios with a higher degree of danger), but also help improve test efficiency without affecting accuracy.
[0130] S303: Taking the Cartesian product of the generalized value sets of each parameter in the n parameters, to obtain an initial parameter set.
[0131] Here, the initial parameter set includes the Cartesian product of the generalized value sets of each parameter in the n parameters.
[0132] Exemplarily, it can also be understood that the initial parameter set includes P parameter groups, each parameter group includes a set of values of n parameters, where P is the product of the number of generalized values of each parameter in the n parameters, and P is an integer greater than M.
[0133] For example, based on the content shown in Table 1, an initial parameter set in the vehicle cut-in scenario can be constructed. In this case, the initial parameter set includes the Cartesian product of the generalized value sets of the four parameters, namely, "main vehicle speed", "speed of the cut-in vehicle relative to the main vehicle", "lane change time of the cut-in vehicle", and "distance of the cut-in vehicle from the main vehicle when cutting in". As can be seen from Table 1, n = 4, and the number of generalized values of the four parameters, namely, "main vehicle speed", "speed of the cut-in vehicle relative to the main vehicle", "lane change time of the cut-in vehicle", and "distance of the cut-in vehicle from the main vehicle when cutting in", are 12, 13, 14, and 18, respectively, so P = 12*13*14*18 = 39312.
[0134] Refer to Table 2, which exemplarily shows the contents of the initial parameter set constructed based on the contents shown in Table 1. Based on Table 2, it can be seen that the initial parameter set includes 39,312 sets of values for the four parameters of "main vehicle speed", "speed of the cutting-in vehicle relative to the main vehicle", "lane changing time of the cutting-in vehicle" and "distance between the cutting-in vehicle and the main vehicle when cutting in". For example, the first parameter group in the initial parameter set is {5,0,1,5}, which means that in the first parameter group, the value of "main vehicle speed" is 5km / h, the value of "speed of the cutting-in vehicle relative to the main vehicle" is 0, the value of "lane changing time of the cutting-in vehicle" is 1s, and the value of "distance between the cutting-in vehicle and the main vehicle when cutting in" is 5m. The remaining parameter groups in the initial parameter set in Table 2 can refer to the description of the first parameter group above, which will not be repeated here.
[0135] Based on Table 2, it can be seen that a parameter group includes a set of values of n parameters, that is, a parameter group includes n values, and each of the n values corresponds to one parameter of the n parameters.
[0136] Table 2
[0137]
[0138] It can be understood that Table 2 is only an example of an initial parameter set constructed based on the content shown in Table 1, and does not limit the content of the initial parameter set to only what is shown in Table 2.
[0139] In some possible embodiments, determining a first parameter set related to a driving scenario includes: determining the first parameter set based on a motion constraint and an initial parameter set, wherein the motion constraint is used to indicate a correlation relationship between multiple parameters among n parameters.
[0140] It can be understood that the first parameter set is included in the above-mentioned initial parameter set, that is, the M parameter groups in the first parameter set are included in the P parameter groups in the initial parameter set.
[0141] Exemplarily, the motion constraint is obtained based on natural driving data statistics in a driving scenario. The motion constraint can refer to the above description of the motion constraint, which will not be repeated here.
[0142] In one implementation, the correlation between the above-mentioned multiple parameters is represented by a two-dimensional joint distribution of the parameters. In this case, the motion constraint includes constraints corresponding to the two-dimensional joint distribution of each parameter in the n parameters and each parameter in the remaining (n-1) parameters. Based on the motion constraint and the initial parameter set, a first parameter set is determined, including: traversing the constraints corresponding to the two-dimensional joint distribution of each parameter in the n parameters in the motion constraint, screening the initial parameter set multiple times, and obtaining the first parameter set.
[0143] Taking the vehicle cut-in scenario as an example, in the vehicle cut-in scenario, the constraints corresponding to the two-dimensional joint distribution of the parameters "main vehicle speed", "speed of the cut-in vehicle relative to the main vehicle", "lane change time of the cut-in vehicle" and "distance between the cut-in vehicle and the main vehicle when cutting in" can be Figure 4 The constraints determined by the upper and lower bounds of the variance shown are, Figure 4 Each graph in corresponds to a constraint, that is, there are 12 constraints in total. The process of obtaining the first parameter set can be, for example, as follows: Use Figure 4 The upper and lower bounds of the variance in (1) are used to select the initial parameter set, since Figure 4 (1) shows the constraints corresponding to the two-dimensional joint distribution of "main vehicle speed" and "speed of the cutting vehicle relative to the main vehicle", so only the values of the parameters "main vehicle speed" and "speed of the cutting vehicle relative to the main vehicle" in the initial parameter set are retained to satisfy Figure 4 The parameter set of the upper and lower bounds of the variance in (1) is obtained to obtain parameter set 1; then use Figure 4 The upper and lower bounds of the variance in (2) filter parameter set 1, and only the values of the parameters "main vehicle speed" and "distance between the cutting vehicle and the main vehicle when cutting in" in parameter set 1 satisfy Figure 4 The parameter set of the upper and lower bounds of the variance in (2) is obtained, and parameter set 2 is obtained; ..., until the use Figure 4 The upper and lower bounds of the variance in (12) are selected based on the previous step Figure 4 The parameter set 11 obtained by the constraints shown in (11) only retains the values of the parameters "lane change time of the cutting-in vehicle" and "distance between the cutting-in vehicle and the main vehicle when cutting-in" in the parameter set 11 that satisfy Figure 4 The parameter set constrained by the upper and lower bounds of the variance in (12) is obtained to obtain parameter set 12, and parameter set 12 can be used as the above-mentioned first parameter set.
[0144] For example, it is also possible to use a parameter unit based on Figure 4 The constraints shown screen the initial parameter set multiple times to obtain the first parameter set. The specific process can refer to the following steps A1-A4:
[0145] A1: First, regarding the parameter of "main vehicle speed", it can be seen that Figure 4 (1), (2), (3), (4), (7) and (10) are related to the parameter "main vehicle speed". Figure 4 The constraints shown in (1), (2), (3), (4), (7) and (10) respectively screen the value of "main vehicle speed" in the initial parameter set multiple times to obtain parameter set one.
[0146] Here, the method for obtaining parameter set 1 can be, for example: first based on Figure 4The constraint shown in (1) determines that the constraint 1 of the “main vehicle speed” is [3,60], and removes the parameter group whose value of the “main vehicle speed” does not meet the constraint 1 from the initial parameter set; then based on Figure 4 The constraint shown in (2) determines the constraint 2 of "main vehicle speed", and continues to remove the parameter group whose value of "main vehicle speed" does not meet the constraint 2 from the initial parameter set, and continues to analyze in the same way Figure 4 (3), (4), (7) and (10), the initial parameter set after the aforementioned multiple elimination operations is the above-mentioned parameter set 1.
[0147] A2: Determine the speed of the cutting vehicle relative to the main vehicle. Figure 4 The graph related to "relative speed" is Figure 4 (1), (4), (5), (6), (8) and (11), traverse Figure 4 The constraints shown in (1), (4), (5), (6), (8) and (11) respectively perform multiple screening on the value of "the speed of the cutting vehicle relative to the main vehicle" in the parameter set 1 obtained in step A1 to obtain the parameter set 2.
[0148] A3: Determine the lane change time for the cutting vehicle. Figure 4 The multiple graphs related to the “lane changing time of the cutting-in vehicle” are shown in the figure, and the constraints shown in each of the multiple graphs are traversed to perform multiple screening on the values of the “lane changing time of the cutting-in vehicle” in the parameter set 2 obtained in step A2 to obtain the parameter set 3.
[0149] A4: Determine the parameter of "the distance between the cutting vehicle and the main vehicle when cutting in" Figure 4 In the above, multiple graphs related to "relative distance" are obtained, and the constraints shown in each of the multiple graphs are traversed to perform multiple screening on the value of "the distance between the cut-in vehicle and the main vehicle when the cut-in vehicle cuts in" in the parameter set 3 obtained in step A3, and the parameter set 4 is obtained. The parameter set 4 can be used as the above-mentioned first parameter set. It can be understood that the acquisition method of the above-mentioned parameter set 2, parameter set 3 and parameter set 4 can refer to the description of the acquisition method of parameter set 1 in step A1, and will not be repeated here.
[0150] It can be understood that based on the first parameter set obtained in any of the above methods, the first parameter group is any parameter group among the M parameter groups in the first parameter set, and each parameter group satisfies the above motion constraint: the value of the i-th parameter in the first parameter group satisfies the constraint corresponding to the two-dimensional joint distribution of the i-th parameter and each parameter in the remaining (n-1) parameters, and i is a positive integer less than or equal to n.
[0151] For example, from the above Figure 4It can be seen that the jth parameter is any one of the remaining (n-1) parameters mentioned above, and the constraints corresponding to the two-dimensional joint distribution of the ith parameter and the jth parameter include: the constraints corresponding to the first joint distribution when the ith parameter is the independent variable and the jth parameter is the dependent variable; and, the constraints corresponding to the second joint distribution when the ith parameter is the dependent variable and the jth parameter is the independent variable; wherein, the constraints corresponding to the first joint distribution include the upper limit of the variance corresponding to the first joint distribution and the lower limit of the variance corresponding to the first joint distribution; the constraints corresponding to the second joint distribution include the upper limit of the variance corresponding to the second joint distribution and the lower limit of the variance corresponding to the second joint distribution. Here, the constraints corresponding to the two-dimensional joint distribution of parameters are not limited to the upper bound of the variance corresponding to the two-dimensional joint distribution of parameters and the lower bound of the variance corresponding to the two-dimensional joint distribution of parameters. In some possible embodiments, the initial parameter set can also be screened by other constraints corresponding to the two-dimensional joint distribution of parameters to obtain the first parameter set. The other constraints corresponding to the two-dimensional joint distribution of parameters may, for example, include the upper bound of the variance corresponding to the two-dimensional joint distribution of parameters and the mean line corresponding to the two-dimensional joint distribution of parameters, or include the mean line corresponding to the two-dimensional joint distribution of parameters and the lower bound of the variance corresponding to the two-dimensional joint distribution of parameters, or other constraints, which are not limited here.
[0152] In some possible embodiments, the correlation between the above-mentioned multiple parameters is represented by a three-dimensional joint distribution of the parameters. In this case, based on the motion constraints and the initial parameter set, a first parameter set is determined, including: traversing the constraints corresponding to the three-dimensional joint distribution of each parameter in the n parameters in the motion constraints to perform multiple screening on the initial parameter set to obtain the first parameter set.
[0153] For example, assuming that a certain driving scenario corresponds to three parameters, namely parameter a, parameter b and parameter c, the three-dimensional joint distribution constraints of these three parameters include: the constraint corresponding to the three-dimensional joint distribution when parameter a and parameter b are both independent variables and parameter c is the dependent variable, the constraint corresponding to the three-dimensional joint distribution when parameter a and parameter c are both independent variables and parameter b is the dependent variable, and the constraint corresponding to the three-dimensional joint distribution when parameter b and parameter c are both independent variables and parameter a is the dependent variable. Here, the constraint corresponding to the three-dimensional joint distribution of the parameters among the n parameters can also be expressed as similar to the above Figure 4 The upper limit of the variance and the upper limit of the variance shown in the figure can be found in the above Figure 4 Here, the embodiment of the present application does not specifically limit the number of parameters involved in the joint distribution.
[0154] S202: Perform a driving simulation test using the first parameter set to obtain a second parameter set, where the second parameter set includes K motion parameter groups out of M motion parameter groups, and the K motion parameter groups meet the test purpose of the driving scenario, and K is a positive integer less than or equal to M.
[0155] In the embodiment of the present application, before executing the driving simulation test, the driver model to be used in the driving simulation test is trained so that the trained driver model has the same driving ability as a human driver. It can be understood that the trained driver model can be applied to the driving simulation test.
[0156] Exemplarily, the driver's ability includes the braking ability of a human driver in response to an emergency scenario (or risk scenario). For example, it is possible to determine whether the driver model has the same driving ability as a human driver by testing whether the risk point identification, reaction time, and braking curve of the driver model in response to an emergency scenario are consistent with the statistical data of the human driver's driving ability. The training of driver models has been widely used in driving simulation tests. For the sake of brevity of the specification, the training method of the driver model will not be repeated here. In addition, the embodiment of the present application does not limit the method used to train the driver model.
[0157] For example, a trained driver model is given as an example. The driver model is applied to the main vehicle to be tested. In the vehicle cut-in scenario, the main vehicle using the driver model can achieve the following effects: after the cut-in vehicle has a lateral deviation of 0.375m, the main vehicle still maintains the original speed within the reaction time of 1.19s, and after the reaction time, the main vehicle can increase its deceleration from 0 to the maximum braking deceleration of 9.14m / s within 0.55s. 2 , then the driver model has the same driving ability as a human driver.
[0158] For another example, a trained driver model is given as an example. The driver model is applied to the main vehicle to be tested. In the vehicle cut-out scenario, the main vehicle using the driver model can achieve the following effects: after the cut-out vehicle is laterally offset by 1 / 4 of the vehicle body (0.45m), the main vehicle still maintains the original speed within the reaction time of 1.25s, and after the reaction time, the main vehicle can increase its deceleration from 0 to the maximum braking deceleration of 8.66m / s within 0.79s. 2 , then the driver model has the same driving ability as a human driver.
[0159] Exemplarily, the trained driver model having the same driving ability as a human driver means that the driving ability possessed by the trained driver model (i.e., the driver model used in the driving simulation test) is the same as or matches the driving ability of a human driver required for the test purpose of the driving scenario.
[0160] Exemplarily, the test objectives of the driving scenario include one or more of: no collision occurs, the distance to the vehicle in front is not less than a preset threshold during braking, and no skidding occurs when exiting a curve.
[0161] Taking the test purpose of no collision as an example, no collision occurs means that no collision occurs for the human driver in the first driving scenario, including no collision occurs in the first driving scenario for at least one of the average driving level, the lowest driving level and the highest driving level of ordinary human drivers, or no collision occurs in the first driving scenario for at least one of the average driving level, the lowest driving level and the highest driving level of cautious and skilled human drivers, or no collision occurs in the first driving scenario for at least one of the average driving level, the lowest driving level and the highest driving level of mature human drivers.
[0162] Here, ordinary human drivers, cautious and skilled human drivers, and mature human drivers are just some examples of human drivers, and the examples of human drivers are not limited to ordinary human drivers, cautious and skilled human drivers, and mature human drivers. For example, ordinary human drivers, cautious and skilled human drivers, and mature human drivers can be classified based on at least one of the parameters such as driving experience, the concentration time of the human driver, the number of historical collision accidents, and the reaction time in emergency scenarios.
[0163] Take a specific example to illustrate that the driving ability of the driver model should be the same as or match the driving ability of the human driver required by the test purpose of the driving scenario: assuming that the test purpose of the driving scenario is to ensure that an ordinary human driver with average driving skills does not collide in the driving scenario, the driving ability of the driver model used in the driving simulation test is the average driving level of an ordinary human driver.
[0164] In an embodiment of the present application, the driving simulation test is performed using the first parameter set, including: using each parameter group in the first parameter set to perform the driving simulation test to determine whether the vehicle to be tested meets the test purpose in the driving scenario, and retaining the parameter group in the first parameter set that meets the test purpose of the driving scenario. The driver model used in the driving simulation test is obtained through driver ability training.
[0165] In some possible embodiments, during the execution of the driving simulation test, the network-side device may also present an interface of the driving simulation test to the user on a display device.
[0166] See also Figure 5 , Figure 5 Schematic diagram of a driving simulation test interface in a vehicle cut-in scenario provided by an embodiment of the present application. Figure 5In the example, the driving scenario is a vehicle cut-in scenario, and the test purpose is to take no collision as an example, wherein vehicle 1 is the main vehicle, i.e., the vehicle to be tested, and vehicle 1 applies the above-mentioned driver model, and vehicle 2 is the cut-in vehicle that will cut in front of vehicle 1. Each parameter group in the first parameter set obtained in the vehicle cut-in scenario is used in turn to perform a driving simulation test to observe whether vehicle 1 collides with vehicle 2, and the parameter group in the first parameter set that is determined by the driving simulation test that vehicle 1 and vehicle 2 do not collide is retained.
[0167] Assume that the first parameter set includes the parameter group {50, -10, 3, 10} shown in Table 2 above, and take the application of the parameter group {50, -10, 3, 10} for driving simulation test as an example, that is, Figure 5 In the example, the speed of vehicle 1 is set to 50km / h, the speed of vehicle 2 relative to vehicle 1 is set to -10km / h, the lane change time of vehicle 2 is set to 3s, and the distance between vehicle 2 and vehicle 1 when cutting in is set to 10m. In this way, observe whether vehicles 1 and 2 will collide. If vehicles 1 and 2 collide, it means that the parameter group {50, -10, 3, 10} does not meet the test purpose; if vehicles 1 and 2 do not collide, it means that the parameter group {50, -10, 3, 10} meets the test purpose, and the parameter group is retained, that is, the parameter group {50, -10, 3, 10} is included in the above second parameter set. In this way, the driving simulation test is performed on each parameter group in the first parameter set in turn, and finally the second parameter set can be obtained. In this case, each parameter group in the second parameter set meets the test purpose of "no collision" in the vehicle cutting-in scenario.
[0168] See also Figure 6 , Figure 6 Schematic diagram of a driving simulation test interface in a vehicle cut-out scenario provided by an embodiment of the present application. Figure 6 In the example, the driving scenario is a vehicle cut-out scenario, and the test purpose is to take no collision as an example, wherein vehicle 1 is the main vehicle, i.e., the vehicle to be tested, and vehicle 1 applies the above-mentioned driver model, vehicle 3 is the cut-out vehicle that is about to cut out in front of vehicle 1, and vehicle 4 is the front vehicle before vehicle 3 cuts out. The following time distance of vehicle 1 can be set to 3s or 1.8s, for example. In this case, each parameter group in the first parameter set obtained in the vehicle cut-out scenario is used in turn to perform a driving simulation test to observe whether vehicle 1 collides with vehicle 4, and the parameter group in the first parameter set that is determined by the driving simulation test that vehicle 1 and vehicle 4 do not collide is retained, so that a second parameter set can be obtained. In this case, each parameter group in the second parameter set meets the test purpose of "no collision" in the vehicle cut-out scenario.
[0169] Here, the acquisition of the first parameter set in the vehicle cut-out scenario can refer to the relevant description of the acquisition of the first parameter set in the vehicle cut-in scenario exemplified in S201, wherein in the vehicle cut-out scenario, the above n parameters include the cut-out vehicle (equivalent to Figure 6 The speed of vehicle 3) in the front vehicle (equivalent to Figure 6 4) The speed of the vehicle cutting out, the lane changing time of the vehicle cutting out, and the distance between the vehicle cutting out and the vehicle in front when cutting out.
[0170] For example, assuming that in the vehicle cut-out scenario, the motion constraint used to obtain the first parameter set is the constraint corresponding to the two-dimensional joint distribution of the parameters. In this case, see Figure 7 , Figure 7 Schematic diagram of constraints corresponding to the two-dimensional joint distribution of parameters in some vehicle cut-out scenarios provided in the embodiments of the present application. Figure 7 It can be seen that in the vehicle cut-out scenario, the parameters include “cut-out vehicle speed (i.e. Figure 7 The speed of the preceding vehicle in the cut-out vehicle ( Figure 7 relative speed in Chinese)”, “lane-changing time of the cutting-out vehicle” and “distance between the cutting-out vehicle and the vehicle in front when cutting out ( Figure 7 These four parameters are combined in pairs to construct a two-dimensional joint distribution. There are 12 combinations in total. For details, see Figure 7 There are 12 pictures listed in it.
[0171] Understandably, Figure 7 The dense data points in each figure are only an example of the distribution form of the corresponding data in the vehicle cut-out scene, and do not limit the distribution form of the data to only Figure 7 As shown. Similarly, Figure 7 The schematic diagram of the constraints corresponding to the two-dimensional joint distribution of parameters in the vehicle cut-out scenario is only given as an example, and does not limit the schematic diagram of the constraints corresponding to the two-dimensional joint distribution of parameters in the vehicle cut-out scenario to only Figure 7 shown.
[0172] by Figure 7 Take (1) as an example to illustrate. Figure 7 (1) shows a schematic diagram of the constraints corresponding to the two-dimensional joint distribution when "front vehicle speed" is the independent variable and "relative speed" is the dependent variable, where the dense data points represent the data distribution of the two parameters "front vehicle speed" and "speed of the cut-in vehicle relative to the main vehicle" in the natural driving data of the vehicle cut-out scene. Figure 7 (1) also shows the Figure 7 The upper and lower bounds of the variance of the data point fitting of (1) are determined by, for example, firstly based on Figure 7The data distribution of the data points shown in (1) is the joint probability distribution of these two parameters, and the mean line ( Figure 7 (1) is not shown), the upper and lower limits of the variance corresponding to the two-dimensional joint distribution are determined by taking the mean as the center and the variance of the upper and lower distances as preset multiples (for example, 3 times), where the upper limit is Figure 7 The upper line segment on the two line segments shown in (1) has a lower boundary of Figure 7 The line segment located above the two line segments shown in (1). Here, Figure 7 You can also refer to other figures in Figure 7 For the sake of brevity, the description of (1) will not be repeated here.
[0173] In some possible embodiments, since the first parameter set will be used for driving simulation testing, in order to improve test efficiency, the first parameter set can also be screened based on driving experience to further reduce the number of parameter groups in the first parameter set. Taking the driving scenario as a vehicle cut-out scenario and the test purpose as no collision, after obtaining the first parameter set in the vehicle cut-out scenario, assuming that the driving experience includes: for the preceding vehicle (e.g. Figure 6 The vehicle 4 in the figure is at rest, that is, the speed of the preceding vehicle is 0, and the preceding vehicle (i.e., the cut-out vehicle, e.g. Figure 6 When the vehicle 3 in the test is cut out, if you want the main vehicle (i.e. the vehicle to be tested, for example Figure 6 The vehicle 1) does not collide with the vehicle in front, and the lane changing time of the vehicle in front cannot be less than 1s when the speed of the vehicle in front is below 20km / h.
[0174] In an embodiment of the present application, the obtained second parameter set can be used for safety testing of the autonomous driving vehicle in a driving scenario. For example, if the driving scenario is a vehicle cut-in scenario, the obtained second parameter set can be used for safety testing of the autonomous driving vehicle in the vehicle cut-in scenario, such as for testing whether the vehicle meets the test purpose of the driving scenario. Further, the vehicle cut-in scenario can be divided into a small car cut-in scenario and a large car cut-in scenario. In this case, the second parameter set obtained in the small car cut-in scenario is used for safety testing of the autonomous driving vehicle in the small car cut-in scenario, and the second parameter set obtained in the large car cut-in scenario is used for safety testing of the autonomous driving vehicle in the large car cut-in scenario.
[0175] Exemplarily, the second parameter set may be provided to developers of vehicles or autonomous driving systems, or to testing organizations of autonomous driving vehicles, etc., without specific limitation herein.
[0176] It can be seen that the method of determining the scene parameters described in the above embodiment shows that: an initial parameter set related to the driving scene can be obtained first, and the initial parameter set can be screened based on the above motion constraints to obtain a first parameter set, and finally a driving simulation test is performed through the first parameter set to obtain a second parameter set. In this way, data screening is performed first and then the driving simulation test is performed, which greatly reduces the workload and improves data processing efficiency.
[0177] It can be seen that by implementing the embodiments of the present application and performing data screening based on motion constraints, the values of parameters in the driving scenario can be made reasonable and predictable in the physical world, and by performing driving simulation tests, the values of parameters that can be avoided by human drivers can be identified. In this way, reasonable values can be provided for the parameters of the driving scenario, which is not only conducive to improving the accuracy of the test results, but also conducive to improving the rationality and effectiveness of the autonomous driving safety test scenarios.
[0178] In some possible embodiments, the embodiments of the present application do not limit the execution steps of the above-mentioned two operations of data screening and driving simulation testing.
[0179] See also Figure 8 , Figure 8 This is another method flow chart of determining scene parameters provided by the embodiment of the present application. This method can be applied to the above Figure 1 The network side device shown. The method includes but is not limited to the following steps:
[0180] S801: Determine an initial parameter set related to a driving scene, where the initial parameter set includes P motion parameter groups, each motion parameter group includes a set of values of n motion parameters, and the n motion parameters are used to describe the state of a subject participating in the driving scene, and P and n are integers greater than 1.
[0181] Here, driving scenes, motion parameters, and motion parameter groups can refer to the above Figure 2 Description of the corresponding contents in the embodiments.
[0182] The process of obtaining the initial parameter set can be found in the above Figure 3 For the sake of brevity of the description, the description of the embodiments will not be repeated here.
[0183] S802: A target parameter set is obtained by performing data screening and driving simulation tests on the initial parameter set, wherein the target parameter set includes K motion parameter groups among the above-mentioned P motion parameter groups, each of the K motion parameter groups satisfies the motion constraints and meets the test purpose of the driving scenario, and K is a positive integer less than P.
[0184] Here, motion constraints, data screening process, and driving simulation tests can refer to the above Figure 2For the sake of brevity of the specification, the description of the corresponding contents in the embodiments will not be repeated here.
[0185] In one implementation, the target parameter set is obtained by performing data screening and driving simulation test on the initial parameter set, including: performing data screening on the initial parameter set to obtain a first parameter set, the first parameter set including M motion parameter groups among P motion parameter groups, each motion parameter group in the M motion parameter groups satisfies the above motion constraint, and M is a positive integer less than P; performing a driving simulation test on the first parameter set to obtain a target parameter set, the above K motion parameter groups are included in the above M motion parameter groups, and M is greater than or equal to K. It can be understood that this implementation is equivalent to the above Figure 2 In the implementation method described in the embodiment, the target parameter set is Figure 2 The second parameter set in S202 of the embodiment.
[0186] In another implementation, a driving simulation test may be performed first and then data screening may be performed based on motion constraints. For example, a target parameter set is obtained by performing data screening and driving simulation tests on the initial parameter set, including: performing a driving simulation test on the initial parameter set to obtain parameter set A1, wherein each parameter group in parameter set A1 meets the test purpose of the driving scenario; performing data screening on parameter set A1 based on motion constraints to obtain a target parameter set, wherein each parameter group in the target parameter set not only meets the test purpose of the driving scenario, but also satisfies the motion constraints. Here, the target parameter set is equivalent to Figure 2 The second parameter set in S202 of the embodiment.
[0187] In another implementation, a driving simulation test and data screening based on motion constraints can also be performed simultaneously. For example, a target parameter set is obtained by performing data screening and driving simulation testing on the initial parameter set, including: performing data screening on the initial parameter set based on the above-mentioned motion constraint to obtain parameter set B1, wherein each parameter group in parameter set B1 satisfies the motion constraint, and performing a driving simulation test on the initial parameter set to obtain parameter set B2, wherein each parameter group in parameter set B2 meets the test purpose of the driving scenario; taking the intersection of parameter set B1 and parameter set B2 to obtain the target parameter set, wherein each parameter group in the target parameter set satisfies both the motion constraint and the test purpose of the driving scenario. Here, the order of obtaining parameter set B1 and parameter set B2 is not limited. The target parameter set is equivalent to Figure 2 The second parameter set in S202 of the embodiment.
[0188] It can be seen that when implementing the embodiments of the present application and determining the parameter values of the driving scenario, not only is it considered that the parameter values in the driving scenario are reasonable and foreseeable, but also the situations that can be avoided by human drivers are considered. It is possible to provide reasonable values for the parameters of the driving scenario, which is not only conducive to improving the accuracy of the test results, but also conducive to improving the rationality and effectiveness of the autonomous driving safety test scenario.
[0189] See also Fig. 9 , Fig. 9 3 is a structural diagram of a data processing device provided in an embodiment of the present application, wherein the data processing device 30 includes a processing unit 310 and a simulation unit 312. The data processing device 30 can be implemented by hardware, software, or a combination of hardware and software.
[0190] For example, the data processing device 30 may be Figure 1 For example, the processing unit 310 is equivalent to the data screening module in the network side device, and the simulation unit 312 is equivalent to the simulation test module in the network side device.
[0191] Among them, the processing unit 310 is used to determine a first parameter set related to the first driving scene, the first parameter set includes M motion parameter groups, each motion parameter group includes a set of values of n motion parameters, the above n motion parameters are used to describe the state of the subject participating in the first driving scene, each motion parameter group in the above M motion parameter groups satisfies the motion constraint, and the motion constraint is used to indicate the correlation between multiple motion parameters in the above n motion parameters, M and n are positive integers greater than 1; the simulation unit 312 is used to perform a driving simulation test through the first parameter set to obtain a second parameter set, the second parameter set includes K motion parameter groups in the above M motion parameter groups, these K motion parameter groups meet the test purpose of the first driving scene, and K is a positive integer less than or equal to M.
[0192] The data processing device 30 can be used to implement Figure 2 The method described in the embodiment. Figure 2 In an embodiment, the processing unit 310 may be used to execute S201 , and the simulation unit 312 may be used to execute S202 .
[0193] In some possible embodiments, the data processing device 30 may also be used to implement Figure 3 The method described in the embodiment. For example, Figure 3 In the embodiment, the processing unit 310 can be used to execute S301-S303. In some possible embodiments, the data processing device 30 can also be used to implement Figure 8 The method described in the embodiment will not be described in detail here.
[0194] It should be understood that the division of the units in the above data processing device 30 is only a division of logical functions, and in actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. In addition, the units in the device can be implemented in the form of a processor calling software; for example, the device includes a processor, the processor is connected to a memory, and instructions are stored in the memory. The processor calls the instructions stored in the memory to implement any of the above methods or realize the functions of the units of the device, wherein the processor is, for example, a general-purpose processor, such as a central processing unit (CPU) or a microprocessor, and the memory is a memory inside the device or a memory outside the device. Alternatively, the units in the device can be implemented in the form of hardware circuits, and the functions of some or all of the units can be realized by designing the hardware circuits, and the hardware circuits can be understood as one or more processors; for example, in one implementation, the hardware circuit is an application-specific integrated circuit (ASIC), and the functions of some or all of the above units are realized by designing the logical relationship of the components in the circuit; for another example, in another implementation, the hardware circuit can be realized by a programmable logic device (PLD), taking a field programmable gate array (FPGA) as an example, which can include a large number of logic gate circuits, and the connection relationship between the logic gate circuits is configured by a configuration file, so as to realize the functions of some or all of the above units. All units of the above devices can be realized in the form of a processor calling software, or in the form of hardware circuits, or in part by a processor calling software, and the rest by hardware circuits.
[0195] In the embodiment of the present application, the processor is a circuit with signal processing capability. In one implementation, the processor can be a circuit with instruction reading and running capability, such as a central processing unit (CPU), a microprocessor, a graphics processing unit (GPU) (which can be understood as a microprocessor), or a digital signal processor (DSP), etc.; in another implementation, the processor can realize certain functions through the logical relationship of the hardware circuit, and the logical relationship of the hardware circuit is fixed or reconfigurable, such as a hardware circuit implemented by a processor as an application-specific integrated circuit (ASIC) or a programmable logic device (PLD), such as an FPGA. In a reconfigurable hardware circuit, the processor loads a configuration document to implement the process of hardware circuit configuration, which can be understood as the process of the processor loading instructions to implement the functions of some or all of the above units. In addition, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as a neural network processing unit (NPU), a tensor processing unit (TPU), a deep learning processing unit (DPU), etc.
[0196] 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.
[0197] In addition, all or part of the units in the above device can be integrated together, or can be implemented independently. In one implementation, these units are integrated together and implemented in the form of a system-on-a-chip (SOC). The SOC may include at least one processor for implementing any of the above methods or implementing the functions of each unit of the device. The type of the at least one processor may be different, for example, including a CPU and an FPGA, a CPU and an artificial intelligence processor, a CPU and a GPU, etc.
[0198] See also Fig.10 , Fig.10 This is a schematic diagram of the structure of a data processing device provided in this application. Fig.10As shown, the data processing device 40 includes: a processor 401, a communication interface 402, a memory 403 and a bus 404. The processor 401, the memory 403 and the communication interface 402 communicate with each other via the bus 404. It should be understood that the present application does not limit the number of processors and memories in the data processing device 40.
[0199] In one implementation, the data processing device 40 may be a network-side device. The network-side device may be, for example, a computing device (e.g., a server) deployed on the network side, or a component or chip in the server, or a system-level device composed of multiple servers. The network-side device may be deployed in a cloud environment or an edge environment, which is not specifically limited herein.
[0200] The bus 404 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig.10 The bus 404 is represented by only one line, but it does not mean that there is only one bus or one type of bus. The bus 404 may include a path for transmitting information between various components of the data processing device 40 (for example, the memory 403, the processor 401, and the communication interface 402).
[0201] The processor 401 can refer to the related description of the processor in the above embodiment, which will not be repeated here.
[0202] The memory 403 is used to provide a storage space, and the storage space can store data such as an operating system and a computer program. The memory 403 can be a random access memory (RAM), an erasable programmable read only memory (EPROM), a read-only memory (ROM), or a portable read-only memory (CD-ROM), etc., or a combination of multiple types. The memory 403 can exist alone or be integrated into the processor 401.
[0203] The communication interface 402 may be used to provide information input or output for the processor 401. Alternatively, the communication interface 402 may be used to receive data sent externally and / or send data externally, and may be a wired link interface such as an Ethernet cable, or a wireless link interface (such as Wi-Fi, Bluetooth, general wireless transmission, etc.). Alternatively, the communication interface 402 may also include a transmitter (such as a radio frequency transmitter, an antenna, etc.) coupled to the interface, or a receiver, etc.
[0204] In some possible embodiments, the data processing device 40 may further include a display 405. The display 405 is connected or coupled to the processor 401 via a bus 404. The display 405 may be used to display an interface of a driving simulation test to a user. The display 405 may be a display screen, and the display screen may be a liquid crystal display (LCD), an organic or inorganic light-emitting diode (OLED), an active matrix organic light-emitting diode panel (AMOLED), etc.
[0205] The processor 401 in the data processing device 40 is used to read the computer program stored in the memory 403 to execute the aforementioned method, for example Figure 2 , Figure 3 or Figure 8 The method described.
[0206] In a possible design, the data processing device 40 may be a processor for executing Figure 2 The processor 401 may be used to read one or more computer programs stored in the memory to perform the following operations:
[0207] Determine a first parameter set related to the driving scene, the first parameter set includes M motion parameter groups, each motion parameter group includes a set of values of n motion parameters, the n motion parameters are used to describe the state of the subject participating in the driving scene, each motion parameter group in the M motion parameter groups satisfies a motion constraint, the motion constraint is used to indicate a correlation relationship between multiple motion parameters in the n motion parameters, and M and n are positive integers greater than 1;
[0208] Through the simulation unit 312, the first parameter set is used to perform a driving simulation test to obtain a second parameter set, where the second parameter set includes K motion parameter groups among the above-mentioned M motion parameter groups. These K motion parameter groups meet the test purpose of the driving scenario, and K is a positive integer less than or equal to M.
[0209] In a possible design, the data processing device 40 may be a processor for executing Figure 8 The processor 401 may be used to read one or more computer programs stored in the memory to perform the following operations:
[0210] Determine an initial parameter set related to the driving scene, the initial parameter set includes P motion parameter groups, each motion parameter group includes a set of values of n motion parameters, the n motion parameters are used to describe the state of the subject participating in the driving scene, and P and n are integers greater than 1;
[0211] By performing data screening and driving simulation tests on the initial parameter set, a target parameter set is obtained, which includes K motion parameter groups among the above-mentioned P motion parameter groups. Each motion parameter group in the K motion parameter groups satisfies the motion constraints and meets the test purpose of the driving scenario, and K is a positive integer less than P.
[0212] In the above-mentioned embodiments of this invention, the description of each embodiment has its own emphasis. For the part that is not described in detail in a certain embodiment, please refer to the relevant description of other embodiments. In addition, in the various embodiments of this application, if there is no special explanation and logical conflict, the terms and / or descriptions between the various embodiments are consistent and can be referenced to each other. The technical features in different embodiments can be combined to form a new embodiment according to their internal logical relationship.
[0213] It should be noted that, a person skilled in the art can see that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electrically-erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.
[0214] The essence of the technical solution of the present application or the part that makes the contribution or all or part of the technical solution can be embodied in the form of a software product. The computer program product is stored in a storage medium and includes a number of instructions for enabling a device (which can be a personal computer, a server, or a network device, a robot, a single-chip microcomputer, a chip, a robot, etc.) to execute all or part of the steps of the method described in each embodiment of the present application.
Claims
1. A method for determining scene parameters, characterized in that: The method comprises: Determine a first parameter set related to a first driving scene, the first parameter set includes M motion parameter groups, each motion parameter group includes a set of values of n motion parameters, the n motion parameters are used to describe a state of a subject participating in the first driving scene, each motion parameter group in the M motion parameter groups satisfies a motion constraint, the motion constraint is used to indicate a correlation relationship between multiple motion parameters in the n motion parameters, and M and n are integers greater than 1; A driving simulation test is performed through the first parameter set to obtain a second parameter set, wherein the second parameter set includes K motion parameter groups among the M motion parameter groups, and the K motion parameter groups meet the test purpose of the first driving scenario, and K is a positive integer less than or equal to M.
2. The method according to claim 1, characterized in that The motion constraint is obtained based on natural driving data statistics in the first driving scenario.
3. The method according to claim 1 or 2, characterized in that: The second parameter set is used for safety testing of the autonomous driving vehicle in the first driving scenario.
4. The method according to any one of claims 1 to 3, characterized in that: The second parameter set is used to test whether the vehicle meets the test purpose.
5. The method according to any one of claims 1 to 4, characterized in that: The first motion parameter group is any one of the M motion parameter groups, and each of the motion parameter groups satisfies the motion constraint, including: The value of the i-th motion parameter in the first motion parameter group satisfies the constraint corresponding to the two-dimensional joint distribution of the i-th motion parameter and each motion parameter in the remaining (n-1) motion parameters, and i is a positive integer less than or equal to n.
6. The method according to any one of claims 1 to 5, characterized in that: The determining a first parameter set related to the first driving scenario includes: Based on the motion constraint and a third parameter set, the first parameter set related to the first driving scenario is determined, wherein the third parameter set includes the Cartesian product of the generalized value sets of the n motion parameters, and the value range corresponding to each of the n motion parameters in the third parameter set at least covers a preset value range of the motion parameter.
7. The method according to claim 6, characterized in that The generalized value set of the i-th motion parameter among the n motion parameters is obtained according to the value range corresponding to the i-th motion parameter and the value step size of the i-th motion parameter.
8. The method according to claim 7, characterized in that The value range corresponding to the i-th motion parameter includes a first interval and a second interval, the value step of the i-th motion parameter includes a first value corresponding to the first interval and a second value corresponding to the second interval, and the first interval and the second interval have no overlap; Among them, when the i-th motion parameter takes a value within the first interval, the danger level of the driving scene is higher than the danger level of the driving scene when the i-th motion parameter takes a value within the second interval, and the first value is smaller than the second value.
9. The method according to any one of claims 1 to 8, characterized in that: The first driving scenario is a vehicle cut-in scenario, the first vehicle is a vehicle to be tested, and the second vehicle is a cut-in vehicle that will cut in front of the first vehicle. The n motion parameters include the speed of the first vehicle, the distance between the second vehicle and the first vehicle when the second vehicle cuts in, the speed of the second vehicle relative to the first vehicle, and the lane changing time of the second vehicle.
10. The method according to any one of claims 1 to 9, characterized in that: The first driving scene is a vehicle cut-out scene, the third vehicle is a vehicle to be tested, the fourth vehicle is a cut-out vehicle in front of the third vehicle, and the fifth vehicle is a front vehicle before the fourth vehicle cuts out. The n motion parameters include the speed of the fourth vehicle, the speed of the fifth vehicle relative to the fourth vehicle, the distance between the fourth vehicle and the fifth vehicle when the fourth vehicle cuts out, and the lane changing time of the fourth vehicle.
11. The method according to any one of claims 1 to 10, characterized in that: The test purpose of the first driving scenario is to avoid collision.
12. A device for determining scene parameters, characterized in that: The device comprises: a processing unit, configured to determine a first parameter set related to a first driving scene, the first parameter set comprising M motion parameter groups, each motion parameter group comprising a set of values of n motion parameters, the n motion parameters being used to describe a state of a subject participating in the first driving scene, each motion parameter group in the M motion parameter groups satisfying a motion constraint, the motion constraint being used to indicate a correlation relationship between multiple motion parameters in the n motion parameters, and M and n being integers greater than 1; A simulation unit is used to perform a driving simulation test through the first parameter set to obtain a second parameter set, wherein the second parameter set includes K motion parameter groups out of the M motion parameter groups, and the K motion parameter groups meet the test purpose of the first driving scenario, and K is a positive integer less than or equal to M.
13. The device according to claim 12, characterized in that The motion constraint is obtained based on natural driving data statistics in the first driving scenario.
14. The device according to claim 12 or 13, characterized in that The second parameter set is used for safety testing of the autonomous driving vehicle in the first driving scenario.
15. The device according to any one of claims 12 to 14, characterized in that: The second parameter set is used to test whether the vehicle meets the test purpose.
16. The device according to any one of claims 12 to 15, characterized in that: The first motion parameter group is any one of the M motion parameter groups, and each of the motion parameter groups satisfies the motion constraint, including: The value of the i-th motion parameter in the first motion parameter group satisfies the constraint corresponding to the two-dimensional joint distribution of the i-th motion parameter and each motion parameter in the remaining (n-1) motion parameters, and i is a positive integer less than or equal to n.
17. The device according to any one of claims 12 to 16, characterized in that: The processing unit is specifically used for: Based on the motion constraint and a third parameter set, the first parameter set related to the first driving scenario is determined, wherein the third parameter set includes the Cartesian product of the generalized value sets of the n motion parameters, and the value range corresponding to each of the n motion parameters in the third parameter set at least covers a preset value range of the motion parameter.
18. The device according to claim 17, characterized in that The generalized value set of the i-th motion parameter among the n motion parameters is obtained according to the value range corresponding to the i-th motion parameter and the value step size of the i-th motion parameter.
19. The device according to claim 18, characterized in that The value range corresponding to the i-th motion parameter includes a first interval and a second interval, the value step of the i-th motion parameter includes a first value corresponding to the first interval and a second value corresponding to the second interval, and the first interval and the second interval have no overlap; Among them, when the i-th motion parameter takes a value within the first interval, the danger level of the driving scene is higher than the danger level of the driving scene when the i-th motion parameter takes a value within the second interval, and the first value is smaller than the second value.
20. The device according to any one of claims 12 to 19, characterized in that The first driving scenario is a vehicle cut-in scenario, the first vehicle is a vehicle to be tested, and the second vehicle is a cut-in vehicle that will cut in front of the first vehicle. The n motion parameters include the speed of the first vehicle, the distance between the second vehicle and the first vehicle when the second vehicle cuts in, the speed of the second vehicle relative to the first vehicle, and the lane changing time of the second vehicle.
21. The device according to any one of claims 12 to 20, characterized in that: The first driving scene is a vehicle cut-out scene, the third vehicle is a vehicle to be tested, the fourth vehicle is a cut-out vehicle in front of the third vehicle, and the fifth vehicle is a front vehicle before the fourth vehicle cuts out. The n motion parameters include the speed of the fourth vehicle, the speed of the fifth vehicle relative to the fourth vehicle, the distance between the fourth vehicle and the fifth vehicle when the fourth vehicle cuts out, and the lane changing time of the fourth vehicle.
22. The device according to any one of claims 12 to 21, characterized in that The test purpose of the first driving scenario is to avoid collision.
23. 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 11.
24. 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 11.
