Scene generalization coverage assessment method, system and device and readable storage medium

Abstract

A scene generalization coverage assessment method, system and device and a readable storage medium relate to the technical field of automobiles, and specifically comprise the steps of determining a target joint probability density function corresponding to a target scene based on probability distribution characteristics of logic scene parameters; and determining a target probability of the target scene according to a preset target interval, the target joint probability density function and a distribution interval corresponding to the target scene, and taking the target probability as an evaluation result of the scene generalization coverage. According to the invention, the coverage degree of the specific scene generated by generalization can be evaluated.

Description

Technical Field

[0001] This application relates to the field of automotive technology, specifically to a method, system, device, and readable storage medium for evaluating scenario generalization coverage. Background Technology

[0002] Currently, with the rapid development of autonomous driving technology, especially the improvement of autonomous driving levels, traditional automotive testing tools and methods can no longer meet the testing needs of autonomous vehicles. Among related technologies, methods that generalize to generate specific scenarios have become an important means of verifying autonomous driving systems efficiently and at low cost, and are therefore crucial for future testing and verification.

[0003] However, existing technologies lack the ability to evaluate the coverage of specific scenarios generated by generalization. Therefore, how to evaluate the coverage of specific scenarios generated by generalization is an urgent problem to be solved. Summary of the Invention

[0004] This application provides a method, system, device, and readable storage medium for evaluating the coverage of generalized scenarios, which can evaluate the coverage of specific scenarios generated by generalization.

[0005] In a first aspect, embodiments of this application provide a method for evaluating scene generalization coverage, the method comprising: The joint probability density function of the target corresponding to the target scene is determined based on the probability distribution characteristics of the logical scene parameters. Based on the preset target interval, the joint probability density function of the targets, and the distribution interval corresponding to the target scenarios, the target probability of the target scenarios is determined, and the target probability is used as the evaluation result of the scenario generalization coverage.

[0006] In conjunction with the first aspect, in one implementation, the logical scene parameters include dynamic parameters and / or state parameters, wherein the dynamic parameters include one or more of the vehicle speed, the vehicle acceleration, and the target vehicle speed, and the state parameters include lighting conditions and / or weather conditions.

[0007] In conjunction with the first aspect, in one implementation, determining the joint probability density function of the target corresponding to the target scene based on the probability distribution characteristics of logical scene parameters includes: If the target scenario includes the vehicle speed and the vehicle acceleration, then the joint probability density function of the target corresponding to the target scenario is determined based on the first normal distribution corresponding to the vehicle speed and the second normal distribution corresponding to the vehicle acceleration. If the target scenario includes the target vehicle speed, lighting conditions, and weather conditions, then the joint probability density function corresponding to the target scenario is determined based on the first equal probability distribution corresponding to the target vehicle speed, the second equal probability distribution corresponding to the lighting conditions, and the third equal probability distribution corresponding to the weather conditions.

[0008] In conjunction with the first aspect, in one implementation, determining the target probability of a target scene based on a preset target interval, a joint target probability density function, and a distribution interval corresponding to the target scene includes: The area of ​​the probability interval is determined based on the preset first target interval corresponding to the vehicle speed and the preset second target interval corresponding to the vehicle acceleration; The target probability of the target scene is determined by the area of ​​the probability interval, the joint probability density function of the target, and the distribution interval corresponding to the target scene.

[0009] In conjunction with the first aspect, in one implementation, determining the target probability of the target scene based on the area of ​​the probability interval, the joint probability density function of the target, and the distribution interval corresponding to the target scene includes: The target probability of the target scene is calculated based on the area of ​​the probability interval, the joint probability density function of the target, and the distribution interval corresponding to the target scene. Specifically:

[0010]

[0011] In the formula, This is the vehicle's speed; For the acceleration of this vehicle; The area of ​​the probability interval; This refers to the distribution range corresponding to the vehicle's speed. This refers to the distribution range corresponding to the vehicle's acceleration. Let P{(x1, x2) be the joint probability density function of the objective. } represents the target probability of the target scenario.

[0012] In conjunction with the first aspect, in one implementation, determining the target probability of a target scene based on a preset target interval, a joint target probability density function, and a distribution interval corresponding to the target scene includes: The volume of the probability interval is determined based on the preset third target interval corresponding to the target vehicle speed, the preset fourth target interval corresponding to the lighting conditions, and the preset fifth target interval corresponding to the weather conditions; The target probability of the target scene is determined based on the volume of the probability interval, the joint probability density function of the target, and the distribution interval corresponding to the target scene.

[0013] In conjunction with the first aspect, in one implementation, determining the target probability of the target scene based on the probability interval volume, the joint probability density function of the target, and the distribution interval corresponding to the target scene includes: The target probability of the target scene is calculated based on the probability interval volume, the joint probability density function of the target, and the distribution interval corresponding to the target scene. Specifically:

[0014]

[0015] In the formula, x3 is the target vehicle speed; x4 is the lighting conditions; and x5 is the weather conditions. Let the volume be the probability interval. Let the joint probability density function be the objective. The distribution range corresponding to the target vehicle speed; The distribution range corresponding to the lighting conditions; Let P{(x3, x4, x5)} represent the distribution intervals corresponding to weather conditions. } represents the target probability of the target scenario.

[0016] Secondly, embodiments of this application provide a system for evaluating scene generalization coverage, the system comprising: The first processing module is used to determine the joint probability density function of the target scene corresponding to the target scene based on the probability distribution characteristics of the logical scene parameters. The second processing module is used to determine the target probability of the target scene based on the preset target interval, the target joint probability density function and the distribution interval corresponding to the target scene, and use the target probability as the evaluation result of the scene generalization coverage.

[0017] Thirdly, embodiments of this application provide a scene generalization coverage evaluation device, the scene generalization coverage evaluation device including a processor, a memory, and a scene generalization coverage evaluation program stored in the memory and executable by the processor, wherein when the scene generalization coverage evaluation program is executed by the processor, it implements the steps of the scene generalization coverage evaluation method as described in any of the preceding claims.

[0018] Fourthly, embodiments of this application provide a computer-readable storage medium storing a scene generalization coverage evaluation program, wherein when the scene generalization coverage evaluation program is executed by a processor, it implements the steps of the scene generalization coverage evaluation method as described in any of the preceding claims.

[0019] The beneficial effects of the technical solutions provided in this application include: Based on the probability distribution characteristics of logical scene parameters, the joint probability density function of the target scene is determined, which can effectively describe the occurrence probability of the target scene. According to the preset target interval, the joint probability density function of the target scene and the distribution interval corresponding to the target scene, the target probability of the target scene is determined. The target probability of the target scene is obtained by discretization, which can quantify the coverage of the target scene in the overall probability distribution. The target probability is used as the evaluation result of scene generalization coverage, so as to better evaluate the scene coverage. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating an embodiment of the method for evaluating the generalization coverage of the scenario in this application. Figure 2 This is a schematic diagram of the target joint probability density function of the vehicle speed and acceleration in an embodiment of the method for evaluating the generalization coverage of the scenario in this application; Figure 3 This is a schematic diagram of the method for determining the preset first target interval in an embodiment of the method for evaluating the generalization coverage of the scenario in this application; Figure 4 This is a schematic diagram of the area of ​​the probability interval in an embodiment of the method for evaluating the generalization coverage of the scenario in this application; Figure 5 This is a schematic diagram of the hardware structure of the scene generalization coverage evaluation device involved in the embodiments of this application. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0022] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0023] Firstly, embodiments of this application provide a method for evaluating scenario generalization coverage.

[0024] In one embodiment, reference is made to Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the method for evaluating the generalization coverage of the scenario in this application. Figure 1 As shown, the evaluation methods for scene generalization coverage include: Step S10: Determine the joint probability density function of the target scene based on the probability distribution characteristics of the logical scene parameters.

[0025] As an example, in this application embodiment, the autonomous driving level corresponding to the tested autonomous driving system is clearly defined in accordance with the provisions of GB / T 40429-2021 "Classification of Driving Automation for Motor Vehicles," and the functional definition of the tested autonomous driving system should meet the corresponding requirements in the standard as a prerequisite for test case design. A functional scenario is an operational scenario at the highest level of semantic description, that is, describing entities within a scenario area and the relationships between entities through linguistic scenario symbols. Functional scenarios are used for project definition, hazard analysis, and risk assessment in the concept phase. They contain an overall dynamic description of the autonomous vehicle and its driving environment components over a period of time, that is, an organic combination of the autonomous vehicle's driving occasions and driving scenarios, with the components determined by the autonomous vehicle functions to be verified.

[0026] It should be noted that the functional scenario description should at least include map information, environmental information, basic information of traffic participants, a description of the dynamic process of the scenario, scenario start conditions, scenario termination conditions, and scenario end conditions. Specifically: 1) Map information clarifies the specific geographical information, roads, facilities, and other static information of the scene. Map information can be provided by map providers based on GNSS information for the corresponding area, or it can be collected by the scene acquisition vehicle or created later according to the actual situation. 2) Environmental information describes the environmental conditions at the time the scene occurred, including light intensity, light direction, weather conditions, and specific time; 3) Basic information of traffic participants describes the basic information of relevant traffic participants in the scene, such as vehicle size and type; 4) The scene dynamic process description includes the scene initial information description, scene process information description, and scene end information description. The scene initial information describes the position, speed, acceleration, and other information of each participant in the initial conditions of the scene. For scenes with complex processes that are not easy to describe, the process can be divided into simple and easy-to-describe sub-processes. For example, the cutting process in the scenario of the vehicle cutting in front can be divided into three sub-processes: longitudinal travel in the original lane, lateral cutting, and longitudinal travel in the new lane. The scene end information describes the position, speed, acceleration, and other information of each traffic participant at the end of the traffic scene. 5) The scenario start conditions describe the external conditions that need to be met before a functional scenario can begin; 6) Scene termination conditions describe the conditions under which this scene cannot proceed smoothly or change to another scene. For example, in a high-speed following scene, the vehicle in front suddenly brakes, causing the following scene to fail and changing to an emergency braking scene. 7) Scene termination conditions describe the conditions that must be met for a scene to end smoothly. For example, in a following scene, the following scene can be considered to have ended when the vehicle in front gradually drives away and the distance between the vehicles reaches a certain level.

[0027] Logical scenarios express entity characteristics and relationships between entities by defining the range of state space variable parameters (i.e., logical scenario parameters). They are based on logical scenario parameters to further describe functional scenarios in detail and are used to generate test requirements during the project development phase. All parameters that need to be logically valued in a functional scenario can be extracted based on the vehicle's ODC (Design Operating Conditions) and test safety requirements. The set of logical scenario parameters constitutes the state space of the logical scenario. A logical scenario with N logical scenario parameters is called an N-dimensional state space. For each logical scenario with a continuous value range, any number of specific scenarios (i.e., target scenarios) can be derived. For each logical scenario with a discrete value range, specific scenarios are derived according to the discrete values.

[0028] Specifically, variable names should consist of letters, numbers, and underscores, with the first letter being an uppercase letter; to facilitate the generation of test cases through programming, the logical relationships between entities should be expressed using clear and easy-to-understand mathematical logic, following the logical expression relationships as follows: a) Variable 1 < Variable 2 means that Variable 1 is less than Variable 2; b) Variable1 = Variable2 means that Variable1 is equal to Variable2; c) Variable 1 > Variable 2 means that Variable 1 is greater than Variable 2; d) Variable 1 ≠ Variable 2, meaning that Variable 1 is not equal to Variable 2.

[0029] It should be noted that logical scenario parameters refer to key variables describing the target scenario, such as vehicle speed, location, road conditions, and weather conditions. These parameters collectively determine the characteristics and evolution of the target scenario. Probability distribution characteristics are obtained through statistical analysis of logical scenario parameters, yielding corresponding mean, standard deviation, correlation coefficient, and probability density function, which reflect the distribution patterns of each logical scenario parameter under different conditions. By combining these probability distribution characteristics, the target joint probability density function corresponding to the target scenario can be determined. This function describes the joint distribution of multiple logical scenario parameters in the target scenario, revealing the interdependencies between different parameters and the probability of their joint occurrence, thus providing a quantitative basis for accurate analysis of the target scenario.

[0030] Step S20: Determine the target probability of the target scene based on the preset target interval, the target joint probability density function, and the distribution interval corresponding to the target scene, and use the target probability as the evaluation result of the scene generalization coverage.

[0031] In this embodiment of the application, the preset target interval refers to the discretization step size set according to the test requirements, which is used to discretize the continuous variables of the target scenario into a finite interval. The specific value can be determined according to the actual needs and is not limited here. The target probability of the target scenario refers to the probability of the scenario occurring in the entire probability distribution. The distribution interval corresponding to the target scenario refers to the actual value range of each parameter in the target scenario. Based on these intervals, the joint probability density function of the target scenario can be evaluated and calculated to obtain the target probability of the target scenario. Finally, the target probability is used as the evaluation result to measure the generalization coverage of the scenario.

[0032] This application determines the joint probability density function of the target scene based on the probability distribution characteristics of the logical scene parameters, which can effectively describe the occurrence probability of the target scene. According to the preset target interval, the joint probability density function of the target scene and the distribution interval corresponding to the target scene, the target probability of the target scene is determined. The target probability of the target scene is obtained by discretization, which can quantify the coverage of the target scene in the overall probability distribution. The target probability is used as the evaluation result of the scene generalization coverage, so as to better evaluate the scene coverage.

[0033] Furthermore, in one embodiment, the logical scene parameters include dynamic parameters and / or state parameters. The dynamic parameters include one or more of the vehicle speed, the vehicle acceleration, and the target vehicle speed. The state parameters include lighting conditions and / or weather conditions.

[0034] In this exemplary embodiment, the target vehicle refers to a vehicle other than the vehicle itself, such as a tricycle; the logical scene state space includes dynamic parameters and / or state parameters. The dynamic parameters include one or more of the vehicle's speed, acceleration, and target vehicle speed, as well as following distance and lane-changing time; the state parameters include lighting conditions (such as cloudy, sunny backlight, sunny shadow, night, dusk) and / or weather conditions (such as sunny, rainy, snowy, foggy), as well as: a) Target object type, such as adults, children, SUVs, trucks, tricycles, and two-wheeled vehicles (bicycles, electric vehicles); b) The target vehicle's movement type, such as cutting in, cutting out, constant speed, acceleration, deceleration, emergency braking, left turn, right turn, U-turn; c) The relative movement of pedestrians and non-motorized vehicles to the main vehicle, such as moving in the same direction or crossing across; d) The target vehicle's position relative to the main vehicle, such as in front, to the left, or to the right; e) The relative positions of pedestrians and non-motorized vehicles to the main vehicle: in front, to the left, and to the right.

[0035] Specifically, as shown in Table 1, logical scenario parameters and their value ranges can be selected from Table 1 according to the test requirements to form a logical scenario parameter space corresponding to the functional scenario.

[0036] Table 1 Logical Scene Parameters and Their Value Ranges

[0037] It should be noted that, to facilitate subsequent probability estimation of the logical scenario parameters, the following assumptions are made: 1) The speed and acceleration distributions of the same vehicle follow a normal distribution and are correlated, i.e. ; 2) The vehicle is unrelated to the target object type, interaction method, and other state variables, i.e. ; 3) The vehicle in question, the target vehicle, weather conditions, and lighting conditions are independent of each other, i.e. .

[0038] Further, in one embodiment, determining the joint probability density function of the target corresponding to the target scene based on the probability distribution characteristics of the logical scene parameters includes: If the target scenario includes the vehicle speed and the vehicle acceleration, then the joint probability density function of the target corresponding to the target scenario is determined based on the first normal distribution corresponding to the vehicle speed and the second normal distribution corresponding to the vehicle acceleration. If the target scenario includes the target vehicle speed, lighting conditions, and weather conditions, then the joint probability density function corresponding to the target scenario is determined based on the first equal probability distribution corresponding to the target vehicle speed, the second equal probability distribution corresponding to the lighting conditions, and the third equal probability distribution corresponding to the weather conditions.

[0039] As an example, in the embodiments of this application, it can be understood that, in order to quantitatively evaluate the coverage of the target scene, the distribution characteristics of the target scene (including the vehicle speed and the vehicle acceleration) should first be determined. According to existing research results, the common distributions that vehicle speed and acceleration follow include: normal distribution, log-normal distribution, logistic distribution, Gamma distribution and Weibull distribution. Among them, the Gamma distribution and Weibull distribution have asymmetry and narrow applicability, and the Logistic distribution is more complex to calculate. In order to simplify the calculation, the embodiments of this application mainly use the normal distribution or the log-normal distribution. The distribution of vehicle speed and acceleration is shown in Table 2.

[0040] Table 2 Normal Distribution and Log-Normal Distribution

[0041] Understandably, referring to Table 2, assuming that both the vehicle speed x1 and the vehicle acceleration x2 follow a normal distribution, then the first normal distribution corresponding to the vehicle speed is: f(x1) = exp(- The second normal distribution corresponding to the acceleration of this vehicle is: f(x2) = exp(- Since the speed and acceleration of the same vehicle are negatively correlated, the joint probability density function of the target corresponding to the two is: f(x1, x2) = exp

[0042] in, = , , , This is the correlation coefficient between the vehicle's speed x1 and its acceleration x2. The specific values ​​can be determined according to actual needs and are not limited here; specifically, as shown in Table 3, the feature distribution (i.e. mean, standard deviation and correlation coefficient) of the logical scenario parameters can be calculated by sampling samples.

[0043] Table 3 Feature distribution of logical scene parameters

[0044] Where 30≤x1≤80, -6≤x2≤2, refer to Figure 2 The target joint probability density function p(x1, x2) is shown in the figure. The dark area represents the vehicle speed and vehicle acceleration with a lower probability in the area, while the light area represents the vehicle speed and vehicle acceleration with a higher probability in the area.

[0045] It should be noted that, in addition to the methods mentioned above, maximum likelihood estimation can also be used to calculate the probability distribution characteristics of the logic scenario parameters. The principle of maximum likelihood estimation is to fix the sample observations (x1, x2, ..., x...). N Select target parameters make L(x1, x2, ..., x N ; =maxL(x1, x2, ..., x N ; ) To obtain Related to sample observations, The maximum likelihood estimate of the parameter is given below, and the expression for the maximum likelihood function is as follows: =arg =arg

[0046] in, This refers to the i-th sample data in the target scene; For target parameters; Let be the maximum likelihood estimate of the target scenario; where is the maximum likelihood estimate of the vehicle speed. The maximum likelihood estimate of the vehicle's acceleration The calculation can be performed using the formulas described above. The principle and implementation process of maximum likelihood estimation are common knowledge in this field, and for the sake of simplicity, they will not be elaborated upon here. Below, referring to Table 4, the probability distribution characteristics of the vehicle speed and acceleration obtained through maximum likelihood estimation are given: Table 4 shows the probability distribution characteristics of the vehicle speed and acceleration obtained through maximum likelihood estimation.

[0047] Therefore, the first normal distribution corresponding to the vehicle speed at this time is: f(x1) = exp(- The second normal distribution corresponding to the acceleration of this vehicle is: f(x2) = exp(- It should be noted that, in addition to the vehicle's speed and acceleration, the probability distribution characteristics of following distance, collision time, and lane change time can be estimated using the kernel density estimation method, ultimately yielding the corresponding joint probability density function of the target. The principle and implementation process of the kernel density estimation method are common knowledge in this field, and will not be elaborated here for the sake of simplicity.

[0048] Understandably, the probability distribution characteristics of weather conditions are difficult to estimate, including: 1) Extreme heavy rainfall is a low-probability event, and the sample size is extremely limited; 2) The collection of sample data is limited by the years of observation data, and the time span is long, making it difficult to carry out in-depth research. As a result, it is still difficult to obtain refined results of the probability density of rainfall, snowfall, and fog, and the accuracy of its function fitting still needs to be improved. 3) The classification of rain, snow, and fog needs to be refined, as the current industry classification method is not suitable for autonomous driving testing; 4) Weather conditions are highly regional and lack universality. Especially for mass-produced automobile products that are required to be applicable to different regions, tests under different weather conditions should be given equal weight and should broadly cover the meteorological environmental characteristics of different regions in China. For coverage tests of autonomous vehicles, different light intensities and different target vehicle speeds should also be given equal weight.

[0049] Based on the above analysis, an equal probability density function is applied as the probability distribution function for weather conditions. For coverage testing of autonomous vehicles, different weather conditions, lighting conditions, and target vehicle speeds are all estimated jointly as equal-probability events. Assuming that x follows a uniform distribution on [a, b], the expression for the probability density function is as follows: f(x) =

[0050] Specifically, if the target scenario includes the target vehicle speed, lighting conditions, and weather conditions, the first equal probability distribution f(x3) corresponding to the target vehicle speed is... The second equal probability distribution f(x4) corresponding to the lighting conditions = The third equal probability distribution corresponding to the weather conditions is f(x5) = These three parameters are unrelated. Therefore, the joint probability density function of the target corresponding to the target scene is f(x3, x4, x5) = f(x3) f(x4) f(x5).

[0051] Similarly, coverage testing for autonomous driving scenarios requires joint probability estimation of different weather and lighting conditions as equal-probability events. Referring to Table 5, maximum likelihood estimation can also be performed for weather and lighting conditions. Table 5. Calculation of Maximum Probability Estimates for Weather and Illumination Conditions

[0052] It should be noted that, in addition to the specific scenarios mentioned above that include the vehicle's speed and acceleration, and those including the target vehicle's speed, lighting conditions, and weather conditions, the specific scenarios can also include the vehicle's speed, acceleration, target vehicle's speed, lighting conditions, and weather conditions. Therefore, for the 5-dimensional logical scenario parameter space Ω5 = (x1, x2, x3, x4, x5), assuming p(x1) is the probability density function of the main vehicle's speed, p(x2) is the probability density function of the main vehicle's acceleration, p(x3) is the probability density function of the target vehicle's speed, p(x4) is the probability density function of weather conditions, and p(x5) is the probability density function of lighting intensity, based on the above assumptions, the formula for calculating the joint probability density function of the target is: p(x)= p(x1,x2,x3,x4,x5)= p(x1,x2)p(x3)p(x4)p(x5) = exp

[0053] in, = p(x3), p(x4), p(x5) are constants. , , , , The specific value can be determined according to actual needs, and is not limited here.

[0054] Further, in one embodiment, determining the target probability of the target scene based on the preset target interval, the joint probability density function of the targets, and the distribution interval corresponding to the target scene includes: The area of ​​the probability interval is determined based on the preset first target interval corresponding to the vehicle speed and the preset second target interval corresponding to the vehicle acceleration; The target probability of the target scene is determined by the area of ​​the probability interval, the joint probability density function of the target, and the distribution interval corresponding to the target scene.

[0055] As an example, in this embodiment, the specific values ​​of the preset first target interval and the preset second target interval can be determined according to actual needs and are not limited here; the preset first target interval corresponding to the vehicle speed and the preset second target interval corresponding to the vehicle acceleration are used to divide the continuous vehicle speed and acceleration into multiple discrete intervals. Specifically, the area of ​​the probability interval is obtained by substituting the first target interval and the second target interval into the following calculation formula, which is:

[0056] In the formula, Let be the area of ​​the probability interval in the i-th row and j-th column; This is the preset first target interval corresponding to the vehicle speed. This is a preset second target interval corresponding to the vehicle's acceleration.

[0057] It should be noted that, in order to reduce the workload of calculations in actual engineering processes, 1, 2. The value should be small enough as needed; however, to more reasonably plan the sampling distribution for specific scenarios, a variable value can also be adopted. 1, Method 2, while also reducing computational load, selects 1, The range of variation for 2 is 1 1, 2 2; with 1 The increment of 1 follows a quadratic polynomial in one variable, such as... Figure 3 As shown, its expression is ,set up For vehicle speed The distribution range, where, a=

[0058] b=

[0059] c=

[0060] in, The same method can be used to calculate 2 2. Distribution method: In practical use, to simplify calculations, a baseline is first selected. 1, 2. Then follow calculate 1, The specific value of 2.

[0061] Specifically, the target joint probability density function describes the joint distribution characteristics between vehicle speed and acceleration. By performing a double integral of this function within their respective distribution intervals, the probability of each parameter occurring jointly within the area of ​​the probability interval is obtained, which is the target probability of the target scene. This provides a quantitative basis for evaluating the coverage of the target scene.

[0062] Further, in one embodiment, determining the target probability of the target scene based on the area of ​​the probability interval, the joint probability density function of the target, and the distribution interval corresponding to the target scene includes: The target probability of the target scene is calculated based on the area of ​​the probability interval, the joint probability density function of the target, and the distribution interval corresponding to the target scene. Specifically:

[0063]

[0064] In the formula, This is the vehicle's speed; For the acceleration of this vehicle; The area of ​​the probability interval; This refers to the distribution range corresponding to the vehicle's speed. This refers to the distribution range corresponding to the vehicle's acceleration. Let P{(x1, x2) be the joint probability density function of the objective. } represents the target probability of the target scenario.

[0065] As an example, in the embodiments of this application, reference is made to Figure 4 As shown, the area of ​​the probability interval Joint probability density function of the target Distribution range corresponding to the vehicle's acceleration Distribution range corresponding to the vehicle speed Substituting into the following formula, we obtain the target probability P{(x1, x2) of the target scene. }:

[0066] .

[0067] Specifically, since Gij is a rectangle, the area of ​​x1 and x2 falling within the rectangular region Gij and the target probability calculation for the specific scene in the i-th row and j-th column can be simplified as follows: P{(x 1,i x 2,j ) = P(x 1,i X x 1,i+1 ,x 2,j X x 2,j+1 ) =F(x 1,i+1 x 2,j+1 )-F(x 1,i+1 x 2,j )- F(x 1,i x 2,i+1 )+ F(x 1,i x 2,j ) Where F(x) 1,i+1 x 2,j+1 ), F(x 1,i+1 x 2,j ), F(x 1,i x 2,i+1 ), F(x 1,i x 2,j This is equivalent to finding the double integral of the joint probability density function, which is common knowledge in this field and will not be elaborated here for the sake of simplicity.

[0068] It should be noted that a probability threshold can be set according to the testing requirements, i.e., a limit can be imposed. (1 <i<m,1<j< To improve testing efficiency, the minimum value of sum{Gij(1) is selected, and values ​​greater than the probability threshold are filtered out. <i<m,1<j< )} is the generalized form (1 <i<m,1<j< The coverage of specific scenarios within the scope.

[0069] Further, in one embodiment, determining the target probability of the target scene based on the preset target interval, the joint probability density function of the targets, and the distribution interval corresponding to the target scene includes: The volume of the probability interval is determined based on the preset third target interval corresponding to the target vehicle speed, the preset fourth target interval corresponding to the lighting conditions, and the preset fifth target interval corresponding to the weather conditions; The target probability of the target scene is determined based on the volume of the probability interval, the joint probability density function of the target, and the distribution interval corresponding to the target scene.

[0070] In this exemplary embodiment, the specific values ​​of the preset third target interval, preset fourth target interval, and preset fifth target interval can be determined according to actual needs and are not limited here; the preset third target interval corresponding to the target vehicle speed, the preset fourth target interval corresponding to the lighting conditions, and the preset fifth target interval corresponding to the weather conditions are used to divide the continuous target vehicle speed, lighting conditions, and weather conditions into multiple discrete intervals; specifically, the preset third target interval, the preset fourth target interval, and the preset fifth target interval are substituted into the following calculation formula to obtain the probability interval volume, the calculation formula being:

[0071] In the formula, The preset third target interval corresponding to the target vehicle speed; This is a preset fourth target interval corresponding to the lighting conditions; The preset fifth target interval corresponding to weather conditions; Let x be the volume of the probability interval in the x-th row, y-th column, and z-th layer.

[0072] Specifically, the joint probability density function describes the joint distribution characteristics of the target vehicle speed, lighting conditions, and weather conditions. By performing a triple integral of this function within their respective distribution intervals, the probability of each parameter occurring jointly within the probability interval volume is obtained, which is the target probability of the target scene. This provides a quantitative basis for evaluating the coverage of the target scene.

[0073] Further, in one embodiment, determining the target probability of the target scene based on the probability interval volume, the target joint probability density function, and the distribution interval corresponding to the target scene includes: The target probability of the target scene is calculated based on the probability interval volume, the joint probability density function of the target, and the distribution interval corresponding to the target scene. Specifically:

[0074]

[0075] In the formula, x3 is the target vehicle speed; x4 is the lighting conditions; and x5 is the weather conditions. Let the volume be the probability interval. Let the joint probability density function be the objective. The distribution range corresponding to the target vehicle speed; The distribution range corresponding to the lighting conditions; Let P{(x3, x4, x5)} represent the distribution intervals corresponding to weather conditions. } represents the target probability of the target scenario.

[0076] As an example, in the embodiments of this application, the probability interval volume is... Joint probability density function of the target Distribution range corresponding to the target vehicle speed Distribution range corresponding to illumination conditions Distribution range corresponding to weather conditions Substituting into the following formula, we obtain the target probability P{(x3, x4, x5) for the target scene. }:

[0077] .

[0078] Secondly, embodiments of this application also provide a system for evaluating scene generalization coverage, the system comprising: The first processing module is used to determine the joint probability density function of the target scene corresponding to the target scene based on the probability distribution characteristics of the logical scene parameters. The second processing module is used to determine the target probability of the target scene based on the preset target interval, the target joint probability density function and the distribution interval corresponding to the target scene, and use the target probability as the evaluation result of the scene generalization coverage.

[0079] Furthermore, in one embodiment, the first processing module is specifically used for: The logical scene parameters include dynamic parameters and / or state parameters. The dynamic parameters include one or more of the vehicle speed, the vehicle acceleration, and the target vehicle speed. The state parameters include lighting conditions and / or weather conditions.

[0080] Furthermore, in one embodiment, the first processing module is specifically used for: If the target scenario includes the vehicle speed and the vehicle acceleration, then the joint probability density function of the target corresponding to the target scenario is determined based on the first normal distribution corresponding to the vehicle speed and the second normal distribution corresponding to the vehicle acceleration. If the target scenario includes the target vehicle speed, lighting conditions, and weather conditions, then the joint probability density function corresponding to the target scenario is determined based on the first equal probability distribution corresponding to the target vehicle speed, the second equal probability distribution corresponding to the lighting conditions, and the third equal probability distribution corresponding to the weather conditions.

[0081] Furthermore, in one embodiment, the second processing module is specifically used for: The area of ​​the probability interval is determined based on the preset first target interval corresponding to the vehicle speed and the preset second target interval corresponding to the vehicle acceleration; The target probability of the target scene is determined by the area of ​​the probability interval, the joint probability density function of the target, and the distribution interval corresponding to the target scene.

[0082] Furthermore, in one embodiment, the second processing module is specifically used for: The target probability of the target scene is calculated based on the area of ​​the probability interval, the joint probability density function of the target, and the distribution interval corresponding to the target scene. Specifically:

[0083]

[0084] In the formula, This is the vehicle's speed; For the acceleration of this vehicle; Let the joint probability density function be the objective. The area of ​​the probability interval; This refers to the distribution range corresponding to the vehicle's speed. This refers to the distribution range corresponding to the vehicle's acceleration. P{(x1, x2)} } represents the target probability of the target scenario.

[0085] Furthermore, in one embodiment, the second processing module is specifically used for: The volume of the probability interval is determined based on the preset third target interval corresponding to the target vehicle speed, the preset fourth target interval corresponding to the lighting conditions, and the preset fifth target interval corresponding to the weather conditions; The target probability of the target scene is determined based on the volume of the probability interval, the joint probability density function of the target, and the distribution interval corresponding to the target scene.

[0086] Furthermore, in one embodiment, the second processing module is specifically used for: The target probability of the target scene is calculated based on the probability interval volume, the joint probability density function of the target, and the distribution interval corresponding to the target scene. Specifically:

[0087] In the formula, x3 is the target vehicle speed; x4 is the lighting conditions; and x5 is the weather conditions. Let the volume be the probability interval. Let the joint probability density function be the objective. The distribution range corresponding to the target vehicle speed; The distribution range corresponding to the lighting conditions; Let P{(x3, x4, x5)} represent the distribution intervals corresponding to weather conditions. } represents the target probability of the target scenario.

[0088] This application determines the joint probability density function of the target scene based on the probability distribution characteristics of the logical scene parameters, which can effectively describe the occurrence probability of the target scene. According to the preset target interval, the joint probability density function of the target scene and the distribution interval corresponding to the target scene, the target probability of the target scene is determined. The target probability of the target scene is obtained by discretization, which can quantify the coverage of the target scene in the overall probability distribution. The target probability is used as the evaluation result of the scene generalization coverage, so as to better evaluate the scene coverage.

[0089] The functions of each module in the above-mentioned scenario generalization coverage evaluation system correspond to the steps in the above-mentioned scenario generalization coverage evaluation method embodiment, and their functions and implementation processes will not be described in detail here.

[0090] Thirdly, embodiments of this application provide a scene generalization coverage evaluation device, which can be a personal computer (PC), laptop computer, server or other device with data processing capabilities.

[0091] Reference Figure 5 , Figure 5 This is a schematic diagram of the hardware structure of the scene generalization coverage evaluation device involved in the embodiments of this application. In the embodiments of this application, the scene generalization coverage evaluation device may include a processor, a memory, a communication interface, and a communication bus.

[0092] The communication bus can be of any type and is used to interconnect the processor, memory, and communication interface.

[0093] The communication interface includes input / output (I / O) interfaces, physical interfaces, and logical interfaces used for interconnecting internal components of the evaluation device to achieve scene generalization coverage, as well as interfaces used for interconnecting the evaluation device with other devices (such as other computing devices or user equipment). Physical interfaces can be Ethernet interfaces, fiber optic interfaces, ATM interfaces, etc.; user equipment can be displays, keyboards, etc.

[0094] Memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0095] The processor can be a general-purpose processor, which can call the scene generalization coverage evaluation program stored in memory and execute the scene generalization coverage evaluation method provided in the embodiments of this application. For example, the general-purpose processor can be a central processing unit (CPU). The method executed when the scene generalization coverage evaluation program is called can refer to the various embodiments of the scene generalization coverage evaluation method of this application, and will not be repeated here.

[0096] Those skilled in the art will understand that Figure 5 The hardware structure shown does not constitute a limitation of this application and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0097] Fourthly, embodiments of this application also provide a readable storage medium.

[0098] The present application stores a scene generalization coverage evaluation program on a readable storage medium, wherein when the scene generalization coverage evaluation program is executed by a processor, the scene generalization coverage evaluation method described above is implemented.

[0099] The method implemented when the scene generalization coverage evaluation procedure is executed can refer to the various embodiments of the scene generalization coverage evaluation method of this application, and will not be repeated here.

[0100] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus. The terms "first," "second," and "third," etc., are used to distinguish different objects, etc., and do not indicate a sequence, nor do they limit "first," "second," and "third" to different types.

[0101] In the description of the embodiments of this application, terms such as "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a concrete manner.

[0102] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.

[0103] In some processes described in the embodiments of this application, multiple operations or steps are included in a specific order. However, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of this application, or they may be executed in parallel. The sequence number of the operation is only used to distinguish different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.

[0104] It should be noted that the sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0105] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device to execute the methods described in the various embodiments of this application.

[0106] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for evaluating scene generalization coverage, characterized in that, The scene generalization coverage evaluation method comprises the following steps: determining a target joint probability density function corresponding to the target scene based on probability distribution characteristics of logical scene parameters; determining a target probability of the target scene according to a preset target interval, the target joint probability density function and a distribution interval corresponding to the target scene, and taking the target probability as an evaluation result of the scene generalization coverage.

2. The method of claim 1, wherein, The logical scene parameters comprise dynamic parameters and / or state parameters, the dynamic parameters comprise one or more of a host vehicle speed, a host vehicle acceleration and a target vehicle speed, and the state parameters comprise an illumination condition and / or a weather condition.

3. The method of claim 2, wherein, The determination of the target joint probability density function corresponding to the target scene based on the probability distribution characteristics of the logical scene parameters comprises the following steps: if the target scene comprises the host vehicle speed and the host vehicle acceleration, determining the target joint probability density function corresponding to the target scene based on a first normal distribution corresponding to the host vehicle speed and a second normal distribution corresponding to the host vehicle acceleration; if the target scene comprises the target vehicle speed, the illumination condition and the weather condition, determining the target joint probability density function corresponding to the target scene based on a first equiprobable distribution corresponding to the target vehicle speed, a second equiprobable distribution corresponding to the illumination condition and a third equiprobable distribution corresponding to the weather condition.

4. The method of claim 3, wherein, The determination of the target probability of the target scene according to the preset target interval, the target joint probability density function and the distribution interval corresponding to the target scene comprises the following steps: determining a probability interval area based on a preset first target interval corresponding to the host vehicle speed and a preset second target interval corresponding to the host vehicle acceleration; determining the target probability of the target scene according to the probability interval area, the target joint probability density function and the distribution interval corresponding to the target scene.

5. The method of claim 4, wherein, The determination of the target probability of the target scene according to the probability interval area, the target joint probability density function and the distribution interval corresponding to the target scene comprises the following steps: calculating the target probability of the target scene according to the probability interval area, the target joint probability density function and the distribution interval corresponding to the target scene, and specifically, In the formula, is the vehicle speed of the host vehicle; is the acceleration of the host vehicle; is the area of the probability interval; is the distribution interval corresponding to the vehicle speed of the host vehicle; is the distribution interval corresponding to the acceleration of the host vehicle; is the target joint probability density function; P{(x1, x2 } is the target probability of the target scenario.

6. The method of claim 3, wherein, The determination of the target probability of the target scene according to the preset target interval, the target joint probability density function and the distribution interval corresponding to the target scene comprises the following steps: determining a probability interval volume based on a preset third target interval corresponding to the target vehicle speed, a preset fourth target interval corresponding to the illumination condition and a preset fifth target interval corresponding to the weather condition; determining the target probability of the target scene according to the probability interval volume, the target joint probability density function and the distribution interval corresponding to the target scene.

7. The method of claim 6, wherein, The determination of the target probability of the target scene according to the probability interval volume, the target joint probability density function and the distribution interval corresponding to the target scene comprises the following steps: calculating the target probability of the target scene according to the probability interval volume, the target joint probability density function and the distribution interval corresponding to the target scene, and specifically, wherein, is the target vehicle speed; is the lighting condition; is the weather condition; is the probability interval volume; is the target joint probability density function; is the distribution interval corresponding to the target vehicle speed; is the distribution interval corresponding to the lighting condition; is the distribution interval corresponding to the weather condition; P{(x3, x4, x5) is the target probability of the target scene.

8. A system for evaluating scene generalization coverage, the system comprising: The scene generalization coverage evaluation system comprises the following steps: a first processing module configured to determine a target joint probability density function corresponding to a target scene based on probability distribution characteristics of logical scene parameters; The second processing module is configured to determine a target probability of the target scene according to a preset target interval, a target joint probability density function, and a distribution interval corresponding to the target scene, and take the target probability as an evaluation result of the scene generalization coverage.

9. An evaluation device for the coverage of a scenario generalization, characterized by The scene generalization coverage evaluation device includes a processor, a memory, and a scene generalization coverage evaluation program stored in the memory and executable by the processor. When the scene generalization coverage evaluation program is executed by the processor, the steps of the scene generalization coverage evaluation method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a scene generalization coverage evaluation program. When the scene generalization coverage evaluation program is executed by the processor, the steps of the scene generalization coverage evaluation method according to any one of claims 1 to 7 are implemented.

Images