Method, device and equipment for determining performance of automatic emergency braking system of automobile and medium

By constructing multiple test scenarios and quantitative evaluation based on real accident case data, combining simulation tests and gray correlation analysis, the problem of incomplete and inaccurate performance testing of automobile automatic emergency braking systems in the existing technology is solved, and the performance determination is improved. Comprehensiveness and accuracy.

CN118939462APending Publication Date: 2024-11-12XIHUA UNIV
View PDF 0 Cites 4 Cited by

Patent Information

Application Number
CN202410996040.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-24
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The performance test of existing automobile automatic emergency braking systems depends on a single working condition or a single dimension artificially established test scenario, which cannot fully reflect the actual road traffic environment, and the evaluation of the importance of the test scenario lacks a quantitative basis, resulting in a comprehensive performance determination. Low in sensitivity and accuracy.

Method used

Test cases are generated by constructing multiple test scenarios based on real accident case data, and quantitative evaluation is performed based on the risk and complexity. Simulation testing and gray correlation analysis algorithms are used to determine the performance evaluation score of the car's automatic emergency braking system.

Benefits of technology

It improves the comprehensiveness and accuracy of the performance determination of the automobile automatic emergency braking system, reduces the dependence on the subjective scores of the assessors, and ensures the objectivity and accuracy of the evaluation results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118939462A_ABST
    Figure CN118939462A_ABST
Patent Text Reader

Abstract

The invention provides an automobile automatic emergency braking system performance determination method and device, equipment and a medium. The method comprises the following steps: acquiring a plurality of test cases generated based on real accident case data and a first weight coefficient determined based on the complexity and risk degree of each test case; performing a simulation test on the to-be-tested automobile by adopting each test case, obtaining target test data of the to-be-tested automobile under each test case, and determining a first evaluation score corresponding to each test case of the automatic emergency braking system based on the target test data and a gray correlation analysis algorithm; determining a second evaluation score of the system based on each first weight coefficient and the first evaluation score; and determining a performance evaluation result of the system based on the second evaluation score. The test cases are generated for testing through the plurality of test scenes constructed based on the real accident case data, and each test case is quantitatively evaluated based on the risk degree and the complexity, so that the comprehensiveness and the accuracy of determining the performance of the automatic emergency braking system of the automobile are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of automobile testing technology, and in particular to a method, device, equipment and medium for determining the performance of an automobile automatic emergency braking system. Background Art

[0002] As an important component of self-driving cars, the Autonomous Emergency Braking (AEB) system can prevent traffic accidents and reduce the severity of accidents to a certain extent. Testing and evaluating the performance of the AEB system is of great significance to ensuring the safety of self-driving cars.

[0003] The performance determination of existing automobile automatic emergency braking systems mostly relies on road testing using a few standardized test scenarios artificially established for a single working condition or a single dimension as test cases, which cannot fully reflect the operating conditions of the automobile in the actual road traffic environment; at the same time, the evaluation of the importance of the test scenario for each test case depends entirely on the subjective scoring of the evaluator, and there is a lack of quantitative evaluation basis for the importance of the test scenario of the test case. The comprehensiveness and accuracy of the performance determination of the automobile automatic emergency braking system are low. Summary of the invention

[0004] In view of this, the embodiments of the present application at least provide a method, device, equipment and medium for determining the performance of an automobile automatic emergency braking system, generate test cases for testing through multiple test scenarios constructed based on real accident case data, and quantitatively evaluate each test case based on the degree of danger and complexity, thereby improving the comprehensiveness and accuracy of determining the performance of an automobile automatic emergency braking system.

[0005] This application mainly includes the following aspects:

[0006] In a first aspect, an embodiment of the present application provides a method for determining performance of an automatic emergency braking system of an automobile, the method comprising:

[0007] Obtain multiple test cases and a first weight coefficient for each test case; any of the test cases is a test case generated based on multiple target test scenarios; any of the target test scenarios is a test scenario determined based on multiple candidate test scenarios generated based on multiple real accident case data; the first weight coefficient for any of the test cases is a weight coefficient determined based on the complexity and riskiness of the test case;

[0008] Using each test case to perform simulation test on the vehicle to be tested, obtaining target test data of the vehicle to be tested under each test case, and determining the first evaluation score of the automatic emergency braking system of the vehicle to be tested corresponding to each test case based on the target test data and a grey correlation analysis algorithm; the target test data is the test data corresponding to each performance evaluation index of the automatic emergency braking system;

[0009] Determining a second evaluation score of the automatic emergency braking system of the vehicle to be tested based on the first weight coefficient and the first evaluation score of each test case;

[0010] Based on the second evaluation score of the automatic emergency braking system, a performance evaluation result of the automatic emergency braking system of the vehicle to be tested is determined.

[0011] In a second aspect, an embodiment of the present application further provides a device for determining the performance of an automatic emergency braking system of an automobile, the device for determining the performance of an automatic emergency braking system of an automobile comprising:

[0012] An acquisition module, used to acquire multiple test cases and a first weight coefficient of each test case; any of the test cases is a test case generated based on multiple target test scenarios; any of the target test scenarios is a test scenario determined based on multiple candidate test scenarios generated based on multiple real accident case data; the first weight coefficient of any of the test cases is a weight coefficient determined based on the complexity and risk of the test case;

[0013] A first determination module is used to perform simulation tests on the vehicle to be tested using various test cases, obtain target test data of the vehicle to be tested under various test cases, and determine a first evaluation score corresponding to each test case of the automatic emergency braking system of the vehicle to be tested based on the target test data and a grey correlation analysis algorithm; the target test data is test data corresponding to each performance evaluation index of the automatic emergency braking system;

[0014] A second determination module, configured to determine a second evaluation score of the automatic emergency braking system of the vehicle to be tested based on the first weight coefficient and the first evaluation score of each test case;

[0015] The third determination module is used to determine the performance evaluation result of the automatic emergency braking system of the vehicle to be tested based on the second evaluation score of the automatic emergency braking system.

[0016] In a third aspect, an embodiment of the present application further provides an electronic device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor to execute the steps of the method for determining the performance of an automatic emergency braking system of an automobile as described above.

[0017] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for determining the performance of an automatic emergency braking system of an automobile as described above are executed.

[0018] The embodiments of the present application provide a method, device, equipment and medium for determining the performance of an automatic emergency braking system of an automobile, wherein the method comprises: obtaining multiple test cases and a first weight coefficient of each test case; any test case is a test case generated based on multiple target test scenarios; any target test scenario is a test scenario determined based on multiple candidate test scenarios generated based on multiple real accident case data; the first weight coefficient of any test case is a weight coefficient determined based on the complexity and danger of the test case; using each test case to perform a simulation test on the automobile to be tested, obtaining target test data of the automobile to be tested under each test case, and determining the first evaluation score of the automatic emergency braking system of the automobile to be tested corresponding to each test case based on the target test data and a grey correlation analysis algorithm; the target test data is the test data corresponding to each performance evaluation indicator of the automatic emergency braking system; based on the first weight coefficient and the first evaluation score of each test case, determining the second evaluation score of the automatic emergency braking system of the automobile to be tested; based on the second evaluation score of the automatic emergency braking system, determining the performance evaluation result of the automatic tight-tie braking system of the automobile to be tested. In this way, test cases are generated based on multiple test scenarios constructed based on real accident case data, and each test case is quantitatively evaluated based on the degree of danger and complexity, which improves the comprehensiveness and accuracy of the performance determination of the automobile automatic emergency braking system.

[0019] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 A flow chart of a method for determining performance of an automatic emergency braking system of an automobile provided in an embodiment of the present application is shown;

[0022] Figure 2 One of the functional module diagrams of a vehicle automatic emergency braking system performance determination device provided by an embodiment of the present application is shown;

[0023] Figure 3 The second functional module diagram of a device for determining the performance of an automatic emergency braking system of an automobile provided in an embodiment of the present application is shown;

[0024] Figure 4 A schematic structural diagram of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0026] To facilitate the understanding of the present application, the technical solution provided by the present application is described in detail below in conjunction with specific embodiments.

[0027] The embodiments of the present application provide a method, apparatus, device and medium for determining the performance of an automobile automatic emergency braking system. Test cases are generated by using multiple test scenarios constructed based on real accident case data for testing, and each test case is quantitatively evaluated based on the degree of danger and complexity, thereby improving the comprehensiveness and accuracy of the performance determination of the automobile automatic emergency braking system.

[0028] See also Figure 1 , Figure 1 This is a flow chart of a method for determining the performance of an automatic emergency braking system of an automobile provided in an embodiment of the present application. Figure 1 As shown, the method for determining the performance of an automatic emergency braking system of an automobile provided in an embodiment of the present application comprises the following steps:

[0029] S101, obtain multiple test cases and a first weight coefficient for each test case; any test case is a test case generated based on multiple target test scenarios; any target test scenario is a test scenario determined based on multiple candidate test scenarios generated based on multiple real accident case data; the first weight coefficient of any test case is a weight coefficient determined based on the complexity and criticality of the test case.

[0030] In this step, since the existing test cases are statically selected and generated based on a few fixed scenes constructed manually, they lack the ability of dynamic generation and real-time update, and cannot fully cover all possible traffic environments and emergency situations, and the determination of the performance of the automobile AEB system is not comprehensive and accurate enough. For example, some test scenes with complex and dangerous road conditions rarely appear on actual roads, but once they appear, accidents are very likely to occur. These test scenes account for a high proportion of accident scenes, but due to their complex road conditions, these uncommon but dangerous test scenes are often ignored when generating test cases based on artificially constructed road scenes, resulting in incomplete and inaccurate evaluation results of the performance determination of the automobile AEB system. Therefore, compared with generating test cases based on artificially constructed test scenes, it is easier to extract dangerous working conditions based on real accident case data. In order to improve the comprehensiveness and accuracy of the performance determination of the automobile AEB system, in the embodiment of the present application, test scenes are extracted based on a large amount of real accident case data to generate test cases to test the ability of the autonomous driving car to deal with dangers.

[0031] Specifically, any test case among the multiple test cases obtained is a test case generated based on multiple target test scenarios; any target test scenario is a test scenario determined by multiple candidate test scenarios generated based on multiple real accident case data. The embodiment of the present application introduces multi-dimensional test scenarios generated based on a large amount of real accident case data, and these scenarios cover a wide range of traffic environments from low complexity to high complexity. Each scenario is carefully constructed based on real accident cases to ensure that the test cases can fully simulate and reflect the performance of the car's AEB under various driving conditions. In this way, not only can the response of the AEB system under normal circumstances be tested, but this scenario-driven testing method makes the evaluation process more in line with actual application needs, and can provide more in-depth and scientific data support for vehicle safety, thereby improving the comprehensiveness of the performance determination of the vehicle's AEB system.

[0032] At the same time, the complexity and criticality of the test cases can also be quantified, and the first weight coefficient of the test cases can be determined based on the complexity and criticality of the test cases. Then, quantitative standards can be introduced in the evaluation process of the test case weights to ensure that the test scenarios corresponding to each test case are reasonably evaluated according to their actual importance and risks, reduce reliance on the subjective scores of evaluators, ensure the objectivity and accuracy of the evaluation results, and thus improve the accuracy of the performance determination of the automotive AEB system.

[0033] S102, using each test case to perform simulation test on the vehicle to be tested, obtaining target test data of the vehicle to be tested under each test case, and determining the first evaluation score of the automatic emergency braking system of the vehicle to be tested corresponding to each test case based on the target test data and a grey correlation analysis algorithm; the target test data is the test data corresponding to each performance evaluation indicator of the automatic emergency braking system.

[0034] In this step, various test cases are used to perform simulation tests on the vehicle to be tested, and target test data of the vehicle to be tested under various test cases is obtained. Among them, the target test data is the test data corresponding to each performance evaluation indicator of the automatic emergency braking system. Specifically, when evaluating the performance of the AEB system, the evaluation indicators should be selected based on its functional purpose. Under the premise of ensuring safety, performance evaluation indicators are selected to comprehensively evaluate the performance of AEB by considering the traffic situation and the feelings of the driver and passengers. In the embodiment of the present application, at least one of the four performance evaluation indicators of the speed reduction of the vehicle to be tested, the relative distance of braking to stop, the AEB triggering moment, and the maximum braking deceleration is selected as the performance evaluation indicator of the automatic emergency braking system.

[0035] Specifically, the speed reduction W of the vehicle to be measured is shown in the following formula:

[0036]

[0037] Where v i Indicates the initial test speed of the vehicle to be tested; v c It indicates the collision speed of the vehicle to be tested. If no collision occurs, the value is 0 or the target vehicle speed. When the speed reduction of the vehicle to be tested is 100%, it means there is no collision.

[0038] The relative braking-stopping distance refers to the distance between the vehicle under test and the target vehicle after the vehicle under test detects a risk of collision with the target vehicle and brakes. If no collision occurs, the larger the value, the earlier the braking time and the better the safety. However, considering the actual driving conditions and traffic capacity, it is usually appropriate to be between 0.5 and 3.0 meters. If the vehicle fails to avoid a collision, the indicator will be directly scored zero.

[0039] The AEB triggering time refers to the moment when the AEB system automatically brakes when the system braking delay time is less than the driver's reaction time and the driver does not receive a collision warning. The triggering time of different AEB control algorithms can be obtained by analyzing the vehicle's speed, relative distance, etc.

[0040] The maximum braking deceleration is an important reflection of the braking performance of AEB. Under emergency conditions, a larger braking deceleration value can better ensure the safety of the driver and passengers, but if the deceleration is too large, it will reduce the driver's comfort and may even cause secondary injuries. Depending on the test speed, the peak braking deceleration is mostly in the range of 4 to 10 m / s2, which is related to the AEB triggering time, the deceleration issued by the system, and the actual braking effect.

[0041] After obtaining the target test data of the vehicle to be tested under each test case, the first evaluation score of the automatic emergency braking system of the vehicle to be tested corresponding to each test case is determined based on the target test data and the grey relational analysis algorithm. The grey relational analysis algorithm is used to judge whether the connection between different sequences is close according to the similarity of the geometric shapes of the sequence curves. In the embodiment of the present application, it is used to determine the performance of the AEB system of the vehicle to be tested based on the geometric similarity between the evaluation sequence of the target test data and the reference sequence, and then determine the first evaluation score of the automatic emergency braking system of the vehicle to be tested corresponding to each test case.

[0042] S103: Determine a second evaluation score of the automatic emergency braking system of the vehicle to be tested based on the first weight coefficient and the first evaluation score of each test case.

[0043] In the embodiment of the present application, the first evaluation score corresponding to each test case is integrated with the weight of the test scenario evaluation of each test case to perform a comprehensive evaluation of AEB, and the second evaluation score of the automatic emergency braking system of the vehicle to be tested is determined based on the first weight coefficient and the first evaluation score of each test case. Specifically, the second evaluation score is shown in the following formula:

[0044] Where M i is the first evaluation score of the i-th test case; α i is the first weight coefficient of the i-th test case.

[0045] Among them, the second evaluation score, as the final evaluation score of the automobile AEB system performance, is a comprehensive indicator that integrates the evaluation score of each test case, the weight of the performance evaluation indicator, and the weight of the test scenario of the test case. It reflects the overall performance in all test scenarios and provides a global and holistic perspective to reflect the performance of the automobile AEB system.

[0046] S104, determining a performance evaluation result of the automatic emergency braking system of the vehicle to be tested based on the second evaluation score of the automatic emergency braking system.

[0047] In this step, the value of the second evaluation score of the AEB system is substituted into the evaluation criteria for the performance evaluation of the AEB system, the score range of the second evaluation score is found, and the evaluation level corresponding to the score range of the second evaluation score is determined as the performance evaluation result of the automatic braking system of the vehicle to be tested.

[0048] In the embodiment of the present application, a 10-point system is used as the evaluation standard, and four evaluation levels, namely excellent, good, average, and poor, are corresponding to four score ranges from high to low as the performance evaluation results of the automatic braking system of the vehicle to be tested.

[0049] Specifically, the evaluation level is excellent, and the corresponding score range is (8.75, 10]: AEB systems of this evaluation level can perform their functions excellently under various weather conditions (sunny, rainy, snowy), different visual field obstructions, and different speed conditions of the test vehicle and the target vehicle. AEB systems of this evaluation level can not only accurately judge the motion state of the vehicle in front and avoid collisions in rear-end collision conditions, but also make correct responses in time according to the relative positions of the test vehicle and the target vehicle in complex intersection conditions. More importantly, in head-on collision conditions, the AEB system can quickly judge and take evasive actions to avoid or mitigate accidents. This means that in various complex driving environments, such as highways, urban roads, rural roads, and in the face of various emergencies, such as pedestrians suddenly crossing and other vehicles suddenly braking, the AEB system can respond accurately and quickly, greatly improving driving safety. For AEB systems of this evaluation level, there are fewer safety hazards in actual driving.

[0050] Good rating, corresponding score range [6.25, 8.75]: AEB systems with this rating can effectively avoid or mitigate accidents in most test scenarios, such as sunny and rainy days, and when the field of vision is generally blocked. However, in some specific scenarios, such as snowy days or when the field of vision is severely blocked, or when the speed of the test vehicle and the target vehicle is high and the relative position is complex, the AEB system may not provide the best performance. Despite this, the overall performance of AEB systems with this rating can still meet most driving needs and greatly improve driving safety.

[0051] The evaluation level is average, with a corresponding score range of (3.75, 6.25]: AEB systems with this evaluation level can effectively avoid or mitigate accidents in more than half of the test scenarios, but may not perform well in some scenarios, such as head-on collisions or when the field of view is severely blocked, or in rainy and snowy weather, or when the speed of the test vehicle and the target vehicle is high and the relative positions are complex. The AEB system may not be able to respond promptly and effectively. Although AEB systems with this evaluation level can provide certain safety guarantees, there is still much room for improvement.

[0052] Poor rating, corresponding score range [0,3.75]: AEB systems with this rating cannot provide adequate protection in most test scenarios, such as in rainy and snowy weather, when the field of vision is severely obstructed, or in head-on collision conditions and complex intersection conditions. For example, when faced with a pedestrian who suddenly rushes out, or a car in front that suddenly brakes, or when the speed of the test car and the target car is high and the relative position is complex, the AEB system cannot respond correctly in time, thereby increasing the risk of an accident. This means that in actual driving, whether in ordinary road environments or in more complex road environments, such as intersections and around schools, the AEB system cannot provide adequate protection. For AEB systems with this rating, there are greater safety risks in actual driving, and further optimization and improvement is needed.

[0053] Further, multiple test cases are determined according to the following steps:

[0054] Step a1, obtaining a candidate test scenario template; the candidate test scenario template includes multiple candidate dimensions, each candidate dimension corresponds to multiple candidate scenario elements.

[0055] In the embodiment of the present application, a structured description is performed on the test scenarios corresponding to multiple real accident case data according to a pre-configured candidate test scenario template, and the candidate test scenario template includes multiple candidate dimensions, and each candidate dimension corresponds to multiple candidate scenario elements. Specifically, the multiple candidate dimensions and each candidate dimension corresponds to multiple candidate scenario elements as shown in Table 1.

[0056] Table 1 Candidate scenario dimensions and candidate scenario elements

[0057]

[0058] Step a2, obtaining multiple real accident case data, extracting candidate dimensions and candidate scenario elements in the candidate test scenario template from each real accident case, and determining the candidate test scenario corresponding to each real accident case.

[0059] In the embodiment of the present application, the multiple real accident case data are real accident case data obtained from the database of the National Accident Investigation System (NAIS). After obtaining the multiple real accident case data, according to the candidate scenario dimensions and candidate scenario elements of the candidate test scenario template as shown in Table 1, the candidate dimensions and candidate scenario elements in the candidate test scenario template are extracted from each real accident case to determine the candidate test scenario corresponding to each real accident case.

[0060] Step a3, classify and sort the candidate test scenarios of each real accident case according to the category of functional scenarios, select the candidate test scenarios ranked in the front preset positions, and determine multiple target test scenarios; the functional scenario is a scenario described based on the preset candidate scenario elements.

[0061] In this step, the candidate test scenarios of each real accident case are divided according to the category of functional scenarios. In the embodiment of the present application, the functional scenario is a scenario described based on the four scenario elements of the motion state of the vehicle to be tested, the motion state of the target vehicle, the relative position, and the road type. The candidate test scenarios with the same four scenario element states of the functional scenario in the candidate test scenarios of each real accident case are divided into one category. For example, a typical functional scenario can be described as follows:

[0062] Road type: straight road section;

[0063] The movement state of the car to be tested: straight;

[0064] Target car motion state: straight;

[0065] Relative position: same direction.

[0066] In this functional scenario, the vehicle under test and the target vehicle are both driving straight on the same straight road section, and the target vehicle is in the same lane in front of the vehicle under test. This scenario usually occurs on main roads in cities or suburbs, and may involve high-speed or medium-speed driving conditions. For the AEB system, it is necessary to accurately identify the speed and position of the target vehicle and respond in time to avoid or mitigate the collision.

[0067] In an embodiment of the present application, each candidate test scenario is classified and counted according to the category of the functional scenario, and the candidate test scenarios corresponding to the functional scenarios ranked in the front preset positions in frequency are selected, for example, the candidate test scenarios corresponding to the functional scenarios ranked in the top 10, and multiple target test scenarios are determined.

[0068] Step a4: Generate multiple test cases based on each target test scenario; the state of at least one scenario element of the target test scenario corresponding to each test case is different.

[0069] In this step, in the pre-set functional scenarios, the internal structure and dynamic behavior of each functional scenario are analyzed, and specific parameter values ​​are sampled for all candidate scenario elements, so as to obtain different numbers of specific test scenarios in each functional scenario. Through the sampling strategy, the comprehensive coverage of candidate scenario elements and parameter combinations is ensured, and multiple test cases are generated.

[0070] Further, the first weight coefficient of each test case is determined according to the following steps:

[0071] Step b1, based on multiple real accident case data, determine the complexity evaluation score and the risk evaluation score of each test case;

[0072] Step b2, based on the complexity evaluation score, risk evaluation score and entropy method of each test case, determine the complexity weight coefficient and risk weight coefficient of each test case.

[0073] In this step, the entropy method is used to determine the complexity weight coefficient and the risk weight coefficient of each test case. The entropy method is based on the principle of information entropy and objectively assigns weights according to the degree of variation of each indicator. If the degree of variation of an indicator is greater and the amount of information provided is more, the smaller the information entropy value is, the greater the weight is, and vice versa.

[0074] Specifically, the entropy method includes the following steps:

[0075] Step h1: data standardization processing.

[0076] Assume there are n test cases and m evaluation indicators, which represents the original data of the i-th test case under the j-th indicator. The original data is X ij , i = 1, ..., n; j = 1, ..., m; before calculating the weight coefficient, the original data needs to be standardized to remove the indicator dimension.

[0077] (x ij is a positive indicator);

[0078] (x ij is a negative indicator);

[0079] Where X max and X min are the maximum and minimum values ​​of the jth indicator respectively. ij For processed data.

[0080] In the embodiment of the present application, the evaluation indicators are the complexity evaluation score and the risk evaluation score of the test case, both of which are negative indicators.

[0081] Step h2, calculate characteristic weight h ij .

[0082] Calculate the feature weight of the i-th test case under the j-th indicator:

[0083]

[0084] In the formula, i=1,…,n; j=1,…,m.

[0085] Step h3, entropy value calculation.

[0086] Calculate the entropy value of the jth indicator:

[0087]

[0088] Where i = 1, ..., n; j = 1, ..., m, 0 ≤ e j ≤1.

[0089] Step h4, calculation of the coefficient of difference.

[0090] Calculate the coefficient of variation of the j-th indicator:

[0091] g j =1-e j ;Wherein, j=1,...,m.

[0092] Step h5, weight calculation.

[0093] Calculate the weight of the jth indicator:

[0094] In the formula, j=1,…,m.

[0095] The complexity weight coefficient and risk weight coefficient of each test case are obtained through the above calculation.

[0096] Step b3, for any test case, determine the first weight coefficient of the test case according to the product of the complexity evaluation score of the test case and the complexity evaluation coefficient, and the product of the risk evaluation score and the risk weight coefficient.

[0097] In the embodiment of the present application, the complexity and danger of the test scenario are comprehensively considered, and the complexity and danger weight coefficients correspond to their respective evaluation scores. The test scenario of each test case is comprehensively scored to characterize the degree of impact on the AEB system, and the first weight coefficient of the test case is determined. Specifically, it is shown in the following formula:

[0098] S scenarioi =CS i CW i +DS i DW i ;

[0099] Where S scenarioi Represents the total evaluation score of the i-th specific test case, CS i DS i are the complexity weight coefficient and risk weight coefficient of the i-th specific test case, CW i , DW i are the complexity evaluation score and risk evaluation score of the i-th specific test case respectively. scenarioi Normalize to get the first weight coefficient α of each test case i .

[0100] Further, the complexity evaluation score of each test case is determined according to the following steps:

[0101] Step c1, based on multiple real accident case data, determine the information entropy of each state of each scene element.

[0102] In the embodiments of the present application, the autonomous driving test scenario covers a rich set of scenario elements. On the basis of the functional scenario, other key elements in the scenario need to be supplemented to form the test scenario elements of the test case. Traditional automobile driving scenarios generally classify the scenario elements of test cases from four dimensions: driver factors, vehicle factors, road factors, and environmental factors. In the autonomous driving test scenario, the driver factor is weakened. According to the scenario definition, the scenario elements can be described from two aspects: the vehicle to be tested and the driving environment. The driving environment includes other vehicles, roads, weather conditions and other elements. Combined with the characteristics of vehicle-to-vehicle accidents, the test scenario requirements of the test case and the nature of the automobile AEB function, the scenario elements of the test case are preliminarily selected from the three dimensions of the vehicle to be tested, the target vehicle and the driving environment, as shown in Table 2.

[0103] Table 2 Scene element selection

[0104]

[0105] According to the information entropy theory, the information content of each state of each scene element is calculated, that is, the calculation formula for the complexity of each state of the scene element is:

[0106] c = lb(ω+1);

[0107] Where ω is the weight of a certain state of different scene elements, and c is the complexity of a certain scene element in that state.

[0108] The logistic regression model is used to effectively calculate the probability of casualties caused by accident scenarios, and its probability is converted into the factor state weight ω.

[0109] Specifically, taking the scenario elements shown in Table 2 as an example, whether the accident caused serious injury or death was taken as the dependent variable, and the independent variables were the 8 test scenario elements. The element status was assigned numbers from 1 to 7, and a binary logistic regression model was established. The probability of causing serious injury or death can be expressed as:

[0110]

[0111] Where: x1, x2, ..., x m is the selected scene element; β0 is a constant term; β1, β2, ..., β m are scene elements x1, x2, ..., x m The regression coefficient represents the correlation between the independent variable and the dependent variable.

[0112] Therefore, the ratio of the probability of causing serious injury or death to the probability of not causing serious injury or death (i.e., the odds ratio) is:

[0113]

[0114] According to the binary logistic regression analysis model, the odds ratio (OR) corresponding to each scene element state is obtained, that is, the probability of serious injury or death caused by the remaining states with the last state of each scene element as a reference is obtained. The sum of the weight values ​​of all states under each scene element is set to 1, and the weight of each state under each scene element is calculated according to the multiple relationship between the states in the OR value, that is, the normalized result, to measure the importance of each scene element state on the scene complexity under the scene element, and calculate the complexity of each state of each scene element.

[0115] Step c2, based on multiple real accident case data and a back propagation neural network model, determine the dimension weight coefficient of each scene dimension and the scene element weight coefficient of each scene element under each scene dimension.

[0116] In order to calculate the complexity evaluation score of each test case, it is also necessary to calculate the dimension weight coefficient of each scene dimension and the scene element weight coefficient of each scene element under each scene dimension. In order to avoid the influence of human factors in the main method on the calculation results, the back propagation (BP) neural network algorithm in the deep learning method is adopted in the embodiment of the present application. The BP neural network is a neural network model of error inverse feedback, which is divided into three levels: input layer, hidden layer and output layer. The neurons between the three levels do not interfere with each other, but each neural unit in the adjacent layer is connected. The data is transferred from the input layer to the hidden layer for processing, and then forward propagated to the output layer. The processed data is compared with the true value, and the error is fed back to the previous level. The error is apportioned to a single neuron by gradient descent, and the weight and threshold are corrected to reduce the error.

[0117] In the embodiment of the present application, the input layer data of the BP neural network is all the data of the scene elements, and the output layer is the data of whether the accident caused serious injury or death. The mean square error, mean absolute error and correlation coefficient are selected as indicators to evaluate the performance of the model; the weight coefficient ω between the input layer and the hidden layer is obtained ki , and the weight coefficient ω between the hidden layer and the output layer jk , the weight coefficients between neurons are converted into corresponding weights, and then after weight influence processing, the element weight μ and the weight σ of each dimension are obtained.

[0118] Specifically, the BP neural network algorithm includes the following steps:

[0119] Step m1, normalize the data and then load it into the BP neural network model, and divide the input and output layers into test sets and training sets respectively.

[0120] Step m2, set the model's initial parameter weight ω and critical value θ, where the system will adjust the initial weight ω, the initial value can be randomly set to a smaller value between [0,1], and the critical value θ is set to a smaller value.

[0121] Step m3: The input data is processed by the activation function to generate the output value y j :

[0122]

[0123] Where x i is the input (i=1,…,m); ω ij is the weight between unit i and unit j between two adjacent layers (j=1,…,n); θ j is the critical value under unit j, y j is the expected output value.

[0124] Step m4, actual output value d j The difference between the expected output value and the error (d j -y j ), the error is fed back to adjust the weight, and the adjustment amount is:

[0125] Δω ij =ηδ j x j ;

[0126] In the formula, η is the learning rate. A learning rate that is too fast will lead to skipping the optimal solution, while a learning rate that is too slow will reduce the learning efficiency. j is the input end of two adjacent layers, that is, for the input layer and the hidden layer x j is the input layer, and for the hidden layer and output layer, x j is the hidden layer, δ j is the input error signal of the jth layer, which is related to the expected output value and the actual output value. For the output:

[0127] δ j =η j (1-y j )(d j -y j );

[0128] The reverse feedback for the hidden layer is:

[0129] δ j =x j (1-x j )∑δ k ω jk ;

[0130] Back propagation is the reverse calculation from the output layer to the hidden layer and from the hidden layer to the input layer, which also ensures that each neuron is covered. Finally, after several corrections, the weights of each neuron meet the termination requirements and are obtained as follows:

[0131] ω ij(t) =ω ij (t-1)+Δω ij ;

[0132] Where t is the number of learning times, and each weight is adjusted by the previous weight and correction coefficient.

[0133] In this way, the weight coefficient ω between the input layer and the hidden layer is obtained ki , and the weight coefficient ω between the hidden layer and the output layer jk , the weight coefficients between neurons are converted into corresponding weights, and then after weight influence processing, the element weight μ and the weight σ of each dimension are obtained.

[0134] Step c3, for any test case, determine the complexity evaluation score of the test case based on the information entropy of the state of each scene element of the test case, the dimension weight coefficient of each scene dimension of the test case, and the scene element weight coefficient of each scene element under each scene dimension.

[0135] In this step, the state of each scene element in the test scenario of the test case corresponds to an entropy value of information entropy, that is, the complexity of the scene element. For the test scenario of a specific test case, a scene element corresponds to only one specific state, that is, each scene element under each scene dimension corresponds to a complexity, and the complexity of a scene dimension is the weighted sum of the complexity of all scene elements under the scene dimension; then the complexity of the three scene dimensions is weighted and summed to obtain the complexity evaluation score of a specific test case; specifically, as shown in the following formula:

[0136]

[0137] In the formula, C i is the complexity of the i-th scene dimension, i = 1, 2, 3; c ij is the complexity of the jth scene element under the i-th scene dimension, μ ij is the weight coefficient of the jth scene element under the i-th scene dimension, n is the number of variables in a certain dimension; σ i is the weight coefficient of the i-th dimension, i = 1, 2, 3; F i Represents the complexity evaluation score of the i-th test case.

[0138] Further, the risk assessment score of each test case is determined according to the following steps:

[0139] Step d1, based on multiple real accident case data, determine the risk index of each real accident case.

[0140] In this step, the danger index of each real accident case is determined by the number of casualties in each real accident case in multiple real accident case data. Specifically, the number of casualties includes the number of deaths, the number of serious injuries and the number of minor injuries, as shown in the following formula: k It can be expressed as:

[0141] R k =α 1k ·D 1k +α 2k ·D 2k +α 3k ·D 3k ;

[0142] Among them, R krepresents the danger index of the kth real accident case; D 1k represents the number of deaths of the kth real case, D 2k represents the number of seriously injured people in the kth real case, D 3k represents the number of minor injuries in the kth real case; α 1k ,α 2k ,α 3k They respectively represent the weights of the number of deaths, serious injuries, and minor injuries in the risk assessment in real accident cases.

[0143] Specifically, α 1k ,α 2k ,α 3k It can be determined through the Analytic Hierarchy Process (AHP).

[0144] Specifically, the AHP includes the following steps:

[0145] Step n1, construct a comparison matrix A.

[0146] Assume there are n indicators, forming D = {D1,…,D n}, the comparison matrix A is:

[0147] A=(a ij ) n×n ; i, j = 1, 2, ..., n;

[0148] where a ij Indicator D i Relative to D j The importance of the indicators was determined by using the three-scale method, and experts compared each indicator pairwise to obtain a relative importance comparison matrix.

[0149] Among them, the scale value 2 represents the indicator D i Ratio Index D j Important; scale value 1 indicator D i With indicator D j Equally important; a scale value of 0 represents an indicator D i No indicator D j important;

[0150] Step n2, calculate the importance ranking index H based on the comparison matrix A i .

[0151]

[0152] Step n3, sort by importance index H i Construct the judgment matrix B.

[0153] B(b ij )n×n ; i, j = 1, 2, ..., n;

[0154] in:

[0155]

[0156] Step n4, based on the judgment matrix B, calculate its antisymmetric matrix C.

[0157] C = lgB;

[0158] c ij =lgb ij ;

[0159] Step n5, calculate the optimal transfer matrix B * .

[0160] B * =(b * ij ) n×n ; i, j = 1, 2, ..., n;

[0161] In the formula,

[0162] Step n6, calculate B * The eigenvector corresponding to the maximum eigenvalue of is the weight of each indicator W′=[w′1 w′2 …w′ n ], and then normalize it to get the standard weight W = [w1 w2 … w n ]. In this way, the weights α of the number of deaths, serious injuries, and minor injuries in the risk assessment in real accident cases are obtained. 1k ,α 2k ,α 3k .

[0163] Step d2, for any test case, based on the similarity of each scenario element between the test case and multiple target accident cases in each scenario dimension, determine the correlation weight coefficients of multiple target accident cases and the test case; the target accident case is a real accident case with the same functional scenario as the test case among multiple real accident cases.

[0164] In this step, a mapping relationship is established to represent the relationship between each specific test case and the real accident case. In the same functional scenario, each specific test case i may correspond to one or more real accident cases k. The weight of this mapping relationship is ω ik , represents the correlation weight coefficient between the i-th test case and the k-th real accident case. Specifically, it is shown in the following formula:

[0165]

[0166] In the formula, ω j is the weight coefficient of scene element j; S ijk is the similarity between the i-th test case and the k-th real accident case in scene element j; p is the number of scene elements j.

[0167] Specifically, regression analysis is used to determine the weight ω of scene element j j , perform one-hot encoding on the data of categorical variables in all scene elements, standardize the numerical variables in all scene elements, use logistic regression or linear regression model, take accident risk as dependent variable, and the value of each scene element as independent variable. Fit the model on the data set to obtain the coefficient of each variable. Normalize the absolute value of the coefficient of each variable to obtain the weight ω j Here, the normalization formula can be expressed as:

[0168]

[0169] Among them, coef j is the regression coefficient of scene element j.

[0170] For the similarity S ijk Different calculation methods are used for categorical variables and numerical variables: For categorical variables, if the i-th test case and the k-th real accident case have the same status on scene element j, S ijk =1; if the status is different, S ijk =0;

[0171] For numerical variables, the inverse of the normalized difference is used to calculate the similarity S ijk :

[0172]

[0173] Among them, v i and v k are the values ​​of the scene elements in the test case and the real accident case, respectively. max(V) and min(V) are the maximum and minimum values ​​of the scene elements in all data.

[0174] Among them, the real accident case with the same functional scenario as the specific test case is a real accident case that is exactly the same as the specific test case in terms of the four scenario elements of the speed of the vehicle to be tested, the movement state of the vehicle to be tested, the movement state of the target vehicle and the road type.

[0175] Step d3, determining the risk assessment score of the test case based on the risk index of multiple target accident cases and the correlation weight coefficient with the test case.

[0176] In the embodiment of the present application, the risk assessment score R of each specific test case is i The hazard level R of m real accident cases similar to the test case can be calculated. k And the correlation weight coefficient ω with the test case ik The weighted average is calculated as shown in the following formula:

[0177]

[0178] Among them, R k is the danger level of the kth real accident case; ω ik is the correlation weight coefficient between the i-th test case and the k-th real accident case; m is the number of real accident cases with the same functional scenario as the i-th specific test case.

[0179] Further, in S102, the first evaluation score corresponding to each test case of the automatic emergency braking system of the vehicle to be tested is determined based on the target test data and the grey correlation analysis algorithm, including:

[0180] Step e1, based on the relative importance of various performance evaluation indicators of the preset automobile emergency braking system and the hierarchical analysis algorithm, determine the evaluation indicator weight coefficient of each performance evaluation indicator of the automatic emergency braking system of the automobile to be tested.

[0181] In the embodiment of the present application, before evaluating the performance of the system according to each performance evaluation index of the automobile automatic emergency braking system, in order to make the evaluation result more comprehensive and accurate, it is necessary to determine the evaluation index weight coefficient of each performance evaluation index. Specifically, the implementation method of determining the evaluation index weight coefficient of each performance evaluation index of the automatic emergency braking system of the automobile to be tested by the hierarchical analysis algorithm in this step is the same as the implementation method of determining the weight coefficient of death, serious injury, and minor injury in the risk assessment in real cases by the hierarchical analysis algorithm in step d1. The specific implementation method is shown in step d1 and will not be repeated here. In this step, the importance of each performance evaluation index of the preset automobile automatic emergency braking system is input into the hierarchical analysis algorithm to obtain the evaluation index weight coefficient of each performance evaluation index of the automatic emergency braking system of the automobile to be tested.

[0182] Step e2, for any test case, based on the target test data and the grey correlation analysis algorithm, determine the grey correlation coefficient of each performance evaluation indicator of the test case, and determine the grey correlation coefficient as the evaluation score of each performance evaluation indicator of the test case.

[0183] In the embodiment of the present application, since the basic principle of the grey correlation analysis algorithm is to compare the geometric similarity between the evaluation sequence representing the target test data of the vehicle to be tested under each test case and the reference sequence representing the ideal test data of the vehicle to be tested under each test case from the aspects of curve change trend, direction, size, etc., the grey correlation coefficient of each performance evaluation index of each test case is obtained. Since the larger the grey correlation coefficient, the higher the similarity between the evaluation sequence curve and the reference sequence curve, then if the grey correlation coefficient of any performance evaluation index under any test case is larger, it means that the test data of the performance evaluation index of the AEB system under the test case is closer to the ideal test data of the performance evaluation index under the test case, the higher the performance of the vehicle AEB system, that is, the size of the grey correlation coefficient is positively correlated with the performance of the vehicle AEB system, and the grey correlation coefficient can be determined as the evaluation score of each performance evaluation index of the test case.

[0184] Step e3, determining a first evaluation score of the automatic emergency braking system of the vehicle to be tested under the test case according to the evaluation index weight coefficients of each performance evaluation index of the automatic emergency braking system of the vehicle to be tested and the evaluation score of each performance evaluation index of the test case.

[0185] In the embodiment of the present application, the performance evaluation indicators of the four systems, namely, the speed reduction of the vehicle to be tested, the relative braking stopping distance, the AEB triggering time and the maximum braking deceleration, are taken as examples, and 10 points are used as the scoring standard for the first evaluation score. The first evaluation score of the automatic emergency braking system of the vehicle to be tested under the test case can be expressed by the following formula:

[0186]

[0187] Where M i is the first evaluation score of the i-th test case, i = 1, 2, ..., n; β i (j) is the evaluation score of the jth performance evaluation indicator of the i-th test case, j = 1, 2, 3, 4; ω j is the evaluation index weight coefficient of the jth performance evaluation index.

[0188] Further, in S102, for any test case, based on the target test data and the grey correlation analysis algorithm, the grey correlation coefficient of each performance evaluation index of the test case is determined, including:

[0189] Step f1, determining the evaluation sequence of the target test data in the grey relational analysis algorithm according to the target test data.

[0190] In the embodiment of the present application, taking n test cases and 4 system performance evaluation indicators as an example, the target test data is processed to obtain the evaluation sequence of the target test data in the grey relational analysis algorithm [C i (j)] n×4 Among them, C i (j) represents the test data of the performance evaluation index of the jth system in the i-th test case.

[0191] Step f2, determining the reference sequence of the target test data in the grey relational analysis algorithm according to the preset ideal test data of the vehicle to be tested under each test case.

[0192] In this step, the reference sequence of the target test data in the grey relational analysis algorithm is determined according to the preset ideal test data of the vehicle under test in each test case. i0 (j)] n×4 ; Among them, X i0 (j) represents the ideal test data of the performance evaluation index of the jth system in the ith test case; the preset ideal test data of the vehicle to be tested under each test case can be obtained based on historical test data statistics, or can be determined by referring to relevant literature or expert recommendations, and no specific restrictions are made here.

[0193] Step f3, determining the grey relational resolution coefficient based on the evaluation sequence and the reference sequence.

[0194] In this step, the test data in the evaluation sequence and the reference sequence are first dimensionally processed according to the mean method. Specifically, the evaluation sequence is shown as follows:

[0195] Where, X i (j) is the value of the performance evaluation index of the jth system in the i-th test case after dimensionless processing; C j is the average value of the test data of the performance evaluation index of the jth system in all i test cases;

[0196] The reference sequence is as follows:

[0197] Where, X i0 (j) is the ideal value of the performance evaluation index of the jth system in the i-th test case after dimensionless processing; C j0 is the average value of the ideal test data of the performance evaluation index of the jth system in all i test cases;

[0198] In the embodiment of the present application, the calculation formula of the grey correlation coefficient is as follows:

[0199]

[0200] Δ min =min i min j |X i0 (j)-X i (j)|;

[0201] Δ max =max i max j |X io (j)-X i (j)|;

[0202] Δ ij =|X io (j)-X i (j)|;

[0203] In the formula, β i (j) is the evaluation score of the performance evaluation index of the jth system of the i-th test case; Δ min Indicates the minimum value of the fluctuation between the value of any performance evaluation indicator of any test case and the ideal value; Δ max Indicates the maximum value of the fluctuation between the value of any performance evaluation indicator of any test case and the ideal value; Δ ij It indicates the degree of fluctuation between the performance evaluation index of the jth system in the i-th test case and the ideal value; ε is the resolution coefficient, which is between (0,1) and usually 0.5.

[0204] The value of the resolution coefficient ε affects the grey relational coefficient β to a certain extent. i (j) is the calculation result. min The value of is usually close to or equal to 0, Δ max The value of greatly affects the grey relational coefficient β i (j), thus affecting the overall evaluation of the performance of the vehicle AEB system. When the evaluation sequence volatility is small, Δ max With Δ ij The values ​​are close. At this time, taking a fixed resolution coefficient of 0.5 will cause the gray correlation coefficient curve to show a trend of gathering as a whole and the value is small; on the contrary, Δ max Much larger than Δ ij When the value of the resolution coefficient ε is dominant, if it is still fixed at 0.5, the correlation coefficient values ​​will all approach 1. Therefore, in order to better adapt to the dynamic fluctuations of the evaluation sequence and make the evaluation results more accurate, the embodiment of the present application introduces the elastic resolution coefficient ε i Substitute the resolution coefficient ε and determine ε in the following formula xi The value of determines the elastic resolution coefficient ε i The value of:

[0205]

[0206] Specifically, when ε xi When the judgment value is greater than 0.5, the volatility of the evaluation sequence is small and the elastic resolution coefficient ε i It can take any value between [0.8,1] to ensure the difference between the grey relational coefficients; when ε xi When the judgment value is less than 0.5, the volatility of the evaluation sequence is large and the elastic resolution coefficient ε i Take ε i =4ε xi 2 To weaken Δ max Impact on the evaluation results of the performance of the automobile AEB system.

[0207] Step f4, for any test case, based on the evaluation sequence, the reference sequence and the grey relational resolution coefficient, determine the grey relational coefficient of each performance evaluation index of the test case.

[0208] In this step, for any test case, based on the evaluation sequence, the reference sequence and the grey relational resolution coefficient, the grey relational coefficient β of each performance evaluation index of the test case is determined according to the calculation formula of the grey relational coefficient as described above. i (j).

[0209] The embodiment of the present application provides a method for determining the performance of an automatic emergency braking system of an automobile, wherein a plurality of test cases and a first weight coefficient of each test case are obtained; any test case is a test case generated based on a plurality of target test scenarios; any target test scenario is a test scenario determined by a plurality of candidate test scenarios generated based on a plurality of real accident case data; the first weight coefficient of any test case is a weight coefficient determined based on the complexity and danger of the test case; each test case is used to perform a simulation test on the automobile to be tested, and target test data of the automobile to be tested under each test case is obtained, and based on the target test data and a grey correlation analysis algorithm, the first evaluation score corresponding to each test case of the automatic emergency braking system of the automobile to be tested is determined; the target test data is the test data corresponding to each performance evaluation index of the automatic emergency braking system; based on the first weight coefficient and the first evaluation score of each test case, the second evaluation score of the automatic emergency braking system of the automobile to be tested is determined; based on the second evaluation score of the automatic emergency braking system, the performance evaluation result of the automatic tight-tie braking system of the automobile to be tested is determined. In this way, by generating test cases based on a plurality of test scenarios constructed based on real accident case data for testing, and quantitatively evaluating each test case based on danger and complexity, the comprehensiveness and accuracy of determining the performance of the automatic emergency braking system of the automobile are improved.

[0210] Based on the same application concept, the embodiments of the present application also provide a vehicle automatic emergency braking system performance determination device corresponding to the vehicle automatic emergency braking system performance determination method provided in the above embodiments. Since the principle of solving the problem by the device in the embodiments of the present application is similar to the vehicle automatic emergency braking system performance determination method in the above embodiments of the present application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0211] See also Figure 2 , Figure 2 This is one of the functional module diagrams of a vehicle automatic emergency braking system performance determination device provided in an embodiment of the present application. Figure 2 As shown, the vehicle automatic emergency braking system performance determination device 200 includes:

[0212] Acquisition module 210 is used to obtain multiple test cases and a first weight coefficient for each test case; any test case is a test case generated based on multiple target test scenarios; any target test scenario is a test scenario determined based on multiple candidate test scenarios generated based on multiple real accident case data; the first weight coefficient of any test case is a weight coefficient determined based on the complexity and criticality of the test case.

[0213] The first determination module 220 is used to use various test cases to perform simulation tests on the vehicle to be tested, obtain target test data of the vehicle to be tested under each test case, and determine the first evaluation score of the automatic emergency braking system of the vehicle to be tested corresponding to each test case based on the target test data and a grey correlation analysis algorithm; the target test data is the test data corresponding to each performance evaluation indicator of the automatic emergency braking system.

[0214] The second determination module 230 is used to determine a second evaluation score of the automatic emergency braking system of the vehicle to be tested based on the first weight coefficient and the first evaluation score of each test case.

[0215] The third determination module 240 is used to determine the performance evaluation result of the automatic emergency braking system of the vehicle to be tested based on the second evaluation score of the automatic emergency braking system.

[0216] For further information, see Figure 3 , Figure 3 This is a second functional module diagram of a vehicle automatic emergency braking system performance determination device provided in an embodiment of the present application. Figure 3 As shown, the vehicle automatic emergency braking system performance determination device 200 also includes:

[0217] The second acquisition module 250 is used to acquire a candidate test scene template; the candidate test scene template includes multiple candidate dimensions, and each candidate dimension corresponds to multiple candidate scene elements.

[0218] The third acquisition module 260 is used to acquire multiple real accident case data, extract candidate dimensions and candidate scenario elements in the candidate test scenario template from each real accident case, and determine the candidate test scenario corresponding to each real accident case.

[0219] The fourth determination module 270 is used to classify and sort the candidate test scenarios of each real accident case according to the category of functional scenarios, select the candidate test scenarios arranged in the front preset positions, and determine multiple target test scenarios; the functional scenario is a scenario described based on the preset candidate scenario elements;

[0220] The test case generation module 280 is used to generate multiple test cases based on each target test scenario; the state of at least one scenario element of the target test scenario corresponding to each test case is different.

[0221] Furthermore, if Figure 3 As shown, the vehicle automatic emergency braking system performance determination device 200 also includes:

[0222] The fifth determination module 290 is used to determine the complexity evaluation score and the risk evaluation score of each test case based on multiple real accident case data.

[0223] The sixth determination module 2100 is used to determine the complexity weight coefficient and the risk weight coefficient of each test case based on the complexity evaluation score, the risk evaluation score and the entropy method of each test case.

[0224] The seventh determination module 2110 is used to determine, for any test case, a first weight coefficient of the test case according to the product of the complexity evaluation score of the test case and the complexity evaluation coefficient, and the product of the risk evaluation score and the risk weight coefficient.

[0225] Furthermore, the fifth determination module 290 determines the complexity evaluation score of each test case according to the following steps:

[0226] Based on the multiple real accident case data, determine the information entropy of each state of each scene element;

[0227] Based on multiple real accident case data and a back propagation neural network model, determine the dimension weight coefficient of each scene dimension and the scene element weight coefficient of each scene element under each scene dimension;

[0228] For any test case, the complexity evaluation score of the test case is determined based on the information entropy of the state of each scene element of the test case, the dimension weight coefficient of each scene dimension of the test case, and the scene element weight coefficient of each scene element under each scene dimension.

[0229] Furthermore, the fifth determination module 290 determines the risk assessment score of each test case according to the following steps:

[0230] Determining a risk index for each real accident case based on the plurality of real accident case data;

[0231] For any of the test cases, based on the similarity of each scenario element between the test case and multiple target accident cases in each scenario dimension, determine the correlation weight coefficients between the multiple target accident cases and the test case; the target accident case is a real accident case with the same functional scenario as the test case among the multiple real accident cases;

[0232] Based on the hazard indexes of the multiple target accident cases and the correlation weight coefficients with the test case, a hazard evaluation score of the test case is determined.

[0233] Further, when the first determination module 420 is used to determine the first evaluation score corresponding to each test case of the automatic emergency braking system of the vehicle to be tested based on the target test data and the grey relational analysis algorithm, the first determination module 420 is specifically used to:

[0234] Based on the relative importance of various performance evaluation indicators of the preset automobile automatic emergency braking system and the hierarchical analysis algorithm, determining the evaluation indicator weight coefficient of each performance evaluation indicator of the automatic emergency braking system of the automobile to be tested;

[0235] For any test case, based on the target test data and the grey correlation analysis algorithm, the grey correlation coefficient of each performance evaluation indicator of the test case is determined, and the grey correlation coefficient is determined as the evaluation score of each performance evaluation indicator of the test case;

[0236] According to the evaluation index weight coefficients of each performance evaluation index of the automatic emergency braking system of the vehicle to be tested and the evaluation score of each performance evaluation index of the test case, the first evaluation score of the automatic emergency braking system of the vehicle to be tested under the test case is determined.

[0237] Further, when the first determination module 420 is used to determine the grey correlation coefficient of each performance evaluation index of a test case based on target test data and a grey correlation analysis algorithm for any test case, the first determination module 420 is specifically used to:

[0238] According to the target test data, determine the evaluation sequence of the target test data in the grey relational analysis algorithm;

[0239] According to the preset ideal test data of the vehicle to be tested under each test case, the reference sequence of the target test data in the grey relational analysis algorithm is determined;

[0240] Based on the evaluation sequence and the reference sequence, the grey relational resolution coefficient is determined;

[0241] For any test case, the grey relational coefficient of each performance evaluation index of the test case is determined based on the evaluation sequence, the reference sequence and the grey relational resolution coefficient.

[0242] The embodiment of the present application provides a device for determining the performance of an automatic emergency braking system of an automobile, comprising: an acquisition module, used to acquire multiple test cases and a first weight coefficient of each test case; any test case is a test case generated based on multiple target test scenarios; any target test scenario is a test scenario determined by multiple candidate test scenarios generated based on multiple real accident case data; the first weight coefficient of any test case is a weight coefficient determined based on the complexity and danger of the test case; a first determination module, used to use each test case to perform a simulation test on the automobile to be tested, obtain target test data of the automobile to be tested under each test case, and determine the first evaluation score of the automatic emergency braking system of the automobile to be tested corresponding to each test case based on the target test data and a grey correlation analysis algorithm; the target test data is the test data corresponding to each performance evaluation index of the automatic emergency braking system; a second determination module, used to determine the second evaluation score of the automatic emergency braking system of the automobile to be tested based on the first weight coefficient and the first evaluation score of each test case; and a third determination module, used to determine the performance evaluation result of the automatic tight-tie braking system of the automobile to be tested based on the second evaluation score of the automatic emergency braking system. In this way, test cases are generated based on multiple test scenarios constructed based on real accident case data, and each test case is quantitatively evaluated based on the degree of danger and complexity, which improves the comprehensiveness and accuracy of the performance determination of the automobile automatic emergency braking system.

[0243] Based on the same application idea, please refer to Figure 4 , Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application is shown in FIG. Figure 4 As shown, the electronic device 400 includes: a processor 410 , a memory 420 and a bus 430 .

[0244] The memory 420 stores machine-readable instructions executable by the processor 410. When the electronic device 400 is running, the processor 410 communicates with the memory 420 through the bus 430. When the processor 410 is running, the machine-readable instructions are executed by the steps of the method for determining the performance of the automatic emergency braking system of an automobile provided in the above embodiment. The specific implementation method can be found in the method embodiment, which will not be repeated here.

[0245] Based on the same application concept, an embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for determining the performance of the automatic emergency braking system of an automobile provided in the above embodiment are executed. The specific implementation method can be found in the method embodiment, which will not be repeated here.

[0246] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0247] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0248] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0249] In addition, each functional unit in the embodiments provided in the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0250] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0251] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and are not to be understood as indicating or implying relative importance.

[0252] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application. They should all be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for determining the performance of an automatic emergency braking system of an automobile, characterized in that: The method comprises: Obtain multiple test cases and a first weight coefficient for each test case; any of the test cases is a test case generated based on multiple target test scenarios; any of the target test scenarios is a test scenario determined based on multiple candidate test scenarios generated based on multiple real accident case data; the first weight coefficient for any of the test cases is a weight coefficient determined based on the complexity and riskiness of the test case; Using each test case to perform simulation test on the vehicle to be tested, obtaining target test data of the vehicle to be tested under each test case, and determining the first evaluation score of the automatic emergency braking system of the vehicle to be tested corresponding to each test case based on the target test data and a grey correlation analysis algorithm; the target test data is the test data corresponding to each performance evaluation index of the automatic emergency braking system; Determining a second evaluation score of the automatic emergency braking system of the vehicle to be tested based on the first weight coefficient and the first evaluation score of each test case; Based on the second evaluation score of the automatic emergency braking system, a performance evaluation result of the automatic emergency braking system of the vehicle to be tested is determined.

2. The method for determining the performance of an automatic emergency braking system of an automobile according to claim 1, characterized in that: The multiple test cases are determined according to the following steps: Obtain a candidate test scenario template; the candidate test scenario template includes multiple candidate dimensions, each candidate dimension corresponds to multiple candidate scenario elements; Acquire multiple real accident case data, extract candidate dimensions and candidate scenario elements in the candidate test scenario template from each real accident case, and determine the candidate test scenario corresponding to each real accident case; Classify and sort the candidate test scenarios of each real accident case according to the category of functional scenario, select the candidate test scenarios ranked in the front preset position, and determine multiple target test scenarios; The functional scenario is a scenario described based on preset candidate scenario elements; Generate multiple test cases based on each target test scenario; The state of at least one scenario element of the target test scenario corresponding to each test case is different.

3. The method for determining the performance of an automobile automatic emergency braking system according to claim 2, characterized in that: After determining the multiple test cases, determine a first weight coefficient for each test case according to the following steps: Based on the multiple real accident case data, determining a complexity evaluation score and a risk evaluation score for each test case; Based on the complexity evaluation score, risk evaluation score and entropy method of each test case, determine the complexity weight coefficient and risk weight coefficient of each test case; For any of the test cases, a first weight coefficient of the test case is determined according to the product of the complexity evaluation score of the test case and the complexity evaluation coefficient, and the product of the risk evaluation score and the risk weight coefficient.

4. The method for determining the performance of an automatic emergency braking system of an automobile according to claim 3, characterized in that: Determine the complexity evaluation score for each test case according to the following steps: Based on the multiple real accident case data, determine the information entropy of each state of each scene element; Based on the multiple real accident case data and the back propagation neural network model, determine the dimension weight coefficient of each scene dimension and the scene element weight coefficient of each scene element under each scene dimension; For any of the test cases, the complexity evaluation score of the test case is determined based on the information entropy of the state of each scene element of the test case, the dimension weight coefficient of each scene dimension of the test case, and the scene element weight coefficient of each scene element under each scene dimension.

5. The method for determining the performance of an automatic emergency braking system of an automobile according to claim 3, characterized in that: Determine the criticality assessment score for each test case by following these steps: Determining a risk index for each real accident case based on the plurality of real accident case data; For any of the test cases, based on the similarity of each scenario element between the test case and multiple target accident cases in each scenario dimension, determine the correlation weight coefficients between the multiple target accident cases and the test case; the target accident case is a real accident case with the same functional scenario as the test case among the multiple real accident cases; Based on the hazard indexes of the multiple target accident cases and the correlation weight coefficients with the test case, a hazard evaluation score of the test case is determined.

6. The method for determining the performance of an automatic emergency braking system of an automobile according to claim 1, characterized in that: The step of determining the first evaluation score of the automatic emergency braking system of the vehicle to be tested corresponding to each test case based on the target test data and the grey correlation analysis algorithm includes: Based on the relative importance of various performance evaluation indicators of the preset automobile automatic emergency braking system and the hierarchical analysis algorithm, determining the evaluation indicator weight coefficient of each performance evaluation indicator of the automatic emergency braking system of the automobile to be tested; For any of the test cases, based on the target test data and the grey relational analysis algorithm, determine the grey relational coefficient of each performance evaluation indicator of the test case, and determine the grey relational coefficient as the evaluation score of each performance evaluation indicator of the test case; According to the evaluation index weight coefficients of each performance evaluation index of the automatic emergency braking system of the vehicle to be tested and the evaluation score of each performance evaluation index of the test case, the first evaluation score of the automatic emergency braking system of the vehicle to be tested under the test case is determined.

7. The method for determining the performance of an automatic emergency braking system of an automobile according to claim 6, characterized in that: The step of determining the grey correlation coefficient of each performance evaluation index of any of the test cases based on the target test data and the grey correlation analysis algorithm comprises: Determining, according to the target test data, an evaluation sequence of the target test data in a grey relational analysis algorithm; Determine the reference sequence of the target test data in the grey relational analysis algorithm according to the preset ideal test data of the vehicle to be tested under each test case; Determining a grey relational resolution coefficient based on the evaluation sequence and the reference sequence; For any test case, based on the evaluation sequence, the reference sequence and the grey relational resolution coefficient, the grey relational coefficient of each performance evaluation indicator of the test case is determined.

8. A device for determining the performance of an automatic emergency braking system of an automobile, characterized in that: The vehicle automatic emergency braking system performance determination device comprises: An acquisition module, used to acquire multiple test cases and a first weight coefficient of each test case; any of the test cases is a test case generated based on multiple target test scenarios; any of the target test scenarios is a test scenario determined based on multiple candidate test scenarios generated based on multiple real accident case data; the first weight coefficient of any of the test cases is a weight coefficient determined based on the complexity and risk of the test case; A first determination module is used to perform simulation tests on the vehicle to be tested using various test cases, obtain target test data of the vehicle to be tested under various test cases, and determine a first evaluation score corresponding to each test case of the automatic emergency braking system of the vehicle to be tested based on the target test data and a grey correlation analysis algorithm; the target test data is test data corresponding to each performance evaluation index of the automatic emergency braking system; A second determination module, configured to determine a second evaluation score of the automatic emergency braking system of the vehicle to be tested based on the first weight coefficient and the first evaluation score of each test case; The third determination module is used to determine the performance evaluation result of the automatic emergency braking system of the vehicle to be tested based on the second evaluation score of the automatic emergency braking system.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor to execute the steps of the method for determining the performance of an automatic emergency braking system of an automobile as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for determining the performance of an automatic emergency braking system for an automobile as claimed in any one of claims 1 to 7 are executed.

Citation Information

Cited By

  • Performance test evaluation method and system for automatic emergency braking system

    CN119469804A

  • A performance test evaluation method and system for an automatic emergency braking system

    CN119469804B

  • Brake detection method of electro-tricycle

    CN119714932A

  • Test case selection method, electronic equipment, storage medium and program product

    CN120994484A