Multi-dimensional evaluation method and device for automatic driving test scene and storage medium

Abstract

The invention discloses an automatic driving test scene multi-dimensional evaluation method, and the method comprises the steps: carrying out the decomposition of each automatic driving test scene based on an evaluation category, and obtaining a dynamic element and a static element corresponding to the evaluation category; the evaluation category comprises risk, complexity and rare degree; determining a subjective weight of each automatic driving test scene by using an analytic hierarchy process, determining an objective weight of each automatic driving test scene by using an entropy weight method, and fusing the subjective weight and the objective weight based on a game theory to obtain a comprehensive weight; inputting an evaluation model based on the comprehensive weight to obtain an evaluation grade corresponding to the evaluation category; the evaluation grades corresponding to the automatic driving scenes are classified based on the clustering algorithm to obtain the corresponding evaluation grades, so that the risk, the complexity and the rare degree of the scenes are comprehensively considered, and the performance of the automatic driving vehicle is comprehensively evaluated.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving, and specifically to a multi-dimensional evaluation method, device and storage medium for autonomous driving test scenarios. Background Art

[0002] Breakthroughs in autonomous driving technology herald a revolution in transportation, which will achieve this change by improving safety, efficiency and convenience. However, ensuring the reliability and safety of autonomous vehicles remains a core challenge. Autonomous vehicles need to be able to handle a variety of driving scenarios and conditions, striving to at least meet the safety standards of human drivers in most cases. This requires comprehensive testing, covering all aspects from common driving scenarios to rare and dangerous events. Although traditional road testing is extremely valuable, it is often limited by safety, cost and the breadth of real-world scenarios that can be simulated.

[0003] In order to improve the efficiency of testing, scenario-based testing technology has received extensive research and attention. Unlike traditional methods that may rely on random test cases, this testing method focuses on specific, predefined scenarios that autonomous vehicles may encounter. By reproducing a wide range of scenarios from daily driving conditions to extreme situations, scenario-based testing can reveal the performance and limitations of autonomous vehicles in more depth. However, how to quantify the difficulty of these test scenarios is a complex issue. Most current studies focus on the risk level of the scenario, but lack quantitative analysis of the multi-dimensional characteristics of the scenario. In view of this, there is an urgent need to develop a comprehensive method to evaluate and quantify the multi-dimensional characteristics of the test scenario. Summary of the invention

[0004] In response to the above technical problems, the present application provides a multi-dimensional evaluation method, device and storage medium for autonomous driving test scenarios, which will comprehensively consider the risk, complexity and rarity of the scenarios, and aim to comprehensively evaluate the performance of autonomous driving vehicles.

[0005] The present application provides a multi-dimensional evaluation method for an autonomous driving test scenario, the method comprising:

[0006] Step S1, decomposing each autonomous driving test scenario based on evaluation categories to obtain dynamic elements and static elements corresponding to the evaluation categories; the evaluation categories include risk, complexity, and rarity;

[0007] Step S2, using the analytic hierarchy process to determine the subjective weight of each autonomous driving test scenario, using the entropy weight method to determine the objective weight of each autonomous driving test scenario, and fusing the subjective weight and the objective weight based on game theory to obtain a comprehensive weight;

[0008] Step S3, inputting an evaluation model based on the comprehensive weight to obtain an evaluation grade corresponding to the evaluation category;

[0009] Step S4: classify the evaluation levels corresponding to each autonomous driving scenario based on a clustering algorithm to obtain a corresponding evaluation level.

[0010] Preferably, the static elements include target vehicle type, lighting conditions, weather, road surface conditions, number of lanes, road type, signal control, number of obstacles, and number of target vehicles.

[0011] Preferably, the dynamic elements are the states of the accident participants immediately before the collision, including the dynamic driving task of the host vehicle before the collision, the relative motion state, the relative position difference, and the behavior of the target vehicle.

[0012] Preferably, the method of determining the subjective weight of each autonomous driving test scenario by using the hierarchical analysis method includes:

[0013] Rating each autonomous driving test scenario according to the nine-point scale and constructing a judgment matrix;

[0014] Normalizing the judgment matrix and calculating the subjective weight of each evaluation category;

[0015] Calculate the maximum eigenvalue (max) coefficient of each subjective weight and perform a consistency check.

[0016] Preferably, the method of determining the objective weight of each autonomous driving test scenario by using the entropy weight method includes:

[0017] Construct an initialization matrix based on each autonomous driving test scenario;

[0018] Normalizing the initialization matrix, performing dimensionless processing based on each evaluation category, and obtaining positive indicators and negative indicators;

[0019] Entropy is calculated based on the positive index and the negative index to determine the objective weight of each evaluation category.

[0020] Preferably, the fusion of the subjective weight and the objective weight based on game theory to obtain a comprehensive weight includes:

[0021] constructing a linear equation based on the subjective weight and the objective weight;

[0022] The weight coefficients obtained from the linear equation are normalized to obtain a comprehensive weight.

[0023] Preferably, the inputting the evaluation model based on the comprehensive weight to obtain the evaluation grade corresponding to the evaluation category includes:

[0024] Based on the comprehensive evaluation corresponding to the evaluation categories of each autonomous driving test scenario, the ideal solution and the negative ideal solution are obtained respectively;

[0025] The ideal solution and the negative ideal solution are input into an evaluation model to obtain an evaluation grade corresponding to the evaluation category.

[0026] The embodiment of the present invention further provides a multi-dimensional evaluation device for an autonomous driving test scenario, comprising:

[0027] A decomposition module, configured to decompose each autonomous driving test scenario based on an evaluation category to obtain dynamic elements and static elements corresponding to the evaluation category; the evaluation category includes risk, complexity, and rarity;

[0028] A first processing module is used to determine the subjective weight of each autonomous driving test scenario by using a hierarchical analysis method, determine the objective weight of each autonomous driving test scenario by using an entropy weight method, and fuse the subjective weight and the objective weight based on game theory to obtain a comprehensive weight;

[0029] A second processing module, configured to input an evaluation model based on the comprehensive weight to obtain an evaluation grade corresponding to the evaluation category;

[0030] The evaluation module is used to classify the evaluation levels corresponding to each autonomous driving scenario based on a clustering algorithm to obtain the corresponding evaluation level.

[0031] The present application also provides a computer device, including: a memory and a processor, wherein the memory stores computer program instructions for execution on the processor, and when the processor executes the computer program instructions, the multi-dimensional evaluation method for autonomous driving test scenarios as described above is implemented.

[0032] The present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the multi-dimensional evaluation method for autonomous driving test scenarios are implemented as described above.

[0033] As described above, each autonomous driving test scenario is decomposed based on the evaluation category to obtain dynamic elements and static elements corresponding to the evaluation category; the evaluation category includes risk, complexity and rarity; the subjective weight of each autonomous driving test scenario is determined by the hierarchical analysis method, the objective weight of each autonomous driving test scenario is determined by the entropy weight method, and the subjective weight and the objective weight are fused based on game theory to obtain a comprehensive weight; the comprehensive weight is input into the evaluation model to obtain the evaluation grade corresponding to the evaluation category; the evaluation grade corresponding to each autonomous driving scenario is classified based on the clustering algorithm to obtain the corresponding evaluation level. In this way, the risk, complexity and rarity of the scenario are comprehensively considered to comprehensively evaluate the performance of the autonomous driving vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application. In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor.

[0035] Figure 1 A schematic diagram of a flow chart of a multi-dimensional evaluation method for an autonomous driving test scenario provided by an embodiment of the present invention;

[0036] Figure 2 A schematic diagram of the structure of a multi-dimensional evaluation device for an autonomous driving test scenario provided in one embodiment of the present invention.

[0037] The realization of the purpose, functional features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. The above-mentioned drawings have shown clear embodiments of this application, which will be described in more detail later. These drawings and textual descriptions are not intended to limit the scope of the concept of this application in any way, but to illustrate the concept of this application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0038] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application.

[0039] Before further describing the present invention in detail, the nouns and terms involved in the embodiments of the present invention are explained. The nouns and terms involved in the embodiments of the present invention are subject to the following explanations.

[0040] See also Figure 1 This embodiment provides a multi-dimensional evaluation method for an autonomous driving test scenario, the method comprising:

[0041] Step S1, decomposing each autonomous driving test scenario based on evaluation categories to obtain dynamic elements and static elements corresponding to the evaluation categories; the evaluation categories include risk, complexity, and rarity;

[0042] Here, by deconstructing the scene and extracting information, we can accurately obtain the key elements required for the test, including static environmental information such as road layout and traffic signals, as well as dynamic parameter distributions such as speed, acceleration, and position of various traffic participants. This refined extraction method makes the test scene construction more flexible and can quickly adjust the scene elements according to specific test requirements, thereby improving the pertinence and efficiency of the test.

[0043] Step S1, decomposing each autonomous driving test scenario based on evaluation categories to obtain dynamic elements and static elements corresponding to the evaluation categories; the evaluation categories include risk, complexity, and rarity;

[0044] Step S2, using the analytic hierarchy process to determine the subjective weight of each autonomous driving test scenario, using the entropy weight method to determine the objective weight of each autonomous driving test scenario, and fusing the subjective weight and the objective weight based on game theory to obtain a comprehensive weight;

[0045] Step S3, inputting an evaluation model based on the comprehensive weight to obtain an evaluation grade corresponding to the evaluation category;

[0046] Step S4: classify the evaluation levels corresponding to each autonomous driving scenario based on a clustering algorithm to obtain a corresponding evaluation level.

[0047] Here, a three-dimensional scene classification method based on the K-means clustering algorithm can be used. The K-means clustering method is used to classify scenes according to risk, complexity, and rarity. The elbow method is used to determine the optimal number of clusters. The clustering results can be used to divide the scenes into several levels and finally confirm the evaluation level.

[0048] The above-mentioned embodiments of the present invention will comprehensively consider the risk, complexity and rarity of the scenarios, aiming to comprehensively evaluate the performance of the autonomous driving vehicle.

[0049] In one embodiment, preferably, the static elements include target vehicle type, lighting conditions, weather, road surface conditions, number of lanes, road type, signal control, number of obstacles, and number of target vehicles.

[0050] In one embodiment, the dynamic elements are the states of the accident participants immediately before the collision, including the dynamic driving task of the host vehicle before the collision, the relative motion state, the relative position difference, and the behavior of the target vehicle.

[0051] Here, in order to more comprehensively and systematically evaluate the autonomous driving test scenarios, the evaluation angles cover three dimensions: complexity, risk, and rarity. Scenario elements are necessary elements for building simulation test scenarios, which are divided into dynamic elements and static elements. Dynamic elements include: the state of the parties involved in the accident at the moment before the collision, specifically including the dynamic driving tasks of the main vehicle before the collision, relative motion state, relative position difference, and the behavior of the target vehicle. Static elements include target vehicle type, lighting conditions, weather, road surface conditions, number of lanes, road type, signal control, number of obstacles, and number of target vehicles. Among them, the road surface conditions are considered when evaluating the risk and rarity of the test scenario.

[0052] In one embodiment, determining the subjective weight of each autonomous driving test scenario using the analytic hierarchy process includes:

[0053] Rating each autonomous driving test scenario according to the nine-point scale and constructing a judgment matrix;

[0054] Normalizing the judgment matrix and calculating the subjective weight of each evaluation category;

[0055] Calculate the maximum eigenvalue (max) coefficient of each subjective weight and perform a consistency check.

[0056] Here, the analytic hierarchy process is a multi-objective decision-making technology that calculates the relative weight of each indicator by breaking down complex problems into multiple specific indicators and comparing these quantitative and qualitative indicators in pairs to meet certain consistency standards. This method relies on the expertise and subjective judgment of domain experts to ensure that the calculated weights can reasonably reflect the actual situation. In the present invention, we can invite six domain experts (including a virtual simulation test expert, two automotive engineering experts, an accident analysis expert, and an autonomous driving algorithm design expert) to participate in the evaluation, and of course, we can only use virtual simulation test experts. These experts will rate each indicator according to Saaty's nine-point scale and construct a series of pairwise comparison matrices to characterize the relative importance between elements at different levels. The process of determining weights by the analytic hierarchy process is as follows:

[0057] Step 1: Construct a judgment matrix;

[0058] Construct a judgment matrix based on the expert scoring results:

[0059] E=[e ij ]

[0060] Among them, e ij represents the comparison result of element i and element j based on the importance level on a scale of 1-9, and e ij =1 / e ji , i, j = 1, 2,…, n.

[0061] Step 2: Normalization;

[0062] Normalize the judgment matrix and calculate the weight of each evaluation index:

[0063]

[0064]

[0065] Among them, x ij represents the value of the i-th j-th column of the identity matrix, is the column-normalized vector, w i for The normalized feature vector, n is the number of indicators.

[0066] Step 3: consistency check;

[0067] Calculate the maximum eigenvalue max coefficient of each indicator and perform consistency test. When the random consistency ratio (CR) is less than 0.10, it indicates that the reliability of the judgment matrix is ​​high. Here, RI represents the random index and CI represents the consistency index:

[0068]

[0069]

[0070]

[0071] Among them, n is the number of indicators, E is the judgment matrix, and w i is the eigenvector, λ max is the maximum eigenvalue of the judgment matrix.

[0072] In one embodiment, the method of determining the objective weight of each autonomous driving test scenario by using the entropy weight method includes:

[0073] Construct an initialization matrix based on each autonomous driving test scenario;

[0074] Normalizing the initialization matrix, performing dimensionless processing based on each evaluation category, and obtaining positive indicators and negative indicators;

[0075] Entropy is calculated based on the positive index and the negative index to determine the objective weight of each evaluation category.

[0076] Here, the entropy weight method is an objective weight allocation technology based on information entropy theory. Its core purpose is to use the concept of information entropy to quantify the amount of information contained in the evaluation indicators and its influence in the decision-making process. This method reveals the importance of indicators through objective data analysis and effectively reduces the influence of subjective bias. The basic principle is that the lower the information entropy of the evaluation indicator, the richer the effective information it contains, and therefore it should be given a higher weight in the evaluation system; conversely, the higher the information entropy, the less information the indicator provides, and its weight should be reduced accordingly. In the present invention, we use the entropy weight method to determine the objective weight of each scene element. The specific calculation steps are as follows:

[0077] Step 1: Construct the initial matrix;

[0078] Based on the initial data, construct the initialization matrix:

[0079] X=[x ij ]

[0080] Among them, x ij is the data value in the i-th row and j-th column.

[0081] Step 2: Normalization;

[0082] Normalize the initialization matrix:

[0083] X1=[x′ ij ]

[0084] Among them, x' ij is the data value of the i-th row and j-th column after normalization.

[0085] Step 3: Dimensionless;

[0086] Positive indicators:

[0087]

[0088] Negative indicators:

[0089]

[0090] Among them, x iminj is the minimum value in the i-th row, x imaxj is the maximum value in the i-th row.

[0091] Step 4: Determine the weight of the jth evaluation indicator of the i-th project;

[0092]

[0093] Among them, r ij Represents the weight of the jth evaluation indicator of the ith project.

[0094] Step 5: Calculate the entropy of each element;

[0095]

[0096]

[0097] Among them, m is the total number of data rows, e j is the entropy of each indicator.

[0098] Step 6: Determine the level of diversification;

[0099] d j =1―e j

[0100] Step 7: Determine the indicator weights;

[0101]

[0102] In one embodiment, the fusion of the subjective weight and the objective weight based on game theory to obtain a comprehensive weight includes:

[0103] constructing a linear equation based on the subjective weight and the objective weight;

[0104] The weight coefficients obtained from the linear equation are normalized to obtain a comprehensive weight.

[0105] Here, the game theory principle is applied to integrate the subjective weight obtained by the hierarchical analysis method with the objective weight obtained by the entropy weight method to establish the comprehensive weight of each indicator and ensure the accuracy of the decision-making results. According to the basic principles of game theory, the weights obtained by the two methods are reasonably distributed, and the subjective weight and the objective weight are regarded as the two participants in the game. This is to minimize the deviation between the comprehensive weight and the subjective and objective weights, thereby improving the scientificity and credibility of the calculation of the comprehensive weight. The specific calculation steps are as follows:

[0106] Step 1: Construct the equation;

[0107] Suppose the weights of n indicators in the evaluation system of m evaluation methods are w k =(w k1 ,w k2 ,…,w kn ), where k = 1, 2, ..., m. Let the linear combination of these weights be w c :

[0108]

[0109] where c k is to make w c With w A and w E The combination coefficient that minimizes the deviation between A and w E are the weight vectors corresponding to the AHP and the entropy weight method respectively. In other words:

[0110] min‖w c ―w k ‖ 2

[0111] This can be formulated as the following system of equations:

[0112]

[0113] Step 2: Normalization;

[0114] Solve the above equation to obtain the weight coefficients and normalize them:

[0115]

[0116] Step 3: Obtain the optimal weight;

[0117]

[0118] In one embodiment, the step of inputting the evaluation model based on the comprehensive weight to obtain the evaluation grade corresponding to the evaluation category includes:

[0119] Based on the comprehensive evaluation corresponding to the evaluation categories of each autonomous driving test scenario, the ideal solution and the negative ideal solution are obtained respectively;

[0120] The ideal solution and the negative ideal solution are input into an evaluation model to obtain an evaluation grade corresponding to the evaluation category.

[0121] Here, the evaluation model can be the AHP-EWN-TOPSIS model; the TOPSIS model is used to calculate the comprehensive evaluation index. TOPSIS is an efficient multi-attribute decision analysis technique that measures the performance of a solution by evaluating the relative proximity of each solution to the ideal solution and the negative ideal solution. The ideal solution refers to the hypothetical best solution that achieves the optimal value in all indicators, while the negative ideal solution is the hypothetical worst solution that performs as the worst value in all indicators. The advantage of the TOPSIS method is that it is simple and practical, can effectively deal with evaluation problems involving quantitative and qualitative indicators, and fully considers the mutual influence of various indicators in the decision-making process. We use the TOPSIS model based on combined weights to calculate the comprehensive scores of dimensions such as risk, complexity, and rarity. The following are the specific steps for calculating the comprehensive evaluation index using the TOPSIS model:

[0122] Step 1: Construct a normalized decision matrix;

[0123] R=[r ij ]

[0124]

[0125] Among them, x ij is the value of the i-th j-th column of the original data, r ij It is the value of the i-th row and j-th column of the normalized decision matrix.

[0126] Step 2: Determine the weighted normalized decision matrix;

[0127] V=[v ij ]

[0128] v ij =w j r ij

[0129] Among them, w j is the weight value of the jth indicator, v ij It is the value of the i-th row and j-th column of the decision matrix after weighted normalization.

[0130] Step 3: Determine the positive and negative ideal solutions;

[0131]

[0132]

[0133] Step 4: Calculate the separation degree;

[0134]

[0135]

[0136] Step 5: Calculate the relative closeness to the ideal solution;

[0137]

[0138] See also Figure 2 , an embodiment of the present invention further provides a multi-dimensional evaluation device for an autonomous driving test scenario, comprising:

[0139] A decomposition module 21, configured to decompose each autonomous driving test scenario based on an evaluation category to obtain dynamic elements and static elements corresponding to the evaluation category; the evaluation category includes risk, complexity, and rarity;

[0140] A first processing module 22 is used to determine the subjective weight of each autonomous driving test scenario by using a hierarchical analysis method, determine the objective weight of each autonomous driving test scenario by using an entropy weight method, and fuse the subjective weight and the objective weight based on game theory to obtain a comprehensive weight;

[0141] A second processing module 23, configured to input an evaluation model based on the comprehensive weight to obtain an evaluation grade corresponding to the evaluation category;

[0142] The evaluation module 24 is used to classify the evaluation levels corresponding to each autonomous driving scenario based on a clustering algorithm to obtain a corresponding evaluation level.

[0143] The present application also provides a computer device, including: a memory and a processor, wherein the memory stores computer program instructions for execution on the processor, and when the processor executes the computer program instructions, the multi-dimensional evaluation method for autonomous driving test scenarios as described above is implemented.

[0144] The present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the multi-dimensional evaluation method for autonomous driving test scenarios are implemented as described above.

[0145] By integrating the functions of these modules, the constructed device can not only achieve efficient and accurate multi-dimensional evaluation of autonomous driving test scenarios, but also improve the accuracy of robustness assessment of autonomous driving vehicles, which helps to improve the safety and reliability of autonomous driving vehicles and promote the development of their practical applications.

[0146] The above examples are only reference examples. In order to avoid redundancy, they are not listed one by one here. In actual development or application, they can be flexibly combined according to actual needs, but any combination belongs to the technical solution of this application and is covered within the scope of protection of this application.

[0147] An embodiment of the present application also provides a computer device, including a processor and a memory, wherein the memory stores computer program instructions for execution on the processor, and when the processor executes the computer program instructions, the multi-dimensional evaluation method for the autonomous driving test scenario as described above is implemented.

[0148] An embodiment of the present application also provides a computer-readable storage medium storing computer instructions, which, when executed by a processor, implement the multi-dimensional evaluation method for the autonomous driving test scenario described above.

[0149] It is understood that the above scenarios are only examples and do not constitute a limitation on the application scenarios of the technical solutions provided in the embodiments of the present application. The technical solutions of the present application can also be applied to other scenarios. For example, it is known to those skilled in the art that with the evolution of the system architecture and the emergence of new business scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0150] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0151] The steps in the method of the embodiment of the present application can be adjusted in order, combined and deleted according to actual needs.

[0152] The units in the device of the embodiment of the present application can be merged, divided and deleted according to actual needs.

[0153] In the present application, the same or similar terminology concepts, technical solutions and / or application scenario descriptions are generally described in detail only the first time they appear. When they appear again later, they are generally not repeated for the sake of brevity. When understanding the technical solutions and other contents of the present application, for the same or similar terminology concepts, technical solutions and / or application scenario descriptions that are not described in detail later, reference can be made to the previous related detailed descriptions.

[0154] In the present application, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0155] The various technical features of the technical solution of the present application can be arbitrarily combined. In order to make the description concise, not all possible combinations of the various technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present application.

[0156] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as above, and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, a controlled terminal, or a network device, etc.) to execute the method of each embodiment of the present application.

[0157] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loading and executing computer program instructions on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. Computer instructions can be stored in a storage medium, or transmitted from one storage medium to another storage medium. For example, computer instructions can be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) mode. The storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integration. Available media can be magnetic media, (e.g., floppy disk, storage disk, tape), optical media (e.g., DVD), or semiconductor media (e.g., solid-state storage disk Solid State Disk (SSD)), etc.

[0158] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A multi-dimensional evaluation method for autonomous driving test scenarios, characterized in that: The method comprises: Step S1, decomposing each autonomous driving test scenario based on evaluation categories to obtain dynamic elements and static elements corresponding to the evaluation categories; the evaluation categories include risk, complexity, and rarity; Step S2, using the analytic hierarchy process to determine the subjective weight of each autonomous driving test scenario, using the entropy weight method to determine the objective weight of each autonomous driving test scenario, and fusing the subjective weight and the objective weight based on game theory to obtain a comprehensive weight; Step S3, inputting an evaluation model based on the comprehensive weight to obtain an evaluation grade corresponding to the evaluation category; Step S4: classify the evaluation levels corresponding to each autonomous driving scenario based on a clustering algorithm to obtain a corresponding evaluation level.

2. The multi-dimensional evaluation method for autonomous driving test scenarios according to claim 1, characterized in that: The static elements include target vehicle type, lighting conditions, weather, road surface conditions, number of lanes, road type, signal control, number of obstacles, and number of target vehicles.

3. The multi-dimensional evaluation method for autonomous driving test scenarios according to claim 1, characterized in that: The dynamic elements are the states of the accident participants immediately before the collision, including the dynamic driving task of the main vehicle before the collision, the relative motion state, the relative position difference, and the behavior of the target vehicle.

4. The multi-dimensional evaluation method for autonomous driving test scenarios according to claim 1, characterized in that: The method of using the hierarchical analysis method to determine the subjective weight of each autonomous driving test scenario includes: Rating each autonomous driving test scenario according to the nine-point scale and constructing a judgment matrix; Normalizing the judgment matrix and calculating the subjective weight of each evaluation category; Calculate the maximum eigenvalue (max) coefficient of each subjective weight and perform a consistency check.

5. The multi-dimensional evaluation method for autonomous driving test scenarios according to claim 1, characterized in that: The method of using the entropy weight method to determine the objective weight of each autonomous driving test scenario includes: Construct an initialization matrix based on each autonomous driving test scenario; Normalizing the initialization matrix, performing dimensionless processing based on each evaluation category, and obtaining positive indicators and negative indicators; Entropy is calculated based on the positive index and the negative index to determine the objective weight of each evaluation category.

6. The multi-dimensional evaluation method for autonomous driving test scenarios according to claim 1, characterized in that: The fusion of the subjective weight and the objective weight based on game theory to obtain a comprehensive weight includes: constructing a linear equation based on the subjective weight and the objective weight; The weight coefficients obtained from the linear equation are normalized to obtain a comprehensive weight.

7. The multi-dimensional evaluation method for autonomous driving test scenarios according to claim 1, characterized in that: The step of inputting the evaluation model based on the comprehensive weight to obtain the evaluation grade corresponding to the evaluation category includes: Based on the comprehensive evaluation corresponding to the evaluation categories of each autonomous driving test scenario, the ideal solution and the negative ideal solution are obtained respectively; The ideal solution and the negative ideal solution are input into an evaluation model to obtain an evaluation grade corresponding to the evaluation category.

8. A multi-dimensional evaluation device for an autonomous driving test scenario, characterized in that: include: A decomposition module, used to decompose each autonomous driving test scenario based on the evaluation category to obtain dynamic elements and static elements corresponding to the evaluation category; Said evaluation categories include risk, complexity, and rarity; A first processing module is used to determine the subjective weight of each autonomous driving test scenario by using a hierarchical analysis method, determine the objective weight of each autonomous driving test scenario by using an entropy weight method, and fuse the subjective weight and the objective weight based on game theory to obtain a comprehensive weight; A second processing module, configured to input an evaluation model based on the comprehensive weight to obtain an evaluation grade corresponding to the evaluation category; The evaluation module is used to classify the evaluation levels corresponding to each autonomous driving scenario based on a clustering algorithm to obtain the corresponding evaluation level.

9. A computer device, characterized in that: It includes a processor and a memory, the memory stores computer program instructions for execution on the processor, and when the processor executes the computer program instructions, it implements the multi-dimensional evaluation method for autonomous driving test scenarios as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: It stores computer instructions, which, when executed by a processor, implement the multi-dimensional evaluation method for autonomous driving test scenarios as described in any one of claims 1 to 7.

Images