A vehicle-road cooperation scene test evaluation method and system

Abstract

The application discloses a kind of vehicle-road cooperation scene test evaluation method and system, it is related to intelligent transportation field, including, through the deployment sensor network capture vehicle, pedestrian and other traffic participants dynamic information, and utilize vehicle self sensor to obtain driving state data, pre-processing is carried out;Image and laser radar data are extracted and fused based on feature;Based on the initial test scene library of fusion data, the difficulty and coverage of scene are evaluated using machine learning algorithm, and the test scene is dynamically adjusted and optimized;According to test plan, execute test, record data and generate analysis report;Finally, based on the test result evaluation vehicle performance and adjust test scene library;Dynamic adjustment mechanism ensures the diversity and representativeness of test scene, improves the comprehensiveness and effectiveness of test. Through machine learning algorithm evaluation, the construction of test scene is more intelligent and automated.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation, and in particular to a vehicle-road collaboration scenario testing and evaluation method and system. Background Art

[0002] In recent years, with the rapid development of autonomous driving technology, intelligent transportation systems have gradually become a research hotspot. The widespread use of autonomous vehicles in urban transportation has placed higher demands on vehicle safety and reliability. Currently, autonomous vehicle testing relies primarily on traditional laboratory simulation tests and open road testing. Laboratory simulation tests typically use virtual simulation environments, which can easily create various traffic scenarios. However, they lack the support of real-world environmental perception data and cannot fully reflect actual road conditions. Existing testing methods often ignore the ability to dynamically adjust test scenarios, resulting in a single test scenario and an inability to fully cover various complex traffic situations. Summary of the Invention

[0003] In view of the above existing problems, the present invention is proposed.

[0004] Therefore, the present invention provides a vehicle-road collaborative scenario testing and evaluation method and system to solve the problems of incomplete feature extraction, single test scenario and difficulty in dynamically adjusting the test difficulty in existing autonomous driving testing methods.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] In a first aspect, embodiments of the present invention provide a vehicle-road collaboration scenario testing and evaluation method and system, which includes deploying a sensor network to capture dynamic information about vehicles, pedestrians, and other traffic participants, obtaining vehicle driving status data through sensors built into the vehicles, and preprocessing the collected data;

[0007] Extract features from image data and lidar data and fuse them;

[0008] Build an initial test scenario library based on the fused data, evaluate the difficulty and coverage of each scenario in the initial scenario library, dynamically adjust the difficulty and complexity of the test scenarios based on the evaluation results, and add the optimized scenarios to the scenario library;

[0009] According to the test plan, select scenarios from the optimized scenario library to perform tests, record and analyze the data during the test process, and generate a test result analysis report;

[0010] Evaluate vehicle performance and adjust the test scenario library based on the test result analysis report.

[0011] As a preferred solution of the vehicle-road collaboration scenario testing and evaluation method and system described in the present invention, the following aspects are provided:

[0012] The sensor network includes cameras, laser radars, and millimeter-wave radars;

[0013] Cameras cover major intersections and crossroads to obtain visual information, lidar obtains three-dimensional information about the vehicle's surroundings, and millimeter-wave radar obtains dynamic information about other traffic participants around the vehicle, including their distance and speed.

[0014] The vehicle's driving status data includes speed, acceleration, steering angle, and braking status;

[0015] The collected data is cleaned, missing data is filled, and high-level noise is removed.

[0016] As a preferred solution of the vehicle-road cooperative scenario test and evaluation method and system of the present invention, the following steps are specifically included: feature extraction and fusion of image data and lidar data:

[0017] Load the image file through image processing software, convert the image into a two-dimensional matrix, and obtain the image pixel intensity value at each position in the matrix;

[0018] Extract scene features from image data and lidar data;

[0019] Fuse the features of image data and lidar data.

[0020] As a preferred solution of the vehicle-road cooperative scenario testing and evaluation method and system described in the present invention, an initial test scenario library is constructed based on the fused data, and a machine learning algorithm is used to evaluate the difficulty and coverage of each scenario in the initial scenario library.

[0021] As a preferred solution of the vehicle-road cooperative scenario testing and evaluation method and system of the present invention, the difficulty and coverage of each scenario are combined to conduct a comprehensive evaluation;

[0022] Set a site scoring threshold and dynamically adjust the difficulty and complexity of the test scenario based on the comprehensive evaluation results;

[0023] The comprehensive scores determine which scenarios are selected for inclusion in the optimized test scenario library.

[0024] As a preferred embodiment of the vehicle-road collaboration scenario test and evaluation method and system of the present invention, a scenario is selected from the optimized scenario library according to the test plan for execution testing, and for each selected test scenario, data during the test is recorded, including vehicle status information, environmental information, and other test-related data;

[0025] After each test scenario, key performance indicators, abnormal conditions and any failure points are recorded and a test result analysis report is generated.

[0026] As a preferred solution of the vehicle-road cooperative scenario test and evaluation method and system of the present invention, wherein: vehicle performance is evaluated and the test scenario library is adjusted based on the test result analysis report;

[0027] Evaluate the overall performance of the vehicle based on the performance indicators in the test result analysis report, and determine the vehicle's performance indicators and performance change trends in different test scenarios;

[0028] Evaluate the effectiveness of the scenarios in the current test scenario library based on the test results, and adjust the scenarios in the test scenario library based on the test result analysis report;

[0029] Feedback the test results and adjusted test scenarios to the relevant teams for the next round of testing. Based on the results of each test, the test scenario library is continuously adjusted and optimized.

[0030] In a second aspect, the present invention provides a vehicle-road collaboration scenario testing and evaluation system, comprising:

[0031] The data collection and preprocessing module is responsible for collecting dynamic information of vehicles, pedestrians and other traffic participants and preprocessing the data;

[0032] Feature extraction and fusion module, responsible for extracting features from images and lidar data and performing fusion processing;

[0033] The scenario library construction and evaluation module builds the initial test scenario library, evaluates the difficulty and coverage of the scenarios, and dynamically adjusts the scenario library;

[0034] Test plan execution and analysis module, executes the test plan, records test data, and generates test result analysis reports;

[0035] The performance evaluation and scenario library adjustment module evaluates vehicle performance and adjusts the test scenario library based on the test results.

[0036] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the vehicle-road collaborative scenario testing and evaluation method and system as described in the first aspect of the present invention is implemented.

[0037] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium on which a computer program is stored, wherein: when the computer program is executed by a processor, any step of the vehicle-road collaborative scenario testing and evaluation method and system as described in the first aspect of the present invention is implemented.

[0038] The beneficial effects of the present invention are as follows: by extracting and fusing features of image data and lidar data through Gaussian kernel functions, key information in traffic scenes can be captured more accurately. The fused feature vector contains complementary information of image and lidar data, which improves the comprehensiveness and descriptive ability of the features, and provides richer data support for the subsequent construction of a test scenario library. By evaluating the difficulty and coverage of each scenario in the initial test scenario library through a machine learning algorithm, and dynamically adjusting the test based on the evaluation results, the test scenario library can be constructed more scientifically. The dynamic adjustment mechanism ensures the diversity and representativeness of the test scenarios, and improves the comprehensiveness and effectiveness of the test. Evaluation through machine learning algorithms makes the construction of test scenarios more intelligent and automated. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0040] Figure 1 This is a flowchart of the vehicle-road collaboration scenario testing and evaluation method and system in Example 1.

[0041] Figure 2 This is a decision diagram for dynamically adjusting the scene difficulty and complexity based on the comprehensive site evaluation results in Example 1. DETAILED DESCRIPTION

[0042] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0043] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0044] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0045] Example 1, with reference to Figure 1 and Figure 2, which is the first embodiment of the present invention, provides a vehicle-road collaboration scenario testing and evaluation method and system, including the following steps:

[0046] S1. The sensor network includes a camera, a lidar, and a millimeter-wave radar;

[0047] Cameras cover major intersections and junctions to acquire visual information. They are primarily used to capture image information of static and dynamic objects, providing rich visual data for subsequent image processing and feature extraction.

[0048] LiDAR obtains three-dimensional information of the vehicle's surrounding environment. By emitting laser beams and receiving reflected signals, LiDAR can accurately measure the distance of the surrounding environment and the position of objects, providing three-dimensional point cloud data.

[0049] Millimeter-wave radar obtains dynamic information of other traffic participants around the vehicle, including the distance and speed of other traffic participants around the vehicle. By transmitting and receiving millimeter-wave signals, millimeter-wave radar can monitor the distance and relative speed of surrounding objects in real time, and can maintain high detection accuracy especially in severe weather conditions.

[0050] S1.1. The vehicle's driving status data includes speed, acceleration, steering angle, and braking status;

[0051] To further clarify, the vehicle's current speed information is used to assess its driving status in different scenarios. The vehicle's acceleration information reflects its acceleration or deceleration, helping to assess its dynamic performance. The vehicle's steering angle information is used to assess its behavior when turning. The vehicle's braking status information, including whether the braking system is activated and the degree of braking, is used to assess the vehicle's braking performance.

[0052] The collected data is cleaned, missing data is filled, and high-level noise is removed. This improves data quality and reliability, ensuring that subsequent feature extraction and fusion processing are based on accurate data. This also enhances system robustness, reduces misjudgments and errors caused by data quality issues, and makes test results more reliable.

[0053] In general, multi-sensor fusion offers many advantages, such as comprehensive perception. By combining cameras, lidar, and millimeter-wave radar, it can acquire information about the vehicle's surroundings from multiple angles and dimensions, providing more comprehensive and accurate perception data. Different sensors complement each other under different conditions. For example, cameras perform well in well-lit environments, while lidar and millimeter-wave radar can still provide stable data at night or in inclement weather. Multi-sensor configuration increases system redundancy. Even if a sensor fails, the others can continue to provide necessary information, improving system reliability.

[0054] Collecting vehicle driving status data is also a crucial step, enabling accurate assessment. By collecting key driving status data such as vehicle speed, acceleration, steering angle, and braking status, we can accurately evaluate the vehicle's behavior and performance in different test scenarios. Comprehensive monitoring not only helps assess the vehicle's basic driving performance but also monitors its dynamic behavior in complex traffic environments, ensuring the comprehensiveness and effectiveness of the test.

[0055] S2, extract features from image data and lidar data, and fuse them;

[0056] The image file is loaded through image processing software and converted into a two-dimensional matrix. In this process, the position (x, y) of each pixel corresponds to an element in the matrix, and the value of this element is the image pixel intensity value. After the image is converted into a two-dimensional matrix, it is convenient for subsequent mathematical processing and feature extraction. The image pixel intensity can accurately reflect the color or grayscale value of each pixel in the image, providing a basis for subsequent feature extraction.

[0057] S2.1. Extract features from image data. The expression is:

[0058]

[0059] Among them, F I Represents image features, I(x,y) represents the image pixel intensity value, which may be a value between 0 and 255 for grayscale images; for color images, it may be an RGB triplet, Ω represents the effective area of ​​the image, that is, the set of all pixels considered, μ x represents the horizontal coordinate of the center of the image, x represents the horizontal coordinate in pixel coordinates, μ y represents the ordinate of the center of the image, y represents the ordinate in pixel coordinates, σ represents the standard deviation of the Gaussian kernel, which is used to control the width of the Gaussian kernel, dx represents the integration of small increments of x, and dy represents the integration of small increments of y;

[0060] Furthermore, weighting the image using the Gaussian kernel function can highlight the central area of ​​the image and suppress the influence of the edge area, thereby better capturing the core features of the image. In addition, the integration operation ensures the continuity and smoothness of feature extraction, improving the robustness of the feature.

[0061] S2.2. Extract the scene features of the lidar data. The expression is:

[0062]

[0063] Among them, F L represents the features of the LiDAR, N represents the number of points in the LiDAR point cloud, and d i Represents the distance of the i-th point in the lidar point cloud, r i represents the reflection intensity of the i-th point in the lidar point cloud, Represents the average value of the reflection intensity of all points, σ r It represents the standard deviation of the reflection intensity and determines the width of the Gaussian function;

[0064] Further explanation: the feature vector F is extracted through the distance and reflection intensity information of the lidar point cloud L This feature extraction method can effectively capture key information in point cloud data, especially the distribution characteristics of reflection intensity, which is very useful for identifying different types of obstacles. The weighted average operation makes feature extraction more accurate.

[0065] S2.3. Fusion of features of image data and lidar data. The expression is:

[0066]

[0067] Among them, G(F I ,F L ) represents the fused feature vector, A represents the image feature F I The number of M represents the laser radar feature F L The number of (F I ) k Represents the kth element in the image feature, (F L ) j Represents the jth element in the lidar feature.

[0068] It is further explained that by fusing the features of image data and lidar data, more comprehensive and accurate environmental information can be obtained. The fused feature vector contains not only the visual information in the image, but also the three-dimensional structure information in the lidar point cloud, thereby improving the descriptive ability of the overall feature.

[0069] S3. Build an initial test scenario library based on the fused data;

[0070] By fusing image data features F I and F L , combined with vehicle driving status data to build an initial test scenario library; by fusing multiple sensor data, the diversity and richness of the test scenarios are ensured, so that the test scenarios can more comprehensively reflect various situations in the actual traffic environment.

[0071] Use a machine learning algorithm to evaluate the difficulty and coverage of each scene in the initial scene library. The expression is:

[0072]

[0073] R(S)=∑ t∈T ω t ·f t (S);

[0074] Where C(S) represents the difficulty of scene S, γ represents the importance coefficient of the optimized feature vector, δ represents the importance coefficient of the driving state data, v represents the speed of the vehicle, max(v) represents the maximum speed of the vehicle in all scenes, a represents the acceleration of the vehicle, max(a) represents the maximum acceleration of the vehicle in all scenes, θ represents the steering angle of the vehicle, max(θ) represents the maximum steering angle of the vehicle in all scenes, b represents the braking state of the vehicle, and the braking state can be a binary variable (for example, 0 represents unbraked and 1 represents braked), dS represents the spatial element of scene S, R(S) represents the coverage of scene S, T represents the set of all possible traffic participant types, t represents an element of the traffic participant type, ω t represents the weight of type t, f t (S) represents the frequency of occurrence of type t in scene S.

[0075] By evaluating the difficulty and coverage, we can quantify the characteristics of each scenario and more scientifically select representative test scenarios, which helps improve the efficiency and accuracy of testing.

[0076] S3.1. Combine the difficulty and coverage of the scenario to conduct a comprehensive evaluation. The expression is:

[0077]

[0078] Where E(S) represents the comprehensive evaluation of scenario S, max(C(S)) represents the maximum difficulty of all scenarios, and max(R(S)) represents the maximum coverage of all scenarios.

[0079] Through comprehensive evaluation, the difficulty and coverage of each scenario can be comprehensively measured to ensure that the selected test scenarios are both challenging and representative;

[0080] Set a site scoring threshold and dynamically adjust the difficulty and complexity of the test scenario based on the comprehensive evaluation results;

[0081] To further explain, scenario A:

[0082] There were only a few vehicles going straight on the road, and no pedestrians or bicycles.

[0083] Vehicle speed: 30km / h.

[0084] Acceleration: 0 m / s 2 (Constant speed driving).

[0085] Steering angle: 0° (straight driving).

[0086] Braking status: Not braked.

[0087] Comprehensive evaluation results show that the difficulty level C(A) = 0.2 and the coverage range R(A) = 0.3;

[0088] Based on the comprehensive evaluation formula:

[0089] Assume that the set field score threshold is 0.7. This means that any scenario with a comprehensive evaluation score below 0.7 needs to be adjusted to increase its difficulty or coverage. Through dynamic adjustment, the difficulty and complexity of the test scenarios are ensured to be moderate, neither too simple nor too complex, so as to better evaluate the performance of the vehicle in different scenarios.

[0090] The comprehensive score determines which test scenario is selected into the optimized test scenario library, and its expression is:

[0091]

[0092] Among them, T represents the optimized test scene library, Sc represents the set of all possible scenes, and S ′ represents the scenarios traversed when calculating the sum of comprehensive evaluation indicators of all scenarios, η represents the scenario importance coefficient, ζ represents the similarity influence coefficient, and sim(S,T) represents the average similarity of all scenarios in the test scenario library T for the current scenario S.

[0093] The value range of T lies between [0,1], where a value close to 1 means that the scenario is more likely to be selected into the test scenario library, while a value close to 0 means that the scenario is more likely to be excluded.

[0094] It is further explained that the optimized test scenario library ensures that each scenario has a certain degree of uniqueness and representativeness, avoiding repeated testing of the same type of scenarios, thereby improving the efficiency and effectiveness of the test.

[0095] S4. Select scenarios from the optimized scenario library according to the test plan to perform tests. For each selected test scenario, record data during the test, including vehicle status information, such as speed, acceleration, steering angle, braking status, environmental information, such as information about traffic participants (e.g., pedestrians, other vehicles), road conditions (e.g., road slipperiness), weather conditions (e.g., sunny, rainy), and other test-related data, such as the vehicle's internal system status, communication data, communication delay, and sensor status.

[0096] After each test scenario, key performance indicators (driving performance, response time, throughput, error rate, vehicle safety distance control, emergency braking performance, vehicle stability, and ride experience), abnormal conditions, and any failure points are recorded;

[0097] Abnormal conditions are any unexpected behaviors that occur during the test, such as the vehicle suddenly leaving the lane or failing to correctly identify traffic signs.

[0098] Failure points are any hardware or software failures that occur during the test, such as sensor failure and control system failure.

[0099] After all tests are completed, a test result analysis report is generated. The report includes test overview, test data summary, key performance indicator analysis, abnormal situation and failure point analysis, conclusions and suggestions.

[0100] S5. Evaluate vehicle performance and adjust the test scenario library based on the test result analysis report;

[0101] Evaluate the overall performance of the vehicle based on the performance indicators in the test result analysis report, and determine the vehicle's performance indicators and performance change trends in different test scenarios;

[0102] Performance indicators include driving ability, decision-making ability, safety and comfort.

[0103] Performance change trends include performance stability, modification effects, and anomaly detection;

[0104] Evaluate the effectiveness of the scenarios in the current test scenario library based on the test results, and adjust the scenarios in the test scenario library based on the test result analysis report;

[0105] Effectiveness includes scenario coverage, scenario difficulty, and scenario diversity. Scenario coverage evaluates whether the test scenarios cover all critical traffic situations and edge cases. Scenario difficulty evaluates whether the test scenarios are appropriately difficult and can effectively differentiate between different vehicle performance levels. Scenario diversity evaluates the diversity of the test scenarios, ensuring that different types of traffic participants and environmental conditions are fully considered.

[0106] Scenario adjustment includes adding new scenarios, adjusting existing scenarios, and deleting invalid scenarios. Adding new scenarios means that if you find that certain key situations are not covered, you can design and add new test scenarios. Adjusting existing scenarios means that if the difficulty of existing scenarios is not appropriate, you can optimize the scenarios by adjusting parameters (such as traffic density, weather conditions, etc.). Deleting invalid scenarios means that if certain scenarios do not bring significant test results, you can consider deleting these scenarios to reduce unnecessary testing burden.

[0107] Feedback the test results and adjusted test scenarios to the relevant teams for the next round of testing. Based on the results of each test, the test scenario library is continuously adjusted and optimized.

[0108] Based on the adjusted test scenario library, organize the next round of testing to ensure that the improvement measures are effectively verified. Based on the results of each test, continuously adjust and optimize the test scenario library to form a continuous improvement cycle.

[0109] This embodiment also provides a vehicle-road collaboration scenario testing and evaluation system, including:

[0110] The data collection and preprocessing module is responsible for collecting dynamic information of vehicles, pedestrians and other traffic participants and preprocessing the data;

[0111] Feature extraction and fusion module, responsible for extracting features from images and lidar data and performing fusion processing;

[0112] The scenario library construction and evaluation module builds the initial test scenario library, evaluates the difficulty and coverage of the scenarios, and dynamically adjusts the scenario library;

[0113] Test plan execution and analysis module, executes the test plan, records test data, and generates test result analysis reports;

[0114] The performance evaluation and scenario library adjustment module evaluates vehicle performance and adjusts the test scenario library based on the test results.

[0115] This embodiment also provides a computer device, which is suitable for the vehicle-road collaborative scenario testing and evaluation method and system, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the vehicle-road collaborative scenario testing and evaluation method and system proposed in the above embodiment.

[0116] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0117] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by the processor, the method and system for implementing the vehicle-road cooperative scenario test and evaluation proposed in the above embodiment are implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, disk or optical disk.

[0118] In summary, the present invention uses Gaussian kernel functions to extract and fuse features of image data and lidar data, which can more accurately capture key information in traffic scenes. The fused feature vector contains complementary information of image and lidar data, which improves the comprehensiveness and descriptive ability of the features, and provides richer data support for the subsequent construction of the test scene library. By evaluating the difficulty and coverage of each scene in the initial test scene library through machine learning algorithms, and dynamically adjusting the difficulty and complexity of the test scenes according to the evaluation results, the test scene library can be constructed more scientifically. The dynamic adjustment mechanism ensures the diversity and representativeness of the test scenes, and improves the comprehensiveness and effectiveness of the test. Evaluation through machine learning algorithms makes the construction of test scenes more intelligent and automated.

[0119] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A vehicle-road collaboration scenario testing and evaluation method, characterized in that: include: Deploy sensor networks to capture dynamic information about vehicles, pedestrians, and other traffic participants, obtain vehicle driving status data through vehicle-mounted sensors, and pre-process the collected data; Feature extraction and fusion of image data and lidar data are performed, which specifically includes the following steps: Load the image file through image processing software, convert the image into a two-dimensional matrix, and obtain the image pixel intensity value at each position in the matrix; Extract scene features from image data and lidar data; Fuse the features of image data and lidar data; Build an initial test scenario library based on the fused data, evaluate the difficulty and coverage of each scenario in the initial scenario library, dynamically adjust the difficulty and complexity of the test scenario according to the evaluation results, and add the optimized scenario to the scenario library. The specific steps include the following: Build an initial test scenario library based on the fused data, and use machine learning algorithms to evaluate the difficulty and coverage of each scenario in the initial scenario library; Combine the difficulty and coverage of each scenario to conduct a comprehensive assessment; Set a site scoring threshold and dynamically adjust the difficulty and complexity of the test scenario based on the comprehensive evaluation results; The selected test scenario library will be determined based on the comprehensive score; According to the test plan, select scenarios from the optimized scenario library to perform tests, record and analyze the data during the test process, and generate a test result analysis report; Evaluate vehicle performance and adjust the test scenario library based on the test result analysis report, specifically including the following steps: Evaluate the overall performance of the vehicle based on the performance indicators in the test result analysis report, and determine the vehicle's performance indicators and performance change trends in different test scenarios; Evaluate the effectiveness of the scenarios in the current test scenario library based on the test results, and adjust the scenarios in the test scenario library based on the test result analysis report; Feedback the test results and adjusted test scenarios to the relevant teams for the next round of testing. Based on the results of each test, the test scenario library is continuously adjusted and optimized.

2. The vehicle-road collaboration scenario testing and evaluation method according to claim 1, characterized in that: The sensor network includes cameras, laser radars, and millimeter-wave radars; Cameras cover major intersections and crossroads to obtain visual information, lidar obtains three-dimensional information about the vehicle's surroundings, and millimeter-wave radar obtains dynamic information about other traffic participants around the vehicle, including their distance and speed. The vehicle's driving status data includes speed, acceleration, steering angle, and braking status; The collected data is cleaned, missing data is filled, and high-level noise is removed.

3. The vehicle-road collaboration scenario testing and evaluation method according to claim 2, characterized in that: Select scenarios from the optimized scenario library according to the test plan to perform tests, and record the data during the test process for each selected test scenario; After each test scenario, key performance indicators, abnormal conditions and any failure points are recorded and a test result analysis report is generated.

4. A vehicle-road cooperative scenario test and evaluation system, based on the vehicle-road cooperative scenario test and evaluation method and system according to any one of claims 1 to 3, characterized in that: include: The data collection and preprocessing module is responsible for collecting dynamic information of vehicles, pedestrians and other traffic participants and preprocessing the data; Feature extraction and fusion module, responsible for extracting features from images and lidar data and performing fusion processing; The scenario library construction and evaluation module builds the initial test scenario library, evaluates the difficulty and coverage of the scenarios, and dynamically adjusts the scenario library; Test plan execution and analysis module, executes the test plan, records test data, and generates test result analysis reports; The performance evaluation and scenario library adjustment module evaluates vehicle performance and adjusts the test scenario library based on the test results.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the vehicle-road cooperative scenario testing and evaluation method and system described in any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the vehicle-road cooperative scenario testing and evaluation method and system described in any one of claims 1 to 3 are implemented.

Images