Automatic driving road test problem analysis method and device, equipment and storage medium

Abstract

The application provides an automatic driving road test problem analysis method and device, equipment and a storage medium, including: integrating the road test problem analysis process to improve the overall analysis efficiency of the road test problem; in addition, the target road test data is subjected to first multi-dimensional problem mining, and the data with problems in the target road test data is accurately given a problem label; then, through the association of the first dimension and the second multi-dimension, the dimension of the related road test data corresponding to the problem label of the first dimension is limited for data mining, invalid mining can be avoided, the depth of the road test data analysis is enhanced, the accuracy and efficiency of the mining are improved, so that the current risk value is more accurately and efficiently obtained; by calculating the risk values of a plurality of second dimensions in the future driving mileage based on the current risk value, potential problems of fine granularity can be accurately and efficiently discovered in advance, and safety is improved.

Description

Technical Field

[0001] This application relates to the field of autonomous vehicle technology, and in particular to an autonomous driving road test problem analysis method, apparatus, device and storage medium. Background Technology

[0002] Autonomous driving road test data is characterized by its large volume. Related technologies perform processes such as tagging problematic data, problem analysis, data cleaning, data analysis, and safety assessment manually and in a decentralized manner, which results in low analysis efficiency and poor timeliness, affecting the safety of road tests. Summary of the Invention

[0003] In view of this, the purpose of this application is to propose a method, apparatus, device and storage medium for analyzing problems in autonomous driving road tests.

[0004] To achieve the above objectives, this application provides a method for analyzing problems in autonomous driving road tests, comprising: performing a first multi-dimensional problem mining based on target road test data to obtain multiple first-dimensional problem labels and corresponding related road test data; the first multi-dimensional includes multiple first dimensions; the target road test data is obtained based on road test data of the current mileage of the target fleet on the target route; performing a second multi-dimensional data mining on the corresponding related road test data according to the multiple first-dimensional problem labels to obtain multiple second-dimensional analyzable information; the second multi-dimensional includes multiple second dimensions; obtaining the current risk value of the multiple second dimensions according to the analyzable information of the multiple second dimensions and the corresponding analysis rules; and predicting the risk value of the multiple second dimensions for future mileage on the target route according to the current risk value of the multiple second dimensions and the corresponding historical risk value.

[0005] Optionally, based on the target road test data, a first multi-dimensional problem mining is performed to obtain multiple first-dimensional problem labels and corresponding related road test data, including: the first dimension includes multiple levels of dimensions; for each first dimension, the process includes: performing problem mining on the target road test data according to the problem mining algorithm associated with the current level dimension to determine whether the target road test data has a problem in the current level dimension; the current level dimension is initially the first-level dimension; when the target road test data has a problem in the current level dimension and the current level dimension has a next-level dimension, the next-level dimension is taken as the current level dimension, and the problem mining process is repeated until the current level dimension has no problem or no next-level dimension exists; based on the problem labels of different level dimensions and the corresponding related road test data, the problem label of the current first dimension and the corresponding related road test data are obtained.

[0006] Optionally, based on the multiple first-dimensional problem labels, second multi-dimensional data mining is performed on the corresponding relevant road test data to determine the analyzable information of multiple second dimensions, including: the first dimension and the second multi-dimensional data are in one-to-one correspondence, and the second dimension includes multi-level indicators; for each second dimension corresponding to each first dimension, the process includes: performing data mining on the relevant road test data of the current first dimension based on the data mining algorithm associated with each level of indicators of the current second dimension to obtain the analyzable information of each level of indicators of the current second dimension.

[0007] Optionally, based on the data mining algorithm for the association of indicators at all levels in the current second dimension, data mining is performed on the relevant road test data of the current first dimension to obtain analyzable information of the indicators at all levels in the current second dimension, including: extracting relevant data from the relevant road test data of the current first dimension according to the data type required in the data mining algorithm for the association of current indicators; and analyzing the extracted data according to the data mining process in the data mining algorithm for the association of current indicators to obtain analyzable information of the current indicators.

[0008] Optionally, obtaining the current risk value of the plurality of second dimensions based on the analyzable information of the plurality of second dimensions and the corresponding analysis rules includes: the second dimension includes a plurality of analyzable information, and for each analyzable information of each second dimension, the process includes: obtaining the current risk value of the current analyzable information based on the current analyzable information and the corresponding analysis rules; the analysis rules include comparing the current analyzable information with a corresponding threshold and calculating the degree to which it exceeds or falls below the threshold; obtaining a score for the current analyzable information based on the degree of influence of the current analyzable information on the current second dimension; and obtaining the current risk value of the current second dimension based on the current risk values ​​of the plurality of analyzable information of the current second dimension and the corresponding scores.

[0009] Optionally, the method further includes: performing data cleaning on the road test data of the target fleet's current mileage on the target route based on the operational design domain of the target route, to obtain target road test data.

[0010] Optionally, the method further includes: statistically analyzing the problem labels of the plurality of first dimensions; in response to the target route input by the user, displaying the statistical results, the current risk values ​​of the plurality of second dimensions, and / or the risk values ​​of future mileage in the target road segment; and / or, based on the current risk values ​​of the plurality of second dimensions and the corresponding risk thresholds, calculating the proportion of the second dimensions exceeding the risk thresholds to obtain the route quality of the target fleet when driving with the current version of the intelligent driving system; and obtaining the adjustment strategy of the current version of the intelligent driving system based on the route quality, the current risk values ​​of the plurality of second dimensions, and the risk values ​​of future mileage in the target road segment.

[0011] This application provides an autonomous driving road test problem analysis device, comprising: a problem mining module, used to perform first multi-dimensional problem mining based on target road test data to obtain multiple first-dimensional problem labels and corresponding related road test data; the first multi-dimensional includes multiple first dimensions; the target road test data is obtained based on road test data of the current driving mileage of the target fleet on the target route; a data mining module, used to perform second multi-dimensional data mining on the corresponding related road test data according to the multiple first-dimensional problem labels to obtain multiple second-dimensional analyzable information; the second multi-dimensional includes multiple second dimensions; and a risk prediction module, used to obtain the current risk value of the multiple second dimensions according to the analyzable information of the multiple second dimensions and the corresponding analysis rules; and to predict the risk value of the multiple second dimensions for future driving mileage on the target route according to the current risk value of the multiple second dimensions and the corresponding historical risk value.

[0012] This application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the aforementioned method for analyzing autonomous driving road test problems.

[0013] This application provides a non-transitory computer-readable storage medium that stores computer instructions for causing a computer to execute the aforementioned autonomous driving road test problem analysis method.

[0014] As can be seen from the above, the autonomous driving road test problem analysis method provided in this application integrates the road test problem analysis process, improving the overall efficiency of road test problem analysis. Furthermore, by performing multi-dimensional problem mining on the target road test data, problem labels are accurately assigned to data containing problems. Then, by associating the first dimension with the second multi-dimensional dimension, the dimensions for data mining of the relevant road test data corresponding to the problem labels in the first dimension are limited, avoiding ineffective mining, enhancing the depth of road test data analysis, and improving the accuracy and efficiency of mining, thereby obtaining the current risk value more accurately and efficiently. By calculating risk values ​​for multiple second dimensions based on the current risk value for future driving mileage, fine-grained potential problems can be accurately and efficiently identified in advance, improving safety. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram illustrating an application scenario of the autonomous driving road test problem analysis method according to an embodiment of this application;

[0017] Figure 2 This is a flowchart illustrating the autonomous driving road test problem analysis method according to an embodiment of this application; Figure 3 This is a schematic diagram of the overall design architecture of the autonomous driving road test problem analysis method according to an embodiment of this application; Figure 4 This is a schematic diagram of the structure of the autonomous driving road test problem analysis device according to an embodiment of this application; Figure 5 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.

[0019] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this application should have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms "first," "second," and similar terms used in the embodiments of this application do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are only used to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0020] It is understood that before using the technical solutions of the various embodiments in this disclosure, users will be informed of the type, scope of use, and usage scenarios of the personal information involved in an appropriate manner, and user authorization will be obtained.

[0021] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose, based on the prompt message, whether to provide personal information to the software or hardware such as the electronic device, application, server, or storage medium performing the operations of this disclosed technical solution.

[0022] As an optional but not limited implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0023] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure.

[0024] Currently, autonomous driving road tests generate massive amounts of data. This data is recorded by testers who sign off on the road test and record the situation at the time. It is then stored on the computer via the Internet. This data is characterized by its large volume and high requirements for timely analysis. If the road test data cannot be analyzed, deeply mined, and safety assessed in a timely manner, it will seriously affect the safety of the road test.

[0025] In related technologies, the processes of tagging problematic data, analyzing problems, cleaning data, conducting data analysis, performing security assessments, and providing feedback on problems in road test data are carried out manually and are scattered across various parts. This results in low analysis efficiency and poor timeliness, and cannot meet the requirement of real-time display of road test problems.

[0026] In addition, the relevant technologies operate in different design domains depending on the product definition, and the cleaning rules for different data also vary. The accuracy of removing some non-design domain scenarios needs to be improved. Furthermore, the accuracy of the relevant technologies in analyzing complex road test problems is also relatively low.

[0027] In view of this, embodiments of this application provide an autonomous driving road test problem analysis method, which integrates the road test problem analysis process and improves the overall analysis efficiency of road test problems. In addition, by performing first multi-dimensional problem mining on the target road test data, problem labels are accurately assigned to the data in the target road test data that have problems. Furthermore, by associating the first dimension with the second multi-dimensional dimension, the dimensions for data mining of the relevant road test data corresponding to the problem labels of the first dimension are limited, which can avoid invalid mining, enhance the depth of road test data analysis, and improve the accuracy and efficiency of mining, thereby obtaining the current risk value more accurately and efficiently. By calculating the risk values ​​of multiple second dimensions in the future driving mileage based on the current risk value, fine-grained potential problems can be accurately and efficiently discovered in advance, thereby improving safety.

[0028] Furthermore, by dividing the first dimension into multiple levels, problems can be explored from shallow to deep, and relevant road test data corresponding to different levels of problems can be identified. Then, based on the second dimension corresponding to different levels of dimensions, problem mining can be performed on the relevant road test data, which can identify analyzable information for problems at different levels and improve the accuracy of analysis of complex problems.

[0029] Furthermore, by dividing the second dimension into multi-level indicators, we can analyze the relevant road test data with problems in the first dimension from shallow to deep, improve the depth of road test problem analysis, and ensure that deep-seated problems in road tests can be reported in a timely manner.

[0030] refer to Figure 1 This is a schematic diagram illustrating an application scenario for the autonomous driving road test problem analysis method provided in this application embodiment. The application scenario includes a road test device 101, a terminal device 102, a server 103, and a data storage system 104. The road test device 101, terminal device 102, server 103, and data storage system 104 can all be connected via wired or wireless communication networks.

[0031] The road testing device 101 may include, but is not limited to, devices such as controllers, multimodal sensors, and vehicle communication modules, for acquiring road testing data and sending the road testing data to the data storage system 104. The terminal device 102 may include, but is not limited to, desktop computers, mobile phones, mobile computers, tablets, media players, smart wearable devices, personal digital assistants (PDAs), or other electronic devices capable of performing the above functions. Both the server 103 and the data storage system 104 may be independent physical servers, server clusters or distributed systems composed of multiple physical servers, or cloud servers providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0032] Server 103 can be used to provide autonomous driving road test problem analysis services to terminal device 101. Terminal device 102 can have a client installed that communicates with server 103. The client can respond to input information and send it to server 103. Server 103 can obtain target road test data from the data storage system based on the input information, and perform first multi-dimensional problem mining based on the target road test data to obtain multiple first-dimensional problem labels and corresponding related road test data. The first multi-dimensional includes multiple first dimensions. The target road test data is obtained based on the current mileage of the target fleet on the target route. Second multi-dimensional data mining is performed on the corresponding related road test data according to the multiple first-dimensional problem labels to obtain multiple second-dimensional analyzable information. The second multi-dimensional includes multiple second dimensions. Based on the analyzable information of the multiple second dimensions and the corresponding analysis rules, the current risk value of the multiple second dimensions is obtained. Based on the current risk value of the multiple second dimensions and the corresponding historical risk value, the risk value of the multiple second dimensions for future mileage on the target route is predicted. The server can send data such as problem labels, current risk values, and / or future mileage risk values ​​to the client to help users intuitively understand the current road test safety information.

[0033] The following is combined Figure 1 The application scenarios described above illustrate the autonomous driving road test problem analysis method according to exemplary embodiments of this application. It should be noted that the above application scenarios are merely shown to facilitate understanding of the spirit and principles of this application, and the embodiments of this application are not limited in any way in this regard. Rather, the embodiments of this application can be applied to any applicable scenario.

[0034] like Figure 2 and Figure 3 As shown in the figure, this application provides a method for analyzing problems in autonomous driving road tests, which may include: S1. Perform first-dimensional problem mining on the target road test data to obtain multiple first-dimensional problem labels and corresponding related road test data; the first multi-dimensional includes multiple first dimensions; the target road test data is obtained based on the road test data of the target fleet's current driving mileage on the target route.

[0035] S2. Based on the multiple first-dimensional problem labels, perform second-dimensional data mining on the corresponding relevant road test data to obtain multiple second-dimensional analyzable information; the second-dimensional data includes multiple second dimensions.

[0036] S3. Based on the analyzable information of multiple second dimensions and the corresponding analysis rules, obtain the current risk values ​​of multiple second dimensions.

[0037] S4. Based on the current risk values ​​and corresponding historical risk values ​​of multiple second dimensions, predict the risk values ​​of the future mileage of multiple second dimensions in the target route.

[0038] In this embodiment, the target road test data refers to the road test data to be processed. This target road test data includes not only the road test data generated by the target vehicle fleet itself, but also the road test data generated by other vehicles on the target route at the current mileage of the target fleet. The target vehicle fleet can refer to a set of one or more groups of autonomous vehicles traveling on the target route and subject to a unified collaborative control strategy. For example, if a user wants to understand the road test safety of the target route, the server can use the name of the target route as a keyword index to automatically search the data storage system to obtain the road test data of the target vehicle fleet at the current mileage on the target route. Then, the target road test data is obtained by preprocessing this road test data.

[0039] First-dimensional problem mining refers to mining problems from target road test data based on problem mining algorithms associated with multiple first dimensions. Different problem mining algorithms are associated with different first dimensions and can be used to mine problems from different aspects of the target road test data. The problem mining algorithm may include the data types and problem mining processes required for the associated first dimensions. In this embodiment, the problem mining processes for multiple first dimensions can be executed in parallel, improving data processing efficiency.

[0040] The first multi-dimensional aspect refers to the various dimensions from which road test data may have problems. For example, the first multi-dimensional aspect may include, but is not limited to, system-related, algorithm-related, hardware-related, software-related, operational, and drive-by-wire chassis aspects. The first-dimensional problem label refers to the label information assigned to data in the target road test data that has problems in the first dimension after first-dimensional problem mining. Relevant road test data refers to road test data in the target road test data that has been assigned the first-dimensional problem label. Different first dimensions correspond to different first-dimensional problem labels.

[0041] For each first dimension, relevant data is extracted from the target road test data according to the data type required by the problem mining algorithm associated with the current first dimension. The extracted data is then processed according to the problem mining process of the algorithm associated with the current first dimension to determine if any problems exist. If problems are found, the extracted data can be used as relevant road test data, and the relevant road test data is labeled with the problem tag of the current first dimension. By pre-setting multiple types of problem mining logic and establishing a one-to-one mapping relationship between road test problems and module tags, the automatic identification, classification, and labeling of complex road test problems is achieved, reducing errors from subjective human judgment and improving the accuracy and consistency of problem analysis and module location.

[0042] Second-dimensional data mining refers to data mining of relevant drive test data based on data mining algorithms that associate multiple second dimensions. In the embodiments of this application, the data mining process for multiple second dimensions corresponding to each first dimension can be executed in parallel, improving data processing efficiency and obtaining analyzable information more quickly.

[0043] The second multi-dimensional approach refers to the multiple dimensions from which road test data can be analyzed. The first dimension can correspond to different second multi-dimensional approaches. For example, if the first dimension is system-related, the second multi-dimensional approach may include system-level, road segment-level, safety-level, and efficiency-level second dimensions; if the first dimension is algorithm-related, the second multi-dimensional approach may include algorithm-level, road segment-level, safety-level, and efficiency-level second dimensions; if the first dimension is hardware-related, the second multi-dimensional approach may include module-level, road segment-level, safety-level, and efficiency-level second dimensions. The data mining algorithms associated with the multiple second dimensions corresponding to the first dimension are different, and can be used to analyze the relevant road test data from different perspectives. The data mining algorithms associated with the same second dimension corresponding to different first dimensions may also be different, and can be used to perform similar analyses on different relevant road test data from different dimensional problem perspectives.

[0044] The data mining algorithm may include the data types and data mining processes required for the associated second dimension. For each second dimension corresponding to each first dimension, relevant data is extracted from the relevant drive test data according to the data types required in the data mining algorithm associated with the current second dimension; the extracted data is analyzed according to the data mining process in the data mining algorithm associated with the current second dimension to obtain the analyzable information of the current second dimension. Based on the analyzable information of the current second dimension and the corresponding analysis rules, the current risk value of the current second dimension is obtained.

[0045] The second-dimensional analyzable information refers to data features that can be used for risk assessment after performing second-dimensional data mining on relevant road test data. Analysis rules can represent rules for evaluating the analyzable information and converting it into a current risk value. This application does not limit the implementation method of the analysis rules. For example, the analysis rule can be to compare the analyzable information with a set threshold as an evaluation process, and determine the corresponding current risk value based on the degree to which it exceeds, approaches, or falls below the threshold.

[0046] Then, based on the risk values ​​of multiple second dimensions and their corresponding historical risk values, the data change trends under each second dimension can be analyzed, thereby predicting the risk values ​​of multiple second dimensions in the future driving mileage of the target route. This enables quantitative assessment of road test problems and prediction of future risk indicators, and has the ability to identify risks proactively.

[0047] In this embodiment of the application, the target road test data can be divided according to vehicle identification information, thereby calculating the risk value of the future driving mileage of each vehicle involved in the target road test data in multiple second dimensions on the target route.

[0048] This application provides a method for analyzing problems in autonomous driving road tests, integrating the road test problem analysis process to improve the overall efficiency of road test problem analysis. Furthermore, by performing multi-dimensional problem mining on the target road test data, problem labels are accurately assigned to data containing problems. Then, by associating the first dimension with the second multi-dimensional dimension, the dimensions for data mining of the relevant road test data corresponding to the problem labels in the first dimension are limited, avoiding invalid mining, enhancing the depth of road test data analysis, and improving the accuracy and efficiency of mining, thereby obtaining the current risk value more accurately and efficiently. By calculating risk values ​​for multiple second dimensions based on the current risk value for future driving mileage, fine-grained potential problems can be accurately and efficiently identified in advance, improving safety.

[0049] In an optional embodiment, before performing the first multi-dimensional problem mining based on the target road test data, the autonomous driving road test problem analysis method may further include: performing data cleaning on the road test data of the target fleet's current driving mileage on the target route based on the operation design domain of the target route to obtain the target road test data.

[0050] In this embodiment, different target routes can correspond to different operational design domains. The operational design domain refers to the specific range of conditions under which the autonomous driving system is designed to operate, which may include, but is not limited to, dimensions such as scenario type, weather, test section, and functional applicability. By using the operational design domain, non-target road test data in the current mileage of the target fleet on the target route is cleaned and removed to ensure the purity of the input data.

[0051] In this embodiment of the application, by establishing differentiated and configurable cleaning rules, the automatic and accurate removal of invalid data outside the design domain is achieved, thereby improving the accuracy of invalid data filtering and ensuring the purity and effectiveness of the data source for subsequent analysis.

[0052] In one optional embodiment, a first multi-dimensional problem mining is performed based on the target road test data to obtain multiple first-dimensional problem labels and corresponding related road test data. This may include: the first dimension includes multiple levels of dimensions; for each first dimension, the process includes: performing problem mining on the target road test data according to the problem mining algorithm associated with the current level dimension to determine whether the target road test data has a problem in the current level dimension; the current level dimension is initially the first-level dimension; when the target road test data has a problem in the current level dimension and the current level dimension has a next-level dimension, the next-level dimension is taken as the current level dimension, and the problem mining process is repeated until the current level dimension has no problem or no next-level dimension exists; based on the problem labels of different levels of dimensions and the corresponding related road test data, the problem label of the current first dimension and the corresponding related road test data are obtained.

[0053] For example, the first-level dimension of a first dimension can be an algorithm class, the next-level dimension of the algorithm class can include a control module, and the next-level dimension of the control module can include a control-horizontal dimension and a control-vertical dimension.

[0054] In this embodiment of the application, by dividing the first dimension into multiple levels, problems can be explored from shallow to deep, and relevant road test data corresponding to different levels of problems can be determined. Then, based on the second dimension corresponding to different levels of dimensions, problems can be explored from the relevant road test data, and analyzable information of problems at different levels can be determined, thereby improving the accuracy of analysis of complex problems.

[0055] In one optional embodiment, second multi-dimensional data mining is performed on the corresponding relevant road test data based on multiple first-dimensional problem labels to determine the analyzable information of multiple second dimensions. This may include: a one-to-one correspondence between the first dimension and the second multi-dimensional data, with the second dimension including multi-level indicators; for each second dimension corresponding to each first dimension, the process includes: performing data mining on the relevant road test data of the current first dimension based on the data mining algorithm associated with each level of indicators of the current second dimension to obtain the analyzable information of each level of indicators of the current second dimension.

[0056] For example, a primary metric of a second dimension can be an algorithm dimension metric, and the next level metric of the algorithm dimension metric can include a control problem metric, and the next level metric of the control problem metric can include a control error metric and a control emergency braking metric.

[0057] In this embodiment of the application, by dividing the second dimension into multi-level indicators, the relevant road test data with problems in the first dimension can be analyzed from shallow to deep, thereby improving the depth of road test problem analysis and ensuring that deep-seated problems in road tests can be reported in a timely manner.

[0058] In one optional embodiment, data mining is performed on the relevant road test data of the current first dimension according to the data mining algorithm associated with each level of indicators in the current second dimension to obtain analyzable information of each level of indicators in the current second dimension. This may include: extracting relevant data from the relevant road test data of the current first dimension according to the data type required in the data mining algorithm associated with the current level indicator; and analyzing the extracted data according to the data mining process in the data mining algorithm associated with the current level indicator to obtain analyzable information of the current level indicator.

[0059] In this embodiment of the application, by mapping each level of indicators to a data mining algorithm, the analyzable information of each level of indicator can be determined quickly and accurately.

[0060] In one optional embodiment, obtaining the current risk value of multiple second-dimensional analyzable information and corresponding analysis rules may include: the second dimension includes multiple analyzable information, and for each analyzable information of each second dimension, the process includes: obtaining the current risk value of the current analyzable information based on the current analyzable information and corresponding analysis rules; the analysis rules include comparing the current analyzable information with a corresponding threshold and calculating the degree to which it exceeds or falls below the threshold; obtaining a score for the current analyzable information based on its influence on the current second dimension; and obtaining the current risk value of the current second dimension based on the current risk values ​​of the multiple analyzable information of the current second dimension and their corresponding scores.

[0061] Different analyzable information can correspond to different thresholds. The current risk value and score of different analyzable information can be obtained by weighted fusion processing to obtain the current risk value of the second dimension. In the embodiments of this application, by considering that the degree of influence of different analyzable information on the second dimension may be different, the accuracy of calculating the current risk value of the second dimension is improved.

[0062] In one optional embodiment, after obtaining multiple problem tags of the first dimension, the multiple problem tags of the first dimension can be statistically analyzed according to different dimensions to determine high-frequency problem tags. For example, the statistical analysis can be performed on problem tags of the same first dimension existing in the target route, or on problem tags of the same first dimension existing in the target fleet. Optionally, an alarm can be triggered for problem tags of the first dimension that exceed a preset frequency. Further, the problem tags of the first dimension can be divided into different levels according to the degree to which they exceed the preset frequency, thereby realizing a graded alarm.

[0063] In an optional embodiment, various data involved in the above process can be displayed in real time or at preset time intervals to intuitively reflect the road test problem situation. This application embodiment does not limit the displayed data or display method. For example, in response to the user's input of a target route, statistical results of problem tags, current risk values ​​of multiple second dimensions, and / or risk values ​​for future mileage can be displayed. Display methods may include, but are not limited to, curves, bar charts, pie charts, graphs, etc.

[0064] In an optional embodiment, the autonomous driving road test problem analysis method may further include: calculating the proportion of the second dimension exceeding the risk threshold based on the current risk values ​​of multiple second dimensions and the corresponding risk thresholds, thereby obtaining the route quality when the target fleet is driving with the current version of the intelligent driving system; and obtaining an adjustment strategy for the current version of the intelligent driving system based on the route quality, the current risk values ​​of multiple second dimensions, and the risk value of future mileage in the target road segment. In this embodiment, the intelligent driving system can be iteratively updated based on the adjustment strategy at preset intervals to improve driving safety.

[0065] In one optional embodiment, multiple alarm points can be set in the second dimension, including safety, efficiency, system, and algorithm. When the current risk value exceeds a risk threshold, the current risk value can be highlighted and an alarm can be triggered. Information related to the current risk value will also be sent to the corresponding decision-makers, forming a closed-loop process of problem discovery, analysis, early warning, and feedback, facilitating rapid decision-making and understanding of the current problem's risk points. Information related to the current risk value may include, but is not limited to, the corresponding vehicle information, the corresponding problem label in the first dimension, the current risk value, and the risk value of future mileage on the target road segment. Furthermore, different risk levels can be set according to the degree to which the current risk value exceeds a preset risk value, and alarms can be triggered according to the risk level.

[0066] In an optional embodiment, prompting services such as road test safety risk alerts, algorithm problem alerts, key concern issue alerts, indicator rollback alerts, and user experience issue alerts may also be provided.

[0067] In summary, this application integrates previously fragmented and manually-driven independent processes—including uploading and indexing autonomous driving road test problem data, cleaning operational design domain data, automated problem mining, automatic module tag mapping, multi-dimensional data statistical analysis, visual dashboard display, and real-time risk alarm exposure—into a complete end-to-end automated process. This significantly improves the overall efficiency of road test problem data analysis, shortens the problem analysis cycle, and solves the problems of process fragmentation and inefficiency caused by manual processing in multiple stages. Secondly, by extracting problem features from multiple second dimensions such as modules, road segments, safety, efficiency, systems, and algorithms, and achieving road test risk prediction through threshold comparison and trend prediction, the depth and foresight of road test data analysis are enhanced, enabling the early detection of potential systemic problems and version quality vulnerabilities. Furthermore, by displaying the distribution and changes of road test problems in a visual dashboard format, and implementing tiered alarms and rapid push notifications for high-risk, high-frequency, and threshold-exceeding problems, the timeliness and intuitiveness of road test problem exposure are greatly improved, facilitating R&D, testing, and decision-making personnel to quickly grasp the version quality status and shorten problem response and iteration decision-making time. In addition, it can realize closed-loop support for road test problems from discovery, analysis to feedback and rectification. Through automated risk warning and accurate problem location, it can effectively improve the iteration efficiency and R&D quality of intelligent driving systems, reduce the risk of missing road test safety hazards, and provide reliable data support and decision-making basis for the continuous optimization of intelligent driving systems.

[0068] It should be noted that the method in this embodiment can be executed by a single device, such as a computer or server. The method can also be applied in a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the method in this embodiment, and the multiple devices will interact with each other to complete the method described.

[0069] It should be noted that the above description describes some embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0070] Based on the same inventive concept, and corresponding to any of the above embodiments, this application also provides an autonomous driving road test problem analysis device.

[0071] refer to Figure 4 The autonomous driving road test problem analysis device 400 may include: The problem mining module 401 is used to perform first-dimensional problem mining based on the target road test data, and obtain multiple first-dimensional problem labels and corresponding related road test data; the first multi-dimensional includes multiple first dimensions; the target road test data is obtained based on the road test data of the target vehicle fleet's current driving mileage on the target route; The data mining module 402 is used to perform second multi-dimensional data mining on the corresponding relevant road test data based on multiple first-dimensional problem labels to obtain multiple second-dimensional analyzable information; the second multi-dimensional includes multiple second dimensions. The risk prediction module 403 is used to obtain the current risk value of multiple second dimensions based on the analyzable information of multiple second dimensions and the corresponding analysis rules; and to predict the risk value of multiple second dimensions in the future mileage of the target route based on the current risk value of multiple second dimensions and the corresponding historical risk value.

[0072] For ease of description, the above devices are described in terms of function, divided into various modules. Of course, in implementing this application, the functions of each module can be implemented in one or more software and / or hardware.

[0073] The apparatus described above is used to implement the corresponding autonomous driving road test problem analysis method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0074] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the autonomous driving road test problem analysis method described in any of the above embodiments.

[0075] Figure 5 This embodiment illustrates a more specific hardware structure of an electronic device. The device may include a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected internally via the bus 1050.

[0076] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0077] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.

[0078] The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.

[0079] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0080] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.

[0081] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0082] The electronic devices described above are used to implement the corresponding autonomous driving road test problem analysis methods in any of the foregoing embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0083] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the autonomous driving road test problem analysis method as described in any of the above embodiments.

[0084] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0085] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute the autonomous driving road test problem analysis method as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0086] It should be noted that the embodiments of this application can also be further described in the following ways: An autonomous driving road test problem analysis method includes: performing a first multi-dimensional problem mining based on target road test data to obtain multiple first-dimensional problem labels and corresponding related road test data; the first multi-dimensional includes multiple first dimensions; the target road test data is obtained based on road test data of the current driving mileage of a target vehicle fleet on a target route; performing a second multi-dimensional data mining on the corresponding related road test data according to the multiple first-dimensional problem labels to obtain multiple second-dimensional analyzable information; the second multi-dimensional includes multiple second dimensions; obtaining the current risk value of the multiple second dimensions according to the analyzable information of the multiple second dimensions and corresponding analysis rules; and predicting the risk value of the multiple second dimensions for future driving mileage on the target route according to the current risk value of the multiple second dimensions and the corresponding historical risk value.

[0087] Optionally, based on the target road test data, a first multi-dimensional problem mining is performed to obtain multiple first-dimensional problem labels and corresponding related road test data, including: the first dimension includes multiple levels of dimensions; for each first dimension, the process includes: performing problem mining on the target road test data according to the problem mining algorithm associated with the current level dimension to determine whether the target road test data has a problem in the current level dimension; the current level dimension is initially the first-level dimension; when the target road test data has a problem in the current level dimension and the current level dimension has a next-level dimension, the next-level dimension is taken as the current level dimension, and the problem mining process is repeated until the current level dimension has no problem or no next-level dimension exists; based on the problem labels of different level dimensions and the corresponding related road test data, the problem label of the current first dimension and the corresponding related road test data are obtained.

[0088] Optionally, based on the multiple first-dimensional problem labels, second multi-dimensional data mining is performed on the corresponding relevant road test data to determine the analyzable information of multiple second dimensions, including: the first dimension and the second multi-dimensional data are in one-to-one correspondence, and the second dimension includes multi-level indicators; for each second dimension corresponding to each first dimension, the process includes: performing data mining on the relevant road test data of the current first dimension based on the data mining algorithm associated with each level of indicators of the current second dimension to obtain the analyzable information of each level of indicators of the current second dimension.

[0089] Optionally, based on the data mining algorithm for the association of indicators at all levels in the current second dimension, data mining is performed on the relevant road test data of the current first dimension to obtain analyzable information of the indicators at all levels in the current second dimension, including: extracting relevant data from the relevant road test data of the current first dimension according to the data type required in the data mining algorithm for the association of current indicators; and analyzing the extracted data according to the data mining process in the data mining algorithm for the association of current indicators to obtain analyzable information of the current indicators.

[0090] Optionally, obtaining the current risk value of the plurality of second dimensions based on the analyzable information of the plurality of second dimensions and the corresponding analysis rules includes: the second dimension includes a plurality of analyzable information, and for each analyzable information of each second dimension, the process includes: obtaining the current risk value of the current analyzable information based on the current analyzable information and the corresponding analysis rules; the analysis rules include comparing the current analyzable information with a corresponding threshold and calculating the degree to which it exceeds or falls below the threshold; obtaining a score for the current analyzable information based on the degree of influence of the current analyzable information on the current second dimension; and obtaining the current risk value of the current second dimension based on the current risk values ​​of the plurality of analyzable information of the current second dimension and the corresponding scores.

[0091] Optionally, the method further includes: performing data cleaning on the road test data of the target fleet's current mileage on the target route based on the operational design domain of the target route, to obtain target road test data.

[0092] Optionally, the method further includes: statistically analyzing the problem labels of the plurality of first dimensions; in response to the target route input by the user, displaying the statistical results, the current risk values ​​of the plurality of second dimensions, and / or the risk values ​​of future mileage in the target road segment; and / or, based on the current risk values ​​of the plurality of second dimensions and the corresponding risk thresholds, calculating the proportion of the second dimensions exceeding the risk thresholds to obtain the route quality of the target fleet when driving with the current version of the intelligent driving system; and obtaining the adjustment strategy of the current version of the intelligent driving system based on the route quality, the current risk values ​​of the plurality of second dimensions, and the risk values ​​of future mileage in the target road segment.

[0093] An autonomous driving road test problem analysis device includes: a problem mining module, used to perform first multi-dimensional problem mining based on target road test data to obtain multiple first-dimensional problem labels and corresponding related road test data; the first multi-dimensional includes multiple first dimensions; the target road test data is obtained based on road test data of the current driving mileage of a target vehicle fleet on a target route; a data mining module, used to perform second multi-dimensional data mining on the corresponding related road test data according to the multiple first-dimensional problem labels to obtain multiple second-dimensional analyzable information; the second multi-dimensional includes multiple second dimensions; and a risk prediction module, used to obtain the current risk value of the multiple second dimensions according to the analyzable information of the multiple second dimensions and corresponding analysis rules; and to predict the risk value of the multiple second dimensions for future driving mileage on the target route according to the current risk value of the multiple second dimensions and corresponding historical risk values.

[0094] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the autonomous driving road test problem analysis method.

[0095] A non-transitory computer-readable storage medium stores computer instructions for causing a computer to execute the autonomous driving road test problem analysis method.

[0096] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of this application (including the claims) is limited to these examples; within the framework of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of this application as described above, which are not provided in the details for the sake of brevity.

[0097] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of this application, the well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of this application, and this also takes into account the fact that the details of the implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of this application will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuits) have been set forth to describe exemplary embodiments of this application, it will be apparent to those skilled in the art that the embodiments of this application can be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0098] Although this application has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0099] The embodiments of this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments of this application should be included within the protection scope of this application.

Claims

1. A method for analyzing problems in autonomous driving road tests, comprising: Based on the target road test data, the first multi-dimensional problem mining is performed to obtain multiple first-dimensional problem labels and corresponding related road test data; The first multi-dimensional dimension includes multiple first dimensions; the target road test data is obtained based on the road test data of the target vehicle fleet's current mileage on the target route; Based on the multiple first-dimensional problem labels, the corresponding relevant road test data are subjected to second-dimensional data mining to obtain multiple second-dimensional analyzable information; the second multi-dimensional includes multiple second dimensions; Based on the analyzable information of the multiple second dimensions and the corresponding analysis rules, the current risk value of the multiple second dimensions is obtained; Based on the current risk values ​​and corresponding historical risk values ​​of the multiple second dimensions, the risk values ​​of the multiple second dimensions for future mileage on the target route are predicted.

2. The method according to claim 1, characterized in that, Based on the target road test data, a first-dimensional problem mining was performed to obtain multiple first-dimensional problem labels and corresponding related road test data, including: The first dimension includes multiple levels of dimensions. For each first dimension, the execution process includes: Based on the problem mining algorithm associated with the current level dimension, problem mining is performed on the target road test data to determine whether there are problems in the target road test data in the current level dimension; the current level dimension is initially the first level dimension. When the target road test data has a problem in the current level dimension and there is a next level dimension in the current level dimension, the next level dimension is taken as the current level dimension, and the problem mining process is repeated until there is no problem in the current level dimension or there is no next level dimension. Based on the issue labels of different levels and dimensions and the corresponding road test data, we obtain the issue label of the current first dimension and the corresponding road test data.

3. The method according to claim 1, characterized in that, Based on the multiple first-dimensional problem labels, second-dimensional data mining is performed on the corresponding relevant road test data to determine multiple analyzable information in the second dimension, including: The first dimension corresponds one-to-one with the second multi-dimensional dimension, and the second dimension includes multi-level indicators; for each second dimension corresponding to each first dimension, the execution process includes: Based on the data mining algorithm that correlates the indicators at each level in the current second dimension, data mining is performed on the relevant road test data in the current first dimension to obtain analyzable information on the indicators at each level in the current second dimension.

4. The method according to claim 3, characterized in that, Based on the data mining algorithm for the correlation of indicators at each level in the current second dimension, data mining is performed on the relevant road test data in the current first dimension to obtain analyzable information for each level of indicators in the current second dimension, including: Based on the data type required in the data mining algorithm associated with the current level indicator, relevant data is extracted from the relevant road test data of the current first dimension. The extracted data is analyzed according to the data mining process in the data mining algorithm associated with the current level indicator to obtain analyzable information of the current level indicator.

5. The method according to claim 1, characterized in that, Based on the analyzable information of the multiple second dimensions and the corresponding analysis rules, the current risk values ​​of the multiple second dimensions are obtained, including: The second dimension includes multiple analyzable pieces of information. For each analyzable piece of information in each second dimension, the execution process includes: Based on the currently analyzable information and the corresponding analysis rules, the current risk value of the currently analyzable information is obtained; the analysis rules include comparing the current analyzable information with the corresponding threshold and calculating the degree to which it exceeds or falls below the threshold; The score of the currently analyzable information is obtained based on the degree of influence of the currently analyzable information on the current second dimension; The current risk value of the current second dimension is obtained based on the current risk value and corresponding score of multiple analyzable information in the current second dimension.

6. The method according to claim 1, characterized in that, The method further includes: Based on the operational design domain of the target route, the road test data of the target vehicle fleet at the current mileage on the target route is cleaned to obtain the target road test data.

7. The method according to any one of claims 1-6, characterized in that, The method further includes: Statistical analysis was performed on the multiple first-dimensional problem labels; In response to the user's input of the target route, display statistical results, the current risk values ​​of the multiple second dimensions, and / or the risk values ​​of future mileage in the target road segment; And / or, Based on the current risk values ​​of the multiple second dimensions and the corresponding risk thresholds, the proportion of the second dimension exceeding the risk thresholds is calculated to obtain the route quality of the target fleet when driving with the current version of the intelligent driving system. The adjustment strategy for the current version of the intelligent driving system is obtained based on the route quality, the current risk values ​​of the multiple second dimensions, and the risk value of future driving mileage in the target road segment.

8. An autonomous driving road test problem analysis device, comprising: The issue mining module is used to perform first-dimensional issue mining based on the target road test data, and obtain multiple first-dimensional issue labels and corresponding related road test data. The first multi-dimensional dimension includes multiple first dimensions; the target road test data is obtained based on the road test data of the target vehicle fleet's current mileage on the target route; The data mining module is used to perform second multi-dimensional data mining on the corresponding relevant road test data based on the multiple first-dimensional problem labels to obtain multiple second-dimensional analyzable information; the second multi-dimensional includes multiple second dimensions; The risk prediction module is used to obtain the current risk value of the multiple second dimensions based on the analyzable information of the multiple second dimensions and the corresponding analysis rules; Based on the current risk values ​​and corresponding historical risk values ​​of the multiple second dimensions, the risk values ​​of the multiple second dimensions for future mileage on the target route are predicted.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method as claimed in any one of claims 1 to 8.

10. A non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the method of any one of claims 1 to 8.

Images