Drive test problem analysis method and device, storage medium and program product
By clustering the road test logs, the target road test logs are output to analyze the road test problems, the problem of low analysis efficiency of road test problems is solved, and a more efficient and comprehensive one-time analysis is achieved.
Patent Information
- Application Number
- CN202411952020.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-27
- Publication Date
- 2025-04-08
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
During the road test, the number of abnormal problems on the vehicle end is huge, resulting in low manual verification efficiency and ineffective improvement of the efficiency of road test analysis.
By obtaining the road test log collection, using the smart driving module and smart driving data to cluster the road test logs, generate target clusters, and output the target road test log to analyze the road test problems.
Through clustering technology, the centralized analysis of similar road test problems has been improved, the efficiency of road test problems has been reduced, and the repeated analysis of similar problems has been achieved, and more comprehensive and efficient road test problems have been achieved.
Smart Images
Figure CN119782857A_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the field of intelligent driving technology, and in particular, relates to a road test problem analysis method, device, storage medium and program product. Background Art
[0002] At present, in vehicle driving, assisted driving, automatic driving and other related technologies, road tests (hereinafter referred to as road tests) can be carried out on a large scale in the field through vehicle driving to discover possible problems of the vehicle while driving on the road (hereinafter referred to as abnormal problems on the vehicle side). The performance of the vehicle can be continuously optimized through the problems fed back during the road test to ensure the driving safety of the vehicle.
[0003] During generalized road testing, a large number of abnormal vehicle-side problems will be fed back. These problems will flood into the background, and each problem will be manually checked, resulting in low efficiency in road test problem analysis. Summary of the invention
[0004] The embodiments of the present application provide a drive test problem analysis method, device, storage medium, and program product, which can solve the problem of low efficiency in drive test problem analysis.
[0005] In a first aspect, an embodiment of the present application provides a method for analyzing a drive test problem, the method comprising:
[0006] Obtain a set of drive test logs in a first area; wherein the set of drive test logs includes at least one drive test log corresponding to a vehicle-side abnormal problem, and each of the drive test logs corresponding to the vehicle-side abnormal problem includes the vehicle-side abnormal problem, first intelligent driving data associated with the vehicle-side abnormal problem, and an intelligent driving module; the intelligent driving module associated with the vehicle-side abnormal problem is determined based on the vehicle-side abnormal problem and the first intelligent driving data associated with the vehicle-side abnormal problem; the first intelligent driving data associated with the vehicle-side abnormal problem includes status data and / or environmental data of the vehicle when the vehicle-side abnormal problem occurs;
[0007] Clustering the drive test logs in the drive test log set according to the intelligent driving module and the first intelligent driving data to obtain at least one target cluster; wherein each of the target clusters includes at least one drive test log in the drive test log set, and the intelligent driving module of the drive test logs in each of the target clusters is the same;
[0008] Outputting a target drive test log in each of the target clusters; wherein the target drive test log is used to analyze the drive test problem in the first area, and the target drive test log includes at least part of the drive test logs in the target cluster.
[0009] In a second aspect, an embodiment of the present application provides a drive test problem analysis device, the device comprising:
[0010] A first acquisition module is used to acquire a set of drive test logs in a first area; wherein the set of drive test logs includes at least one drive test log corresponding to a vehicle-side abnormal problem, and each of the drive test logs corresponding to the vehicle-side abnormal problem includes the vehicle-side abnormal problem, first intelligent driving data associated with the vehicle-side abnormal problem, and an intelligent driving module; the intelligent driving module associated with the vehicle-side abnormal problem is determined based on the vehicle-side abnormal problem and the first intelligent driving data associated with the vehicle-side abnormal problem; the first intelligent driving data associated with the vehicle-side abnormal problem includes status data and / or environmental data of the vehicle when the vehicle-side abnormal problem occurs;
[0011] a second acquisition module, configured to cluster the drive test logs in the drive test log set according to the intelligent driving module and the first intelligent driving data to obtain at least one target cluster; wherein each of the target clusters includes at least one drive test log in the drive test log set, and the intelligent driving module of the drive test logs in each of the target clusters is the same;
[0012] An output module is used to output a target drive test log in each of the target clusters; wherein the target drive test log is used to analyze the drive test problem in the first area, and the target drive test log includes at least part of the drive test logs in the target cluster.
[0013] In a third aspect, an embodiment of the present application provides a drive test problem analysis device, the device comprising:
[0014] a processor and a memory storing computer program instructions;
[0015] When the processor executes the computer program instructions, the drive test problem analysis method as described in the first aspect is implemented.
[0016] In a fourth aspect, an embodiment of the present application provides a computer storage medium, wherein the computer readable storage medium stores computer program instructions, and when the computer program instructions are executed by a processor, the drive test problem analysis method as described in the first aspect is implemented.
[0017] In a fifth aspect, an embodiment of the present application provides a computer program product. When instructions in the computer program product are executed by a processor of an electronic device, the electronic device executes the drive test problem analysis method as described in the first aspect.
[0018] In an embodiment of the present application, after obtaining a set of drive test logs in the first area, the drive test logs in the set of drive test logs can be clustered according to the first intelligent driving data and intelligent driving module associated with the vehicle-side abnormal problems in the drive test logs corresponding to each vehicle-side abnormal problem in the set of drive test logs to obtain at least one target cluster, wherein the intelligent driving module of the drive test logs in each target cluster is the same; then the target drive test logs in each target cluster are output for analyzing the drive test problems in the first area, and the target drive test logs include at least part of the drive test logs of the target cluster. In this way, by clustering the drive test logs in the set of drive test logs, and then selecting the target drive test logs from each cluster for analysis, the concentration of the analysis of similar drive test problems can be enhanced compared to analyzing one by one without rules, thereby improving the efficiency of the analysis of the drive test problems. Furthermore, if the target drive test log only includes part of the drive test logs of the target cluster, the embodiment of the present application can not only realize the comprehensive analysis of the drive test problems, but also reduce the repeated analysis of similar drive test problems, thereby further improving the efficiency of the analysis of the drive test problems. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] In order to more clearly illustrate the technical solution of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.
[0020] Figure 1 This is one of the flow charts of the drive test problem analysis method provided in the embodiment of the present application;
[0021] Figure 2 is a flowchart of processing vehicle-side abnormal information provided by an embodiment of the present application;
[0022] Figure 3 This is the second flow chart of the drive test problem analysis method provided in the embodiment of the present application;
[0023] Figure 4a It is one of the schematic diagrams of the intelligent driving module provided in the embodiment of the present application;
[0024] Figure 4b This is the second schematic diagram of the intelligent driving module provided in the embodiment of the present application;
[0025] Figure 4c This is the third schematic diagram of the intelligent driving module provided in the embodiment of the present application;
[0026] Figure 5 is a structural diagram of a drive test problem analysis device provided in an embodiment of the present application;
[0027] Figure 6 It is a structural diagram of a drive test problem analysis device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0028] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by illustrating the examples of the present application.
[0029] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0030] The drive test problem analysis method provided by the embodiment of the present application is described in detail below through some embodiments and their application scenarios in conjunction with the accompanying drawings.
[0031] The drive test problem analysis method of the embodiment of the present application can be applied to any one or more execution entities in a terminal, a server, a service platform, a cloud, a distributed system, an Internet of Things, an Internet of Vehicles system, etc. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, or a desktop computer, etc. For the convenience of description, the following example is taken as an example of executing the drive test problem analysis method of the embodiment of the present application in the cloud, but the execution entity of the drive test problem analysis method is not limited thereby.
[0032] See also Figure 1 , Figure 1 This is one of the flow charts of the drive test problem analysis method provided in the embodiment of the present application.
[0033] like Figure 1 As shown, the drive test problem analysis method may include the following steps:
[0034] Step 101. Obtain a set of road test logs in the first area; wherein the set of road test logs includes at least one road test log corresponding to a vehicle-side abnormal problem, and each road test log corresponding to the vehicle-side abnormal problem includes the vehicle-side abnormal problem, the first intelligent driving data associated with the vehicle-side abnormal problem, and an intelligent driving module; the intelligent driving module associated with the vehicle-side abnormal problem is determined based on the vehicle-side abnormal problem and the first intelligent driving data associated with the vehicle-side abnormal problem; the first intelligent driving data associated with the vehicle-side abnormal problem includes status data and / or environmental data of the vehicle when the vehicle-side abnormal problem occurs.
[0035] The first area may be any drive test area, which is an area that needs to be tested (i.e., an area used for drive testing). The embodiment of the present application does not limit the division range of the drive test area. In some embodiments, the first area may be a country, a province, a city, a county, etc., which may be determined based on actual needs.
[0036] It should be noted that the embodiments of the present application do not limit the driving type of the vehicle undergoing the road test. In some embodiments, the driving type of the vehicle may be automatic driving, assisted driving, or remote control driving, etc. If the driving type of the vehicle is automatic driving, it may be referred to as an automatic driving vehicle.
[0037] When the vehicle is conducting a road test in the first area, if a vehicle-side abnormality occurs, the vehicle-side abnormality information can be recorded and uploaded to the cloud, so that the cloud can store all the vehicle-side abnormality information that occurs during the road test within the first area, so that relevant personnel can analyze the road test problems in the first area based on the stored information, thereby optimizing the performance of the vehicle and improving the driving safety of the vehicle.
[0038] The above-mentioned vehicle-side abnormal information may include but is not limited to at least one of the following: abnormal problems on the vehicle side (also referred to as abnormal performance on the vehicle side); intelligent driving data of the vehicle when the vehicle-side abnormality is discovered, wherein the intelligent driving data of the vehicle may include vehicle status data and / or vehicle environmental data.
[0039] In some embodiments, if the vehicle exhibits any of the following performances or any combination of performances during road testing, it can be regarded as a vehicle-side abnormality: driving on the line, abnormal lane change, abnormal steering, function exit, abnormal speed, abnormal longitudinal braking, abnormal lateral swing, collision between the vehicle and other dynamic traffic participants, collision between the vehicle and static road components, and lane change across a lane line that cannot be crossed. In other words, the vehicle-side abnormality problem may include but is not limited to at least one of the following: driving on the line, abnormal lane change, abnormal steering, function exit, abnormal speed, abnormal longitudinal braking, abnormal lateral swing, collision between the vehicle and other dynamic traffic participants, collision between the vehicle and static road components, and lane change across a lane line that cannot be crossed.
[0040] In some embodiments, the vehicle status data may include, but is not limited to, at least one of the following data of the vehicle when the vehicle-side abnormality occurs: recorded timestamp, location coordinates (Loc), vehicle-side, vehicle speed (velocity), direction (head), posture, matched map element (MapID), etc. In some optional implementations, the vehicle speed, direction, posture, etc. when the vehicle-side abnormality occurs can be extracted based on the timestamp and / or location coordinates recorded by the vehicle when the vehicle-side abnormality occurs, but is not limited to this.
[0041] In some embodiments, the environmental data of the vehicle may include, but is not limited to, at least one of the following data of the vehicle when a vehicle-side abnormality occurs: the road scene (also referred to as an intelligent driving scene, a traffic scene or a scene), and the relationship with other vehicles. In some optional implementations, the road scene in which the vehicle is located when a vehicle-side abnormality occurs can be determined based on a query map (which may be, but is not limited to, a high-precision (HD) map) based on the position coordinates of the vehicle when the vehicle-side abnormality occurs, such as: the road type of the road to which this position coordinate belongs in the map can be determined as the road scene in which the vehicle is located. In some optional implementations, the relationship between the vehicle and other vehicles can be determined based on a query map based on the position coordinates of the vehicle when the vehicle-side abnormality occurs, or it can be determined by the vehicle's perception module.
[0042] The embodiments of the present application do not limit the form of presentation of the road scene. In some embodiments, the road scene may include but is not limited to at least one of the following: a road under an elevated road, a non-layered road, a road on an elevated road, an underground tunnel, a mountain tunnel, an above-ground tunnel, a cross-river tunnel, a main road to a ramp, a ramp merging into the main road, a road within a ramp, etc.
[0043] In addition, the embodiments of the present application do not limit the way in which the vehicle detects vehicle-side abnormalities. In some embodiments, the vehicle can autonomously detect vehicle-side abnormalities, such as: driving on the line, collision, changing lanes across a lane line that cannot be crossed, and other vehicle-side abnormalities can be discovered through the vehicle's own location information and the layout relationship of the map; speed abnormalities, longitudinal braking abnormalities, lateral swing abnormalities, and other vehicle-side abnormalities can be discovered through the detection parameters of the vehicle's own sensors. In other embodiments, the vehicle can discover vehicle-side abnormalities by interacting with testers, such as: testers can pay attention to the self-driving behavior of the vehicle during the road test. If the self-driving behavior is found to be abnormal (i.e., a vehicle-side abnormality occurs), they can interact with the vehicle and inform the vehicle of the vehicle-side abnormality, so that the vehicle can record the vehicle-side abnormal information in a timely manner and upload the vehicle-side abnormal information to the cloud.
[0044] It should be noted that the embodiment of the present application does not limit the reporting format of the above-mentioned vehicle-side abnormal information.
[0045] In some embodiments, the above-mentioned vehicle-side abnormal information can be reported to the cloud in an information format. In these embodiments, after receiving the vehicle-side abnormal information, the cloud can organize the above-mentioned vehicle-side abnormal information and generate a road test log based on each vehicle-side abnormal information. It can be understood that each vehicle-side abnormal information corresponds to a road test log, and each road test log corresponds to a vehicle-side abnormal problem. In actual applications, the road test log corresponding to the vehicle-side normal information may include part or all of the information in the vehicle-side abnormal information, which can be determined according to actual needs, and the embodiments of the present application are not limited to this.
[0046] In other embodiments, the above-mentioned vehicle-side abnormal information can be reported to the cloud in a log format. In these embodiments, the vehicle-side abnormal information recorded by the vehicle can be first converted into a road test log and then reported to the cloud. In this way, the operating burden of the cloud can be reduced.
[0047] In an embodiment of the present application, the road test log corresponding to the vehicle-side abnormal problem includes the vehicle-side abnormal problem and its associated intelligent driving data (i.e., the first intelligent driving data) and the intelligent driving module.
[0048] The intelligent driving data in the road test log corresponding to the vehicle-side abnormal problem may include part or all of the intelligent driving data in the vehicle-side abnormal information corresponding to the road test log.
[0049] The intelligent driving module associated with the abnormal problem on the vehicle side can be understood as: the intelligent driving module that causes the abnormal problem on the vehicle side, that is, the preliminary inferred cause of the abnormality on the vehicle side. The intelligent driving module can also be called intelligent driving behavior or intelligent driving action, which can be understood as: a functional point in intelligent driving. The embodiments of the present application do not limit the specific form of expression of the intelligent driving module. In some embodiments, the intelligent driving module may, but is not limited to, include at least one of the following: high-precision positioning, trajectory planning, speed planning, behavior decision-making, fusion perception, lateral control, longitudinal control, inter-module signal transmission, map broadcasting, map reconstruction, etc.
[0050] In an embodiment of the present application, the intelligent driving module associated with the vehicle-side abnormal problem can be determined based on the vehicle-side abnormal problem and its associated intelligent driving data. For details, please refer to the following relevant description and will not be described here.
[0051] It is worth noting that the embodiments of the present application do not limit the execution device of the determination step of the intelligent driving module associated with the abnormal problem on the vehicle side.
[0052] In some embodiments, the intelligent driving module associated with the vehicle-side abnormal problem can be determined by the vehicle based on the vehicle-side abnormal problem and its associated intelligent driving data. In these embodiments, the vehicle-side abnormal information can further include the intelligent driving module associated with the vehicle-side abnormal problem.
[0053] In other embodiments, the intelligent driving module associated with the vehicle-side abnormal problem can be determined by the cloud based on the vehicle-side abnormal problem and its associated intelligent driving data. In these embodiments, the intelligent driving module in the drive test log can be added to the drive test log by the cloud.
[0054] Step 102: Cluster the drive test logs in the drive test log set according to the intelligent driving module and the first intelligent driving data to obtain at least one target cluster; wherein each of the target clusters includes at least one drive test log in the drive test log set, and the intelligent driving module of the drive test logs in each of the target clusters is the same.
[0055] After the cloud obtains the set of drive test logs in the first area, it can cluster (also called classify) the drive test logs in the set of drive test logs based on the intelligent driving modules and the first intelligent driving data included in each road test log in the set of drive test logs to obtain at least one target cluster, wherein the intelligent driving modules in the road test logs included in each target cluster are the same, and further, the first intelligent driving data can be similar. Since the intelligent driving modules included in each target cluster are the same, that is, each target cluster corresponds to only one intelligent driving module, it can be understood that the target cluster corresponds to the intelligent driving module one-to-one, that is, one intelligent driving module corresponds to one target cluster, and therefore, at least one target cluster can be understood as: a target cluster corresponding to at least one intelligent driving module.
[0056] The embodiment of the present application does not limit the method of clustering the drive test logs based on the intelligent driving module and the first intelligent driving data.
[0057] In some embodiments, a clustering algorithm may be used to cluster the road test log based on the intelligent driving module and the first intelligent driving data. Further, the clustering algorithm may be used for classification using any of the following: Support Vector Machines (SVM), random forest, or other deep learning networks.
[0058] In other embodiments, the similarity of the first intelligent driving data in different road test logs can be calculated, and then the first intelligent driving data whose similarity is greater than a preset threshold and the road test logs with the same intelligent driving module can be grouped into a target cluster.
[0059] Step 103: output the target drive test log in each of the target clusters; wherein the target drive test log is used to analyze the drive test problem in the first area, and the target drive test log includes at least part of the drive test logs in the target cluster.
[0060] For road test logs with the same intelligent driving module and similar intelligent driving data, it can be considered that the road test problems they report are similar or identical. In this way, by clustering the road test logs in the road test log set through the intelligent driving modules and intelligent driving data of the road test logs, the same or similar road test problems (i.e., the same type of road test problems) can be integrated into a target cluster. In this way, for each target cluster, even if only part of the test logs are analyzed, accurate analysis of such road test problems can be achieved, thereby ensuring the reliability of road test information processing.
[0061] Based on this, in an embodiment of the present application, after acquiring at least one target cluster, for each target cluster, a target drive test log output may be selected for manual verification and analysis.
[0062] In some embodiments, the target drive test log may include all drive test logs of the target cluster. In these embodiments, although all drive test logs in the drive test log set are still analyzed, compared with analyzing one by one without rules, outputting the drive test logs in the target cluster as a unit for drive test problem analysis can enhance the centralization of the analysis of similar drive test problems, thereby improving the efficiency of drive test problem analysis.
[0063] In other embodiments, the target drive test log may include a portion of the drive test logs of the target cluster. In these embodiments, compared with the embodiment in which the target drive test log includes all the drive test logs of the target cluster, it is further possible to achieve a comprehensive analysis of the drive test problem and reduce repeated analysis of similar drive test problems, thereby improving the efficiency of drive test information processing.
[0064] In some embodiments, when outputting the drive test log, the output mode of the drive test log may be determined based on the problem verification division mode. The problem verification division mode may include but is not limited to at least one of the following:
[0065] Problem verification is divided according to the task method, that is, during problem verification, the road test logs corresponding to different intelligent driving modules are assigned to different personnel for verification and analysis. It can be understood that in this method, one person is responsible for the verification of the road test logs corresponding to one intelligent driving module, and these road test logs may come from one road test area or from different road test areas.
[0066] Problem verification is divided by region, that is, during problem verification, the drive test logs corresponding to different regions are assigned to different personnel for verification and analysis. It can be understood that in this method, one person is responsible for the verification of the drive test logs in one region, and the intelligent driving modules corresponding to these drive test logs can be the same or different.
[0067] In these embodiments, if the problem verification division method is that the problem verification is divided according to the task method, a small amount of drive test logs extracted from different target clusters are respectively output to different verification personnel. If the problem verification division method is that the problem verification is divided according to the area method, a small amount of drive test logs extracted from different target clusters will be output to the same verification personnel. In this way, the flexibility of the drive test problem verification can be improved.
[0068] In some embodiments, the verification of the road test log by the verification personnel may specifically include: verifying whether the abnormal problems on the vehicle side, the first intelligent driving data, and the intelligent driving module in the road test log match. If they match, the verification passes; if they do not match, the verification fails, and the intelligent driving module in the road test log can be manually corrected so that the corrected intelligent driving module matches the abnormal problems on the vehicle side and the first intelligent driving data in the road test log. Afterwards, the verified road test log can be used to analyze the road test problems, thereby optimizing the performance of the vehicle and improving the driving safety of the vehicle.
[0069] In some optional implementations, the intelligent driving modules included in the verified road test logs can be regarded as intelligent driving modules with defects that can be optimized, and feedback can be given to developers to optimize the functions of these intelligent driving modules and thus optimize the performance of the vehicle.
[0070] Furthermore, when the number of these intelligent driving modules is greater than 1, the optimization priority of each of these intelligent driving modules can be further determined, and the optimization priority can be fed back to the developer, so that the developer can determine the optimization order of each intelligent driving module based on the optimization priority of each intelligent driving module. In some implementations, the optimization priority of each intelligent driving module can be positively correlated with the degree of influence of each intelligent driving module on safe driving. The higher the degree of influence of the intelligent driving module on safe driving, the higher the urgency of its optimization, so its optimization priority can be set higher, and vice versa. In other implementations, the optimization priority of each intelligent driving module can be positively correlated with the number of road test logs in the target cluster corresponding to each intelligent driving module. The more road test logs in the target cluster corresponding to the intelligent driving module, the more serious the problem of the intelligent driving module, so the optimization priority can be set higher, and vice versa. In this way, by providing a priority processing mechanism, the effectiveness of performance optimization of autonomous driving measurement can be improved.
[0071] The method for analyzing the drive test problems of this embodiment, after obtaining the set of drive test logs in the first area, can cluster the drive test logs in the set of drive test logs according to the first intelligent driving data and intelligent driving module associated with the abnormal vehicle-side problems in the drive test logs corresponding to each abnormal vehicle-side problem in the set of drive test logs, and obtain at least one target cluster, wherein the intelligent driving module of the drive test logs in each target cluster is the same; then output the target drive test logs in each target cluster for analyzing the drive test problems in the first area, and the target drive test logs include at least part of the drive test logs of the target cluster. In this way, by clustering the drive test logs in the set of drive test logs, and then selecting the target drive test logs from each cluster for analysis, in this way, compared with analyzing one by one without rules, the concentration of the analysis of the same type of drive test problems can be enhanced, thereby improving the efficiency of the analysis of the drive test problems. Furthermore, if the target drive test log only includes part of the drive test logs of the target cluster, the embodiment of the present application can not only realize the comprehensive analysis of the drive test problems, but also reduce the repeated analysis of the same type of drive test problems, thereby further improving the efficiency of the analysis of the drive test problems.
[0072] In an embodiment of the present application, when clustering the drive test logs to obtain target clusters, it is possible but not limited to performing secondary clustering on the drive test logs based on the intelligent driving modules and intelligent driving data in the drive test logs to obtain target clusters.
[0073] In some embodiments, clustering the drive test logs in the drive test log set according to the intelligent driving module and the first intelligent driving data to obtain at least one target cluster includes:
[0074] According to the intelligent driving module, a first clustering task is performed on the drive test logs in the drive test log set to obtain at least one initial cluster; wherein each of the initial clusters includes at least one drive test log in the drive test log set, and the intelligent driving module of the drive test logs in each of the initial clusters is the same;
[0075] For each initial cluster in the at least one initial cluster, a second clustering task is performed on the drive test logs in the initial cluster according to the first intelligent driving data to obtain at least one target cluster corresponding to the initial cluster.
[0076] In these embodiments, the first clustering task (i.e., initial classification) can be performed on the road test log based on the intelligent driving module to obtain the initial clusters corresponding to each intelligent driving module; then, for each initial cluster, the second clustering task (i.e., secondary classification) can be performed on the road test log based on the first intelligent driving data to obtain the sub-clusters under each initial cluster (i.e., the target clusters corresponding to each intelligent driving module). In this way, the road test logs corresponding to the vehicle-side abnormal problems associated with the same intelligent data and similar intelligent driving data can be accurately divided into a target cluster, thereby improving the reliability of the road test information processing.
[0077] In other embodiments, the road test log may be initially classified based on the first intelligent driving data, and then for each classification in the initial classification result, the road test log may be secondary classified based on the intelligent driving module to obtain the target cluster corresponding to each intelligent driving module. In this way, the method of obtaining the target cluster can be enriched.
[0078] The determination of the intelligent driving module in the embodiment of the present application is described in detail below.
[0079] In some embodiments, the intelligent driving module associated with the vehicle-side abnormal problem is: an intelligent driving module determined by using a problem analysis model (also referred to as a problem classification model) based on the vehicle-side abnormal problem and the first intelligent driving data associated with the vehicle-side abnormal problem;
[0080] Among them, the problem analysis model represents the corresponding relationship between the problem feedback label and the intelligent driving module (also known as a mapping relationship); the problem feedback label includes the associated vehicle-side abnormal problem and the second intelligent driving data, and the second intelligent driving data includes the state data and / or environmental data of the vehicle when the vehicle-side abnormal problem associated with the second intelligent driving data occurs. In other words, the problem feedback label can include m labels, where m is an integer greater than 1, and each data in the vehicle-side abnormal problem and the second intelligent driving data can be regarded as a label.
[0081] The second intelligent driving data has at least some of the same data as the first intelligent driving data. In actual applications, the data types included in the second intelligent driving data can be exactly the same as the data types included in the first intelligent driving data, or they can be partially the same. The specific settings can be based on actual needs, and the embodiments of the present application do not limit this. In this way, it can be ensured that the intelligent driving module associated with the abnormal problem on the vehicle side in the test log can be accurately determined based on the problem analysis model.
[0082] In addition, the embodiments of the present application do not limit the specific data type of the same intelligent driving data included in the first intelligent driving data and the second intelligent driving data.
[0083] In some embodiments, the first intelligent driving data and the second intelligent driving data may both include road scenes associated with vehicle-side abnormal problems. In these embodiments, the determination of the intelligent driving module associated with the vehicle-side abnormal problems in the test log can be achieved based on the vehicle-side abnormal problems in the test log and their corresponding road scenes. Since the intelligent driving modules enabled in different road scenes may be different, the reliability of the determination of the intelligent driving module can be improved by locating the intelligent driving module associated with the vehicle-side abnormal problems through the vehicle-side abnormal problems in the test log and their corresponding road scenes.
[0084] In other embodiments, the first intelligent driving data and the second intelligent driving data may both include the posture / orientation of the vehicle associated with the vehicle-side abnormal problem, the relationship with other vehicles, etc. In this way, the determination method of the intelligent driving module can be enriched, thereby improving the determination flexibility of the intelligent driving module.
[0085] It should be noted that the embodiments of the present application do not limit the specific form of expression of the problem analysis model.
[0086] In some embodiments, the problem analysis model may be expressed as a file, table, etc. that stores the correspondence between the problem feedback label and the intelligent driving module. In these embodiments, the problem analysis model may store at least one problem feedback label and its corresponding intelligent driving module. Based on the vehicle-side abnormal problem and the first intelligent driving data associated with the vehicle-side abnormal problem, the intelligent driving module determined using the problem analysis model can be specifically expressed as: the intelligent driving module queried from the problem analysis model based on the vehicle-side abnormal problem and the first intelligent driving data associated with the vehicle-side abnormal problem. In specific implementation, the vehicle-side abnormal problem and the first intelligent driving data associated with the vehicle-side abnormal problem in the test log can be matched with the problem classification label in the problem analysis model respectively, and the target problem classification label that matches the test log is found, and then the intelligent driving module corresponding to the target problem classification label is determined as the intelligent driving module associated with the vehicle-side abnormal problem in this test log.
[0087] In other embodiments, the problem analysis model may be a machine learning model that has learned the correspondence between the problem feedback label and the intelligent driving module. In these embodiments, based on the vehicle-side abnormal problem and the first intelligent driving data associated with the vehicle-side abnormal problem, the intelligent driving module determined by the problem analysis model may be specifically expressed as: the vehicle-side abnormal problem and the first intelligent driving data associated with the vehicle-side abnormal problem are input into the problem analysis model, and the problem analysis model outputs the intelligent driving module.
[0088] The embodiments of the present application do not limit the acquisition method of the problem analysis model, and the problem analysis model can be received, configured, or determined autonomously. For the autonomously determined acquisition method, in some embodiments, the problem analysis model can be determined based on the vehicle-side abnormal information set of the second area;
[0089] Among them, the vehicle-side abnormal information set includes at least one vehicle-side abnormal information, and each vehicle-side abnormal information includes an associated vehicle-side abnormal problem and the second intelligent driving data.
[0090] The second area can be any road test area. In practical applications, the second area can be completely identical or partially identical to the first area, or completely different areas, which can be determined according to actual conditions, and the embodiments of the present application do not limit this. It can be understood that if the second area and the first area are the same area, the problem analysis model can be determined based on the historical vehicle-side abnormal information of the first area, and then the intelligent driving module corresponding to the current vehicle-side abnormal information of the first area can be determined based on the problem analysis model.
[0091] If the problem analysis model is the aforementioned file, table, etc. that stores the correspondence between the problem feedback label and the intelligent driving module, it can be obtained by manually analyzing the vehicle-side abnormal information set in the second area.
[0092] If the problem analysis model is the aforementioned machine learning model that has learned the correspondence between the problem feedback label and the intelligent driving module, it can be obtained by training the machine learning model using the vehicle-side abnormal information set of the second area as training data. In some embodiments, the problem analysis model is obtained by performing the following processing:
[0093] Obtain a training sample set; the training sample set includes a plurality of training samples, each of the training samples includes one vehicle-side abnormal information in the vehicle-side abnormal information set and an intelligent driving module label corresponding to the vehicle-side abnormal information;
[0094] The machine learning model is trained using the training sample set until a training stop condition is met to obtain the problem analysis model.
[0095] In these embodiments, the vehicle-side abnormal information set of the second area can be manually analyzed to determine the intelligent driving module label corresponding to each vehicle-side abnormal information in the abnormal information set. Then, each vehicle-side abnormal information and its corresponding intelligent driving module label are used as a training sample to generate a training sample set. Afterwards, the training sample set can be used to train the machine learning model until the training stop condition is met to obtain a problem analysis model.
[0096] In specific implementation, for each training sample, the following steps can be performed respectively: input the training sample into the machine learning model to obtain the predictive intelligent driving module corresponding to the training sample; determine the loss function value of the machine learning model according to the predictive intelligent driving module and the intelligent driving module label of the training sample; when the loss function value does not meet the training stop condition, adjust the model parameters of the machine learning model to obtain an updated machine learning model, and return to input the training sample into the machine learning model to obtain the predictive intelligent driving module corresponding to the training sample, until the training stop condition is met to obtain the problem analysis model.
[0097] It should be noted that the problem analysis model, which is a machine learning model that has learned the correspondence between problem feedback labels and intelligent driving modules, can also be iteratively updated in subsequent applications. For example, after manually verifying the test log, if the verification finds abnormal problems on the vehicle side in the test log, and the first intelligent driving data and the intelligent driving module do not match, the intelligent driving module in the test log can be manually corrected and the corrected test logs can be used as new training samples to continue to iterate the problem analysis model. In this way, the accuracy of determining the intelligent driving module can be further improved, thereby reducing the amount of manual corrections and improving the efficiency and reliability of road test information processing.
[0098] By the above method, the problem analysis model can be determined based on the vehicle-side abnormal information set of the second area, thereby improving the determination autonomy of the problem analysis model. In addition, the determination method of the problem analysis model can be determined based on its corresponding expression form, which can further improve the determination reliability of the problem analysis model.
[0099] In an embodiment of the present application, in addition to determining the intelligent driving module associated with the abnormal problem on the vehicle side through the above-mentioned problem analysis model, the intelligent driving module associated with the abnormal problem on the vehicle side can also be determined manually, or the intelligent driving module associated with the abnormal problem on the vehicle side can be determined through a pre-set intelligent driving module determination rule. The specific determination can be made based on actual needs, and the embodiment of the present application is not limited to this.
[0100] The embodiment of the present application does not limit the number of drive test logs extracted from each target cluster. In some embodiments, the number of drive test logs extracted can be randomly determined. In other embodiments, the outputting of part of the drive test logs in each target cluster includes:
[0101] For each of the target clusters, k drive test logs are extracted from the target cluster; wherein k is preset, or k is determined based on the ratio of the number of drive test logs included in the target cluster to the total number of drive test logs included in the at least one target cluster; k is a positive integer;
[0102] Output the k drive test logs.
[0103] When k is preset, the number of drive test logs extracted from each target cluster is equal, namely k. k can be set based on demand, and this embodiment of the present application does not limit this.
[0104] For the case where k is determined based on the ratio of the number of drive test logs included in the target cluster to the total number of drive test logs included in at least one target cluster, the number of drive test logs extracted from each target cluster may be equal or different. For each target cluster, its corresponding k can be the ratio of the number of drive test logs it includes to the number of drive test logs included in all target clusters rounded up or rounded down. In this way, the number extracted from each target cluster can match the number of drive test logs it includes, thereby ensuring the verification amount of each target cluster.
[0105] By determining the number of drive test logs extracted from each target cluster in the above manner, the verification amount of each target cluster can be guaranteed and the flexibility of determining the number of drive test logs extracted from each target cluster can be improved.
[0106] It should be noted that the various optional embodiments or implementations introduced in the embodiments of the present application can be implemented in combination with each other or can be implemented separately if they do not conflict with each other, and the embodiments of the present application are not limited to this.
[0107] For ease of understanding, a specific embodiment is used as an example for illustration:
[0108] In this embodiment, analysis / model training can be first performed based on the problem feedback records of a certain area (such as City A) (i.e. the aforementioned vehicle-side abnormal information, which can be expressed as feedback problem log) to obtain multiple factors (causes, such as the following scenario + vehicle-side abnormal performance) that can reflect the feedback problems and the problem analysis model of intelligent driving behavior / functional point / module / intelligent driving module; and then, combined with the problem analysis model and log feature (i.e. the aforementioned first intelligent driving data) clustering method, the feedback problem log of the target area (such as City A / other areas outside City A) is subdivided to obtain sub-clusters under each intelligent driving module. The intelligent driving modules in each sub-cluster are the same and the multiple factors of the problem feedback are similar. Then, a small amount of problem feedback logs can be extracted from each sub-cluster for manual verification, thereby saving a lot of manpower.
[0109] The specific implementation may include the following steps:
[0110] Step 1: Before generalizing the drive test, design a general drive test plan and break down the problems that may be found in the drive test as detailed as possible according to the dimensions of intelligent driving modules and scenarios;
[0111] Step 2: Design a set of automated testing software and hardware equipment based on the overall test plan to facilitate the preliminary recording of problems during the drive test;
[0112] Step 3: Automatically transmit the problems recorded during the drive test to the internal anti-problem background (i.e., the cloud);
[0113] Step 4: Filter the feedback questions through the filtering method designed in this paper to reduce the repeated verification of repeated problem feedback by manual work, thereby improving the efficiency of overall manual verification;
[0114] The specific instructions are as follows:
[0115] 1. Test plan design.
[0116] The following table is based on the traffic scenario in which the test occurred, the abnormal performance of the vehicle, and the reasons or causes based on preliminary judgment based on experience: based on a target city / region, each problem log is analyzed to obtain a problem analysis model with the corresponding relationship between m labels (such as the following scenario classification and abnormal performance of the vehicle) and the intelligent driving module; the actual content of the intelligent driving module is similar to intelligent driving behavior / action. Finally, the problems caused by different intelligent driving behaviors (such as whether the abnormal performance of the vehicle and the scenario classification match the intelligent driving behavior) need to be manually verified. Intelligent driving behavior can be a functional point in intelligent driving, such as high-precision positioning, trajectory planning, speed planning, etc.
[0117] Among them, the aforementioned problem analysis model (as shown in Table 1 below) can be obtained by manually analyzing the data, or the obtained original data can be used as training data of the machine learning model to train the machine learning model to obtain the aforementioned problem analysis model.
[0118] Table 1: Problem analysis model obtained through data analysis methods.
[0119]
[0120]
[0121] The above three dimensions are defined as:
[0122] "Scene Classification: SceneClass(SC)",
[0123] "Vehicle-side abnormal behavior: VehicleAbnormalBehaviorType(VABT)",
[0124] "Autonomous Driving Module (ADM)",
[0125] "Record location: LocationOfAbnormalBehavior(LOAB)".
[0126] For example, if a vehicle suddenly sways sideways on a main road heading to a ramp in city A, the preliminary classification based on the above test is defined as: {SC-VABT-ADM-LOAB:"Main road heading to ramp (scenario)-abnormal lateral swing (abnormal performance on the vehicle side)-lateral control (intelligent driving module)-coordinate position"}.
[0127] The data format of the recorded mapping relationship can be: "scenario-abnormal performance on the vehicle side-intelligent driving module / behavior-vehicle position (coordinates)".
[0128] When used for preliminary classification in the future, the corresponding "intelligent driving mode / behavior" can be queried through the "scenario + abnormal performance on the vehicle side" in the log information.
[0129] The vehicle's driving log records time + location. "Time + location" can be used to extract the vehicle's posture / speed, etc. The location can be combined with a high-precision map to obtain the vehicle's current scene.
[0130] It should be noted that in the aforementioned problem analysis model, in addition to scene classification and abnormal performance on the vehicle side, other problem feedback labels can also be set as needed, such as the vehicle's posture / orientation, and its relationship with other vehicles.
[0131] 2. Automated drive test software design, for details, please refer to Figure 2 , Figure 2 The empirical model in is the aforementioned problem analysis model.
[0132] 3. Use the inverse problem to filter and clean. For details, please refer to Figure 3 In step 3, the steps after using the anti-question filtering module can be understood as the output processing of the aforementioned partial test log.
[0133] With the anti-filtering module, you can do the following:
[0134] 1) Initial log (i.e. the aforementioned drive test log) records:
[0135] f(X0)=LogPrintModule{loc0,velocity0,heading0,MapID0,SC0,VABT0,...,
[0136] ADM0};
[0137] f(X1)=LogPrintModule{loc1,velocity1,heading1,MapID1,SC1,VABT1,...,
[0138] ADM1};
[0139] f(X i)=LogPrintModule{loc i , velocity i , heading i , MapID i , S.C. i , VABT i , ...,
[0140] ADM i}.
[0141] in:
[0142] f(X i ) represents a log coordinate printed by the automated road test software for the road test, as well as certain vehicle parameters extracted from the point and preliminary cause classification results;
[0143] X i : Indicates the descriptive characteristics of log;
[0144] loc i : Indicates the coordinates of the log record;
[0145] velocity i : Indicates the speed of the vehicle at the time when the log is recorded;
[0146] heading i : Indicates the direction the vehicle was traveling when the leg was recorded;
[0147] MapID i : Indicates the map element matched by the log position;
[0148] SC i : Indicates the scene classification extracted from the log position through the map;
[0149] VABT i : Indicates the abnormal performance classification of the vehicle during log recording;
[0150] ADM i : Indicates the preliminary cause classification inferred by the empirical model using the above parameters when recording the drive test log.
[0151] 2) Clustering.
[0152] The initial classification is performed based on the category of the preliminary cause classification result ADM in the log; that is, the information contained in each is processed to obtain the aforementioned M labels (such as scene classification, abnormal performance on the vehicle side, etc.), and the initial classification is performed based on the problem analysis model obtained by the aforementioned test scheme design to obtain the intelligent driving module corresponding to the m labels extracted from each log; Figure 4aIt shows the category of the intelligent module corresponding to each log position after the initial classification of each log.
[0153] And then Figure 1 Each category of intelligent driving module performs the following steps S31, S32 and S33:
[0154] Step S31: Based on the information items (such as location, timestamp, etc.) contained in each log and combined with the HD map (the location can be combined with the HD map to locate the location of the vehicle corresponding to the log and its relationship with other vehicles, etc.), extract the log features of each log in this class (such as extracting the vehicle's posture, direction, speed, distance, and relative position relationship with other vehicles). Figure 4b , illustrating the feature extraction of intelligent driving modules for the type of “map current changes”.
[0155] Step S32: Based on the log features of each log extracted in step S31, the log features of each class (classification of intelligent driving modules) are clustered (i.e., secondary classification). The clustering algorithm can use SVM, random forest or other deep learning networks for further classification. Different clustering methods have different feature distance calculation logics, such as Figure 4c , and gives the result of further feature clustering based on the same ADM for intelligent driving behaviors such as map broadcasting. Figure 4c The area covered by each color is a classification result of step S32. The causes of map broadcasting problems in each classification are similar, so only some log problem feedbacks in each classification need to be manually verified.
[0156] Step S33: Perform manual verification on the logs of each category after the secondary classification in step S32. The sampling method is flexible. For example, the logs of each category can be filtered according to fixed set parameters (quantity or other sampling logic) and assigned to professional modules for verification and investigation; or sampling can be performed based on the number of logs contained in different categories according to a preset proportion of the total amount for manual verification.
[0157] Through the above steps, a small part of each secondary log clustering result is extracted for manual problem verification. Therefore, overall, only a small part of the data needs to be verified manually, saving a lot of manpower costs. This not only realizes the comprehensive exposure of the problems, but also enables analysts to obtain more top-level problems, reducing the waste of time and manpower costs of repeated analysis.
[0158] Based on the drive test problem analysis method provided in the above embodiment, the present application also provides a specific implementation of a drive test problem analysis device. Please refer to the following embodiment.
[0159] See also Figure 5 , the drive test problem analysis device provided in the embodiment of the present application may include:
[0160] A first acquisition module 501 is used to acquire a set of drive test logs in a first area; wherein the set of drive test logs includes at least one drive test log corresponding to a vehicle-side abnormal problem, and each of the drive test logs corresponding to the vehicle-side abnormal problem includes the vehicle-side abnormal problem, the first intelligent driving data associated with the vehicle-side abnormal problem, and an intelligent driving module; the intelligent driving module associated with the vehicle-side abnormal problem is determined based on the vehicle-side abnormal problem and the first intelligent driving data associated with the vehicle-side abnormal problem; the first intelligent driving data associated with the vehicle-side abnormal problem includes the status data and / or environmental data of the vehicle when the vehicle-side abnormal problem occurs;
[0161] A second acquisition module 502 is used to cluster the drive test logs in the drive test log set according to the intelligent driving module and the first intelligent driving data to obtain at least one target cluster; wherein each of the target clusters includes at least one drive test log in the drive test log set, and the intelligent driving module of the drive test logs in each of the target clusters is the same;
[0162] The output module 503 is used to output the target drive test log in each of the target clusters; wherein the target drive test log is used to analyze the drive test problem in the first area, and the target drive test log includes at least part of the drive test logs in the target cluster.
[0163] The drive test problem analysis device provided in the embodiment of the present application can implement each process in the method embodiment, and will not be described again here to avoid repetition.
[0164] Figure 6 A schematic diagram of the hardware structure of the drive test information processing provided in an embodiment of the present application is shown.
[0165] The drive test problem analysis device may include a processor 601 and a memory 602 storing computer program instructions.
[0166] Specifically, the processor 601 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.
[0167] The memory 602 may include a large capacity memory for data or instructions. By way of example and not limitation, the memory 602 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In appropriate cases, the memory 602 may include a removable or non-removable (or fixed) medium. In appropriate cases, the memory 602 may be inside or outside the integrated gateway disaster recovery device. In a specific embodiment, the memory 602 is a non-volatile solid-state memory.
[0168] The memory may include a read-only memory (ROM), a random access memory (RAM), a magnetic disk storage medium device, an optical storage medium device, a flash memory device, an electrical, optical or other physical / tangible memory storage device. Therefore, generally, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to one aspect of the present disclosure.
[0169] The processor 601 implements any one of the drive test problem analysis methods in the above embodiments by reading and executing computer program instructions stored in the memory 602 .
[0170] In one example, the drive test problem analysis device may further include a communication interface 606 and a bus 610. Figure 6 As shown, the processor 601, the memory 602, and the communication interface 606 are connected via a bus 610 and communicate with each other.
[0171] The communication interface 606 is mainly used to implement communication between various modules, devices, units and / or equipment in the embodiments of the present application.
[0172] Bus 610 includes hardware, software, or both, coupling components of the drive test problem analysis device to each other. By way of example and not limitation,
[0173] The bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnection, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnection, a Low Pin Count (LPC) bus, a Memory Bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus or other suitable bus or a combination of two or more of these. Where appropriate, bus 610 may include one or more buses. Although the present application describes and illustrates a specific bus, the present application considers any suitable bus or interconnection.
[0174] In addition, in combination with the insulation resistance detection method in the above embodiment, the embodiment of the present application can provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when the computer program instructions are executed by a processor, any insulation resistance detection method in the above embodiment is implemented.
[0175] It should be clear that the present application is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present application.
[0176] The functional blocks shown in the above-described block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application-specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present application are programs or code segments used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link by a data signal carried in a carrier. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM, floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency links, etc. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.
[0177] It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, this application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiment, or in a different order from the embodiment, or several steps can be performed simultaneously.
[0178] Aspects of the present disclosure are described above with reference to the flowchart and / or block diagram of the method, device (system) and computer program product according to the embodiment of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device to produce a machine so that these instructions executed by the processor of the computer or other programmable data processing device enable the implementation of the function / action specified in one or more boxes of the flowchart and / or block diagram. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field programmable logic circuit. It can also be understood that each box in the block diagram and / or flowchart and the combination of boxes in the block diagram and / or flowchart can also be implemented by dedicated hardware that performs a specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions.
[0179] The above is only a specific implementation of the present application. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the protection scope of the present application is not limited to this. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in this application, and these modifications or replacements should be included in the protection scope of this application.
Claims
1. A method for analyzing a drive test problem, characterized in that: include: Obtain a set of drive test logs in a first area; wherein the set of drive test logs includes at least one drive test log corresponding to a vehicle-side abnormal problem, and each of the drive test logs corresponding to the vehicle-side abnormal problem includes the vehicle-side abnormal problem, first intelligent driving data associated with the vehicle-side abnormal problem, and an intelligent driving module; the intelligent driving module associated with the vehicle-side abnormal problem is determined based on the vehicle-side abnormal problem and the first intelligent driving data associated with the vehicle-side abnormal problem; the first intelligent driving data associated with the vehicle-side abnormal problem includes status data and / or environmental data of the vehicle when the vehicle-side abnormal problem occurs; Clustering the drive test logs in the drive test log set according to the intelligent driving module and the first intelligent driving data to obtain at least one target cluster; wherein each of the target clusters includes at least one drive test log in the drive test log set, and the intelligent driving module of the drive test logs in each of the target clusters is the same; Outputting a target drive test log in each of the target clusters; wherein the target drive test log is used to analyze the drive test problem in the first area, and the target drive test log includes at least part of the drive test logs in the target cluster.
2. The method according to claim 1, characterized in that The clustering the drive test logs in the drive test log set according to the intelligent driving module and the first intelligent driving data to obtain at least one target cluster includes: According to the intelligent driving module, a first clustering task is performed on the drive test logs in the drive test log set to obtain at least one initial cluster; wherein each of the initial clusters includes at least one drive test log in the drive test log set, and the intelligent driving module of the drive test logs in each of the initial clusters is the same; For each initial cluster in the at least one initial cluster, a second clustering task is performed on the drive test logs in the initial cluster according to the first intelligent driving data to obtain at least one target cluster corresponding to the initial cluster.
3. The method according to claim 1, characterized in that The intelligent driving module associated with the vehicle-side abnormal problem is: an intelligent driving module determined by using a problem analysis model based on the vehicle-side abnormal problem and the first intelligent driving data associated with the vehicle-side abnormal problem; Among them, the problem analysis model characterizes the correspondence between the problem feedback label and the intelligent driving module; the problem feedback label includes the associated vehicle-side abnormal problem and the second intelligent driving data, the second intelligent driving data includes the status data and / or environmental data of the vehicle when the vehicle-side abnormal problem associated with the second intelligent driving data occurs, and the second intelligent driving data and the first intelligent driving data have at least some data in common.
4. The method according to claim 3, characterized in that Both the first intelligent driving data and the second intelligent driving data include road scenes associated with abnormal problems on the vehicle side.
5. The method according to claim 3, characterized in that: The problem analysis model is determined based on a set of vehicle-side abnormal information in the second area; Among them, the vehicle-side abnormal information set includes at least one vehicle-side abnormal information, and each vehicle-side abnormal information includes an associated vehicle-side abnormal problem and the second intelligent driving data.
6. The method according to claim 5, characterized in that The problem analysis model is obtained by performing the following processing: Obtain a training sample set; the training sample set includes a plurality of training samples, each of the training samples includes one vehicle-side abnormal information in the vehicle-side abnormal information set and an intelligent driving module label corresponding to the vehicle-side abnormal information; The machine learning model is trained using the training sample set until a training stop condition is met to obtain the problem analysis model.
7. The method according to claim 1, characterized in that The outputting a portion of the drive test logs in each of the target clusters includes: For each of the target clusters, k drive test logs are extracted from the target cluster; wherein k is preset, or k is determined based on the ratio of the number of drive test logs included in the target cluster to the total number of drive test logs included in the at least one target cluster; k is a positive integer; Output the k drive test logs.
8. A drive test problem analysis device, characterized in that: The device comprises: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, the drive test problem analysis method according to any one of claims 1 to 7 is implemented.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed by a processor, the drive test problem analysis method according to any one of claims 1 to 7 is implemented.
10. A computer program product, characterized in that When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device executes the drive test problem analysis method according to any one of claims 1 to 7.