Abnormality detection method, device, vehicle, storage medium, and program product

CN122967371APending Publication Date: 2026-10-23CHERY INTELLIGENT VEHICLE TECH (HEFEI) CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610989474.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-03
Publication Date
2026-10-23

AI Technical Summary

Technical Problem

[0004]本申请提供一种异常的检测方法、装置、车辆、存储介质及程序产品,以解决相关技术中,人工分析效率低易漏信息,通用AI工具不适配车机日志、精度差且外网上传存在数据泄露风险,云端AI分析需搭建车联网与云端服务器,落地难度高等问题

Benefits of technology

[0020]本申请实施例可以先确定车辆运行过程中异常日志的实际文件体积和用户访问权限的实际校验结果,并依据实际文件体积和实际校验结果确定目标日志推理组件,进而提取业务特征信息,并根据业务特征信息确定异常成因、异常过程信息和/或异常佐证信息,从而生成对应的异常分析报告,通过结合日志文件体积与用户权限筛选适配推理组件,针对性提取日志业务特征推导车辆异常相关信息并生成报告,可适配不同大小日志、保障数据访问安全,精准高效完成车辆故障分析。由此,解决了相关技术中,人工分析效率低易漏信息,通用AI工具不适配车机日志、精度差且外网上传存在数据泄露风险,云端AI分析需搭建车联网与云端服务器,落地难度高等问题。

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122967371A_ABST
    Figure CN122967371A_ABST
Patent Text Reader

Abstract

The application relates to the technical field of data analysis, in particular to an abnormality detection method and device, a vehicle, a storage medium and a program product. The method comprises the following steps: collecting an abnormal log, and determining an actual file volume of the abnormal log and an actual verification result of user access authority; based on the actual file volume and the actual verification result, a target log reasoning component is selected from a plurality of candidate log reasoning components contained in a log analysis engine; business feature information of the abnormal log is extracted, abnormal cause information, abnormal process information and / or abnormal corroboration information of the vehicle are determined, and based on the abnormal cause information, the abnormal process information and / or the abnormal corroboration information, an abnormality analysis report is generated. Therefore, the problems in the prior art that manual analysis is low in efficiency and easy to miss information, a general AI tool is not suitable for vehicle log, precision is poor, data leakage risk exists in external network uploading, cloud AI analysis needs to be built on a vehicle network and a cloud server, and landing difficulty is high are solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of data analysis technology, and in particular to an anomaly detection method, device, vehicle, storage medium, and program product. Background Technology

[0002] In related technologies, log analysis methods are mainly divided into three categories: manual analysis, cloud-based AI (Artificial Intelligence) analysis, and general-purpose AI log parsing tools. Manual analysis involves technicians meticulously reviewing raw logs, combining their experience with vehicle operating conditions to manually analyze the fault sequence and root causes, and then identifying and writing log fault reports. Cloud-based AI analysis relies on dedicated large-scale models deployed in the cloud to receive uploaded logs in batches, perform structured transformation, intelligent event analysis, cross-validation of supplier reports, and output standardized fault analysis reports. General-purpose AI log parsing uses a general large-scale model interface to parse log text, enabling basic log content extraction and simple classification.

[0003] However, among the relevant technologies, manual analysis is inefficient and prone to missing information, general AI tools are not compatible with vehicle logs, have poor accuracy, and there is a risk of data leakage when uploading to external networks, while cloud-based AI analysis requires the construction of vehicle networks and cloud servers, which is difficult to implement and urgently needs improvement. Summary of the Invention

[0004] This application provides an anomaly detection method, device, vehicle, storage medium, and program product to solve the problems in related technologies, such as low efficiency and easy omission of information by manual analysis, unsuitability of general AI tools for vehicle logs, poor accuracy and risk of data leakage when uploaded to external networks, and high difficulty in implementation of cloud AI analysis which requires the construction of vehicle network and cloud server.

[0005] The first aspect of this application provides a method for detecting vehicle anomalies, comprising the following steps: collecting anomaly logs generated during vehicle operation, and determining the actual file size of the anomaly logs and the actual verification results of user access permissions; based on the actual file size and the actual verification results, selecting a target log inference component suitable for analyzing the anomaly logs from a plurality of candidate log inference components included in the log analysis engine; extracting business feature information of the anomaly logs using the target log inference component, determining the cause of the vehicle anomaly, anomaly process information and / or anomaly supporting information based on the business feature information, and generating an anomaly analysis report of the vehicle based on the cause of the anomaly, anomaly process information and / or anomaly supporting information.

[0006] Optionally, in one embodiment of this application, before selecting a target log inference component suitable for analyzing the abnormal log from a plurality of candidate log inference components included in the log analysis engine, the method further includes: obtaining the candidate file size of the abnormal log and the candidate verification result of the user access permissions; if the candidate file size is less than a first preset file size threshold and the candidate verification result is a successful verification, then the deployment environment of the log analysis engine is determined to be an intranet deployment environment, and first component information is determined based on the intranet deployment environment, and the candidate log inference component is determined based on the first component information; if the candidate file size is greater than or equal to the first preset file size threshold, or if the candidate verification result is not a successful verification, then the deployment environment of the log analysis engine is determined to be a local deployment environment, and second component information is determined based on the local deployment environment, and the candidate log inference component is determined based on the second component information.

[0007] Optionally, in one embodiment of this application, the step of extracting the business feature information of the abnormal log using the target log inference component includes: if the component information of the target log inference component is a first component information, and the actual file size is greater than a second preset file size threshold, then based on the abnormal log, determining at least one abnormal log sample fragment; extracting keywords corresponding to the abnormal log sample fragment, and constructing a knowledge base based on the keywords; determining a first actual keyword of the abnormal log based on the knowledge base, and determining the business feature information based on the first actual keyword.

[0008] Optionally, in one embodiment of this application, the step of extracting the business feature information of the abnormal log using the target log inference component includes: when the component information of the target log inference component is second component information, determining a target knowledge base matching the log feature based on at least one log feature of the abnormal log; determining a second actual keyword of the abnormal log based on the target knowledge base, and determining the business feature information based on the second actual keyword.

[0009] Optionally, in one embodiment of this application, before extracting the business feature information of the abnormal log using the target log inference component, the method further includes: determining the actual format of the abnormal log based on the log data of the abnormal log; if the actual format does not match the preset format, determining a conversion instruction to convert the abnormal log to the preset format based on the preset format, and converting the abnormal log to the preset format based on the conversion instruction.

[0010] Optionally, in one embodiment of this application, the method further includes: calculating the similarity between the anomaly analysis report and the target anomaly analysis report; if the similarity is greater than a preset similarity threshold, determining the initial anomaly handling strategy of the vehicle based on the target anomaly handling strategy of the target anomaly analysis report; determining a correction instruction for the initial anomaly handling strategy based on the anomaly analysis report, and correcting the initial anomaly handling strategy according to the correction instruction to obtain the actual anomaly handling strategy of the vehicle.

[0011] A second aspect of this application provides a vehicle anomaly detection device, comprising: a collection module for collecting anomaly logs generated during vehicle operation and determining the actual file size of the anomaly logs and the actual verification result of user access permissions; a selection module for selecting a target log inference component suitable for analyzing the anomaly logs from a plurality of candidate log inference components included in the log analysis engine based on the actual file size and the actual verification result; and a generation module for extracting business feature information from the anomaly logs using the target log inference component, determining the cause of the vehicle anomaly, anomaly process information, and / or anomaly supporting information based on the business feature information, and generating an anomaly analysis report of the vehicle based on the cause of the anomaly, anomaly process information, and / or anomaly supporting information.

[0012] Optionally, in one embodiment of this application, it further includes: an acquisition module, configured to acquire the candidate file size of the abnormal log and the candidate verification result of the user access permissions before selecting a target log inference component suitable for analyzing the abnormal log from a plurality of candidate log inference components included in the log analysis engine; a first determination module, configured to determine that the deployment environment of the log analysis engine is an intranet deployment environment when the candidate file size is less than a first preset file size threshold and the candidate verification result is a verification pass, and determine first component information based on the intranet deployment environment, and determine the candidate log inference component based on the first component information; a second determination module, configured to determine that the deployment environment of the log analysis engine is a local deployment environment when the candidate file size is greater than or equal to the first preset file size threshold, or when the candidate verification result is not a verification pass, and determine second component information based on the local deployment environment, and determine the candidate log inference component based on the second component information.

[0013] Optionally, in one embodiment of this application, the generation module includes: a first determining unit, configured to determine at least one abnormal log sample fragment based on the abnormal log if the actual file size is greater than a second preset file size threshold when the component information of the target log inference component is first component information; an extraction unit, configured to extract keywords corresponding to the abnormal log sample fragment and construct a knowledge base based on the keywords; and a second determining unit, configured to determine a first actual keyword of the abnormal log based on the knowledge base and determine the business feature information based on the first actual keyword.

[0014] Optionally, in one embodiment of this application, the generation module includes: a third determining unit, configured to determine a target knowledge base matching the log features based on at least one log feature of the abnormal log when the component information of the target log inference component is the second component information; and a fourth determining unit, configured to determine the second actual keyword of the abnormal log based on the target knowledge base, and determine the business feature information based on the second actual keyword.

[0015] Optionally, in one embodiment of this application, it further includes: a third determining module, configured to determine the actual format of the abnormal log based on the log data of the abnormal log before extracting the business feature information of the abnormal log using the target log inference component; and a conversion module, configured to determine a conversion instruction to convert the abnormal log to the preset format based on the preset format if the actual format does not match the preset format, and convert the abnormal log to the preset format based on the conversion instruction.

[0016] Optionally, in one embodiment of this application, it further includes: a calculation module, used to calculate the similarity between the anomaly analysis report and the target anomaly analysis report; a fourth determination module, used to determine the initial anomaly handling strategy of the vehicle based on the target anomaly handling strategy of the target anomaly analysis report when the similarity is greater than a preset similarity threshold; and a correction module, used to determine the correction instruction of the initial anomaly handling strategy based on the anomaly analysis report, and correct the initial anomaly handling strategy according to the correction instruction to obtain the actual anomaly handling strategy of the vehicle.

[0017] A third aspect of this application provides a vehicle, 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 vehicle anomaly detection method as described in the above embodiments.

[0018] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the vehicle anomaly detection method described above.

[0019] A fifth aspect of this application provides a computer program product, including a computer program that, when executed, implements the vehicle anomaly detection method described above.

[0020] This application's embodiments can first determine the actual file size of abnormal logs during vehicle operation and the actual verification results of user access permissions. Based on the actual file size and verification results, a target log inference component is determined. Then, business feature information is extracted, and based on this information, the cause of the anomaly, the process information, and / or supporting evidence information are determined, thereby generating a corresponding anomaly analysis report. By combining log file size and user permissions to select and adapt the inference component, targeted extraction of log business features is used to deduce vehicle anomaly-related information and generate reports. This approach can adapt to logs of different sizes, ensure data access security, and accurately and efficiently complete vehicle fault analysis. Therefore, it solves the problems in related technologies, such as low efficiency and easy information omission in manual analysis, incompatibility and poor accuracy of general AI tools with vehicle logs, and the risk of data leakage when uploading to external networks. Cloud-based AI analysis requires the construction of a vehicle network and cloud servers, making implementation difficult.

[0021] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0022] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart of a vehicle anomaly detection method according to an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a tool for parsing exception logs based on CheryGPT according to an embodiment of this application; Figure 3 This is a flowchart illustrating an anomaly log parsing method based on CheryGPT according to an embodiment of this application; Figure 4 This is a schematic diagram of the structure of a tool for parsing anomaly logs based on a general AI model according to an embodiment of this application; Figure 5 This is a flowchart illustrating an anomaly log parsing method based on a general AI model, according to one embodiment of this application. Figure 6This is a block diagram of a vehicle anomaly detection device provided according to an embodiment of this application; Figure 7 This is a structural schematic diagram of a vehicle provided according to an embodiment of this application. Detailed Implementation

[0023] The embodiments of this application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0024] The following description, with reference to the accompanying drawings, describes an anomaly detection method, apparatus, vehicle, storage medium, and program product according to embodiments of this application. Addressing the issues mentioned in the background art—such as low efficiency and easy information omission in manual analysis, poor accuracy of general AI tools due to incompatibility with vehicle logs and the risk of data leakage when uploaded to external networks, and the high difficulty of implementation of cloud-based AI analysis requiring the construction of a vehicle network and cloud servers—this application provides a vehicle anomaly detection method. In this method, the actual file size of the anomaly log during vehicle operation and the actual verification results of user access permissions are first determined. Based on the actual file size and verification results, a target log inference component is determined, and then business feature information is extracted. Based on the business feature information, the cause of the anomaly, the process information of the anomaly, and / or supporting evidence information are determined, thereby generating a corresponding anomaly analysis report. By combining log file size and user permissions to select and adapt the inference component, targeted extraction of log business features is used to deduce vehicle anomaly-related information and generate a report. This method can adapt to logs of different sizes, ensure data access security, and accurately and efficiently complete vehicle fault analysis. This solves the problems in related technologies, such as low efficiency and easy omission of information in manual analysis, unsuitability of general AI tools for vehicle logs, poor accuracy and risk of data leakage when uploaded to external networks, and high difficulty in implementation of cloud-based AI analysis due to the need to build vehicle network and cloud server.

[0025] Specifically, Figure 1 This is a flowchart of a vehicle anomaly detection method provided according to an embodiment of this application.

[0026] like Figure 1 As shown, the method for detecting vehicle anomalies includes the following steps: In step S101, abnormal logs generated during vehicle operation are collected, and the actual file size of the abnormal logs and the actual verification results of user access permissions are determined.

[0027] It is understood that, in the embodiments of this application, the abnormal log can be understood as text / structured data automatically recorded during the operation of the vehicle's on-board controller, vehicle system, sensors, and electronic control unit, including abnormal events such as equipment failure, communication errors, hardware abnormalities, program crashes, and signal exceeding limits. It may include, but is not limited to, timestamps, module names, event descriptions, parameter values, etc., and this application does not impose specific limitations. The actual file size can be understood as the actual storage space occupied by the abnormal log file. User access permissions may include, but are not limited to, read, write, view, download, and delete permissions for the abnormal log, and this application does not impose specific limitations.

[0028] In some embodiments, the present application can collect abnormal logs generated during vehicle operation in real time, measure the actual file size of the abnormal logs in real time, determine the actual file volume, and verify whether the current user has user access permissions such as read, write, view, download, and delete, thereby obtaining the corresponding actual verification result.

[0029] Optionally, in one embodiment of this application, before extracting the business feature information of the abnormal log using the target log inference component, the method further includes: determining the actual format of the abnormal log based on the log data of the abnormal log; if the actual format does not match the preset format, determining a conversion instruction to convert the abnormal log to the preset format based on the preset format, and converting the abnormal log to the preset format based on the conversion instruction.

[0030] It is understood that in the embodiments of this application, the preset format can be understood as a preset standard unified log format used to adapt to the target log inference component. It may include, but is not limited to, UTF-8 encoded text format and preprocessed structured log file format. The preset format can be set by those skilled in the art according to the actual situation, and this application does not impose any specific restrictions.

[0031] The conversion instructions from the actual format to the preset format may include, but are not limited to, binary to text conversion, character encoding conversion, data trimming, field removal, etc., and this application does not impose specific restrictions.

[0032] In some embodiments, the present application can identify the actual format of the abnormal log based on the log data in the abnormal log, and generate a corresponding conversion instruction when the actual format does not match the preset format, run the conversion instruction, and convert the abnormal log into the preset format in a unified manner.

[0033] For example, in the case where the target log inference component is CheryGPT, this application embodiment determines that the preset format is UTF-8 encoded text. Further, this application embodiment uses a Java-developed format conversion algorithm. When the actual format is .log, it reads the byte stream of the .log file, splits the log content according to the structure of "timestamp-module name-event description-parameter," and converts special characters and binary data into UTF-8 encoded text format. For instance, when processing the car radio anomaly log A.log, this application embodiment converts TUNER.RadioPlayFragment:initview,listisempty,sostartscan in A.log into structured text containing "timestamp: 07:38:33.602, module: TUNER.RadioPlayFragment, event: initview, parameter: listisempty, action: startscan," ensuring that CheryGPT can recognize key information.

[0034] CheryGPT is a privately-owned generative language reasoning system adapted to vehicle operation and maintenance scenarios. Unlike general AI models, it relies on a full range of vehicle fault samples for targeted training. It is fine-tuned based on professional datasets of vehicle electronic control fault logs and vehicle fault codes. It supports local deployment on the intranet. The input is abnormal logs in UTF-8 encoded text format, and the output is fault classification, abnormal source tracing, and fault cause judgment results. It is specifically used for intelligent parsing of vehicle abnormal logs.

[0035] It only processes vehicle operation anomaly logs in a uniform format, and completes log semantic parsing and intelligent vehicle fault diagnosis reasoning.

[0036] Furthermore, in this embodiment of the application, when the target log inference component is a general AI model, the preset format is determined to be a preprocessed structured log file format. This is achieved using Python, regular expressions to remove invalid characters (such as special symbols and repeated spaces) from the abnormal logs, deduplication based on timestamps (e.g., retaining the first record of the same event), and extraction of key fields (such as timestamp, module name, event description, parameter values, etc.) to generate the preprocessed structured log file. For example, in processing B.log, this embodiment of the application can remove approximately 3000 invalid characters and retain 8620 valid records.

[0037] Among them, the general AI model is a large-scale intelligent log analysis model trained based on vehicle abnormal log samples. The input is abnormal logs in preprocessed structured log file format, and the output is vehicle fault identification results and abnormal root cause judgment information.

[0038] Optionally, in one embodiment of this application, before selecting a target log inference component suitable for analyzing abnormal logs from a plurality of candidate log inference components included in the log analysis engine, the method further includes: obtaining candidate file sizes and candidate verification results of user access permissions for the abnormal logs; if the candidate file size is less than a first preset file size threshold and the candidate verification result is a pass, then the deployment environment of the log analysis engine is determined to be an intranet deployment environment, and first component information is determined based on the intranet deployment environment, and candidate log inference components are determined based on the first component information; if the candidate file size is greater than or equal to the first preset file size threshold, or if the candidate verification result is not a pass, then the deployment environment of the log analysis engine is determined to be a local deployment environment, and second component information is determined based on the local deployment environment, and candidate log inference components are determined based on the second component information.

[0039] It is understood that, in the embodiments of this application, the intranet deployment environment can be understood as deploying the log analysis engine on the intranet server of the car manufacturer's cloud, which has sufficient computing power and is suitable for batch parsing of small-scale logs with compliant permissions; the local deployment environment can be understood as running the log analysis engine locally on the vehicle terminal, without external network or intranet transmission, and is used for scenarios with excessively large logs or abnormal permissions; and the first component information and the second component information may include, but are not limited to, program component configurations, computing power parameters, and interface parameters adapted to the intranet and local deployment environments respectively, and this application does not impose specific limitations.

[0040] Furthermore, as can be seen from the above analysis, the candidate log inference component may include, but is not limited to, CheryGPT, general AI models, etc., and this application does not impose any specific limitations.

[0041] In some embodiments, if the candidate file size is less than a first preset file size threshold and the candidate verification result is passed, it can be determined that the log analysis engine is deployed in an intranet environment, and then the first component information corresponding to the intranet can be retrieved to match the corresponding candidate log inference component. The first preset file size threshold can be set by those skilled in the art according to actual conditions, and this application does not impose specific limitations.

[0042] It can be understood that, in the case of CheryGPT as the candidate log inference component, this application embodiment determines that its deployment environment is an intranet deployment environment, and then constructs CheryGPT based on the first component information of the intranet deployment environment.

[0043] In some embodiments, if the file size is greater than or equal to a first preset file size threshold, or if the candidate verification result is not a verification pass, it can be determined that the log analysis engine is deployed in a local environment, and the local supporting second component information can be retrieved to match the corresponding candidate log inference component.

[0044] It can be understood that, in the case of a candidate log inference component being a general AI model, this application embodiment determines that its deployment environment is a local deployment environment, and then constructs a general AI model based on the second component information of the local deployment environment.

[0045] In step S102, based on the actual file size and the actual verification results, a target log inference component suitable for analyzing abnormal logs is selected from multiple candidate log inference components included in the log analysis engine.

[0046] In some embodiments, this application can select a target log inference component that meets the parsing requirements from several candidate log inference components of the log analysis engine, based on the actual file size of the abnormal log and the actual verification results of user access permissions.

[0047] Optionally, in one embodiment of this application, extracting business feature information of abnormal logs using the target log inference component includes: if the component information of the target log inference component is the first component information, and the actual file size is greater than a second preset file size threshold, then determining at least one abnormal log sample fragment based on the abnormal log; extracting keywords corresponding to the abnormal log sample fragment, and constructing a knowledge base based on the keywords; determining the first actual keyword of the abnormal log based on the knowledge base, and determining business feature information based on the first actual keyword.

[0048] It is understood that, in the embodiments of this application, the service characteristic information may include, but is not limited to, the type of faulty service, the scope of service impact, the level of service anomaly, etc., and this application does not impose specific limitations.

[0049] In some embodiments, when the component information corresponding to the target log inference component is the first component information, the target log inference component is determined to be CheryGPT, and the deployment environment is an intranet deployment environment. Further, when the actual file size exceeds a second preset file size threshold, the abnormal log is segmented into one or more abnormal log sample segments according to time segments, and keywords are extracted from the abnormal log sample segments. A retrieval knowledge base is built using all keywords, and then the first actual keyword of the abnormal log is obtained by matching the knowledge base. The business characteristic information of the abnormal log is then output based on the first actual keyword. The second preset file size threshold can be set by those skilled in the art according to actual conditions, and this application does not impose specific limitations.

[0050] In addition, in this embodiment of the application, when the actual file size is less than or equal to the second preset file size threshold, the corresponding business feature information can be directly output using CheryGPT.

[0051] For example, in the embodiment of this application, when the target log inference component is CheryGPT, for abnormal logs with an actual file size greater than 10M, the logs are automatically split into time segments, such as each abnormal log sample segment being 2M, resulting in multiple abnormal log sample segments. The core keywords of each sample segment, such as "location failure," "weak signal," and "path planning," are extracted using the TF-IDF algorithm, and then a dedicated knowledge base is constructed based on these keywords. Furthermore, when analyzing abnormal logs, this embodiment first locates the abnormal log sample segments related to the problem using the knowledge base, and then calls CheryGPT for in-depth analysis.

[0052] For example, in this embodiment of the application, a 50M vehicle navigation anomaly log is divided into 25 anomaly log sample segments, and keywords such as "positioning failure", "weak signal", and "path planning" are extracted from each sample segment to build a knowledge base. Each sample segment is then matched with the knowledge base to quickly locate the relevant segment where the anomaly actually occurred, thereby obtaining the corresponding first actual keyword. The corresponding business feature information is then output in combination with the first actual keyword, such as the fault business type being navigation anomaly, the business impact range being general, and the business anomaly level being medium. This application does not impose any specific limitations.

[0053] Optionally, in one embodiment of this application, extracting business feature information of abnormal logs using a target log inference component includes: when the component information of the target log inference component is second component information, determining a target knowledge base matching the log features based on at least one log feature of the abnormal log; determining a second actual keyword of the abnormal log based on the target knowledge base, and determining business feature information based on the second actual keyword.

[0054] It is understood that, in this embodiment, log features can be understood as basic shallow features directly extracted from abnormal logs, which may include, but are not limited to, module identifiers, error fields, operation types, device numbers, etc., and this application does not impose specific limitations. The target knowledge base refers to a static knowledge base that is pre-deployed and permanently stored on the local terminal / edge device. It can filter out a dedicated knowledge base corresponding to the current log scenario based on the extracted log features. For example, the local device may have pre-installed static knowledge bases such as a radio module knowledge base, a payment module knowledge base, a device maintenance knowledge base, and a user login knowledge base. If the extracted log features contain the module identifier "pay_service" and the error field "insufficient balance," then the payment module knowledge base is determined as the target knowledge base used for this parsing through log feature retrieval and matching.

[0055] In some embodiments, when the component information of the target log inference component is the second component information, that is, when the target log inference component is a general AI model, the log features of the abnormal log are extracted, and a target knowledge base adapted to the current scenario is obtained by matching the log features. The second actual keyword of the abnormal log is parsed with the help of the target knowledge base, and then the corresponding business feature information is output based on the keyword.

[0056] For example, in this application embodiment, abnormal logs can be imported into the AnythingLLM interface, triggering Qwen4B model analysis to extract corresponding log features. Based on the log features, the target knowledge base, such as the radio module knowledge base, can be retrieved. Then, based on the built-in car radio frequency index mapping rules: index 0=87.5MHz, index 30=107.5MHz, index 31=108.0MHz, etc., the second actual keywords of the abnormal logs can be identified, such as the trigger condition "listisempty", the scanning action "startscan", the storage event "autostore", etc., and the automatic station search process from FM87.5 to FM108.0 can be sorted out according to the timestamp, thereby obtaining the corresponding business feature information.

[0057] In step S103, the business feature information of the abnormal log is extracted using the target log inference component. Based on the business feature information, the cause of the vehicle's abnormality, the abnormal process information, and / or the supporting information of the abnormality are determined. Based on the cause of the abnormality, the abnormal process information, and / or the supporting information of the abnormality, an abnormal analysis report of the vehicle is generated.

[0058] It is understood that, in this embodiment, the cause of vehicle anomalies can be understood as the root cause of vehicle malfunctions and error logs, such as sensor drift, poor wiring harness contact, and battery overcharge protection triggering; the anomaly process information can be understood as the complete time-series behavior chain of the fault from triggering, development to log reporting, such as vehicle power-on → abnormal sensor signal acquisition → verification failure → reporting of a level 1 alarm → triggering a fault shutdown after 30 seconds, fully reconstructing the entire fault occurrence process; the anomaly corroborating information can be understood as supporting evidence that can corroborate and support the anomaly cause and fault process, such as synchronous error logs of other controllers during the same time period, historical records of similar faults, vehicle driving condition data, voltage and current curves, fault code matching knowledge base entries, and other corroborating materials, improving the credibility of the analysis results; the anomaly analysis report refers to the final standardized output product of this embodiment, which is a structured document carrier that integrates business characteristic information, anomaly causes, anomaly process information, and anomaly corroborating information to form a readable, exportable, and callable standardized report for maintenance or operation and maintenance platforms, supporting automatic archiving and push to the operation and maintenance end.

[0059] In some embodiments, the present application can extract business feature information corresponding to abnormal logs through the target log inference component, and analyze any one or more of the abnormal cause, abnormal process information, and abnormal supporting information based on the business feature information, and then automatically output a standardized abnormal analysis report in combination with the obtained various types of vehicle abnormal information.

[0060] For example, in the case where the target log inference component is CheryGPT, based on the business feature information returned by CheryGPT, the anomaly cause, anomaly process information, and anomaly supporting information are determined according to the structure of "core conclusion - triggering condition - event timeline - key evidence - difference comparison". Specifically, the event timeline automatically extracts the timestamps from the anomaly logs and arranges them in chronological order; key evidence directly quotes excerpts from the original anomaly logs; and the difference comparison supports user-uploaded supplier reports, using text similarity algorithms to identify differences and mark uncovered root causes of problems.

[0061] Furthermore, in this embodiment of the application, based on the cause of the anomaly, the process of the anomaly, and the supporting evidence of the anomaly, the corresponding anomaly analysis report is exported in PDF or Excel format, and a visual display is provided in the tool interface, including timeline charts, keyword cloud maps, and difference comparison tables.

[0062] Furthermore, in embodiments of this application, when the target log inference component is a general AI model, a local visualization interface can be built using the Flask framework to output an anomaly analysis report, as shown in Table 1. Table 1 is a schematic table of anomaly analysis reports provided according to one embodiment of this application.

[0063] Table 1

[0064] Optionally, in one embodiment of this application, the method further includes: calculating the similarity between the anomaly analysis report and the target anomaly analysis report; if the similarity is greater than a preset similarity threshold, determining the initial anomaly handling strategy of the vehicle based on the target anomaly handling strategy of the target anomaly analysis report; determining the correction instruction of the initial anomaly handling strategy based on the anomaly analysis report, and correcting the initial anomaly handling strategy according to the correction instruction to obtain the actual anomaly handling strategy of the vehicle.

[0065] It is understood that in the embodiments of this application, similarity is used as a numerical indicator to quantify the degree of matching between the anomaly analysis report and the target anomaly analysis report in terms of anomaly cause, anomaly process information, and anomaly supporting information. It can be calculated by text vector matching, keyword overlap, fault code matching, and business feature similarity algorithm. The higher the value, the more similar the current fault is to the historical standard fault.

[0066] The target anomaly handling strategy can be understood as a standardized fault handling scheme that has been historically verified and is tied to the target anomaly analysis report; it is a general reference scheme, such as standard maintenance steps like sensor recalibration, wiring harness replacement, electronic control firmware upgrade, and battery equalization. The initial anomaly handling strategy can be understood as directly reusing the target anomaly handling strategy when the similarity between two reports exceeds a preset similarity threshold. The correction instruction can be understood as an adjustment instruction generated based on the differentiated fault information in the anomaly analysis report. For example, if the current vehicle has additional evidence of communication message loss, the correction instruction would be "add a bus wiring harness continuity detection step," which is not specifically limited in this application. The preset similarity threshold can be set by those skilled in the art according to actual conditions, and is not specifically limited in this application.

[0067] In some embodiments, the present application can calculate the similarity between the anomaly analysis report and the target anomaly analysis report. When the similarity is greater than a preset similarity threshold, an initial anomaly handling strategy for the vehicle is first generated based on the target anomaly handling strategy in the target anomaly analysis report. Then, a correction instruction is generated in conjunction with the anomaly analysis report. The initial anomaly handling strategy is then corrected using the correction instruction to obtain the actual anomaly handling strategy for the vehicle.

[0068] The working principle of the vehicle anomaly detection method proposed in this application will be described below with reference to several embodiments.

[0069] Example 1: Using CheryGPT within the intranet to parse exception logs.

[0070] like Figure 2 As shown in the figure, this application provides a tool structure for parsing abnormal logs based on CheryGPT, which may include, but is not limited to, a user terminal module 201, an access control module 202, a log format adaptation module 203, a CheryGPT interface call module 204, a large file knowledge base module 205, a log analysis engine 206, a supplier result comparison module 207, and a result output module 208.

[0071] The user terminal module 201 is used to import exception logs and target exception analysis reports.

[0072] The access control module 202 is used to deploy the parsing tool in the enterprise intranet environment, set user access permissions, ensure that log data is not uploaded to the external network, and verify whether the actual verification result of user access permissions is successful.

[0073] The log format adaptation module 203 is used to convert .log format exception logs into a text format that CheryGPT can recognize, such as UTF-8 encoded text format, through a format conversion algorithm, while retaining key data such as timestamps, module names, event descriptions, and parameter values ​​in the logs.

[0074] CheryGPT interface call module 204 is used to open the development port of CheryGPT to achieve deep integration between the tool and CheryGPT and remove the token limit for single log analysis.

[0075] The large file knowledge base module 205 is used to automatically split and build a dedicated knowledge base for abnormal logs that are larger than the second preset file size threshold, and to achieve complete analysis through knowledge base association queries.

[0076] Log analysis engine 206 is used to call CheryGPT to extract business characteristic information (such as autostore, startscan, listisempty, etc.).

[0077] The results output module 207 is used to sort out the event timeline and triggering conditions, and generate an anomaly analysis report containing the cause of the anomaly, information on the anomaly process, and supporting information on the anomaly.

[0078] The supplier results comparison module 208 is used to compare the anomaly analysis report obtained by CheryGPT with the target anomaly analysis report provided by the supplier, calculate the similarity, cross-validate the target anomaly handling strategy provided by the target anomaly analysis report, mark the differences, generate correction instructions, and thus determine the actual anomaly handling strategy.

[0079] like Figure 3 As shown in the figure, this application provides a process for parsing exception logs based on CheryGPT, the main contents of which include: In this embodiment of the application, the operating environment for log parsing is an enterprise intranet server with a CPU of ≥16 cores and a memory of ≥32G; the software depends on CheryGPTV version 2.0, a Java development environment, and a MySQL database (used to store analysis reports and knowledge base); the log adaptation scope is: vehicle system .log format logs, including logs generated by modules such as radio, navigation, and entertainment systems.

[0080] Step S301: Log import and format conversion.

[0081] In this embodiment, a Java-developed format conversion algorithm is used to read the byte stream of the .log file, split the log content according to the structure of "timestamp-module name-event description-parameter", and convert special characters and binary data into UTF-8 encoded text format. For example, when processing the car radio abnormal log A.log, this embodiment converts TUNER.RadioPlayFragment:initview,listisempty,sostartscan in A.log into structured text containing "timestamp: 07:38:33.602, module: TUNER.RadioPlayFragment, event: initview, parameter: listisempty, action: startscan", ensuring that CheryGPT can recognize key information.

[0082] Step S302: Permission verification.

[0083] In this embodiment, CheryGPT can be deployed on an enterprise intranet Tomcat server, and user permission management can be implemented through the Spring Security framework, with three levels of permissions: administrator, developer, and tester. Administrators can configure the log access scope, while developers and testers can only view the log analysis results of the modules they are responsible for. The log files are stored in an intranet database with no external network access interface.

[0084] Step S303: Obtain the actual file size of the exception log.

[0085] In this embodiment, the tool can establish a long-lived connection with CheryGPT via the open development port provided by CheryGPT using the HTTP protocol, setting up a dedicated call channel and removing the token restriction for single log analysis. During the call, the structured log text is passed to the model in batches, with each batch of data controlled within 5000 characters to ensure fast analysis response.

[0086] Furthermore, in this embodiment of the application, the 50M vehicle navigation anomaly log is divided into 25 anomaly log sample segments, and keywords such as "positioning failure", "weak signal", and "path planning" are extracted from each sample segment to construct a knowledge base. Each sample segment is then matched with the knowledge base to quickly locate the relevant segment where the anomaly actually occurred, thereby obtaining the corresponding first actual keyword. The corresponding business feature information is then output in combination with the first actual keyword, such as the fault business type being navigation anomaly, the business impact range being general, and the business anomaly level being medium. This application does not impose specific limitations.

[0087] Step S304: Call CheryGPT.

[0088] Step S305: Generate an anomaly analysis report.

[0089] In this embodiment, based on the business feature information returned by CheryGPT, the anomaly cause, anomaly process information and anomaly supporting information can be determined according to the structure of "core conclusion - triggering condition - event timeline - key evidence - difference comparison". Based on the anomaly cause, anomaly process information and anomaly supporting information, the corresponding anomaly analysis report can be exported in PDF or Excel format, and a visual display can be provided on the tool interface.

[0090] Step S306: Compare target anomaly analysis reports.

[0091] Step S307: Output the actual exception handling strategy.

[0092] In this embodiment, the similarity between the anomaly analysis report and the target anomaly analysis report can be calculated first. When the similarity is greater than a preset similarity threshold, an initial anomaly handling strategy for the vehicle is generated based on the target anomaly handling strategy in the target anomaly analysis report. A correction instruction is generated in conjunction with the anomaly analysis report, and the initial anomaly handling strategy is corrected using the correction instruction to obtain the actual anomaly handling strategy.

[0093] Option 2: Locally call a general AI model to perform log parsing.

[0094] like Figure 4 As shown, this application embodiment provides a tool structure for parsing abnormal logs based on a general AI model, which may include, but is not limited to, a log file import module 401, a log preprocessing module 402, a local high-performance server 403, a general AI model deployment module 404, a target knowledge base 405, a model tuning module 406, a general AI model analysis module 407, a visualization output module 408, and a user operation module 409.

[0095] Among them, the log file import module 401 supports users to directly import local vehicle log files (supporting files ≥16.05MB).

[0096] The log preprocessing module 402 is used to convert the actual format to a preset format, such as deduplication, noise reduction, and key information extraction.

[0097] The local high-performance server 403 is equipped with a high-performance graphics card.

[0098] The general AI model deployment module 404 is used to deploy a combined model, including Ollama+AnythingLLM+Qwen4B, on a local server.

[0099] The Target Knowledge Base 405 provides exclusive keywords for in-vehicle systems.

[0100] The model tuning module 406 is used to adjust model parameters for the Ollam+AnythingLLM+Qwen4B combined model deployed on the local server based on business feature information, thereby optimizing keyword recognition accuracy and event correlation analysis capabilities.

[0101] The general AI model analysis module 407 is used to identify event triggering conditions, sort out timelines, and use the target knowledge base to assist in inference to determine the mechanism of the problem.

[0102] The visualization output module 408 is used to output anomaly analysis reports in a visual manner, including anomaly causes, anomaly process information, and anomaly supporting information.

[0103] User operation module 409 is used to analyze the differences between the anomaly analysis report and the target anomaly analysis report to determine the actual anomaly handling strategy.

[0104] like Figure 5 As shown in the figure, this application provides a process for parsing abnormal logs based on a general AI model, the main contents of which include: In this embodiment, log parsing is performed on a local server with a high-performance graphics card to meet the resource requirements for running a general AI model, such as an Intel i9-13900K CPU, an NVIDIA RTX 4090 graphics card (24GB VRAM), 64GB of RAM, and 1TB SSD storage. The software environment is Ubuntu 22.04 operating system, Ollama v0.1.28, Anything LLM v0.8.0, and the Qwen4B model (quantization version q4_0). The system is compatible with various local log files of the vehicle system, supporting a maximum single file size of ≥50MB.

[0105] Step S501: Import logs locally.

[0106] Step S502: Log preprocessing.

[0107] In this embodiment, Python can be used to remove invalid characters (such as special symbols, duplicate spaces, etc.) from the exception log using regular expressions, deduplicate based on timestamps (e.g., retaining the first record of the same event), extract key fields (such as timestamp, module name, event description, parameter values, etc.), and generate a preprocessed structured log file. For example, when processing B.log, this embodiment can remove approximately 3000 invalid characters and retain 8620 valid records.

[0108] Step S503: Model deployment and optimization.

[0109] In this embodiment, the Qwen4B model can be deployed on a local server using the Ollam tool, with AnythingLLM configured as the model management interface. Model inference parameters are set as follows: context window size 8192, inference temperature 0.3 (balancing accuracy and flexibility). For in-vehicle infotainment scenarios, a target knowledge base (containing over 200 exclusive keywords such as autostore, startscan, and dabinfo) is imported to fine-tune the model and improve keyword recognition accuracy.

[0110] Step S504: General AI model analysis.

[0111] Step S505: Target knowledge base-assisted inference.

[0112] In this embodiment, abnormal logs can be imported into the AnythingLLM interface, triggering Qwen4B model analysis to extract corresponding log features. Based on these log features, a target knowledge base, such as the radio module knowledge base, is retrieved. Then, based on the built-in car radio frequency index mapping rules (index 0 = 87.5MHz, index 30 = 107.5MHz, index 31 = 108.0MHz, etc.), the second actual keywords of the abnormal logs are identified, such as the "listisempty" trigger condition, the "startscan" scanning action, and the "autostore" storage event. The automatic station search process from FM87.5 to FM108.0 is then sorted out according to the timestamps, thereby obtaining the corresponding business feature information.

[0113] Step S506: Generate an anomaly analysis report.

[0114] Step S507: Visualize the results.

[0115] In this embodiment of the application, a local visualization interface and anomaly analysis report can be built using the Flask framework, and the results are shown in Table 1.

[0116] The vehicle anomaly detection method proposed in this application first determines the actual file size of the anomaly log during vehicle operation and the actual verification results of user access permissions. Based on the actual file size and verification results, a target log inference component is determined. Then, business feature information is extracted, and the anomaly cause, anomaly process information, and / or anomaly supporting information are determined based on this information, thereby generating a corresponding anomaly analysis report. By combining log file size and user permissions to select and adapt the inference component, targeted extraction of log business features is used to deduce vehicle anomaly-related information and generate a report. This method can adapt to logs of different sizes, ensure data access security, and accurately and efficiently complete vehicle fault analysis. Therefore, it solves the problems in related technologies, such as low efficiency and easy information omission in manual analysis, incompatibility and poor accuracy of general AI tools with vehicle logs, and the risk of data leakage when uploading to external networks, and the high difficulty of implementation of cloud-based AI analysis requiring the construction of vehicle networks and cloud servers.

[0117] Next, with reference to the accompanying drawings, a vehicle anomaly detection device according to an embodiment of this application is described.

[0118] Figure 6 This is a block diagram of a vehicle anomaly detection device provided according to an embodiment of this application.

[0119] like Figure 6 As shown, the vehicle anomaly detection device 10 includes: a data acquisition module 100, a selection module 200, and a generation module 300.

[0120] The acquisition module 100 is used to collect abnormal logs generated during vehicle operation and determine the actual file size of the abnormal logs and the actual verification results of user access permissions.

[0121] Selection module 200 is used to select a target log inference component suitable for analyzing abnormal logs from multiple candidate log inference components included in the log analysis engine, based on the actual file size and actual verification results.

[0122] The generation module 300 is used to extract business feature information from abnormal logs using the target log inference component, determine the cause of the vehicle's abnormality, abnormal process information and / or abnormal supporting information based on the business feature information, and generate an abnormality analysis report for the vehicle based on the cause of the abnormality, abnormal process information and / or abnormal supporting information.

[0123] Optionally, in one embodiment of this application, it further includes: an acquisition module, a first determination module, and a second determination module.

[0124] The acquisition module is used to acquire candidate file sizes and candidate verification results of user access permissions for abnormal logs before selecting a target log inference component suitable for analyzing abnormal logs from multiple candidate log inference components included in the log analysis engine.

[0125] The first determining module is used to determine that the deployment environment of the log analysis engine is an intranet deployment environment when the candidate file size is less than a first preset file size threshold and the candidate verification result is verification passed, and to determine the first component information based on the intranet deployment environment, and to determine the candidate log inference component based on the first component information.

[0126] The second determining module is used to determine the deployment environment of the log analysis engine as a local deployment environment when the candidate file size is greater than or equal to the first preset file size threshold, or when the candidate verification result is not a verification pass, and to determine the second component information based on the local deployment environment, and to determine the candidate log inference component based on the second component information.

[0127] Optionally, in one embodiment of this application, the generation module 300 includes: a first determining unit, an extraction unit, and a second determining unit.

[0128] The first determining unit is used to determine at least one abnormal log sample fragment based on the abnormal log if the actual file size is greater than a second preset file size threshold, when the component information of the target log inference component is the first component information.

[0129] The extraction unit is used to extract keywords corresponding to abnormal log sample fragments and build a knowledge base based on the keywords.

[0130] The second determining unit is used to determine the first actual keyword of the abnormal log based on the knowledge base, and to determine the business characteristic information based on the first actual keyword.

[0131] Optionally, in one embodiment of this application, the generation module 300 includes a third determining unit and a fourth determining unit.

[0132] The third determining unit is used to determine a target knowledge base that matches the log features based on at least one log feature of the abnormal log, when the component information of the target log inference component is the second component information.

[0133] The fourth determining unit is used to determine the second actual keyword of the abnormal log based on the target knowledge base, and to determine the business feature information based on the second actual keyword.

[0134] Optionally, in one embodiment of this application, it further includes: a third determining module and a conversion module.

[0135] The third determining module is used to determine the actual format of the abnormal log based on the log data of the abnormal log before extracting the business feature information of the abnormal log using the target log inference component.

[0136] The conversion module is used to determine the conversion instruction to convert the exception log to the preset format when the actual format does not match the preset format, and to convert the exception log to the preset format based on the conversion instruction.

[0137] Optionally, in one embodiment of this application, it further includes: a calculation module, a fourth determination module, and a correction module.

[0138] The calculation module is used to calculate the similarity between the anomaly analysis report and the target anomaly analysis report.

[0139] The fourth determination module is used to determine the initial anomaly handling strategy for the vehicle based on the target anomaly handling strategy in the target anomaly analysis report when the similarity is greater than a preset similarity threshold.

[0140] The correction module is used to determine the correction instructions for the initial anomaly handling strategy based on the anomaly analysis report, and correct the initial anomaly handling strategy according to the correction instructions to obtain the actual anomaly handling strategy for the vehicle.

[0141] It should be noted that the explanation of the above-described method for detecting vehicle anomalies also applies to the vehicle anomaly detection device of this embodiment, and will not be repeated here.

[0142] The vehicle anomaly detection device proposed in this application can first determine the actual file size of the anomaly log during vehicle operation and the actual verification results of user access permissions. Based on the actual file size and verification results, a target log inference component is determined, and then business feature information is extracted. Based on this business feature information, the cause of the anomaly, the process information, and / or supporting evidence information are determined, thereby generating a corresponding anomaly analysis report. By combining log file size and user permissions to select and adapt the inference component, targeted extraction of log business features is used to deduce vehicle anomaly-related information and generate a report. This device can adapt to logs of different sizes, ensure data access security, and accurately and efficiently complete vehicle fault analysis. Therefore, it solves the problems in related technologies, such as low efficiency and easy information omission in manual analysis, incompatibility and poor accuracy of general AI tools with vehicle logs, and the risk of data leakage when uploading to external networks, and the high difficulty of implementation of cloud-based AI analysis requiring the construction of a vehicle network and cloud servers.

[0143] Figure 7 This is a schematic diagram of the structure of a vehicle according to an embodiment of this application. The vehicle may include: The memory 701, the processor 702, and the computer program stored on the memory 701 and capable of running on the processor 702.

[0144] When the processor 702 executes the program, it implements the vehicle anomaly detection method provided in the above embodiments.

[0145] Furthermore, the vehicle also includes: Communication interface 703 is used for communication between memory 701 and processor 702.

[0146] The memory 701 is used to store computer programs that can run on the processor 702.

[0147] The memory 701 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0148] If the memory 701, processor 702, and communication interface 703 are implemented independently, then the communication interface 703, memory 701, and processor 702 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized into address buses, data buses, control buses, etc. For ease of representation, Figure 7 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0149] Optionally, in a specific implementation, if the memory 701, processor 702, and communication interface 703 are integrated on a single chip, then the memory 701, processor 702, and communication interface 703 can communicate with each other through an internal interface.

[0150] The processor 702 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0151] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described method for detecting vehicle anomalies.

[0152] This application also provides a computer program product, including a computer program that, when executed, implements the above-described method for detecting vehicle anomalies.

[0153] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0154] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0155] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0156] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0157] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, it can be implemented using any one or more of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0158] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0159] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0160] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A method for detecting vehicle anomalies, characterized in that, Includes the following steps: Collect abnormal logs generated during vehicle operation, and determine the actual file size of the abnormal logs and the actual verification results of user access permissions; Based on the actual file size and the actual verification result, a target log inference component suitable for analyzing the abnormal log is selected from multiple candidate log inference components included in the log analysis engine. The target log inference component is used to extract the business feature information of the abnormal log, and the abnormal cause, abnormal process information and / or abnormal supporting information of the vehicle are determined based on the business feature information. Based on the abnormal cause, abnormal process information and / or abnormal supporting information, an abnormal analysis report of the vehicle is generated.

2. The method according to claim 1, characterized in that, Before selecting a target log inference component suitable for analyzing the abnormal logs from a plurality of candidate log inference components included in the log analysis engine, the method further includes: Obtain the candidate file size of the exception log and the candidate verification results of the user access permissions; If the candidate file size is less than the first preset file size threshold and the candidate verification result is a successful verification, then the deployment environment of the log analysis engine is determined to be an intranet deployment environment, and the first component information is determined based on the intranet deployment environment, and the candidate log inference component is determined based on the first component information. If the candidate file size is greater than or equal to the first preset file size threshold, or if the candidate verification result is not a verification pass, then the deployment environment of the log analysis engine is determined to be a local deployment environment, and the second component information is determined based on the local deployment environment, and the candidate log inference component is determined based on the second component information.

3. The method according to claim 2, characterized in that, The step of extracting business feature information from the abnormal logs using the target log inference component includes: If the component information of the target log inference component is the first component information, and the actual file size is greater than the second preset file size threshold, then at least one abnormal log sample segment is determined based on the abnormal log. Extract keywords corresponding to the abnormal log sample fragments, and construct a knowledge base based on the keywords; Based on the knowledge base, the first actual keyword of the abnormal log is determined, and the business feature information is determined based on the first actual keyword.

4. The method according to claim 2, characterized in that, The step of extracting business feature information from the abnormal logs using the target log inference component includes: When the component information of the target log inference component is the second component information, a target knowledge base matching the log feature is determined based on at least one log feature of the abnormal log. The second actual keyword of the abnormal log is determined based on the target knowledge base, and the business feature information is determined based on the second actual keyword.

5. The method according to claim 1, characterized in that, Before using the target log inference component to extract the business feature information of the abnormal log, the method further includes: Based on the log data of the anomaly log, determine the actual format of the anomaly log; If the actual format does not match the preset format, a conversion instruction to convert the exception log to the preset format is determined based on the preset format, and the exception log is converted to the preset format based on the conversion instruction.

6. The method according to claim 1, characterized in that, Also includes: Calculate the similarity between the anomaly analysis report and the target anomaly analysis report; If the similarity is greater than a preset similarity threshold, then the initial anomaly handling strategy for the vehicle is determined based on the target anomaly handling strategy in the target anomaly analysis report. Based on the anomaly analysis report, a correction instruction for the initial anomaly handling strategy is determined, and the initial anomaly handling strategy is corrected according to the correction instruction to obtain the actual anomaly handling strategy for the vehicle.

7. A vehicle anomaly detection device, characterized in that, include: The data acquisition module is used to collect abnormal logs generated during vehicle operation and determine the actual file size of the abnormal logs and the actual verification results of user access permissions. The selection module is used to select a target log inference component suitable for analyzing the abnormal log from multiple candidate log inference components included in the log analysis engine, based on the actual file size and the actual verification result. The generation module is used to extract business feature information from the abnormal log using the target log inference component, determine the cause of the abnormality, abnormal process information and / or abnormal supporting information of the vehicle based on the business feature information, and generate an abnormality analysis report of the vehicle based on the cause of the abnormality, abnormal process information and / or abnormal supporting information.

8. A vehicle, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the vehicle anomaly detection method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the vehicle anomaly detection method as described in any one of claims 1-6.

10. A computer program product, characterized in that, Includes a computer program, which, when executed, is used to implement the vehicle anomaly detection method as described in any one of claims 1-6.