Vehicle function fault intelligent troubleshooting method, device and equipment and storage medium

Abstract

The invention discloses an intelligent troubleshooting method, device and equipment for vehicle function faults and a storage medium, and relates to the cross technical field of natural language processing and vehicle fault troubleshooting, and the method comprises the following steps: retrieving a knowledge base according to fault description to obtain technical data associated with the fault description, the knowledge base is constructed based on a vehicle function system design document; based on the technical data and in combination with a big language model, generating a fault reason and a fault checking strategy which corresponds to the fault reason and is used for indicating a signal needing to be detected and detection logic; and analyzing the vehicle-mounted protocol and the vehicle-mounted data to obtain a signal of a corresponding fault moment, sequentially detecting related signals according to the troubleshooting strategy, determining a fault signal, and realizing vehicle function troubleshooting. According to the invention, the efficiency and the accuracy of vehicle function troubleshooting can be effectively improved.

Description

Technical Field

[0001] This application relates to the cross - technical field of natural language processing and vehicle fault troubleshooting, and specifically relates to an intelligent vehicle function fault troubleshooting method, device, equipment, and storage medium. Background Art

[0002] The RAG (Retrieval - Augmented Generation) technology is an advanced question - answering system that combines information retrieval and content generation. It can integrate and process various formats of files and data, build a comprehensive knowledge base, and on this basis, achieve efficient information retrieval and flexible content generation. The core of the RAG technology lies in combining the generation ability of the large language model (LLM) with a precise information retrieval system. By dynamically accessing external knowledge bases and real - time data, it significantly enhances the quality and comprehensiveness of the model's answers. The RAG technology can dynamically retrieve the latest information during the generation process, support real - time updates, ensure that the answers are always based on the latest data, thus effectively solving the drawback of the slow knowledge update process of traditional pre - trained question - answering models. At the same time, by referring to the retrieved context information, the answers of the RAG technology are more interpretable and traceable, and users can clearly understand the source and basis of the answers.

[0003] With the development of automotive technology, the electronic control systems of vehicles have fully moved towards intelligence and networking. The service signals involved in vehicle functions are increasing, and fault troubleshooting has become increasingly difficult. The traditional vehicle function fault troubleshooting method mainly relies on engineers manually consulting the function design documents related to faults. Only after consulting and understanding the function logic can they combine personal experience to locate problems, with low efficiency and prone to misjudgment. Especially in the process of parsing in - vehicle signals in complex power systems, engineers in a single function field often cannot effectively cross multiple technical fields for function diagnosis, resulting in a long cycle for vehicle function fault troubleshooting. In addition, although some auxiliary tools can be used for vehicle function fault troubleshooting currently, most of these tools only conduct simple fault - cause analysis based on automotive fault codes and cannot accurately locate vehicle function faults. Summary of the Invention

[0004] This application provides an intelligent vehicle function fault troubleshooting method, device, equipment, and storage medium, which can effectively improve the efficiency and accuracy of vehicle function fault troubleshooting.

[0005] In a first aspect, an embodiment of this application provides an intelligent vehicle function fault troubleshooting method, and the intelligent vehicle function fault troubleshooting method includes: Retrieve technical materials associated with the fault description from the knowledge base according to the fault description, where the knowledge base is constructed based on the vehicle function system design document; Based on the technical materials and combined with a large language model, generate the fault cause and a fault troubleshooting strategy for the corresponding fault cause to indicate the signals to be detected and the detection logic; Parse the in-vehicle protocol and in-vehicle data to obtain the signals at the corresponding fault moment, and sequentially detect the relevant signals according to the fault troubleshooting strategy to determine the fault signal, thereby realizing the troubleshooting of vehicle function faults.

[0006] Combined with the first aspect, in one implementation, the step of generating the fault cause and a fault troubleshooting strategy for the corresponding fault cause to indicate the signals to be detected and the detection logic based on the technical materials and combined with a large language model specifically includes: Generate a fault cause query description based on the technical materials and input it into the large language model to generate the fault cause; Input the fault description and the generated fault cause into the large language model to generate the fault troubleshooting strategy; Wherein, the fault troubleshooting strategy includes the signals to be detected, the associated in-vehicle protocol and in-vehicle data of the signals to be detected, the expected conditions of the signals to be detected, the signal detection sequence, and the jump logic.

[0007] Combined with the first aspect, in one implementation, the step of parsing the in-vehicle protocol and in-vehicle data to obtain the signals at the corresponding fault moment, and sequentially detecting the relevant signals according to the fault troubleshooting strategy to determine the fault signal specifically includes: Parse the in-vehicle protocol and in-vehicle data to obtain the signals at the corresponding fault moment, and filter out the signals to be detected according to the signals to be detected shown in the fault troubleshooting strategy; According to the signal detection sequence and jump logic of the fault troubleshooting strategy, and the expected conditions of the signals to be detected, sequentially detect the signals to be detected to determine the fault signal.

[0008] Combined with the first aspect, in one implementation, the step of parsing the in-vehicle protocol and in-vehicle data to obtain the signals at the corresponding fault moment, and filtering out the signals to be detected according to the signals to be detected shown in the fault troubleshooting strategy specifically includes: Parse the various in-vehicle protocols and in-vehicle data of the vehicle to obtain all signal information, and filter out the signals to be detected according to the signals to be detected shown in the fault troubleshooting strategy; According to the pre-operation of the fault trigger, determine the fault period based on the pre-operation as a reference; Obtain the signal value information of the signals to be detected within the fault period time segment.

[0009] Combined with the first aspect, in one implementation, Detect the signal value of the signal to be detected within the fault cycle time period, and detect whether the signal to be detected meets the expected conditions, so as to determine the fault signal.

[0010] Combined with the first aspect, in one implementation, For the detection of the detection signal, it includes single signal detection and verification of the logical relationship between signals; The single signal detection is to detect the numerical value, the interval where it is located, and the corresponding state transition of the signal value of the current signal; The verification of the logical relationship between signals is to detect whether the relationship between the signal values of different signals meets the expected conditions based on the finite state machine and the graph structure according to the fault troubleshooting strategy.

[0011] Combined with the first aspect, in one implementation, the intelligent troubleshooting method for vehicle function faults further includes: Generate troubleshooting evidence data for vehicle function fault troubleshooting according to the execution situation of the fault troubleshooting strategy; Among them, the troubleshooting evidence data includes the names of the signals participating in the detection, the actual signal values of the fault signals at the fault moment, the expected conditions of the signals participating in the detection, and the detection results of the signals participating in the detection.

[0012] Combined with the first aspect, in one implementation, for the construction of the knowledge base, it specifically includes: Obtain the vehicle function system design document and the vehicle function defect analysis document, and perform formatting processing to obtain a formatted document; Parse the formatted document to obtain text data, and add the vehicle model project, function component name, and title to the text data; Based on the recursive character text segmentation technology, split the text data into multiple text blocks, and based on the semantic representation model, extract the semantic features of the text blocks and convert the extracted semantic features into text vectors to obtain the final text data; Based on the text vector indexing technology, index and store the final text data to realize the construction of the knowledge base.

[0013] Combined with the first aspect, in one implementation, the parsing of the formatted document to obtain text data specifically includes: According to the keyword matching method, perform logical chunking on the formatted document to obtain logical chunks; Based on the optical character recognition technology, recognize the text in the logical chunks to obtain text data, based on the structured parsing method, extract the table content in the logical chunks and convert it into a computable data format, based on the computer vision technology, recognize the logic block diagrams and flowcharts in the logical chunks, extract key information and combine it with the text data to form a knowledge index.

[0014] In combination with the first aspect, in one embodiment, the intelligent vehicle function fault troubleshooting method further includes: Analyze the historical data of vehicle function fault troubleshooting to determine whether there are abnormal troubleshooting situations, and actively add new fault troubleshooting strategies for abnormal troubleshooting situations; Verify the newly added fault troubleshooting strategy, and after the verification passes, add the newly added fault troubleshooting strategy to the knowledge base.

[0015] In combination with the second aspect, an embodiment of the present application provides an intelligent vehicle function fault troubleshooting device, and the intelligent vehicle function fault troubleshooting device includes: A retrieval module, which is used to retrieve technical materials associated with the fault description from the knowledge base according to the fault description, wherein the knowledge base is constructed based on the vehicle function system design document; A generation module, which is used to generate fault causes and fault troubleshooting strategies for the corresponding fault causes to indicate the signals to be detected and the detection logic based on the technical materials and in combination with a large language model; A troubleshooting module, which is used to analyze the in-vehicle protocol and in-vehicle data to obtain the signals at the corresponding fault moment, and sequentially detect the relevant signals according to the fault troubleshooting strategy to determine the fault signal and realize the vehicle function fault troubleshooting.

[0016] In combination with the third aspect, an embodiment of the present application provides an intelligent vehicle function fault troubleshooting device, and the intelligent vehicle function fault troubleshooting device includes a processor, a memory, and a vehicle function fault troubleshooting program stored on the memory and executable by the processor. When the vehicle function fault troubleshooting program is executed by the processor, the steps of the above-mentioned vehicle function fault troubleshooting method are realized.

[0017] In combination with the fourth aspect, an embodiment of the present application provides a computer-readable storage medium, and a vehicle function fault troubleshooting program is stored on the computer-readable storage medium. When the vehicle function fault troubleshooting program is executed by a processor, the steps of the above-mentioned vehicle function fault troubleshooting method are realized.

[0018] The beneficial effects brought by the technical solutions provided by the embodiments of the present application include: (1) Identify vehicle function faults in an automated and intelligent manner. Not only can it give fault troubleshooting strategies, but further, it can also analyze the in-vehicle signal data at the fault moment and directly locate the specific fault signal, greatly reducing the workload and time cost of manual processing; (2) By using the RAG technology, it can process more complex fault descriptions and retrieve professional technical documents, improving the accuracy and reliability of the output of fault troubleshooting strategies and fault location; (3) By generating through understanding the fault description and the description of the fault cause inquiry, it can capture more comprehensive professional knowledge information, enhance the professionalism of fault troubleshooting, and the fault troubleshooting results can be presented to relevant personnel in the vehicle function domain, helping function product personnel, system designers, and function testers quickly identify the fault cause and improve the efficiency and quality of fault regression resolution. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a schematic flowchart of the vehicle function fault intelligent troubleshooting method of this application; Figure 2 It is a schematic flowchart of the construction process of the knowledge base; Figure 3 It is a schematic flowchart of obtaining text data by parsing a formatted document; Figure 4 It is a schematic flowchart of the generation of fault troubleshooting strategies; Figure 5 It is a schematic flowchart of fault signal determination; Figure 6 It is a schematic diagram of the functional modules of the vehicle function fault intelligent troubleshooting device of this application; Figure 7 It is a schematic diagram of the hardware structure of the vehicle function fault intelligent troubleshooting equipment of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] In order to enable those skilled in the art to better understand the solution of this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0021] To make the purpose, technical solutions, and advantages of this application clearer, the embodiments of this application will be further described in detail below in conjunction with the accompanying drawings.

[0022] In a first aspect, the embodiments of this application provide a vehicle function fault intelligent troubleshooting method, which improves the efficiency and accuracy of vehicle function fault troubleshooting by combining the retrieval and generation capabilities of the RAG technology and the automated analysis of in-vehicle signals. [[ID=^{37}]]

[0023] In one embodiment, referring to Figure 1 , Figure 1 It is a schematic flowchart of the vehicle function fault intelligent troubleshooting method of this application. As Figure 1 shown, the vehicle function fault intelligent troubleshooting method includes: S1: Retrieve technical materials associated with the fault description from the knowledge base according to the fault description, where the knowledge base is constructed based on vehicle functional system design documents; That is, based on vehicle-related documents such as vehicle functional system design documents, a knowledge base is constructed, and the documents are structurally segmented and knowledge indexes are established. Through relevant document segmentation strategies, information such as tabular signal data and logic block diagrams in the documents is automatically parsed to generate a vectorized and easily retrievable knowledge base; In practical applications, according to the fault description input by the user, retrieving the knowledge base can obtain technical materials associated with the fault description; S2: Based on the technical materials and in combination with a large language model, generate fault causes and a fault troubleshooting strategy for the corresponding fault causes to indicate the signals to be detected and the detection logic; That is, according to the retrieval results, through a large language model, a fault troubleshooting strategy for indicating the fault troubleshooting idea can be automatically analyzed; S3: Analyze in-vehicle protocols and in-vehicle data to obtain signals at the corresponding fault moment, and sequentially detect relevant signals according to the fault troubleshooting strategy to determine the fault signal, realizing vehicle functional fault troubleshooting.

[0024] That is, by analyzing in-vehicle protocols and in-vehicle data, signals corresponding to the fault moment and associated with the fault troubleshooting strategy are obtained. By gradually verifying the state changes of the signals through a finite state machine model and performing conditional checks, the fault signal can be located. Further, during the fault troubleshooting process, information such as troubleshooting steps, signal expected values and actual values, and signal verification conclusions is automatically recorded to generate a complete troubleshooting evidence chain for the user to view the troubleshooting results of each step and ensure the traceability of the fault location process.

[0025] Further, in one embodiment, as shown in Figure 2 For the construction of the knowledge base, it specifically includes: S101: Obtain vehicle functional system design documents and vehicle functional defect analysis documents, and perform formatting processing to obtain formatted documents; Specifically, collect vehicle functional system design documents and vehicle functional defect analysis documents from diverse data sources (such as documents in PDF, Word, HTML, etc. formats). The documents include signal lists, principle block diagrams, logical descriptions, and functional decomposition content of each functional component. In practical applications, the documents usually contain structured and unstructured data, such as information in tables, pictures, and logical block diagrams. To ensure the integrity and usability of the data, formatting processing is required. Specifically, optimize the format of the charts and adjust the layout of the vehicle functional system design documents. For the vehicle functional defect analysis documents, eliminate irrelevant display information to avoid introducing excessive data noise during document parsing; S102: Parse the formatted document to obtain text data, and add the vehicle model project, functional component name, and title to the text data; Specifically, for the text data obtained after parsing the formatted document, irrelevant characters, special symbols, and redundant information are removed through data cleaning techniques. At the same time, relevant information is added to the text data to improve the efficiency and accuracy of content screening and condition filtering. Specifically, after normalizing the signal names, parameter values, etc. in the text data, the vehicle model project and functional component name involved in the text data are added to the text data, and a title is added to the text data for subsequent efficient retrieval and analysis positioning; S103: Split the text data into multiple text blocks based on the recursive character text segmentation technique, and perform semantic feature extraction on the text blocks based on the semantic representation model and convert the extracted semantic features into text vectors to obtain the final text data; Specifically, for the text data, the recursive character text segmentation technique is adopted. According to delimiters such as spaces, line breaks, and punctuation marks, the long text is split into multiple structured text blocks. At the same time, according to the context window limit of the large language model, the size of a single text block is dynamically adjusted to balance the retrieval accuracy and the integrity of context information; in addition, a sliding window can also be used for text slicing to enhance the context understanding ability and avoid hard cuts in paragraphs; After the above processing is completed, based on an efficient semantic representation model (such as BERT, RoBERTa, DeBERTa, etc.), semantic feature extraction is performed on the text blocks, and the extracted semantic features are converted into text vectors (for example, for a text data, it can be split into 10 text blocks, and after semantic feature extraction for each text block, vector conversion is performed. Subsequently, when retrieving the knowledge base, calculate the similarity between the semantic features and the semantic features of the input fault description. When the similarity calculation result is greater than the preset value, the current text block is used as an alternative) to capture deep semantic information for subsequent retrieval; S104: Based on the text vector indexing technique, index and store the final text data to implement knowledge base construction. <>

[0026] Specifically, use the text vector indexing technique (such as the IVF technique or the Ball-Tree technique) to index and store the text data, and establish an efficient semantic retrieval mechanism to support fast query and knowledge recall. Optionally, expert Q&A knowledge can be collected, and the knowledge graph is used to construct the association relationship between fault troubleshooting and vehicle functional system design documents. When retrieving relevant questions subsequently, the Q&A knowledge is efficiently retrieved and extracted based on the knowledge graph to enhance the professionalism and accuracy of question responses and improve the answering effect of fault troubleshooting ideas.

[0027] Further, in one embodiment, refer to Figure 3As shown, text data is obtained by parsing the formatted document, specifically including: S1021: Logically divide the formatted document into logical blocks according to the keyword matching method to obtain logical blocks; S1022: Recognize the text in the logical block based on optical character recognition technology to obtain text data, extract the table content in the logical block based on the structured parsing method and convert it into a computable data format, recognize the logical block diagrams and flowcharts in the logical block based on computer vision technology, extract key information and combine it with the text data to form a knowledge index.

[0028] Specifically, for the document obtained after formatting and processing the collected vehicle function system design documents and vehicle function defect analysis documents, the document structure is logically divided into multiple logical blocks according to the keyword matching method. The logical blocks include signal logical blocks and function logical blocks. The signal logical blocks specifically include service signal logical blocks, CAN signal logical blocks, hardwired switch signal logical blocks, etc. The function logical blocks include terminal APP requirement logical blocks, central control screen display requirement logical blocks, VIU (vehicle domain control unit) logical requirement logical blocks, actuator controller requirement logical blocks, etc.

[0029] Next, the content of the logical block is identified and processed. According to the content data form of the logical block, optical character recognition technology is used to perform text conversion on formats such as PDF and Word to obtain text data; for table data, the structured parsing method is used to extract the table content and convert it into a computable data format (such as JSON, CSV, etc.); for the embedded logical block diagrams and flowcharts, computer vision technology is used to extract the key information in the diagrams and combine it with the text data to form a knowledge index.

[0030] Further, in one embodiment, for retrieving the technical materials associated with the fault description from the knowledge base in the present application, specifically: a: By analyzing the problem input by the user, that is, the fault description, identify the problem intention. The problem intention includes the fault type, the pre-operations for the fault occurrence, the vehicle sub-function domain system modules involved in the fault (such as the cockpit domain, vehicle control domain, and chassis domain, etc.), and the functional components that may be affected; by identifying the problem intention, ensure the pertinence and accuracy of the problem analysis; b: Based on the problem intention obtained from the fault description and identification, expand the problem (for example, for "the vehicle cannot go on high voltage", expand it to "what are the conditions for the vehicle to go on high voltage"), generate multiple equivalent query questions, perform hybrid retrieval in the knowledge base, call multiple retrieval methods such as keyword retrieval, sparse retrieval, and dense retrieval, search for knowledge cases and design documents related to the fault to obtain more comprehensive relevant information, sort the retrieved results according to relevance, and take the top preset number of retrieved results as the technical materials related to the fault description in the order of decreasing relevance.

[0031] It should be noted that if the retrieved result hits the expert Q&A knowledge, a higher sorting priority will be given to the current retrieved result, and the current retrieved result will be marked for subsequent use by the large language model.

[0032] Further, in one embodiment, as shown in Figure 4 Based on the technical materials and in combination with the large language model, generate a fault troubleshooting strategy for the fault cause and the corresponding fault cause to indicate the signals to be detected and the detection logic, specifically including: S201: Generate a fault cause inquiry description based on the technical materials and input it into the large language model to generate a fault cause; S202: Input the fault description and the generated fault cause into the large language model to generate a fault troubleshooting strategy; Among them, the fault troubleshooting strategy includes the signals to be detected, the associated vehicle protocols and vehicle data of the signals to be detected, the expected conditions of the signals to be detected, the signal detection order, and the jump logic.

[0033] Specifically, according to the retrieved technical materials, based on the knowledge items to which the technical materials belong, generate a fault cause inquiry description for asking questions to the large language model, and input the generated fault cause inquiry description into the large language model. The large language model generates text content that meets the fault cause inquiry description, so as to summarize the fault cause.

[0034] Exemplarily, the large language model can be DeepSeek-R1 or QWen series; the fault cause inquiry description can be "the XX function of the vehicle cannot be used / regulated / opened, please analyze the possible functional component modules involved, and analyze the possible causes of the fault from the perspective of the fault signal logic or the signal transmission link".

[0035] After obtaining the fault cause, input the fault cause and the previously input fault description into the large language model again to generate a fault troubleshooting strategy for fault diagnosis. The fault troubleshooting strategy can be stored in a structured configuration file.

[0036] Exemplarily, the configuration file adopts the JSON format, and the troubleshooting strategy in the configuration file includes but is not limited to the following fields: signals to be detected, associated vehicle-mounted protocols and vehicle-mounted data of the signals to be detected, expected conditions of the signals to be detected, signal detection sequence, and jump logic. For the jump logic, for example, there are multiple signals to be detected, and there is an association between the multiple signals. When the current signal detection passes, the jump logic is used to determine the next signal to be detected.

[0037] It should be added that the reason for storing the troubleshooting strategy in the configuration file is to convert the troubleshooting strategy into a format that can be understood by the program, which is convenient for the subsequent automation of detection.

[0038] Combining the obtained fault causes and troubleshooting strategies above, a detailed fault analysis report can be provided to the user, including possible fault causes and recommended troubleshooting steps, and the fault analysis report can be displayed through an intelligent interaction interface. The overall function supports API interface calls and standardized configuration file outputs to ensure seamless docking with subsequent signal parsing and detection, thereby completing the automated intelligent fault troubleshooting function.

[0039] Further, in one embodiment, as shown in Figure 5 Parse the vehicle-mounted protocol and vehicle-mounted data to obtain the signals corresponding to the fault moment, and sequentially detect the relevant signals according to the troubleshooting strategy to determine the fault signal, which specifically includes: S301: Parse the vehicle-mounted protocol and vehicle-mounted data to obtain the signals corresponding to the fault moment, and filter out the signals to be detected according to the signals to be detected shown in the troubleshooting strategy; S302: According to the signal detection sequence and jump logic of the troubleshooting strategy, and the expected conditions of the signals to be detected, sequentially detect the signals to be detected to determine the fault signal.

[0040] Specifically, parsing the vehicle-mounted protocol and vehicle-mounted data to obtain the signals corresponding to the fault moment, and filtering out the signals to be detected according to the signals to be detected shown in the troubleshooting strategy specifically includes: S3011: Parse the various vehicle-mounted protocols and vehicle-mounted data of the vehicle to obtain all signal information, and filter out the signals to be detected according to the signals to be detected shown in the troubleshooting strategy; S3012: Determine the fault cycle based on the pre-operation of the fault trigger, with the pre-operation as the reference; S3013: Obtain the signal value information of the signals to be detected during the fault cycle time period.

[0041] In practical applications, the vehicle-mounted protocols in formats such as ARXML and DBC, as well as vehicle-mounted data, can be parsed based on the CANTools tool to obtain information such as the frame ID, position, accuracy, offset, etc. of the signals. Then, the configuration file is read to export the fault troubleshooting strategies in the configuration file. According to the signal names involved in the fault troubleshooting strategies, the parsed signals are filtered to obtain the signals to be detected. Based on the pre-operations for fault triggering (since the exact moment of fault occurrence cannot be determined, the pre-operations for fault triggering are introduced), the fault period is determined with the pre-operations as the reference (the preset time length segments before and after the corresponding moments of the pre-operations are used as the fault period). According to the parsing results, the signal value information of the signals to be detected within the fault period is obtained. Further, the signal values can be preprocessed, including time synchronization and signal physical value parsing.

[0042] In this application, to detect the signals to be detected and determine the fault signals, specifically, the signal values of the signals to be detected within the fault period are detected to check whether the signals to be detected meet the expected conditions, thereby determining the fault signals. For example, when detecting signal A, the signal values of signal A within the fault period are compared and detected with the expected conditions, and based on whether the detection passes, it is judged whether signal A is a fault signal. The signal values of the signals to be detected within the fault period may be fixed values or variable values. The expected conditions are used to indicate the correct numerical values and the relationships between the numerical values of the signals within the fault period.

[0043] In this application, the detection of the signals includes single-signal detection and verification of the logical relationships between signals; single-signal detection is to compare the signal values of a single signal within the fault period with the numerical values in the expected conditions; the logical relationships between signals are to compare the signal values of related signals within the fault period with the set relationships in the expected conditions.

[0044] Single-signal detection is to detect the numerical magnitude, the interval it is in, and the corresponding state transitions of the signal value of the current signal. The numerical magnitude is to detect whether the signal value is greater than, less than, or equal to a specific value; the interval it is in is to detect whether the signal value is within a specific interval range; the corresponding state transition is to detect whether the signal jumps within the judgment interval based on the state quantity corresponding to the signal value, such as the suspension function state changing from the non-activated state to the activated state.

[0045] Verification of the logical relationships between signals is to detect whether the relationships between the signal values of different signals meet the expected conditions based on the finite state machine and the graph structure according to the fault troubleshooting strategy. Specifically, according to the fault troubleshooting strategy, it is gradually verified whether the relationships between the signal values of related signals meet the expected conditions. For example, based on the signal values, it is judged whether the relationships between the signal values of multiple signals, such as the magnitude or the linkage on / off relationship, meet the expected conditions.

[0046] Further, in one embodiment, the intelligent vehicle function fault troubleshooting method further includes: Generating troubleshooting evidence data for vehicle function fault troubleshooting according to the execution of the fault troubleshooting strategy; Wherein, the troubleshooting evidence data includes the name of the signal participating in the detection, the actual signal value of the fault signal at the fault moment, the expected conditions of the signal participating in the detection, and the detection result of the signal participating in the detection.

[0047] Specifically, during the fault troubleshooting process, record the execution of each troubleshooting step, including the detection result of the signal value, the comparison situation, and the troubleshooting conclusion. For the multi-step verification process, gradually store the determination results of each signal, and output the signals involved in the troubleshooting process and their corresponding names, that is, output the troubleshooting evidence data.

[0048] Further, in one embodiment, the intelligent vehicle function fault troubleshooting method of the present application further includes: S401: Analyze the historical data of vehicle function fault troubleshooting to determine whether there are abnormal troubleshooting situations, and actively add new fault troubleshooting strategies for abnormal troubleshooting situations; S402: Verify the newly added fault troubleshooting strategy and add the newly added fault troubleshooting strategy to the knowledge base after the verification is passed.

[0049] That is, in order to ensure that the knowledge base can be continuously optimized and expanded, and improve the accuracy of troubleshooting, a dynamic knowledge management mechanism for expert feedback collection and question-and-answer knowledge incremental update is introduced. Specifically, during the actual troubleshooting process, expert engineers will discover new fault modes, optimize existing troubleshooting logics, and supplement missing signal correlation relationships based on experience. For this reason, the present application provides an expert feedback mechanism to support experts in submitting improvement opinions, automatically recording and processing them, including: (1) Feedback collection: Experts can submit new signal comparison logics (such as the jump condition and timing correlation of a certain signal) or new fault mode descriptions based on the historical data of vehicle function fault troubleshooting after discovering abnormal situations, that is, actively add new fault troubleshooting strategies or modify existing fault troubleshooting strategies; (2) Rule optimization: For the newly added or corrected fault troubleshooting strategies, perform consistency checks to ensure that the optimized content will not affect the correctness of the existing troubleshooting logic; (3) Incremental update: When the newly added or corrected fault troubleshooting strategy passes expert review or is verified through data analysis, incorporate it into the knowledge base and associate it with the existing knowledge graph to improve the professionalism and accuracy of subsequent question-and-answer troubleshooting.

[0050] The intelligent troubleshooting method for vehicle function failures in this application constructs a semi-structured document parsing and vectorization strategy, extracts the relationships of functional components in the vehicle system field based on knowledge graph technology, and promotes knowledge fusion between different functional domains; based on the powerful content understanding ability of the large language model, it combines semantic retrieval and structured label retrieval to achieve precise positioning of knowledge base documents related to the mentioned function failures and output troubleshooting ideas. At the same time, interactive prompt words are designed to enable the large language model to further automatically generate troubleshooting strategies according to the troubleshooting ideas; by parsing in-vehicle protocols and in-vehicle data, it collects the specific parameter values of in-vehicle signals within a specific time window, matches and detects them with the signals involved in the troubleshooting strategy, and performs troubleshooting jumps between different trigger conditions based on the finite state machine to achieve automatic troubleshooting of function failure signals; a complete fault cause analysis and recording mechanism is constructed, and the entire automatic troubleshooting process is seamlessly recorded to ensure full process traceability of the process, which is conducive to engineers intuitively restoring the troubleshooting process and accurately positioning the fault source; the knowledge base is incrementally updated, and the knowledge base is continuously updated and optimized according to the actual troubleshooting process and expert feedback to improve the accuracy and efficiency of problem solving.

[0051] In a second aspect, an embodiment of this application also provides an intelligent troubleshooting device for vehicle function failures.

[0052] In one embodiment, referring to Figure 6 , Figure 6 is a schematic diagram of the functional modules of the intelligent troubleshooting device for vehicle function failures in this application. As Figure 6 shown, the intelligent troubleshooting device for vehicle function failures includes: a retrieval module, a generation module, and a troubleshooting module.

[0053] The retrieval module is used to retrieve technical materials related to the fault description from the knowledge base according to the fault description, where the knowledge base is constructed based on the vehicle function system design document; the generation module is used to generate fault causes and troubleshooting strategies for the corresponding fault causes indicating the signals to be detected and the detection logic based on the technical materials and in combination with the large language model; the troubleshooting module is used to parse the in-vehicle protocol and in-vehicle data to obtain the signals at the corresponding fault moment, and sequentially detect the relevant signals according to the troubleshooting strategy to determine the fault signal, so as to achieve vehicle function fault troubleshooting.

[0054] In a third aspect, an embodiment of this application provides an intelligent troubleshooting device for vehicle function failures. The intelligent troubleshooting device for vehicle function failures can be a device with data processing functions such as a personal computer (PC), a laptop computer, a server, etc.

[0055] Referring to Figure 7 , Figure 7It is a schematic diagram of the hardware structure of the intelligent vehicle function fault troubleshooting device involved in the solution of the embodiment of the present application. In the embodiment of the present application, the intelligent vehicle function fault troubleshooting device may include a processor, a memory, a communication interface, and a communication bus.

[0056] Among them, the communication bus can be of any type and is used to interconnect the processor, the memory, and the communication interface.

[0057] The communication interface includes interfaces such as input / output (I / O) interfaces, physical interfaces, and logical interfaces for interconnecting components inside the intelligent vehicle function fault troubleshooting device, as well as interfaces for interconnecting the intelligent vehicle function fault troubleshooting device with other devices (such as other computing devices or user devices). The physical interface can be an Ethernet interface, an optical fiber interface, an ATM interface, etc.; the user device can be a display (Display), a keyboard (Keyboard), etc.

[0058] The memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical memory, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0059] The processor can be a general-purpose processor, and the general-purpose processor can call the intelligent vehicle function fault troubleshooting program stored in the memory and execute the intelligent vehicle function fault troubleshooting method provided by the embodiment of the present application. For example, the general-purpose processor can be a central processing unit (CPU), a graphics processing unit (GPU). Among them, the method executed when the intelligent vehicle function fault troubleshooting program is called can refer to each embodiment of the intelligent vehicle function fault troubleshooting method of the present application, which will not be elaborated here.

[0060] Those skilled in the art can understand that Figure 7 the hardware structure shown in does not constitute a limitation to the present application, and it may include more or fewer components than shown in the figure, or combine some components, or have different component arrangements.

[0061] In a fourth aspect, the embodiment of the present application also provides a computer-readable storage medium.

[0062] A vehicle function fault intelligent troubleshooting program is stored on a computer-readable storage medium of the present application. When the vehicle function fault intelligent troubleshooting program is executed by a processor, the steps of the vehicle function fault intelligent troubleshooting method as described above are implemented.

[0063] Among them, the method implemented when the vehicle function fault intelligent troubleshooting program is executed can refer to each embodiment of the vehicle function fault intelligent troubleshooting method of the present application, which will not be elaborated here.

[0064] In the description of the embodiments of the present application, the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products, or devices. The descriptions of terms such as "first", "second", and "third" are used to distinguish different objects, etc., and do not represent a sequential order, nor do they limit that "first", "second", and "third" are different types.

[0065] In the description of the embodiments of the present application, terms such as "exemplary", "for example", or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design solution described as "exemplary", "for example", or "for instance" in the embodiments of the present application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of terms such as "exemplary", "for example", or "for instance" is intended to present relevant concepts in a specific manner.

[0066] In the description of the embodiments of the present application, unless otherwise specified, " / " means "or". For example, A / B can mean A or B; "and / or" in the text is merely a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of the present application, "a plurality of" means two or more than two.

[0067] In some processes described in the embodiments of the present application, a plurality of operations or steps appear in a specific order. However, it should be understood that these operations or steps may not be executed in the order in which they appear in the embodiments of the present application or may be executed in parallel. The serial numbers of the operations are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed in sequence or in parallel, and these operations or steps may be combined.

[0068] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-described embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. This computer software product is stored in a storage medium as described above (such as ROM / RAM, magnetic disk, optical disc), and includes several instructions for causing a terminal device to execute the methods described in various embodiments of the present application.

[0069] The above are only the preferred embodiments of the present application, and do not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present application, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present application.

Claims

1. An intelligent troubleshooting method for vehicle functional failures, characterized in that, The intelligent troubleshooting method for vehicle function failures includes: Retrieving technical materials associated with the fault description from a knowledge base according to the fault description, where the knowledge base is constructed based on vehicle function system design documents; Generating fault causes and a fault troubleshooting strategy for the corresponding fault causes indicating the signals to be detected and the detection logic based on the technical materials and in combination with a large language model; Analyzing in-vehicle protocols and in-vehicle data to obtain signals at the corresponding fault moment, and sequentially detecting relevant signals according to the fault troubleshooting strategy to determine the fault signal, thereby realizing the troubleshooting of vehicle function failures.

2. The intelligent troubleshooting method for vehicle functional failures according to claim 1, wherein, The generating of fault causes and a fault troubleshooting strategy for the corresponding fault causes indicating the signals to be detected and the detection logic based on the technical materials and in combination with a large language model specifically includes: Generating a fault cause query description based on the technical materials and inputting it into the large language model to generate fault causes; Inputting the fault description and the generated fault causes into the large language model to generate a fault troubleshooting strategy; Among them, the fault troubleshooting strategy includes signals to be detected, associated in-vehicle protocols and in-vehicle data of the signals to be detected, expected conditions of the signals to be detected, signal detection order, and jump logic.

3. The intelligent troubleshooting method for vehicle functional failures according to claim 2, wherein The analyzing in-vehicle protocols and in-vehicle data to obtain signals at the corresponding fault moment, and sequentially detecting relevant signals according to the fault troubleshooting strategy to determine the fault signal specifically includes: Analyzing in-vehicle protocols and in-vehicle data to obtain signals at the corresponding fault moment, and screening out signals to be detected according to the signals to be detected shown in the fault troubleshooting strategy; Sequentially detecting the signals to be detected according to the signal detection order and jump logic of the fault troubleshooting strategy and the expected conditions of the signals to be detected to determine the fault signal.

4. The intelligent troubleshooting method for vehicle functional failures according to claim 3, characterized in that, The analyzing in-vehicle protocols and in-vehicle data to obtain signals at the corresponding fault moment, and screening out signals to be detected according to the signals to be detected shown in the fault troubleshooting strategy specifically includes: Analyzing each in-vehicle protocol and in-vehicle data of the vehicle to obtain all signal information, and screening out signals to be detected according to the signals to be detected shown in the fault troubleshooting strategy; Determining the fault cycle based on the pre-operation before the fault trigger, with the pre-operation as the reference; Obtaining the signal value information of the signals to be detected within the fault cycle time period.

5. The intelligent troubleshooting method for vehicle function failures according to claim 4, wherein: Detecting the signal values of the signals to be detected within the fault cycle time period, and detecting whether the signals to be detected meet the expected conditions, so as to determine the fault signal.

6. The intelligent troubleshooting method for vehicle function failures according to claim 1, wherein: The detection of the signals to be detected includes single signal detection and verification of the logical relationship between signals; The single signal detection is to detect the numerical value, the interval where it is located, and the corresponding state transition of the signal value of the current signal; The verification of the logical relationship between signals is to detect whether the relationship between the signal values of different signals meets the expected conditions based on a finite state machine and a graph structure according to the fault troubleshooting strategy.

7. The intelligent troubleshooting method for vehicle function failures according to claim 1, characterized in that, The intelligent troubleshooting method for vehicle function failures further includes: Generating troubleshooting evidence data for vehicle function failures according to the execution situation of the fault troubleshooting strategy; Among them, the troubleshooting evidence data includes the name of the signal participating in the detection, the actual signal value of the fault signal at the fault moment, the expected conditions of the signal participating in the detection, and the detection result of the signal participating in the detection.

8. The intelligent troubleshooting method for vehicle functional failures according to claim 1, characterized in that, For the construction of the knowledge base, it specifically includes: Obtain the vehicle function system design document and the vehicle function defect analysis document, and perform formatting processing to obtain a formatted document; Parse the formatted document to obtain text data, and add the vehicle model project, function component name, and title to the text data; Based on the recursive character text segmentation technology, segment the text data to obtain multiple text blocks, and based on the semantic representation model, extract the semantic features of the text blocks and convert the extracted semantic features into text vectors to obtain the final text data; Based on the text vector indexing technology, perform index storage on the final text data to realize the construction of the knowledge base.

9. The intelligent troubleshooting method for vehicle function failures according to claim 8, wherein, The parsing of the formatted document to obtain text data specifically includes: According to the keyword matching method, perform logical chunking on the formatted document to obtain logical chunks; Based on the optical character recognition technology, recognize the text in the logical chunks to obtain text data, based on the structured parsing method, extract the table content in the logical chunks and convert it into a computable data format, and based on the computer vision technology, recognize the logic block diagrams and flowcharts in the logical chunks, extract key information and combine it with the text data to form a knowledge index.

10. The intelligent troubleshooting method for vehicle functional failures according to claim 1, wherein, The vehicle function fault intelligent troubleshooting method further includes: Analyze the historical data of vehicle function fault troubleshooting to determine whether there are abnormal troubleshooting situations, and actively add fault troubleshooting strategies for abnormal troubleshooting situations; Verify the newly added fault troubleshooting strategy, and after the verification is passed, add the newly added fault troubleshooting strategy to the knowledge base.

11. An intelligent troubleshooting device for vehicle function failures, characterized in that, The vehicle function fault intelligent troubleshooting device includes: A retrieval module, which is used to retrieve the knowledge base according to the fault description to obtain technical materials associated with the fault description, wherein the knowledge base is constructed based on the vehicle function system design document; A generation module, which is used to generate the fault cause and the fault troubleshooting strategy indicating the signals to be detected and the detection logic corresponding to the fault cause based on the technical materials and in combination with the large language model; 12. An intelligent vehicle function fault troubleshooting device, characterized in that, A troubleshooting module, which is used to parse the in-vehicle protocol and in-vehicle data to obtain the signals at the corresponding fault moment, and sequentially detect the relevant signals according to the fault troubleshooting strategy to determine the fault signal, so as to realize the troubleshooting of vehicle function faults.

13. A computer-readable storage medium, characterized in that, The vehicle function fault intelligent troubleshooting device includes a processor, a memory, and a vehicle function fault intelligent troubleshooting program stored on the memory and executable by the processor. When the vehicle function fault intelligent troubleshooting program is executed by the processor, the steps of the vehicle function fault intelligent troubleshooting method according to any one of claims 1 to 10 are realized. A vehicle function fault intelligent troubleshooting program is stored on the computer-readable storage medium, and when the vehicle function fault intelligent troubleshooting program is executed by the processor, the steps of the vehicle function fault intelligent troubleshooting method according to any one of claims 1 to 10 are realized.

Images