Ai-based vehicle diagnosis method and system, electronic device and storage medium
Patent Information
- Application Number
- PCT/CN2025/094729
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-03-24
- Filing Date
- 2025-05-14
- Publication Date
- 2026-10-01
Smart Images
Figure CN2025094729_01102026_PF_FP_ABST
Abstract
Description
AI-based automotive diagnostic methods, systems, electronic devices, and storage media Technical Field
[0001] This application relates to the field of vehicle diagnostics, and more particularly to an AI-based vehicle diagnostic method, system, electronic device, and storage medium. Background Technology
[0002] As automotive electronics become increasingly sophisticated, vehicle menus become more and more complex, making it difficult for users to fully describe problems when using diagnostic equipment.
[0003] Currently, traditional automotive intelligent diagnostic assistants generally rely solely on specific data such as fault codes input by the user for diagnosis, and cannot analyze other input data from the user, making it difficult for the user to operate. Therefore, how to reduce the operational complexity of vehicle diagnosis has become an urgent problem to be solved. Summary of the Invention
[0004] This application provides an AI-based vehicle diagnostic method, system, electronic device, and storage medium. By introducing AI-assisted controls, users can activate the AI function for vehicle diagnostics with just one click, eliminating the need for complex operating procedures and thus reducing the operational complexity of vehicle diagnostics.
[0005] In a first aspect, embodiments of this application provide an AI-based vehicle diagnostic method, applied to diagnostic equipment, comprising:
[0006] The vehicle diagnostic main page of the diagnostic device is displayed, wherein the vehicle diagnostic main page includes AI-assisted controls;
[0007] The target working state of the diagnostic device is obtained; the target working state includes one of the following: normal state and diagnostic abnormal state; when the target working state includes the diagnostic abnormal state, in response to the trigger operation of the AI-assisted control, an AI diagnostic interaction page is displayed; the AI diagnostic interaction page includes: a first information display area, a second information display area, and an information input window;
[0008] In response to an information input operation on the information input window, the target question text corresponding to the information input operation is determined, and the target question text is displayed in the first information display area;
[0009] The target problem text is analyzed by a preset target AI model in the diagnostic device to obtain target diagnostic suggestions;
[0010] The target diagnostic recommendations are displayed in the second information display area.
[0011] Secondly, embodiments of this application provide an AI-based automotive diagnostic system applied to diagnostic equipment. The system includes: a display unit, a response unit, and a diagnostic unit, wherein:
[0012] The display unit is used to display the vehicle diagnostic main page of the diagnostic device, wherein the vehicle diagnostic main page includes AI-assisted controls;
[0013] The response unit is configured to acquire the target operating state of the diagnostic device; the target operating state includes one of the following: normal state and diagnostic abnormal state; when the target operating state includes the diagnostic abnormal state, in response to a trigger operation on the AI-assisted control, an AI diagnostic interaction page is displayed; the AI diagnostic interaction page includes: a first information display area, a second information display area, and an information input window; in response to an information input operation on the information input window, the target problem text corresponding to the information input operation is determined, and the target problem text is displayed in the first information display area;
[0014] The diagnostic unit is used to analyze the target problem text using a preset target AI model in the diagnostic device to obtain target diagnostic suggestions;
[0015] The display unit is also used to display the target diagnostic suggestion in the second information display area.
[0016] Thirdly, embodiments of this application provide an electronic device, including: a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the programs include instructions for performing the steps in the first aspect of embodiments of this application.
[0017] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program for electronic data interchange, wherein the computer program causes a computer to perform some or all of the steps described in the first aspect of embodiments of this application.
[0018] Fifthly, embodiments of this application provide a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps described in the first aspect of embodiments of this application. The computer program product may be a software installation package.
[0019] Implementing this application will have the following beneficial effects:
[0020] As can be seen, the AI-based vehicle diagnostic method described in this application displays the vehicle diagnostic main page of the diagnostic device, wherein the vehicle diagnostic main page includes AI-assisted controls; it obtains the target working state of the diagnostic device; the target working state includes one of the following: normal state, diagnostic abnormal state; when the target working state includes the diagnostic abnormal state, in response to the trigger operation of the AI-assisted controls, it displays an AI diagnostic interaction page; the AI diagnostic interaction page includes: a first information display area, a second information display area, and an information input window; in response to the information input operation of the information input window, it determines the target problem text corresponding to the information input operation and displays the target problem text in the first information display area; it analyzes the target problem text through a preset target AI model in the diagnostic device to obtain a target diagnostic suggestion; and it displays the target diagnostic suggestion in the second information display area. Thus, by introducing AI-assisted controls, users only need to click once to activate the AI function (i.e., the target AI model) for vehicle diagnostics, without complicated operation procedures, thereby reducing the operational complexity of vehicle diagnostics. Attached Figure Description
[0021] Figure 1 is a schematic diagram of the structure of a diagnostic device provided in an embodiment of this application;
[0022] Figure 2 is a schematic diagram of an application scenario of an AI-based vehicle diagnostic method provided in an embodiment of this application;
[0023] Figure 3 is a flowchart illustrating an AI-based vehicle diagnostic method provided in an embodiment of this application;
[0024] Figure 4 is a schematic diagram of a vehicle diagnostic main page provided in an embodiment of this application;
[0025] Figure 5 is a schematic diagram of an AI diagnostic interactive page provided in an embodiment of this application;
[0026] Figure 6 is a flowchart illustrating an abnormal state resolution method provided in an embodiment of this application;
[0027] Figure 7 is a flowchart illustrating a model optimization method provided in an embodiment of this application;
[0028] Figure 8 is a functional unit block diagram of an AI-based vehicle diagnostic system provided in an embodiment of this application;
[0029] Figure 9 is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation
[0030] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.
[0031] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.
[0032] It should be understood that the term "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this document indicates that the preceding and following related objects are in an "or" relationship. In the embodiments of this application, "multiple" refers to two or more.
[0033] In the embodiments of this application, "at least one item" or its similar expression refers to any combination of these items, including any combination of a single item or a plurality of items. "One or more" means one or more, while "multiple" means two or more. For example, "at least one item" of a, b, or c can represent the following seven cases: a, b, c; a and b; a and c; b and c; a, b, and c. Each of a, b, and c can be an element or a set containing one or more elements.
[0034] In this application, the term "connection" refers to various connection methods, such as direct connection or indirect connection, to achieve communication between devices. This application does not impose any limitations on this.
[0035] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0036] The electronic devices described in the embodiments of this application may include diagnostic devices, smartphones (such as Android phones, iOS phones, Windows Phones, etc.), tablet computers, handheld computers, laptops, video matrices, monitoring platforms, mobile internet devices (MIDs), or wearable devices, etc. The above are merely examples and not exhaustive, and include but are not limited to the above devices.
[0037] Of course, the aforementioned electronic devices can also be servers, such as cloud servers.
[0038] The following describes the relevant content, concepts, meanings, technical issues, technical solutions, and beneficial effects involved in the embodiments of this application.
[0039] First, let me explain some of the technical terms used in this application:
[0040] Vehicle diagnostics refers to the process of inspecting, testing, analyzing, and judging the technical condition, performance indicators, and fault status of automobiles and other vehicles using professional technical means, tools, and methods. Through vehicle diagnostics, it is possible to determine whether a vehicle has a fault, its specific location and cause, and to predict potential problems, providing a basis for vehicle repair, maintenance, and normal operation. For example, diagnostic equipment is used to read vehicle fault codes, check various engine parameters, and test the operating status of the vehicle's electrical system.
[0041] Large Language Models (LLMs) are deep learning-based artificial intelligence models that learn and train on massive amounts of text data to understand the semantics, syntax, and context of natural language and generate natural and fluent text responses. LLMs typically have a large parameter scale, such as the GPT (Generative Pretrained Transformer) series of models. They can be used for various natural language processing tasks, such as text generation, question answering systems, machine translation, and text summarization. In the field of vehicle diagnostics, LLMs can generate relevant diagnostic suggestions, fault cause analyses, and other text content based on user input describing vehicle malfunctions.
[0042] Transfer learning is a machine learning technique that aims to apply knowledge learned from one or more source tasks to a target task. In transfer learning, the source and target tasks typically have some correlation. By leveraging knowledge such as feature representations and model parameters already learned from the source task, the amount of training data and training time for the target task can be reduced, thereby improving the model's performance on the target task. For example, in vehicle diagnostics, a model can be pre-trained on large-scale general natural language data, and then transfer learning can be used to transfer the knowledge of the pre-trained model to a specific task in the field of vehicle fault diagnosis, enabling the model to learn and process text data related to vehicle fault diagnosis more quickly and effectively.
[0043] RAG (Retrieval Augmented Generation) service is a technology service that combines information retrieval and text generation. In RAG services, based on the user's input question or command, relevant information is first retrieved from a pre-built knowledge base (such as a document library or database). Then, the retrieved information, along with the user's input, is fed into a text generation model (such as a large language model). The model then generates a corresponding response or text based on this information. By incorporating retrieved external knowledge, RAG services can make the generated text more accurate, reliable, and targeted. In vehicle diagnostics, RAG services can retrieve information related to vehicle faults, such as repair manuals, fault cases, and technical documents, to assist the large language model in generating more accurate diagnostic suggestions and solutions.
[0044] With the increasing complexity of automotive electronic systems, traditional diagnostic equipment suffers from significant shortcomings in real-time performance, interactivity, and self-learning capabilities. This makes it difficult for users to obtain immediate help and support when encountering problems, and when the diagnostic process gets stuck or reaches a dead end, users are often unable to resolve the issue independently, forcing the diagnostic work to be interrupted. This application proposes an AI-based automotive diagnostic method. By introducing AI-assisted controls into the diagnostic equipment, users can activate the AI interactive function of the diagnostic equipment with a single click, allowing them to input problem descriptions in natural language. The AI-assisted module in the diagnostic equipment incorporates multiple advanced AI models to process data from user operation records, log information, and the current vehicle status. It uses deep learning algorithms and fault knowledge graphs for comprehensive analysis, providing highly targeted diagnostic suggestions. This method not only continuously monitors subsequent user operations and updates model parameters based on new data to improve the accuracy and efficiency of future diagnostic tasks, but also possesses strong self-learning capabilities, continuously optimizing its own model parameters. This enhances the ability to identify and handle new types of faults, significantly improving the user experience. It not only reduces the operational complexity of vehicle diagnostics but also reduces the number of interruptions during the diagnostic process, thereby improving overall work efficiency.
[0045] Please refer to Figure 1, which is a schematic diagram of the structure of a diagnostic device provided in an embodiment of this application. It can be seen that the diagnostic device may include: a control module, a user interaction module, an AI-assisted module, a data storage module, etc., which are not limited here.
[0046] The control module, acting as the "brain" of the diagnostic device, is responsible for coordinating and managing the operation of the entire device. It can schedule the work of various modules, ensuring smooth collaboration between them. For example, after the user inputs information through the user interaction module, the control module coordinates the AI-assisted module to process the input information and transmits the processing results to the corresponding modules for display or storage.
[0047] The user interaction module provides an input interface for users, allowing them to input various information into the diagnostic equipment, such as vehicle fault descriptions, vehicle model, and mileage. It also collects user behavior data, such as the frequency of user input and the duration of viewing diagnostic results. This data can be used to optimize the user experience and evaluate the diagnostic system. Furthermore, it displays the results and information generated by the diagnostic equipment to the user, including various content on the vehicle diagnostic main page and the AI diagnostic interaction page, such as the target problem text and target diagnostic suggestions. In addition, it provides intuitive feedback to the user on the system's status, such as processing, completion, and error messages, allowing the user to understand the progress of the diagnostic process.
[0048] The AI-assisted module incorporates various AI models, such as a target AI model, responsible for processing and analyzing the user-input target question text and data acquired from other modules. In automotive diagnostics, it can analyze potential causes and locations of vehicle faults based on the description of the fault, using deep learning algorithms and existing fault knowledge graphs, and recommend corresponding diagnostic methods and solutions. Furthermore, it possesses self-learning and optimization capabilities, adjusting and optimizing its model parameters based on continuously input new data and actual diagnostic results to improve diagnostic accuracy and efficiency. For example, when encountering new types or situations of vehicle faults, the AI-assisted module can learn relevant data and update its knowledge base to more accurately handle similar problems in subsequent diagnoses.
[0049] The data storage module is responsible for storing various data generated and used by the diagnostic equipment during operation, including the vehicle's historical diagnostic records, user-inputted problem texts and feedback information, AI model training data, and externally acquired automotive technical data and fault cases. It effectively organizes and manages this data through a well-structured data system and database management system, facilitating data querying, updating, and retrieval.
[0050] Please refer to Figure 2, which is a schematic diagram of an application scenario of an AI-based vehicle diagnostic method provided in an embodiment of this application. As can be seen, the user is the main entity initiating the diagnostic request. By interacting with the diagnostic device, the user inputs the problem or information of the vehicle (i.e., the target vehicle in Figure 2) to the diagnostic device, such as describing the fault phenomenon or operating status of the vehicle. The diagnostic device is in a core position. On the one hand, it receives the information input by the user. On the other hand, the diagnostic device can also connect with the target vehicle to obtain relevant vehicle data. Based on these two, it performs AI diagnosis and finally feeds back the diagnostic results to the user. The target vehicle is the car that needs to be diagnosed. It communicates with the diagnostic device and provides the diagnostic device with its own vehicle data so that the diagnostic device can analyze the vehicle's condition and find the cause of the fault.
[0051] Please refer to Figure 3, which is a flowchart illustrating an AI-based vehicle diagnostic method provided in an embodiment of this application. This method is applied to diagnostic equipment and may include, but is not limited to, the following steps:
[0052] S301. Display the vehicle diagnostic main page of the diagnostic device, wherein the vehicle diagnostic main page includes AI-assisted controls.
[0053] In this application embodiment, the diagnostic device may include at least one of the following: smart mobile devices (e.g., mobile phones, computers), portable diagnostic instruments, professional repair shop diagnostic equipment, etc., which are not limited here.
[0054] In a specific embodiment, when the diagnostic device is started, the vehicle diagnostic main page can be displayed on the diagnostic device's page. Users can perform operations on the vehicle diagnostic main page, such as clicking the AI-assisted control to enter the AI diagnostic interactive page.
[0055] In one embodiment, please refer to Figure 4, which is a schematic diagram of a vehicle diagnostic main page provided in an embodiment of this application. It can be seen that the vehicle diagnostic main page may include: general diagnostics, vehicle connection, diagnostic history, AI-assisted controls, etc., which are not limited here. Among them, general diagnostics: This is a function option. After clicking, the diagnostic device can perform diagnostics on the vehicle according to the traditional diagnostic process, by reading the vehicle's fault codes, detecting sensor data, etc., suitable for situations with a clear diagnostic direction or where in-depth AI analysis is not required.
[0056] Connect to Vehicle: This option establishes a connection between the diagnostic device and the target vehicle. Clicking this option allows the diagnostic device to connect to the user's vehicle via OBD interface, Bluetooth, or other wireless connection methods. This enables the diagnostic device to obtain real-time data and information from the vehicle, preparing for subsequent diagnostic operations.
[0057] Diagnostic History: Clicking this option allows you to view previous diagnostic records for the vehicle, including the diagnosis time, problems found, and measures taken. This helps users and repair personnel understand the vehicle's past condition and conduct comparative analysis.
[0058] AI-assisted control: Presented in a circle, this is the entry point for vehicle diagnostics using AI technology. When the user triggers this control, they will enter the AI diagnostic interaction page, where they can interact with the AI to obtain more intelligent and comprehensive diagnostic suggestions. For example, after entering a description of the vehicle fault, the AI will provide analysis results based on a large amount of data and algorithms.
[0059] These functional modules together form the main page of vehicle diagnostics, providing users with diverse access points for vehicle diagnostic operations.
[0060] S302. Obtain the target working state of the diagnostic device; the target working state includes one of the following: normal state, diagnostic abnormal state; when the target working state includes the diagnostic abnormal state, in response to the trigger operation of the AI auxiliary control, display the AI diagnostic interaction page; the AI diagnostic interaction page includes: a first information display area, a second information display area, and an information input window.
[0061] In this embodiment, the target operating state of the diagnostic equipment can be obtained first. Specifically, the diagnostic equipment can have a built-in status monitoring system. If it is a hardware problem, various parameters of the diagnostic equipment's operation, such as temperature, voltage, current, and speed, can be collected in real time through sensors. The control module of the diagnostic equipment will compare these real-time parameters with preset normal ranges. If the parameters are all within the normal range and all functions of the equipment are operating normally, then the equipment can be determined to be in a normal state. If one or more parameters exceed the normal range, or if the diagnostic equipment displays fault codes or alarm prompts, it means that the diagnostic equipment is in a diagnostic abnormal state. If it is a software problem, the status monitoring system can analyze the current operating data of the diagnostic equipment to determine whether there are potential faults or abnormalities. If so, the diagnostic equipment is determined to be in a diagnostic abnormal state; otherwise, the diagnostic equipment is in a normal state. For example, whether the diagnostic process has reached a dead end, or whether the user cannot continue after reaching a certain step, causing the diagnostic work to be interrupted, etc. When the target working state includes a diagnostic abnormal state, the user can click the AI-assisted control to trigger the AI diagnostic function of the diagnostic device. Specifically, the diagnostic device can respond to the user's trigger operation on the AI-assisted control (i.e., the user's click operation) and display the AI diagnostic interaction page. The user can perform operations on the AI diagnostic interaction page to make the diagnostic device exit the diagnostic abnormal state and return to the normal state.
[0062] Please refer to Figure 5, which is a schematic diagram of an AI diagnostic interaction page provided in an embodiment of this application. As can be seen, the AI diagnostic interaction page may include: a first information display area, a second information display area, an information input window, clear, submit, etc., which are not limited here. The first information display area is located at the top of the page. When the diagnostic device has not received information from the user, it prompts "No input content available. Please click the 'Information Input Window' to input content for analysis~" to guide the user's input. After the user inputs content, this area is used to display the user's input content (i.e., the target problem text), allowing the user to confirm the entered information.
[0063] The second information display area is located in the middle of the page. When the diagnostic device has not received any information from the user, it is displayed as a blank area. This area is used to display the target diagnostic suggestions derived by the AI model after analyzing the user's input.
[0064] Information Input Window: Located at the bottom of the page, this is where users enter information such as a description of the vehicle problem. It can also be accompanied by "Clear" and "Submit" buttons. Clicking the "Clear" button will clear the content of the information input window, while clicking the "Submit" button will send the input information to the target AI model for analysis and processing. The overall page layout is simple, guiding users through the information input window and operation buttons to complete the interactive diagnostic process with AI.
[0065] Optionally, in one embodiment, please refer to Figure 6, which is a flowchart illustrating an abnormal state resolution method provided by an embodiment of this application. The specific steps are as follows:
[0066] Start: The point where the process begins;
[0067] Check if the diagnostic equipment is in an abnormal diagnostic state: Determine the current working status of the diagnostic equipment. If it is normal (N), end the process directly; if it is abnormal (Y), proceed to the next step.
[0068] Determine the target cause of the abnormal state: find out the specific reason for the equipment malfunction.
[0069] Generate target solutions based on the causes of target anomalies: Develop corresponding solutions based on the identified causes.
[0070] Implement the target solution: Implement the established solution.
[0071] Check if the diagnostic equipment is in an abnormal diagnostic state: Check the equipment status again. If it is still abnormal (Y), return to re-determine the cause of the abnormality and resolve it; if it returns to normal (N), the process ends.
[0072] This process involves continuously detecting and identifying the cause, generating and implementing solutions until the abnormal state of the equipment is eliminated, thereby ensuring the normal operation of the diagnostic equipment.
[0073] For example, when the target operating state is in a diagnostic anomaly state, the device operation log of the diagnostic device can be obtained. The target AI model can then analyze the device operation log to determine the cause of the diagnostic device anomaly and obtain the target anomaly cause. Suppose the target anomaly cause is "diagnostic process stuck". Then, a stuck prompt can be displayed in the first information display area and analyzed. For example, suppose the first information display area prompts "During the data acquisition process, the device is stuck due to insufficient memory". The user can understand the problem based on this prompt.
[0074] The target solution is displayed in the second information display area. For example, the second information display area could display "Please try closing other unnecessary programs to free up memory space." Following this target solution, the user can close currently running unnecessary programs through the diagnostic device's task manager or related operation interface. Alternatively, the user can authorize the target AI model to automatically execute the target solution. After executing the target solution, the user can provide feedback on the operation in the information input window. For example, the user can enter "[specific program name] etc. closed, requesting an attempt to resume the process." Upon receiving feedback, the diagnostic device automatically attempts to restart the stalled data acquisition process. If there is sufficient memory space, the process can resume normally, and the device status returns to normal. For another example, assuming the target anomaly is "diagnostic process interruption," the interruption reason (i.e., the target anomaly reason) can be displayed in the first information display area. For example, the first information display area could display "Network connection interruption caused diagnostic process interruption." This clearly indicates a network-related problem.
[0075] The second information display area shows the target solution, for example, suggesting solutions like "Check network device connectivity, restart the router, or try switching networks." Users can first check if the connection between their device and network device (such as the router) is loose, and then restart the router. If the problem persists, they can try switching to another available network in the device's network settings, or the target AI model can automatically execute the target solution. After executing the target solution, the user can enter the processing result in the information input window, for example, "Network connectivity checked, router restarted and switched to a backup network, requesting resumption of the diagnostic process." Upon receiving this information, the diagnostic device will attempt to reconnect to the network and resume the interrupted diagnostic process. If the network connection is successful, the process continues, and the diagnostic device returns to normal operation.
[0076] S303. In response to an information input operation on the information input window, determine the target question text corresponding to the information input operation, and display the target question text in the first information display area.
[0077] In this embodiment of the application, in response to an information input operation on the information input window, the question content input by the information input operation is received, and then the question can be used as the target question text and displayed in the first information display area.
[0078] Optionally, determining the target question text corresponding to the information input operation in response to the information input window may include the following steps:
[0079] S31. In response to an information input operation on the information input window, receive the question text input by the information input operation to obtain a first question text; the first question text is a natural language expression text;
[0080] S32. Extract the first keyword from the first question text to obtain m first keywords; m is a positive integer;
[0081] S33. Determine the target communication intent based on the m first keywords; the target communication intent includes one of the following: chat intent, inquiry intent;
[0082] S34. Based on the target communication intent and the m first keywords, convert the first question text into text in a preset format to obtain the target question text.
[0083] In this embodiment of the application, the preset format can be preset in advance or defaulted.
[0084] In a specific embodiment, in response to an information input operation on the information input window, the system can receive the question text entered by the information input operation to obtain the first question text. For example, if a user clicks on the information input window and then starts typing "My car's engine malfunction light came on while driving, and it was shaking violently," then "My car's engine malfunction light came on while driving, and it was shaking violently" is the obtained first question text. Next, the first keywords in the first question text can be extracted to obtain m first keywords. Specifically, a word segmentation tool can be used to split the first question text into individual words or phrases to obtain m first keywords. For example, for Chinese text, tools such as jieba word segmentation can be used, and for English text, word segmentation functions in NLTK (Natural Language Toolkit) can be used for word segmentation. Assuming that jieba word segmentation is used to extract keywords from "My car's engine malfunction light came on while driving, and it was shaking violently," the keywords obtained are ['car', 'driving', 'engine', 'malfunction light', 'on', 'shaking']. These keywords can accurately summarize the core information described in the text, namely, the fault phenomenon that the engine malfunction light comes on and the vehicle shakes while driving.
[0085] Furthermore, the target communication intent can be determined based on m primary keywords. Specifically, the m primary keywords can be categorized to identify keywords with obvious semantic biases. The target communication intent can then be determined based on these keywords with obvious biases. For example, keywords such as "fault," "cause," "solution," and "what to do" are usually related to asking for specific information, while keywords such as "feeling," "interesting," and "casual chat" are more inclined towards chat scenarios. Finally, the first question text can be converted into a preset format text based on the target communication intent and the m primary keywords to obtain the target question text. For example, the preset format can be "[communication intent]: [communication topic]: [keyword 1], [keyword 2]... [question text]", converting the first question text into the preset format text, which is the target question text.
[0086] Thus, by extracting the m primary keywords from the first question text, the core content of the text can be quickly grasped, and irrelevant information can be filtered out. For example, when processing a large number of user inquiries, quickly locating keywords such as "car," "fault," and "repair" allows for direct focus on the key issues, eliminating the need to repeatedly search for crucial information within lengthy texts, significantly improving the speed and accuracy of information processing. Furthermore, converting the question text into a pre-formatted target question text makes the data more standardized and structured. This facilitates both statistical analysis of user questions and evaluation of the system's response effectiveness. For instance, it's easy to statistically analyze the frequency of questions with different inquiries, understand the hot topics users are concerned about, and provide data support for system optimization.
[0087] S304. The target problem text is analyzed by the preset target AI model in the diagnostic device to obtain target diagnostic suggestions.
[0088] In this embodiment of the application, the target problem text can be input into the target AI model, and the target AI model can analyze and diagnose it to obtain target diagnostic suggestions.
[0089] Optionally, the method may further include the following steps:
[0090] A1. Obtain a preset first model and a pre-trained model; both the first model and the pre-trained model are large language models.
[0091] A2. Obtain the target task corresponding to the first model;
[0092] A3. Based on the pre-trained model and the target task, perform transfer learning on the first model to obtain the second model;
[0093] A4. Obtain the first historical vehicle diagnostic data; the first historical vehicle diagnostic data includes the problem text and its corresponding diagnostic suggestions.
[0094] A5. Preprocess the first historical vehicle diagnostic data to obtain the second historical vehicle diagnostic data;
[0095] A6. Divide the second historical vehicle diagnostic data into a training set and a test set according to a preset ratio;
[0096] A7. Train the second model using the training set to obtain the third model;
[0097] A8. Determine whether the third model meets the preset conditions;
[0098] A9. If the third model meets the preset conditions, then the third model is determined to be the target AI model;
[0099] A10. If the third model does not meet the preset conditions, then the training of the third model continues until the trained third model meets the preset conditions, and the third model that meets the preset conditions is taken as the target AI model.
[0100] In this embodiment of the application, the first model, the pre-trained model, the preset ratio, and the preset conditions can all be preset in advance or defaulted.
[0101] In a specific embodiment, a preset first model and a pre-trained model can be obtained first; then, the target task corresponding to the first model can be obtained. Since the first model is ultimately to be pre-installed on the diagnostic device, it can be determined that the target task of the first model is "vehicle diagnosis". Among them, the pre-trained model is a large language model that has been trained on large-scale general text data.
[0102] Next, transfer learning can be performed on the first model based on the pre-trained model and the target task to obtain the second model. Specifically, the structure of the first model and the pre-trained model can be appropriately adjusted according to the characteristics of the target task. For example, specific task layers can be added to the first model, such as a classification layer for classifying fault types and a decoding layer for generating diagnostic steps. Then, the corresponding layers of the first model can be initialized using the weights of the pre-trained model, and the general language features learned by the pre-trained model can be transferred to the first model to obtain the second model. Then, the first historical vehicle diagnostic data can be obtained. Specifically, vehicle diagnostic data can be queried from the data storage module of the diagnostic equipment to obtain the first historical vehicle diagnostic data, or the diagnostic data can be downloaded from the website of the user's vehicle manufacturer. Next, the first historical vehicle diagnostic data can be preprocessed to obtain the second historical vehicle diagnostic data. Specifically, data cleaning, data labeling, and data transformation can be performed on the first historical vehicle diagnostic data in sequence to obtain the second historical vehicle diagnostic data.
[0103] Furthermore, the second historical vehicle diagnostic data can be divided into training and testing sets according to a preset ratio. For example, the order of the second historical vehicle diagnostic data can be shuffled first, and then divided according to the preset ratio and the shuffled data order. Assuming the preset ratio is 8:2, that is, 80% of the data is used as the training set and 20% of the data is used as the testing set. If there are 1000 records in total, then the first 800 records can be divided into the training set and the last 200 records into the testing set. Then, the second model can be trained using the training set to obtain the third model. Next, it can be determined whether the third model meets the preset conditions. If the third model meets the preset conditions, then the third model is determined to be the target AI model.
[0104] If the third model does not meet the preset conditions, training of the third model continues until the trained third model meets the preset conditions. The third model that meets the preset conditions is then used as the target AI model. It should be explained that the preset conditions can be: the model's accuracy is greater than a preset accuracy (e.g., 80%), or the number of training iterations is greater than a preset number (e.g., 10 times).
[0105] In this way, by performing transfer learning on the first model, the capabilities of the pre-trained model are transferred to the first model, avoiding the need for the first model to be trained from scratch. This greatly reduces training time and computational resource consumption, and enables the training process on the target task to be started quickly, thus improving the efficiency of model training.
[0106] Optionally, step S304, which involves analyzing the target problem text using a preset target AI model in the diagnostic device to obtain target diagnostic suggestions, may include the following steps:
[0107] B1. Obtain the historical vehicle diagnostic data of the diagnostic device to obtain the third historical vehicle diagnostic data;
[0108] B2. Determine whether the target problem text is contained in the third historical vehicle diagnostic data;
[0109] B3. If so, then determine the diagnostic suggestion corresponding to the target problem text in the third historical vehicle diagnostic data as the target diagnostic suggestion;
[0110] B4. If not, then the target AI model generates diagnostic suggestions corresponding to the target question text based on the target communication intent, thus obtaining the first diagnostic suggestion;
[0111] B5. Obtain the operating data and log data of the target vehicle corresponding to the target problem text, and obtain the target operating data and target log data;
[0112] B6. Determine the target diagnostic suggestion based on the target operating data, the target log data, and the first diagnostic suggestion.
[0113] In this embodiment of the application, the third historical vehicle diagnostic data includes problem text and its corresponding diagnostic suggestions.
[0114] In a specific embodiment, historical vehicle diagnostic data from the diagnostic device can be obtained first to obtain third historical vehicle diagnostic data. Specifically, historical vehicle diagnostic data can be queried from the data storage module of the diagnostic device to obtain the third historical vehicle diagnostic data. Next, it can be determined whether the third historical vehicle diagnostic data contains the target question text. Specifically, keywords can be extracted from the target question text to obtain n keywords, where n is a positive integer. Then, for each question text record in the third historical vehicle diagnostic data, the keyword extraction operation is also performed to obtain p keyword sets, where p is a positive integer greater than 1. Next, the n keywords can be matched with each keyword set in the p keyword sets to obtain p matching results. Each matching result includes the proportion of keywords that are successfully matched. Based on the p matching results, it is determined whether the third historical vehicle diagnostic data contains the target question text. If there are matching results in the p matching results that are greater than the preset keyword proportion (e.g., more than 50%), then the third historical vehicle diagnostic data is considered to contain the target question text. Conversely, if there are no matching results in the p matching results that are greater than the preset keyword proportion (e.g., more than 50%), then the third historical vehicle diagnostic data is considered not to contain the target question text.
[0115] If so, the diagnostic suggestion corresponding to the target problem text in the third historical vehicle diagnostic data is determined as the target diagnostic suggestion. Specifically, the matching result with the highest proportion of successfully matched keywords among p matching results can be determined to obtain the target matching result, and the diagnostic suggestion corresponding to the target matching result in the third historical vehicle diagnostic data is determined as the target diagnostic suggestion.
[0116] If not, the target AI model generates diagnostic suggestions corresponding to the target problem text based on the target communication intent, thus obtaining the first diagnostic suggestion. Specifically, the target communication intent and the target problem text can be input into the target AI model, which generates the first diagnostic suggestion. Further, the target vehicle's operating data and log data corresponding to the target problem text can be obtained to obtain target operating data and target log data. Specifically, the diagnostic device can communicate or physically connect with the target vehicle. For example, the diagnostic device's connection device (e.g., OBD-II Bluetooth adapter, USB adapter) can be inserted into the target vehicle's corresponding interface (e.g., OBD-II interface), and a connection can be established with the diagnostic device via Bluetooth or USB to ensure a stable connection for successful data acquisition. Then, the target vehicle's operating data and log data can be collected, thus obtaining target operating data and target log data. Finally, the target diagnostic suggestion can be determined based on the target operating data, target log data, and the first diagnostic suggestion.
[0117] Thus, by determining whether the target problem text is contained in the third historical vehicle diagnostic data, it is possible to quickly determine whether ready-made diagnostic suggestions can be obtained directly from historical data. If it is contained, there is no need for complex model generation operations; historical experience can be directly invoked, greatly improving the efficiency of diagnostic suggestion generation and saving time and computing resources. When the third historical vehicle diagnostic data does not contain the target problem text, diagnostic suggestions can be generated based on the target communication intent using the target AI model, demonstrating the flexibility in handling new problems. The AI model can analyze and reason about new and unseen problems by learning patterns and rules in large amounts of data, generating reasonable diagnostic suggestions, compensating for the limitations of historical data, and enabling the system to adapt to constantly changing vehicle fault conditions.
[0118] Optionally, determining the target diagnostic suggestion based on the target runtime data, the target log data, and the first diagnostic suggestion may include the following steps:
[0119] C1. Determine the fault code data in the target log data;
[0120] C2. Determine the target fault type based on the target operating data and the fault code data;
[0121] C3. Verify the first diagnostic suggestion according to the target fault type to obtain the target verification result; the target verification result includes verification success or verification failure;
[0122] C4. If the target verification result includes verification success, then the target diagnostic suggestion is determined based on the first diagnostic suggestion;
[0123] C5. If the target verification result includes verification failure, then generate a diagnostic suggestion corresponding to the target fault type to obtain the target diagnostic suggestion.
[0124] In this embodiment of the application, the target operating data may include at least one of the following: engine speed, vehicle speed, water temperature, oil pressure, throttle opening, etc., which are not limited here.
[0125] In a specific embodiment, fault code data in the target log data can be identified. Specifically, fault code data in log files typically have a specific format and identifier. Common fault code formats, such as those in the OBD-II standard, begin with "P", "C", "B", or "U", followed by four digits. Fault code data can be located by searching for codes starting with these specific characters in the log file. Furthermore, based on information such as timestamps and event descriptions in the log file, combined with the time and symptoms of the vehicle fault, relevant fault code data can be further identified. Then, the target fault type can be determined based on the target operating data and fault code data. Specifically, the fault code manual, online database, or professional diagnostic software provided by the vehicle manufacturer can be used to find the specific meaning of the obtained fault code data and obtain multiple possible fault types. Each fault code corresponds to a specific fault range or possible fault cause. For example, the P0171 fault code usually indicates that the fuel system is too lean, which may be related to components such as the air flow meter, fuel injectors, and oxygen sensor. Further, the most likely fault type can be selected from multiple fault types based on the target operating data, which is the target fault type. For example, if the target operating data shows that the air flow meter shows an intake air volume that is 20% higher than the standard value at idle speed, while the data of the fuel injectors and oxygen sensor are within the normal range, then the target fault type can be determined to be an air flow meter fault.
[0126] Next, the first diagnostic suggestion can be verified based on the target fault type to obtain the target verification result. Specifically, the operating data of the target vehicle can be collected before and after the user implements the first diagnostic suggestion. The operating data after implementation is compared with the standard values of normal operating data and the data before implementation. If the abnormal operating data caused by the target fault type is significantly improved after implementing the diagnostic suggestion and approaches the standard value, then the diagnostic suggestion is effective, that is, the target verification result is successful. For example, if the fault type is transmission shift jerking, the first diagnostic suggestion is to change the transmission fluid. After the change, the operating data such as the speed fluctuation during vehicle shifting returns to normal, which verifies the correctness of the diagnostic suggestion. Otherwise, the target verification result is a failure. If the target verification result includes successful verification, the first diagnostic suggestion can be directly used as the target diagnostic suggestion. If the target verification result includes failure, a diagnostic suggestion corresponding to the target fault type is generated to obtain the target diagnostic suggestion. Specifically, the target fault type and target problem text can be input into the target AI model, and the target AI model outputs the target diagnostic suggestion.
[0127] Thus, identifying fault codes in the target log data provides a clear direction for fault diagnosis, as fault codes are a direct identifier of a malfunction in a vehicle system. Combining this with target operational data to determine the target fault type allows for a comprehensive analysis of the fault from multiple dimensions, avoiding misjudgments that might occur based on a single data point, and thereby more accurately locating the vehicle's fault.
[0128] Optionally, in one embodiment, the server of the diagnostic device is equipped with a RAG service, and the step of analyzing the target problem text using a preset target AI model in the diagnostic device to obtain target diagnostic suggestions may include the following steps:
[0129] D1. Transmit the target question text to the server, and retrieve data related to the target question text through the RAG service to obtain the first retrieval data;
[0130] D2. Merge the first search data and the target question text to obtain the target fused text;
[0131] D3. Analyze the target fused text using the target AI model to obtain the target diagnostic suggestions.
[0132] In this embodiment, the diagnostic device can transmit the target problem text to the server and retrieve data related to the target problem text through the RAG service on the server to obtain the first retrieval data. Specifically, key features, such as keywords and subject terms, can be extracted from the target problem text. Based on the extracted key features, keyword matching retrieval can be performed in a preset database stored on the server (e.g., a fault code repair document database or a repair document vector knowledge base). The database query function can be used, such as using the LIKE statement for fuzzy matching in a relational database, or using the corresponding query method in a non-relational database, thereby obtaining the first retrieval data. Then, the server can send the first retrieval data to the diagnostic device. Next, the diagnostic equipment can fuse the first search data and the target question text to obtain the target fused text. Specifically, the first search data and the target question text can be concatenated in a certain order to obtain the target fused text. For example, if the target question text is "What causes car engine vibration?" and the first search data contains "Engine vibration may be due to spark plug failure", then the two can be directly concatenated to "What causes car engine vibration? Engine vibration may be due to spark plug failure". Finally, the target fused text can be analyzed by the target AI model to obtain target diagnostic suggestions.
[0133] Thus, by fusing the initial search data with the target question text, the AI model can better understand the background and context of the question. The fused target text contains more semantic information, enabling the model to more accurately grasp the key points and intent of the question, thereby improving the accuracy and depth of the model's analysis.
[0134] S305. Display the target diagnostic suggestion in the second information display area.
[0135] In this embodiment, target diagnostic suggestions can be displayed in the second information display area of the AI diagnostic interaction page to prompt the user to perform operations based on the target diagnostic suggestions.
[0136] Optionally, please refer to Figure 7, which is a flowchart illustrating a model optimization method provided in an embodiment of this application. After displaying the target diagnostic suggestion in the second information display area, the method may further include the following steps:
[0137] E1. Obtain user feedback data regarding the target diagnostic suggestions;
[0138] E2. Determine the target feedback type corresponding to the target feedback data; the target feedback type includes one of the following: positive feedback type, negative feedback type, and opinion feedback type;
[0139] E3. When the target feedback type includes the opinion feedback type, optimize the target AI model based on the target feedback data to obtain the optimized target AI model.
[0140] In this embodiment, target feedback data from the user regarding the target diagnostic suggestion can be obtained. Specifically, the user can click the information input window to directly input the target feedback data into the diagnostic device. Then, the target feedback type corresponding to the target feedback data can be determined. Specifically, a preset sentiment dictionary, such as CNKI sentiment dictionary or HowNet, can be used to match the words in the target feedback data with sentiment words in the dictionary. The sentiment tendency score of the entire text is calculated based on the polarity (positive or negative) and intensity of the sentiment words to obtain the target sentiment score. If the target sentiment score is greater than a first preset score threshold (e.g., 0.5), it is judged as positive feedback; if the target sentiment score is lower than a second preset score threshold (e.g., -0.5), it is judged as negative feedback; if the target sentiment score is between the first and second preset score thresholds, it can be judged as opinion feedback.
[0141] When the target feedback type includes opinion feedback, the target AI model is optimized based on the target feedback data to obtain the optimized target AI model. Specifically, semantic analysis can be performed on the target feedback data to clarify the problems pointed out by users or the directions for improvement proposed. The parameters of the target AI model can then be adjusted according to the problems or directions for improvement. For example, if the diagnostic results of the user feedback model are too conservative or aggressive, the parameters related to the decision threshold in the model can be adjusted. If it is found that the weights of certain features in the model are unreasonable, resulting in inaccurate diagnosis, the weights of these features need to be adjusted to obtain the optimized target AI model.
[0142] In summary, the AI-based vehicle diagnostic method described in this application displays the vehicle diagnostic main page of the diagnostic device, wherein the vehicle diagnostic main page includes AI-assisted controls; it acquires the target working state of the diagnostic device; the target working state includes one of the following: normal state and diagnostic abnormal state; when the target working state includes the diagnostic abnormal state, in response to a trigger operation on the AI-assisted controls, it displays an AI diagnostic interaction page; the AI diagnostic interaction page includes: a first information display area, a second information display area, and an information input window; in response to an information input operation on the information input window, it determines the target problem text corresponding to the information input operation and displays the target problem text in the first information display area; it analyzes the target problem text through a preset target AI model in the diagnostic device to obtain a target diagnostic suggestion; and it displays the target diagnostic suggestion in the second information display area. Thus, by introducing AI-assisted controls, users only need to click once to activate the AI function (i.e., the target AI model) for vehicle diagnostics, without complicated operation procedures, thereby reducing the operational complexity of vehicle diagnostics.
[0143] Please refer to Figure 8, which is a functional unit block diagram of an AI-based automotive diagnostic system 800 provided in an embodiment of this application. Applied to diagnostic equipment, the AI-based automotive diagnostic system 800 includes: a display unit 801, a response unit 802, and a diagnostic unit 803, wherein:
[0144] The display unit 801 is used to display the vehicle diagnostic main page of the diagnostic device, wherein the vehicle diagnostic main page includes AI-assisted controls;
[0145] The response unit 802 is used to acquire the target working state of the diagnostic device; the target working state includes one of the following: normal state and diagnostic abnormal state; when the target working state includes the diagnostic abnormal state, in response to a trigger operation on the AI-assisted control, an AI diagnostic interaction page is displayed; the AI diagnostic interaction page includes: a first information display area, a second information display area, and an information input window; in response to an information input operation on the information input window, the target problem text corresponding to the information input operation is determined, and the target problem text is displayed in the first information display area;
[0146] The diagnostic unit 803 is used to analyze the target problem text through a preset target AI model in the diagnostic device to obtain target diagnostic suggestions;
[0147] The display unit 801 is also used to display the target diagnostic suggestion in the second information display area.
[0148] Optionally, the AI-based vehicle diagnostic system 800 is also specifically used for:
[0149] Obtain a preset first model and a pre-trained model; both the first model and the pre-trained model are large language models.
[0150] Obtain the target task corresponding to the first model;
[0151] Based on the pre-trained model and the target task, the first model is transferred to obtain the second model;
[0152] Obtain the first historical vehicle diagnostic data; the first historical vehicle diagnostic data includes the problem text and its corresponding diagnostic suggestions;
[0153] The first historical vehicle diagnostic data is preprocessed to obtain the second historical vehicle diagnostic data.
[0154] The second historical vehicle diagnostic data is divided into a training set and a test set according to a preset ratio;
[0155] The second model is trained using the training set to obtain the third model;
[0156] Determine whether the third model meets the preset conditions;
[0157] If the third model meets the preset conditions, then the third model is determined to be the target AI model;
[0158] If the third model does not meet the preset conditions, then the training of the third model continues until the trained third model meets the preset conditions, and the third model that meets the preset conditions is taken as the target AI model.
[0159] Optionally, in determining the target question text corresponding to the information input operation in response to the information input window, the response unit 802 is specifically configured to:
[0160] In response to an information input operation on the information input window, the system receives the question text input by the information input operation and obtains a first question text; the first question text is a natural language expression text.
[0161] Extract the first keyword from the first question text to obtain m first keywords; m is a positive integer.
[0162] The target communication intent is determined based on the m first keywords; the target communication intent includes one of the following: chat intent, inquiry intent;
[0163] Based on the target communication intent and the m first keywords, the first question text is converted into text in a preset format to obtain the target question text.
[0164] Optionally, in the aspect of analyzing the target problem text through a preset target AI model in the diagnostic device to obtain target diagnostic suggestions, the diagnostic unit 803 is specifically used for:
[0165] Obtain historical vehicle diagnostic data from the diagnostic device to obtain third historical vehicle diagnostic data.
[0166] Determine whether the target problem text is contained in the third historical vehicle diagnostic data;
[0167] If so, then the diagnostic suggestion corresponding to the target problem text in the third historical vehicle diagnostic data is determined to be the target diagnostic suggestion;
[0168] If not, then the target AI model generates diagnostic suggestions corresponding to the target question text based on the target communication intent, thus obtaining the first diagnostic suggestion;
[0169] Obtain the operating data and log data of the target vehicle corresponding to the target problem text to obtain the target operating data and target log data;
[0170] The target diagnostic recommendation is determined based on the target runtime data, the target log data, and the first diagnostic recommendation.
[0171] Optionally, in determining the target diagnostic suggestion based on the target runtime data, the target log data, and the first diagnostic suggestion, the diagnostic unit 803 is specifically used for:
[0172] Identify the fault code data in the target log data;
[0173] The target fault type is determined based on the target operating data and the fault code data;
[0174] The first diagnostic suggestion is verified based on the target fault type to obtain a target verification result; the target verification result includes verification success or verification failure.
[0175] If the target verification result includes verification success, then the target diagnostic recommendation is determined based on the first diagnostic recommendation;
[0176] If the target verification result includes verification failure, then a diagnostic suggestion corresponding to the target fault type is generated, and the target diagnostic suggestion is obtained.
[0177] Optionally, the server of the diagnostic device is equipped with a RAG service. In the aspect of analyzing the target problem text using a preset target AI model in the diagnostic device to obtain target diagnostic suggestions, the diagnostic unit 803 is specifically used for:
[0178] The target question text is transmitted to the server, and data related to the target question text is retrieved through the RAG service to obtain the first retrieval data;
[0179] The first retrieved data and the target question text are fused to obtain the target fused text;
[0180] The target fusion text is analyzed by the target AI model to obtain the target diagnostic suggestions.
[0181] Optionally, after displaying the target diagnostic suggestion in the second information display area, the AI-based vehicle diagnostic system 800 is specifically used for:
[0182] Obtain user feedback data regarding the target diagnostic suggestions;
[0183] Determine the target feedback type corresponding to the target feedback data; the target feedback type includes one of the following: positive feedback type, negative feedback type, and opinion feedback type.
[0184] When the target feedback type includes the opinion feedback type, the target AI model is optimized based on the target feedback data to obtain the optimized target AI model.
[0185] In specific implementations, the AI-based vehicle diagnostic system 800 described in this embodiment of the invention can also execute other implementations described in any of the above method embodiments, which will not be repeated here.
[0186] Please refer to Figure 9, which is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include a processor, a memory, a communication interface, and one or more programs. The processor, memory, and communication interface can be interconnected via a bus. The one or more programs are stored in the memory and configured to be executed by the processor. In this embodiment, the electronic device includes a diagnostic device, and the programs include instructions for performing the following steps:
[0187] The vehicle diagnostic main page of the diagnostic device is displayed, wherein the vehicle diagnostic main page includes AI-assisted controls;
[0188] The target working state of the diagnostic device is obtained; the target working state includes one of the following: normal state and diagnostic abnormal state; when the target working state includes the diagnostic abnormal state, in response to the trigger operation of the AI-assisted control, an AI diagnostic interaction page is displayed; the AI diagnostic interaction page includes: a first information display area, a second information display area, and an information input window;
[0189] In response to an information input operation on the information input window, the target question text corresponding to the information input operation is determined, and the target question text is displayed in the first information display area;
[0190] The target problem text is analyzed by a preset target AI model in the diagnostic device to obtain target diagnostic suggestions;
[0191] The target diagnostic recommendations are displayed in the second information display area.
[0192] Optionally, the above procedure may also include instructions for performing the following steps:
[0193] Obtain a preset first model and a pre-trained model; both the first model and the pre-trained model are large language models.
[0194] Obtain the target task corresponding to the first model;
[0195] Based on the pre-trained model and the target task, the first model is transferred to obtain the second model;
[0196] Obtain the first historical vehicle diagnostic data; the first historical vehicle diagnostic data includes the problem text and its corresponding diagnostic suggestions;
[0197] The first historical vehicle diagnostic data is preprocessed to obtain the second historical vehicle diagnostic data.
[0198] The second historical vehicle diagnostic data is divided into a training set and a test set according to a preset ratio;
[0199] The second model is trained using the training set to obtain the third model;
[0200] Determine whether the third model meets the preset conditions;
[0201] If the third model meets the preset conditions, then the third model is determined to be the target AI model;
[0202] If the third model does not meet the preset conditions, then the training of the third model continues until the trained third model meets the preset conditions, and the third model that meets the preset conditions is taken as the target AI model.
[0203] Optionally, in response to an information input operation on the information input window, the above program includes instructions for performing the following steps:
[0204] In response to an information input operation on the information input window, the system receives the question text input by the information input operation and obtains a first question text; the first question text is a natural language expression text.
[0205] Extract the first keyword from the first question text to obtain m first keywords; m is a positive integer.
[0206] The target communication intent is determined based on the m first keywords; the target communication intent includes one of the following: chat intent, inquiry intent;
[0207] Based on the target communication intent and the m first keywords, the first question text is converted into text in a preset format to obtain the target question text.
[0208] Optionally, in the step of analyzing the target problem text using a preset target AI model in the diagnostic device to obtain target diagnostic suggestions, the above procedure includes instructions for performing the following steps:
[0209] Obtain historical vehicle diagnostic data from the diagnostic device to obtain third historical vehicle diagnostic data.
[0210] Determine whether the target problem text is contained in the third historical vehicle diagnostic data;
[0211] If so, then the diagnostic suggestion corresponding to the target problem text in the third historical vehicle diagnostic data is determined to be the target diagnostic suggestion;
[0212] If not, then the target AI model generates diagnostic suggestions corresponding to the target question text based on the target communication intent, thus obtaining the first diagnostic suggestion;
[0213] Obtain the operating data and log data of the target vehicle corresponding to the target problem text to obtain the target operating data and target log data;
[0214] The target diagnostic recommendation is determined based on the target runtime data, the target log data, and the first diagnostic recommendation.
[0215] Optionally, in determining the target diagnostic recommendation based on the target runtime data, the target log data, and the first diagnostic recommendation, the above procedure includes instructions for performing the following steps:
[0216] Identify the fault code data in the target log data;
[0217] The target fault type is determined based on the target operating data and the fault code data;
[0218] The first diagnostic suggestion is verified based on the target fault type to obtain a target verification result; the target verification result includes verification success or verification failure.
[0219] If the target verification result includes verification success, then the target diagnostic recommendation is determined based on the first diagnostic recommendation;
[0220] If the target verification result includes verification failure, then a diagnostic suggestion corresponding to the target fault type is generated, and the target diagnostic suggestion is obtained.
[0221] Optionally, the server of the diagnostic device is equipped with a RAG service. Regarding the step of analyzing the target problem text using a preset target AI model in the diagnostic device to obtain target diagnostic suggestions, the above procedure includes instructions for performing the following steps:
[0222] The target question text is transmitted to the server, and data related to the target question text is retrieved through the RAG service to obtain the first retrieval data;
[0223] The first retrieved data and the target question text are fused to obtain the target fused text;
[0224] The target fusion text is analyzed by the target AI model to obtain the target diagnostic suggestions.
[0225] Optionally, after displaying the target diagnostic suggestion in the second information display area, the above procedure further includes instructions for performing the following steps:
[0226] Obtain user feedback data regarding the target diagnostic suggestions;
[0227] Determine the target feedback type corresponding to the target feedback data; the target feedback type includes one of the following: positive feedback type, negative feedback type, and opinion feedback type.
[0228] When the target feedback type includes the opinion feedback type, the target AI model is optimized based on the target feedback data to obtain the optimized target AI model.
[0229] In specific implementations, the electronic devices described in the embodiments of the present invention may also execute other implementation methods described in any of the above method embodiments, which will not be repeated here.
[0230] This application also provides a computer-readable storage medium storing a computer program for electronic data interchange, which causes a computer to perform some or all of the steps of any of the methods described in the above method embodiments, wherein the computer includes an electronic device.
[0231] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps of any of the methods described in the above method embodiments. The computer program product may be a software installation package, and the computer may include an electronic device.
[0232] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.
[0233] In the above embodiments, the descriptions of each embodiment have their own emphasis. Parts not described in detail in a certain embodiment can be referred to in the relevant descriptions of other embodiments. It should be understood that the disclosed apparatus can be implemented in other ways in the several embodiments provided in this application. For example, the apparatus embodiments described above are merely illustrative; for example, the division of the above units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.
[0234] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.
[0235] The steps of the methods or algorithms described in the embodiments of this application can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in RAM, flash memory, ROM, EPROM, electrically erasable programmable read-only memory (EEPROM), registers, hard disk, portable hard disk, read-only optical disk (CD-ROM), or any other form of storage medium well known in the art. An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and storage medium can reside in an ASIC. Furthermore, the ASIC can reside in a terminal device or management device. Alternatively, the processor and storage medium can exist as discrete components in the terminal device or management device.
[0236] Those skilled in the art will recognize that, in one or more of the examples above, the functions described in the embodiments of this application can be implemented, in whole or in part, by software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, in the form of a computer program product. This computer program product includes one or more computer instructions. When these computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated.
[0237] The aforementioned computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access, or a data storage device such as a server or data center that integrates one or more available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., digital video discs (DVDs)), or semiconductor media (e.g., solid-state disks (SSDs)).
[0238] The modules / units included in the various devices and products described in the above embodiments can be software modules / units, hardware modules / units, or a combination of both. For example, for devices and products applied to or integrated into a chip, all modules / units can be implemented using hardware methods such as circuits, or at least some modules / units can be implemented using software programs that run on a processor integrated within the chip, while the remaining (if any) modules / units can be implemented using hardware methods such as circuits. For devices and products applied to or integrated into a chip module, all modules / units can be implemented using hardware methods such as circuits. Different modules / units can be located in the same component (e.g., chip, circuit module, etc.) or different components of the chip module, or at least some modules / units can be implemented using hardware methods such as circuits. The implementation is achieved through a software program that runs on a processor integrated within the chip module. The remaining modules / units (if any) can be implemented using hardware methods such as circuits. For various devices and products applied to or integrated into terminal equipment, each of their modules / units can be implemented using hardware methods such as circuits. Different modules / units can be located in the same component (e.g., chip, circuit module, etc.) or different components within the terminal equipment. Alternatively, at least some modules / units can be implemented using a software program that runs on a processor integrated within the terminal equipment, while the remaining modules / units (if any) can be implemented using hardware methods such as circuits.
[0239] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the embodiments of this application. It should be understood that the above descriptions are merely specific embodiments of the embodiments of this application and are not intended to limit the protection scope of the embodiments of this application. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solutions of the embodiments of this application should be included within the protection scope of the embodiments of this application.
Claims
1. An AI-based vehicle diagnostic method, characterized in that, Used in diagnostic equipment, including: The vehicle diagnostic main page of the diagnostic device is displayed, wherein the vehicle diagnostic main page includes AI-assisted controls; The target working state of the diagnostic device is obtained; the target working state includes one of the following: normal state and diagnostic abnormal state; when the target working state includes the diagnostic abnormal state, in response to the trigger operation of the AI-assisted control, an AI diagnostic interaction page is displayed; the AI diagnostic interaction page includes: a first information display area, a second information display area, and an information input window; In response to an information input operation on the information input window, the target question text corresponding to the information input operation is determined, and the target question text is displayed in the first information display area; The target problem text is analyzed by a preset target AI model in the diagnostic device to obtain target diagnostic suggestions; The target diagnostic recommendations are displayed in the second information display area.
2. The method as described in claim 1, characterized in that, The method further includes: Obtain a preset first model and a pre-trained model; both the first model and the pre-trained model are large language models. Obtain the target task corresponding to the first model; Based on the pre-trained model and the target task, the first model is transferred to obtain the second model; Obtain the first historical vehicle diagnostic data; the first historical vehicle diagnostic data includes the problem text and its corresponding diagnostic suggestions; The first historical vehicle diagnostic data is preprocessed to obtain the second historical vehicle diagnostic data. The second historical vehicle diagnostic data is divided into a training set and a test set according to a preset ratio; The second model is trained using the training set to obtain the third model; Determine whether the third model meets the preset conditions; If the third model meets the preset conditions, then the third model is determined to be the target AI model; If the third model does not meet the preset conditions, then the training of the third model continues until the trained third model meets the preset conditions, and the third model that meets the preset conditions is taken as the target AI model.
3. The method as described in claim 1 or 2, characterized in that, The step of determining the target question text corresponding to the information input operation in response to the information input window includes: In response to an information input operation on the information input window, the system receives the question text input by the information input operation and obtains a first question text; the first question text is a natural language expression text. Extract the first keyword from the first question text to obtain m first keywords; m is a positive integer. The target communication intent is determined based on the m first keywords; the target communication intent includes one of the following: chat intent, inquiry intent; Based on the target communication intent and the m first keywords, the first question text is converted into text in a preset format to obtain the target question text.
4. The method as described in claim 3, characterized in that, The step of analyzing the target problem text using a preset target AI model in the diagnostic device to obtain target diagnostic suggestions includes: Obtain historical vehicle diagnostic data from the diagnostic device to obtain third historical vehicle diagnostic data. Determine whether the target problem text is contained in the third historical vehicle diagnostic data; If so, then the diagnostic suggestion corresponding to the target problem text in the third historical vehicle diagnostic data is determined to be the target diagnostic suggestion; If not, then the target AI model generates diagnostic suggestions corresponding to the target question text based on the target communication intent, thus obtaining the first diagnostic suggestion; Obtain the operating data and log data of the target vehicle corresponding to the target problem text to obtain the target operating data and target log data; The target diagnostic recommendation is determined based on the target runtime data, the target log data, and the first diagnostic recommendation.
5. The method as described in claim 4, characterized in that, The step of determining the target diagnostic suggestion based on the target operational data, the target log data, and the first diagnostic suggestion includes: Identify the fault code data in the target log data; The target fault type is determined based on the target operating data and the fault code data; The first diagnostic suggestion is verified based on the target fault type to obtain a target verification result; the target verification result includes verification success or verification failure. If the target verification result includes verification success, then the target diagnostic recommendation is determined based on the first diagnostic recommendation; If the target verification result includes verification failure, then a diagnostic suggestion corresponding to the target fault type is generated, and the target diagnostic suggestion is obtained.
6. The method as described in claim 1 or 2, characterized in that, The diagnostic device's server is equipped with a RAG service. The analysis of the target problem text using a pre-set target AI model within the diagnostic device to obtain target diagnostic suggestions includes: The target question text is transmitted to the server, and data related to the target question text is retrieved through the RAG service to obtain the first retrieval data; The first retrieved data and the target question text are fused to obtain the target fused text; The target fusion text is analyzed by the target AI model to obtain the target diagnostic suggestions.
7. The method as described in claim 1 or 2, characterized in that, After displaying the target diagnostic suggestion in the second information display area, the method further includes: Obtain user feedback data regarding the target diagnostic suggestions; Determine the target feedback type corresponding to the target feedback data; the target feedback type includes one of the following: positive feedback type, negative feedback type, and opinion feedback type. When the target feedback type includes the opinion feedback type, the target AI model is optimized based on the target feedback data to obtain the optimized target AI model.
8. An AI-based vehicle diagnostic system, characterized in that, The system, applied to diagnostic equipment, includes: a display unit, a response unit, and a diagnostic unit, wherein: The display unit is used to display the vehicle diagnostic main page of the diagnostic device, wherein the vehicle diagnostic main page includes AI-assisted controls; The response unit is configured to acquire the target operating state of the diagnostic device; the target operating state includes one of the following: normal state and diagnostic abnormal state; when the target operating state includes the diagnostic abnormal state, in response to a trigger operation on the AI-assisted control, an AI diagnostic interaction page is displayed; the AI diagnostic interaction page includes: a first information display area, a second information display area, and an information input window; in response to an information input operation on the information input window, the target problem text corresponding to the information input operation is determined, and the target problem text is displayed in the first information display area; The diagnostic unit is used to analyze the target problem text using a preset target AI model in the diagnostic device to obtain target diagnostic suggestions; The display unit is also used to display the target diagnostic suggestion in the second information display area.
9. An electronic device, characterized in that, include: Processor, memory, communication interface, and one or more programs; The one or more programs are stored in the memory and configured to be executed by the processor, the programs including instructions for performing the steps of the method as described in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, A computer program for storing electronic data interchange, wherein the computer program causes a computer to perform the method as described in any one of claims 1-7.