Configuration Information Extraction Method and Device

Abstract

This application discloses a method and apparatus for extracting configuration information. A first intelligent agent acquires preprocessed information from a user-input technical specification document and determines the target automotive open system architecture associated with the technical specification document. A second intelligent agent extracts target configuration information from the preprocessed information based on the target automotive open system architecture. The target configuration information includes the value of at least one configuration item of the automotive electronic control unit (ECU). A third intelligent agent generates a configuration file based on the target configuration information. According to embodiments of this application, through the collaboration of multiple intelligent agents and utilizing a large language model, the target configuration information of the ECU is intelligently extracted based on the technical specification document and the target AUTOSAR associated with the technical specification document, thereby improving the efficiency and accuracy of configuration information extraction.

Description

Technical Field

[0001] This application belongs to the field of automotive electronic system development and intelligent document processing technology, and particularly relates to a configuration information extraction method and apparatus. Background Technology

[0002] Currently, in the development of automotive electronic control unit (ECU) configurations, OEMs mainly extract ECU configuration information from the OEM's technical specification documents through methods such as manual document parsing, keyword extraction, and static knowledge base mapping.

[0003] However, due to the complexity of technical specification documents, the different terminology used by different car manufacturers, and the differences in the definition of configuration items, the above-mentioned scheme for extracting configuration information suffers from low efficiency and inaccuracy. Summary of the Invention

[0004] This application provides a method and apparatus for extracting configuration items, which can improve the efficiency and accuracy of extracting ECU configuration information from technical specification documents.

[0005] On the one hand, embodiments of this application provide a configuration information extraction method, the method including: The first intelligent agent obtains preprocessed information from the technical specification document input by the user and determines the target automotive open system architecture associated with the technical specification document. Through the second intelligent agent, a pre-trained large language model is used to extract target configuration information that matches the target vehicle open system architecture from the pre-processed information. The large language model is trained based on knowledge in the automotive electronics field. The target configuration information includes the value of at least one configuration item of the automotive electronic control unit. A configuration file is generated based on the target configuration information through a third intelligent agent.

[0006] On the other hand, embodiments of this application provide a configuration information extraction device, the device comprising: The document parsing module is used to obtain preprocessed information of the technical specification document input by the user through the first intelligent agent, and to determine the target automotive open system architecture associated with the technical specification document; The configuration item mapping module is used by a second intelligent agent to extract target configuration information that matches the target vehicle open system architecture from preprocessed information using a pre-trained large language model. The large language model is trained based on knowledge in the automotive electronics field, and the target configuration information includes the value of at least one configuration item of the automotive electronic control unit. The delivery module is used to generate configuration files based on target configuration information through a third-party intelligent agent.

[0007] In another aspect, embodiments of this application provide an electronic device, the device comprising: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, it implements the configuration information extraction method as described in the first aspect.

[0008] In another aspect, embodiments of this application provide a computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the configuration information extraction method as described in the first aspect.

[0009] In another aspect, embodiments of this application provide a computer program product in which instructions, when executed by a processor of an electronic device, cause the electronic device to perform the configuration information extraction method as described in the first aspect.

[0010] The configuration information extraction method, apparatus, device, storage medium, and program product of this application, through the collaboration of multiple intelligent agents and utilizing a large language model, intelligently extracts the target configuration information of the ECU based on technical specification documents and the target AUTOSAR associated with the technical specification documents. This achieves automatic extraction of configuration information, improves the efficiency of configuration information extraction, and shortens the manual mapping process that originally required hours or even days to minutes, significantly accelerating the ECU development cycle. Moreover, by using a large language model in the automotive electronics field combined with AUTOSAR specifications to extract configuration information, configuration errors caused by terminology ambiguity can be reduced, improving the accuracy of configuration information. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart illustrating the configuration information extraction method provided in some embodiments of this application; Figure 2 This is a flowchart illustrating a configuration information extraction method provided in some other embodiments of this application; Figure 3 This is a flowchart illustrating a configuration information extraction method provided in some other embodiments of this application; Figure 4 This is a flowchart illustrating the configuration information extraction method provided in some other embodiments of this application; Figure 5 This is a schematic diagram of data interaction within a configuration information extraction device provided in some embodiments of this application; Figure 6 This is a schematic diagram of the configuration information extraction device provided in some embodiments of this application; Figure 7 These are schematic diagrams of the structure of electronic devices provided in some embodiments of this application. Detailed Implementation

[0013] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0014] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

[0015] Before providing a further detailed description of the embodiments of this application, the nouns and terms involved in the embodiments of this application will be explained, and the nouns and terms involved in the embodiments of this application shall be interpreted as follows.

[0016] AUTOSAR (Automotive Open System Architecture) is a software architecture framework jointly established by global automakers, component suppliers, and other electronics, semiconductor, and software systems companies, with each member maintaining a development partnership. AUTOSAR aims to develop open and standardized software architectures for the automotive industry.

[0017] Agent: An intelligent agent refers to an entity in artificial intelligence and machine learning that is able to perceive its environment and make decisions.

[0018] ECU (electronic control unit): The electronic control unit of an automobile is a comprehensive control device for the engine.

[0019] Before providing a more detailed description of the embodiments of this application, the methods for extracting configuration items in related technologies will be introduced. As mentioned above, related technologies mainly rely on the following three technical means to extract ECU configuration information: 1. Manual document analysis: Engineers read hundreds of pages of technical specification documents and manually extract configuration items and their specific values ​​from the AUTOSAR specifications. However, this method requires a high level of expertise from engineers, and because the technical specification documents are extensive and some manuals are written entirely in English, the process from document reading to AUTOSAR configuration extraction is time-consuming and inefficient.

[0020] 2. Keyword Extraction: A broad-based keyword full-text search is used, or configuration items and values ​​are extracted from the document using simple regular expressions and template matching tools. This method struggles to handle multilingual expressions and complex semantic scenarios.

[0021] 3. Static Knowledge Base: Some tools build configuration item mapping tables based on fixed versions of AUTOSAR metadata, and extract configuration items and values ​​from these tables. However, vehicle manufacturer specifications are mostly described in natural language and lack a unified data format, making it difficult for automated tools to parse them directly. Moreover, some vehicle manufacturer specifications do not directly contain configuration items or words similar to configuration items, requiring professional knowledge for identification. Mapping errors are prone to occur for some configuration items with similar meanings. In addition, configuration item mapping tables are usually built based on a single version of AUTOSAR, and the definitions of configuration items differ between different AUTOSAR versions (such as AR4.2, R23-11, and R24-11). Tools cannot dynamically adapt to these differences, leading to deviations in parsing results.

[0022] In summary, the configuration information extraction methods in related technologies suffer from low efficiency and inaccuracy due to the complexity of technical specification documents and the different terminology used by different car manufacturers.

[0023] Therefore, to improve the efficiency and accuracy of configuration information extraction, this application provides a configuration information extraction method, apparatus, device, storage medium, and program product. It intelligently extracts configuration information based on AUTOSAR specification knowledge and vehicle manufacturer technical specifications.

[0024] The configuration information extraction method provided in this application will be described below. This method is applicable to ECU software development, configuration management, and toolchain integration. It should be noted that the execution entity of the configuration information extraction method provided in this application embodiment can be a configuration information extraction device. The configuration information extraction device can adopt distributed control, which may include multiple intelligent agents and a knowledge base. During the extraction of configuration information, different agents can play different roles and perform different functions. Each agent makes independent decisions, resulting in high robustness. The knowledge base can store the intermediate results generated by each agent during the configuration information extraction process, enabling flexible expansion and dynamic adaptation. This application embodiment uses the execution of the configuration information extraction method by a configuration information extraction device as an example to illustrate the configuration information extraction method provided in this application embodiment.

[0025] See Figure 1 The above is a flowchart illustrating the configuration information extraction method provided in some embodiments of this application, such as... Figure 1 As shown, the method includes the following steps 110-130.

[0026] Step 110. The first intelligent agent obtains the preprocessed information of the technical specification document input by the user and determines the target automotive open system architecture associated with the technical specification document.

[0027] In this embodiment of the application, the first intelligent agent acts as a "project manager". The responsibilities of the first intelligent agent mainly include: uploading and preprocessing technical specification documents, and determining the target automotive open system architecture associated with the technical specification documents.

[0028] In practical applications, users can upload technical specification documents in PDF or Word format to the first intelligent agent via a web interface. Here, the technical specification document input by the user is the original document containing the configuration information to be extracted; the purpose is to extract the required configuration information from this document. For example, the user is a project manager from a car manufacturer, and the technical specification document can be a formal document developed and published by the car manufacturer, clearly defining the technical requirements, standards, methods, and processes that the vehicles it produces must meet in terms of components, systems, materials, processes, testing, and quality.

[0029] Preprocessed information for technical specification documents refers to the information obtained after preprocessing the technical specification document. This information may include editable text, structured tabular data, and metadata (such as page numbers, paragraph numbers, table coordinates, etc.). Technical specification documents are mostly described in natural language and lack a standardized data format, making them difficult for computers to process and understand. After receiving the technical specification document input by the user, the first intelligent agent obtains the preprocessed information from the document to facilitate subsequent processing.

[0030] In some embodiments of this application, to reduce unnecessary processing and improve the efficiency of configuration information extraction, the first intelligent agent supports a cache skipping mechanism. That is, if the technical specification document input by the user is a technical specification document previously preprocessed by the first intelligent agent, the previously preprocessed information is used as the preprocessed information for the current technical specification document, eliminating the need to reprocess the currently input technical specification document to obtain preprocessed information. Based on this, the preprocessed information of the currently input technical specification document can be obtained through the following steps 210-240.

[0031] Step 210. Generate the document identifier for the technical specification document.

[0032] In this embodiment of the application, the document identifier is the content fingerprint of the technical specification document. Documents with the same content correspond to the same document identifier, and documents with different content correspond to different document identifiers.

[0033] In some embodiments of this application, the SHA-256 hash algorithm can be used to perform a digest operation on the original byte stream of the technical specification document to generate a unique identifier for the technical specification document, and this identifier can be used as the text identifier of the technical specification document.

[0034] Step 220. Match the document identifier with the document identifier stored in the knowledge base to obtain the matching result. The knowledge base is used to store preprocessing information and document identifiers of historical technical specification documents.

[0035] After obtaining the document identifier of the technical specification document currently input by the user, the document identifier is matched with the document identifiers of historical technical specification documents stored in the knowledge base to determine whether the technical specification document currently input by the user is a technical specification document that has been processed by the first intelligent agent before.

[0036] In this embodiment of the application, the knowledge base stores the document identifiers and preprocessing information of the technical specification documents processed by the first intelligent agent.

[0037] In some embodiments of this application, the knowledge base may maintain a document index table, which records the document identifiers of technical specification documents that the first intelligent agent has processed. In this way, after the first intelligent agent obtains the document identifier of the technical specification document currently input by the user, it can compare the document identifier with the document index table to determine whether the technical specification document currently input by the user has been preprocessed and stored.

[0038] Step 230. If the matching result indicates that the knowledge base contains a target historical technical specification document whose document identifier matches the document identifier of the technical specification document, obtain the preprocessing information of the target historical technical specification document as the preprocessing information of the technical specification document.

[0039] If, through matching, it is determined that the knowledge base contains a target historical technical specification document whose document identifier matches that of the technical specification document, then it is determined that the technical specification document currently entered by the user has already been preprocessed and stored, and no further preprocessing is required. The preprocessing information of the target historical technical specification document can be directly obtained from the knowledge base as the preprocessing information of the technical specification document.

[0040] In some embodiments of this application, when it is determined that the knowledge base contains a target historical technical specification document, before obtaining the preprocessing information of the target historical technical specification document, the validity of the preprocessing information of the target historical technical specification document can be determined first. If it is determined that the preprocessing information of the target historical technical specification document is still valid, the preprocessing information of the target historical technical specification document is then obtained as the preprocessing information of the technical specification document. If the preprocessing information of the target historical technical specification document is invalid, the preprocessing information of the technical specification document is obtained by preprocessing the technical specification document currently input by the user.

[0041] Step 240. If the matching result indicates that the knowledge base does not contain a target historical technical specification document that matches the document identifier of the technical specification document, preprocess the technical specification document to obtain the preprocessed information of the technical specification document, and store the document identifier and preprocessed information of the technical specification document in the knowledge base.

[0042] If the matching process determines that the knowledge base does not contain a target historical technical specification document whose document identifier matches the document identifier of the technical specification document, then the technical specification document currently entered by the user is determined to be a new document that has not been processed before. In this case, the technical specification document is preprocessed to obtain preprocessed information, and the preprocessed information and the corresponding document identifier are stored in the knowledge base for future reuse of the same document.

[0043] In some embodiments of this application, preprocessing the technical specification document may include the following steps 310-330.

[0044] Step 310. Convert the images in the technical specification document into editable text.

[0045] Technical specification documents may contain images, such as scanned PDF files. For images, OCR (Optical Character Recognition) tools can be used to convert the image content into editable text.

[0046] Step 320. Parse the electronic documents in the technical specification document to obtain the structured tabular data of the electronic documents.

[0047] Technical specification documents may exist as electronic documents such as PDF documents and Word documents. For electronic documents, their text, paragraphs, tables and annotations and other structured elements can be directly parsed to obtain structured tabular data.

[0048] Step 330. Generate a unified text stream from the editable text and structured tabular data, while preserving the metadata of the technical specification document.

[0049] Metadata refers to information in technical specification documents that indicates the location of content within the document, such as page numbers, paragraph numbers, and table coordinates. Retaining metadata facilitates subsequent content traceability.

[0050] The cache skipping mechanism enables "process once, benefit multiple times". When the same document is used repeatedly, the existing preprocessed results can be reused directly, saving document processing time and thus improving the efficiency of configuration information extraction.

[0051] In some embodiments of this application, the first intelligent agent supports forced parsing. Upon receiving a user-triggered "forced re-parsing," even if the knowledge base contains target historical technical specification documents, it obtains preprocessed information of the technical specification documents by preprocessing the documents. This is mainly because previous preprocessed information may be inaccurate, and the accuracy of the preprocessed information can be improved by regenerating the preprocessed information.

[0052] In this embodiment, configuration information is extracted based on the AUTOSAR specification. Considering the differences in configuration item definitions between different AUTOSAR versions (such as AR4.2, R23-11, and R24-11), in order to extract configuration information that meets actual needs, before extracting configuration information, a first intelligent agent first determines the target AUTOSAR associated with the technical specification document input by the current user. Subsequently, configuration information is extracted from the technical specification document based on the target AUTOSAR to improve the accuracy of the extracted configuration information.

[0053] In some embodiments of this application, the target AUTOSAR associated with the user-input technical specification document can be determined through the following steps 410-430.

[0054] Step 410. Receive the first input from the user specifying the version of the automotive open system architecture.

[0055] In some embodiments of this application, the first intelligent agent may provide an interface to the user for inputting the AUTOSAR version to be used, through which the user may input a first input specifying the Automotive Open System Architecture version to be used.

[0056] In some embodiments of this application, multiple AUTOSAR version options can be displayed to the user through a web interface. The first input mentioned above can be the user's selection of the AUTOSAR version option to be used in the web interface.

[0057] In some embodiments of this application, an input box may be provided to the user, through which the user can input the AUTOSAR version to be used. Based on this, the aforementioned first input may be the user's input operation in the input box.

[0058] Step 420. Upon receiving the first input, determine the automotive open system architecture corresponding to the automotive open system architecture version specified in the first input as the target automotive open system architecture associated with the technical specification document.

[0059] If the user's first input is received, the automotive open system architecture corresponding to the first specified automotive open system architecture version is determined as the target automotive open system architecture associated with the technical specification document, and the association between the technical specification document and the target automotive open system architecture is established for subsequent use.

[0060] In some embodiments of this application, when the first input is a user's selection of the AUTOSAR version option to be used in the web interface, the AUTOSAR corresponding to the AUTOSAR version option selected by the user is determined as the target AUTOSAR.

[0061] In some embodiments of this application, when the first input is a user's input operation in the input box, the AUTOSAR corresponding to the AUTOSAR version input by the user is determined as the target AUTOSAR.

[0062] This method allows users to determine the target AUTOSAR to use based on their actual needs, improving the flexibility of configuration information extraction.

[0063] Step 430. In the absence of the first input, the latest version of the stored automotive open system architecture is identified as the target automotive open system architecture associated with the technical specification document.

[0064] If no initial input is received, meaning the user does not actively select an AUTOSAR version, the latest version of AUTOSAR will be used as the target AUTOSAR by default. This ensures that subsequent processing is always based on the correct specification version, avoiding configuration errors caused by version mismatch.

[0065] After obtaining the preprocessed information and identifying the target AUTOSAR associated with the technical specification document, the first intelligent agent transmits the preprocessed information and the association between the technical specification document and the target AUTOSAR to the second intelligent agent, thereby enabling the second intelligent agent to extract configuration information based on the aforementioned information.

[0066] Step 120. Using a second intelligent agent and a pre-trained large language model, extract target configuration information that matches the target vehicle open system architecture from the pre-processed information.

[0067] In this embodiment, the second intelligent agent integrates a pre-trained large language model with enhanced knowledge in the automotive electronics field. This model incorporates automotive electronics terminology and semantic knowledge during the training or architecture design phase, enabling it to accurately understand specific technical expressions such as "information security enablement" and "CAN communication cycle," and adapt to the parsing tasks of different car manufacturers' specification documents without additional fine-tuning.

[0068] In this embodiment, the second intelligent agent acts as a "technical expert," whose main responsibility is to extract the target configuration information of the ECU from the preprocessed information of the technical specification document based on the target AUTOSAR. The second intelligent agent can input the preprocessed information and the target AUTOSAR into a large language model, thereby intelligently extracting the target configuration information conforming to the target AUTOSAR from the preprocessed information through the large language model. The target configuration information includes the value of at least one configuration item of the automotive electronic control unit.

[0069] After obtaining the target configuration information, the second intelligent agent can transmit the target configuration information to the third intelligent agent, so that the third intelligent agent can generate the final configuration file.

[0070] Step 130. Generate a configuration file based on the target configuration information through a third intelligent agent.

[0071] In this embodiment of the application, the third intelligent agent acts as a "delivery specialist", whose main responsibilities include generating configuration files in a specified format according to user requirements.

[0072] In some embodiments of this application, the third agent can generate a structured JSON-formatted configuration file. This format facilitates subsequent program parsing and automated processing, and can be used as the system's default output option. The output content may include, but is not limited to, configuration item names and values, and may also include the units of the values, the target AUTOSAR version, metadata of the original technical specification document, etc., to ensure information integrity and traceability.

[0073] For example, the final configuration file includes the following: { "ComTimeout": { "value": "50", "unit": "ms", "version": "R24-11", "ref": "xxx" }, "PduPeriodTime": { "value": "100", "unit": "ms", "version": "R24-11", "ref": "xxx" } }” In this configuration, "ComTimeout" and "PduPeriodTime" are the names of the configuration items. "value" indicates the numerical value of the configuration item. For example, in the example above, the value of "ComTimeout" is 50 and the value of "PduPeriodTime" is 100. "unit" indicates the unit of the numerical value of the configuration item. For example, in the example above, the unit of the numerical value of "ComTimeout" is ms and the unit of the numerical value of "PduPeriodTime" is also ms. "version" indicates the version of the target AUTOSAR. For example, in the example above, the version of the target AUTOSAR is R24-11. "ref" indicates the source of the numerical value of the configuration item, such as page number, paragraph number, table coordinates, etc.

[0074] Based on user needs, configuration files conforming to the target AUTOSAR standard can be generated. After simple processing, they can be imported into mainstream development tools such as DaVinci Configurator and EAS Configurator for seamless integration.

[0075] In this embodiment, multiple intelligent agents collaborate, utilizing a large language model and based on technical specification documents and the associated target AUTOSAR, to intelligently extract the target configuration information of the ECU. This achieves automatic extraction of configuration information, lowers the professional threshold, reduces reliance on senior experts, and improves the efficiency of configuration information extraction. The manual mapping process, which originally took hours or even days, is shortened to minutes, significantly accelerating the ECU development cycle. Furthermore, extracting configuration information using a large language model from the automotive electronics field combined with AUTOSAR specifications reduces configuration errors caused by terminology ambiguity and improves the accuracy of the configuration information.

[0076] In some embodiments of this application, the ECU typically performs a variety of functions, such as enabling information security and CAN communication cycles. Different functions usually correspond to different configuration items or configuration item values. When configuring the ECU, the user may only need to configure one or a few functions. In this case, it is only necessary to obtain the configuration information corresponding to the target function that the user needs to configure, without needing to obtain the full configuration information of the ECU.

[0077] In view of this, see Figure 2 The flowchart shows a configuration information extraction method provided in some other embodiments of this application, such as... Figure 5 As shown, in order to obtain configuration information more accurately, steps 510-520 can be executed before step 120 above.

[0078] Step 510. Receive the target function description information input by the user through the second intelligent agent.

[0079] The target function description information is used to describe the target function that needs to be configured.

[0080] In some embodiments of this application, the second intelligent agent may provide an interface for the user to input information. Through this interface, the user may input target function description information to the second intelligent agent by means of voice input, text input, etc.

[0081] Step 520. Determine the user intent based on the target function description information using the large language model in the second intelligent agent.

[0082] After receiving the user's input description of the target function, the second agent uses a built-in augmented model of automotive electronics to semantically parse the user's input description, thereby obtaining the user's intent. The user intent is used to instruct the user on the target function to be configured.

[0083] Based on this, step 120 above can be implemented as follows: using the large language model in the second intelligent agent, based on the user intent, extract target configuration information that matches the target vehicle open system architecture from the preprocessed information.

[0084] The above methods can accurately extract target configuration information that matches the user's intent, adapt to the diverse needs of users, and achieve precise extraction of configuration information.

[0085] In some embodiments of this application, the target configuration information that matches the target vehicle open system architecture is extracted from preprocessed information based on user intent using a large language model, which may include the following steps 1201-1204.

[0086] Step 1201. Based on the user intent and the configuration structure and naming rules in the target vehicle open system architecture, determine the candidate configuration item information that conforms to the user intent. The candidate configuration item information includes the name and semantics of the candidate configuration item related to the target function.

[0087] The second intelligent agent can automatically identify and map candidate configuration item information related to the target function through a large language model, based on user intent and combined with the configuration structure and naming rules in the target AUTOSAR. The candidate configuration item information includes the name and semantics of the candidate configuration item related to the target function, as well as the functional semantics of the target function and the contextual meaning of the candidate configuration item. Based on the candidate configuration item information, a preliminary configuration item proposal is formed.

[0088] Step 1202. Retrieve the preprocessed information based on the candidate configuration item information to obtain the retrieval results, which include the context content of each candidate configuration item.

[0089] After obtaining the candidate configuration item information, the second agent can perform related information retrieval in the preprocessed information based on the name and semantics of the candidate configuration items to obtain retrieval results. The retrieval results include the context content of each candidate configuration item. The context content of a candidate configuration item refers to the content in the preprocessed information related to that candidate configuration item.

[0090] Step 1203. For each candidate configuration item, determine the value of the candidate configuration item under the target function based on the context content of the candidate configuration item using the large language model.

[0091] After obtaining the search results, the second agent can submit the context of the candidate configuration item to the large language model for comprehensive analysis. Combining the configuration patterns under different functional scenarios, it can infer and generate a reasonable value for the candidate configuration item under the target function.

[0092] In some embodiments of this application, considering that a candidate configuration item may retrieve multiple related contextual contents, in order to improve processing efficiency, for each candidate configuration item, if the number of entries in its corresponding contextual contents is greater than N, the relevance score between each contextual content and the candidate configuration item is calculated. The contextual contents with the top N relevance scores are used as the target contextual contents. In step 1203 above, only the target contextual contents are input into the large language model, so that the large language model determines the value of the candidate configuration item under the target function based on the target contextual contents of the candidate configuration item. The relevance score can be calculated using similarity algorithms such as TF / IDF. N is a positive integer, and the value of N can be set according to actual needs. For example, if N is 5, then for each candidate configuration item, the top 5 related entries with the top relevance scores can be selected from its corresponding contextual contents as the target contextual contents input into the large language model.

[0093] Step 1204. Based on the names and values ​​of the candidate configuration items, construct the target configuration information.

[0094] The above approach can accurately extract target configuration information and reduce configuration errors caused by ambiguity in terminology.

[0095] In some embodiments of this application, see Figure 3 This is a flowchart illustrating a configuration information extraction method provided in some embodiments of this application. To further improve the accuracy of the configuration information, please refer to... Figure 3 Before step 130 above, step 600 can be performed first.

[0096] Step 600. The fourth intelligent agent verifies each candidate configuration item in the target configuration information to obtain the target configuration item that passes the verification.

[0097] In this embodiment, the fourth intelligent agent acts as a "quality auditor," whose main responsibilities include verifying candidate configuration items of the target configuration information. Verification may include at least one of the following: numerical validity verification and automotive open system architecture version compatibility verification.

[0098] In some embodiments of this application, numerical validity verification is used to verify whether the values ​​of each candidate configuration item in the target configuration information conform to the configuration constraint specifications (such as data type, value range, and unit) defined by the target AUTOSAR. When performing numerical validity verification, the validity of the value of each candidate configuration item in the target configuration information can be verified one by one based on the configuration constraint specifications defined in the target vehicle open system architecture. For each candidate configuration item, if the value of the candidate configuration item is valid, it is determined that the candidate configuration item has passed the numerical validity verification; otherwise, it is determined that the candidate configuration item has failed the numerical validity verification. For example, for ComTimeout, version R24-11 specifies that its minimum value is 10ms and its maximum value is 1000ms. If ComTimeout is assigned the value "5ms" in the target configuration information, the fourth agent can determine that ComTimeout is "numerically invalid".

[0099] In some embodiments of this application, when the target configuration information contains invalid values, that is, when there are candidate configuration items with invalid values, the fourth intelligent agent can automatically trigger a feedback mechanism to return the candidate configuration item to the second intelligent agent for re-analysis and correction of the values.

[0100] In some embodiments of this application, the automotive open system architecture version compatibility check is used to verify whether each candidate configuration item in the target configuration information is defined in the target AUTOSAR. When performing the automotive open system architecture version compatibility check, each candidate configuration item in the target configuration information can be verified one by one to see if it is defined in the target AUTOSAR. For each candidate configuration item, if it is defined in the target AUTOSAR, then the candidate item has passed the automotive open system architecture version compatibility check; otherwise, it has failed. For example, ComTimeout is not defined in AUTOSAR version AR4.2. If the user specifies the target AUTOSAR version as AR4.2, then the ComTimeout configuration item in the target configuration information can be determined to be "incompatible".

[0101] In some embodiments of this application, configuration items determined to be incompatible can be returned to the second intelligent agent to re-evaluate alternative configuration items or adjust semantic understanding.

[0102] In some embodiments of this application, if the verification only includes numerical validity verification, then candidate configuration items in the target configuration information that have passed the numerical validity verification are determined as target configuration items. If the verification only includes automotive open system architecture version compatibility verification, then candidate configuration items in the target configuration information that have passed the automotive open system architecture version compatibility verification are determined as target configuration items. If the verification includes both numerical validity verification and automotive open system architecture version compatibility verification, then candidate configuration items in the target configuration information that have passed both numerical validity verification and automotive open system architecture version compatibility verification are determined as target configuration items.

[0103] Accordingly, step 130 above can be implemented as follows: by a third intelligent agent, an ECU configuration file is generated based on the information corresponding to the target configuration item in the target configuration information.

[0104] In the above scheme, the fourth intelligent agent performs numerical validity verification on the target configuration information, which can restrict and filter illegal values ​​and reduce ECU function abnormalities caused by incorrect configuration. By performing version compatibility verification of the automotive open system architecture, it can support users to specify the target AUTOSAR version, ensuring that the parsing results are consistent with the project requirements and reducing configuration errors caused by version differences.

[0105] In some embodiments of this application, after the entire verification process is completed, the intermediate results generated in this step (including identified configuration items, verification status, context analysis basis, etc.) can be persistently saved. When the user subsequently performs the same or similar configuration tasks, they can choose to load the historical intermediate results, skip the repeated identification and verification process, significantly improve system response efficiency, and reduce user waiting time.

[0106] In related technologies, the disordered order of configuration items in the extracted configuration information often leads to import failures in downstream toolchains or generates numerous warnings. See also Figure 4 The flowchart illustrates a configuration information extraction method provided in some other embodiments of this application, such as... Figure 4 As shown, in order to improve the standardization of the final generated configuration file and facilitate its import by downstream toolchains, the following step 700 can be executed before step 130 above.

[0107] Step 700. Using the fourth intelligent agent, based on the configuration item dependency graph corresponding to the target vehicle open system architecture, sort the configuration items in the target configuration information to obtain the sorting result.

[0108] In some embodiments of this application, a configuration item dependency graph corresponding to the target AUTOSAR can be pre-constructed based on the target AUTOSAR specification document, and the configuration item dependency graph can be stored in a knowledge base. The configuration item dependency graph is used to indicate the dependencies between configuration items, and the dependencies include one of the following: inclusion, reference, or initialization. Based on this, the fourth agent can obtain the pre-constructed configuration item dependency graph from the knowledge base.

[0109] In some embodiments of this application, target configuration information can be mapped to the target AUTOSAR container structure. A topological sorting algorithm is used to process the dependencies of each configuration item in the target configuration information based on the configuration item dependency graph, and then a sorting result is obtained based on the dependencies. The sorting result meets sorting conditions. The sorting conditions include at least one of the following three conditions: 1. The referenced item is placed before the referencing item; that is, the referenced item comes first, and the referencing item comes after.

[0110] 2. The parent container appears after its child elements.

[0111] 3. Configuration items at the same level are arranged in the recommended order according to the target vehicle's open system architecture.

[0112] Accordingly, step 130 above can be implemented as follows: generating a configuration file based on the sorting results through a third intelligent agent.

[0113] In this way, configuration items can be automatically sorted based on the dependencies of the target AUTOSAR, reducing the workload of manual adjustments and making it easier for downstream toolchains to import them directly.

[0114] In some embodiments of this application, considering that not all users require configuration items in the configuration file to be arranged according to dependencies, the fourth intelligent agent can provide users with multiple configuration file generation modes to choose from in order to meet different user needs. Different modes correspond to different configuration information processing methods. Among the multiple modes is a first mode for indicating that configuration items are sorted before generating the configuration file. Based on this, before performing the above step 700, a second input from the user to select the configuration file generation mode can be received first, wherein the second input is used to select the configuration file generation mode to be used by the user. When the configuration file generation mode selected by the second input is the first mode, the configuration items in the target configuration information are sorted based on the configuration item dependency graph.

[0115] In some embodiments of this application, the configuration file generation mode available to the user may further include a second mode, which indicates that configuration items are grouped before generating the configuration file. Based on this, the configuration information extraction method may further include: When the configuration file generation mode selected in the second input is the second mode, the configuration items in the target configuration information are grouped based on the configuration item dependency graph corresponding to the target AUTOSAR, and the configuration file is generated based on the grouping results. The grouping strategy can be set according to actual needs; for example, configuration items at the same level are grouped together, or configuration items referencing the same configuration item are grouped together, etc., without specific limitations.

[0116] The configuration items are grouped but not sorted, making the generated configuration files easier for humans to read.

[0117] In some embodiments of this application, the configuration file generation mode available to the user may further include a third mode, which indicates that the configuration file is generated based on the original order in the target configuration information. Based on this, the configuration information extraction method may further include: if the configuration file generation mode selected by the second input is the third mode, directly executing step 130.

[0118] The above approach allows users to determine whether to sort or group configuration items based on their actual needs, thereby meeting diverse user requirements and improving the flexibility of configuration information extraction.

[0119] In some embodiments of this application, the configuration information extraction device can also provide an intuitive web front-end interface, where engineers can view all extracted configuration items and verification results, and manually input correction values ​​for items that failed verification or need adjustment. The review process supports batch processing to improve processing efficiency. The confirmed results will be fed back to the system as the "gold standard" for subsequent knowledge base updates.

[0120] In some embodiments of this application, the knowledge base can be continuously updated to optimize it and improve the accuracy of configuration information extraction. The configuration information extraction device can record the rules and context of each successful mapping, dynamically optimizing the mapping rule base within the second intelligent body. This mapping rule base is used to indicate the mapping relationship between functions and configuration items. For example, if the "communication cycle" function is repeatedly confirmed to correspond to the PduPeriodTime configuration item, the priority of this rule can be increased to reduce ambiguity in similar future scenarios.

[0121] In addition, the information extraction device can also periodically clean up expired or infrequently accessed document caches to ensure the efficiency of knowledge base storage.

[0122] The configuration information extraction device can also support manually marking "important documents" to retain their preprocessed results for a long time.

[0123] The following example, using a configuration information extraction device comprising a first intelligent agent, a second intelligent agent, a third intelligent agent, a fourth intelligent agent, and a knowledge base, illustrates the data interaction between the entities within the configuration information extraction device during the configuration information extraction process.

[0124] See Figure 5 The configuration information extraction device includes a first intelligent agent 501, a second intelligent agent 502, a third intelligent agent 503, a fourth intelligent agent 504, and a knowledge base 505, such as... Figure 5 As shown, the first intelligent agent 501 has functions such as uploading documents, calculating document content fingerprints, querying caches, preprocessing, and loading document metadata. The second intelligent agent 502 integrates an enhanced large model in the field of automotive electronics, and has functions such as recognizing user intent and configuration items, context-aware matching of configuration items and values, and receiving and correcting verification feedback. The third intelligent agent 503 has functions such as generating JSON format configuration files, generating ARXML configuration files, adapting to the target AUTOSAR version standard, and outputting metadata. The fourth intelligent agent 504 has functions such as numerical validity verification, version compatibility verification, dependency resolution and topological sorting, and saving intermediate results. The knowledge base 505 is used to store the configuration item mapping rule base, configuration item dependency graph, processed document index (including content fingerprints), structured intermediate result cache, and document cache index.

[0125] During the configuration information extraction process, the knowledge base 505 can provide the first intelligent agent 501 with a structured intermediate result cache and a document cache index, so that the first intelligent agent 501 can obtain the preprocessing information of the technical specification document uploaded by the current user. The first intelligent agent 501 provides the preprocessing information and association relationships of the technical specification document to the second intelligent agent 502. The association relationships are used to indicate the target AUTOSAR associated with the technical specification document. The knowledge base 505 can provide the second intelligent agent 502 with a configuration item mapping rule base. The second intelligent agent 502 can determine a candidate configuration item list based on the large language model and the configuration item mapping rule base, and transmit the candidate configuration item list to the fourth intelligent agent 504. The candidate configuration item list may include the numerical value and context content of the candidate configuration item. After verifying the candidate configuration item list, the fourth intelligent agent 504 can return invalid configuration items to the second intelligent agent 502 and request the second intelligent agent 502 to correct the invalid configuration items. The knowledge base 505 can provide a configuration item dependency graph to the fourth intelligent agent 504. After verifying the candidate configuration item list, the fourth intelligent agent 504 can sort the verified configuration items based on the configuration item dependency graph and transmit the verified and sorted configuration items to the third intelligent agent 503. The third intelligent agent 503 can generate a configuration file and update the mapping rules to the knowledge base.

[0126] The configuration information extraction method provided in this application reduces the manual mapping process, which originally took hours or even days, to minutes, significantly accelerating the ECU development cycle and improving configuration efficiency. It utilizes an enhanced large language model analysis of automotive electronics specifications, combined with AUTOSAR standards, to reduce configuration errors caused by terminological ambiguity. A user verification mechanism ensures data consistency with the original documents; numerical validity restrictions filter illegal values, reducing ECU malfunctions caused by incorrect configurations; it supports user-specified AUTOSAR versions, ensuring the parsing results match project requirements, reducing configuration errors caused by version differences, and improving configuration accuracy; new engineers can quickly complete complex configurations with system assistance, reducing reliance on senior experts and lowering the professional threshold; furthermore, extraction rules and mapping relationships can be accumulated as organizational assets, supporting reuse across multiple projects, achieving knowledge accumulation and reuse; it employs multi-agent collaborative operation for distributed control, with each agent making independent decisions, resulting in high system robustness; the knowledge base can be continuously optimized through user feedback, achieving flexible expansion and dynamic adaptation; and automatic sorting based on AUTOSAR dependencies reduces manual adjustment workload and improves the standardization of the final output configuration information.

[0127] Based on the configuration information extraction method provided in the above embodiments, this application also provides specific implementation methods of the configuration information extraction device. Please refer to the following embodiments.

[0128] Figure 6 The diagram illustrates a configuration information extraction device provided in some embodiments of this application, such as... Figure 6 As shown, the configuration information extraction device 600 provided in this application embodiment includes the following modules: The document parsing module 601 is used to obtain preprocessed information of the technical specification document input by the user through the first intelligent agent, and determine the target automotive open system architecture associated with the technical specification document; The configuration item mapping module 602 is used to extract target configuration information that matches the target vehicle open system architecture from preprocessed information through a second intelligent agent using a pre-trained large language model. The large language model is trained based on knowledge in the automotive electronics field. The target configuration information includes the value of at least one configuration item of the automotive electronic control unit. Delivery module 603 is used to generate a configuration file based on the target configuration information through a third intelligent agent.

[0129] In this embodiment, multiple intelligent agents collaborate, utilizing a large language model and based on technical specification documents and the associated target AUTOSAR, to intelligently extract the target configuration information of the ECU. This achieves automatic extraction of configuration information, improving extraction efficiency and reducing the manual mapping process, which originally took hours or even days, to minutes, significantly accelerating the ECU development cycle. Furthermore, extracting configuration information using a large language model from the automotive electronics field combined with AUTOSAR specifications reduces configuration errors caused by terminology ambiguity, improving the accuracy of the configuration information.

[0130] In some embodiments of this application, in order to save document processing time, the document parsing module 601 can be used for: The document identifier for generating the aforementioned technical specification document; The document identifier is matched with the document identifiers stored in the knowledge base to obtain the matching result. The knowledge base is used to store preprocessing information and document identifiers of historical technical specification documents. If the matching result indicates that the knowledge base contains a target historical technical specification document whose document identifier matches the document identifier of the technical specification document, the preprocessing information of the target historical technical specification document is obtained as the preprocessing information of the technical specification document. If the matching result indicates that the knowledge base does not contain a target historical technical specification document whose document identifier matches the document identifier of the technical specification document, the technical specification document is preprocessed to obtain preprocessed information of the technical specification document, and the document identifier and preprocessed information of the technical specification document are stored in the knowledge base.

[0131] In some embodiments of this application, to ensure that subsequent processing is always based on the correct specification version and to avoid configuration errors caused by version mismatch, the document parsing module 601 can be used for: Receive the first input from the user specifying the version of the automotive open system architecture; Upon receiving the first input, the automotive open system architecture corresponding to the automotive open system architecture version specified by the first input is determined as the target automotive open system architecture associated with the technical specification document; If the first input is not received, the latest version of the stored automotive open system architecture is identified as the target automotive open system architecture associated with the technical specification document.

[0132] In some embodiments of this application, in order to achieve accurate extraction of configuration information, the configuration item mapping module 602 is further configured to: Before the second intelligent agent extracts target configuration information matching the target vehicle open system architecture from the pre-processed information using a pre-trained large language model, the second intelligent agent receives target function description information input by the user; based on the target function description information, the large language model determines the user's intent, which is used to instruct the user on the target function to be configured. Accordingly, the step of extracting target configuration information matching the target vehicle open system architecture from preprocessed information using a pre-trained large language model via a second intelligent agent includes: The second intelligent agent, using a pre-trained large language model, extracts target configuration information that matches the target vehicle open system architecture from pre-processed information based on the user intent.

[0133] In some embodiments of this application, in order to adapt to the technical specification documents of different car manufacturers and accurately extract configuration information from them, the configuration item mapping module 602 is used for: Based on the user intent and the configuration structure and naming rules in the target vehicle open system architecture, candidate configuration item information that conforms to the user intent is determined. The candidate configuration item information includes the name and semantics of the candidate configuration item related to the target function. The preprocessed information is retrieved based on the candidate configuration item information to obtain retrieval results, which include the context content of each candidate configuration item. For each candidate configuration item, the value of the candidate configuration item under the target function is determined based on the context content of the candidate configuration item using a large language model. The target configuration information is constructed based on the name and value of the candidate configuration items.

[0134] In some embodiments of this application, to improve the validity of configuration information and reduce abnormal data, the device 600 further includes: a verification module, used for: Before generating the configuration file based on the target configuration information through the third intelligent agent, the fourth intelligent agent verifies each candidate configuration item in the target configuration information to obtain the target configuration item that passes the verification. The verification includes at least one of the following: numerical validity verification and automotive open system architecture version compatibility verification. Delivery module 603 is used for: An ECU configuration file is generated by a third intelligent agent based on the information corresponding to the target configuration item in the target configuration information.

[0135] In some embodiments of this application, the verification module is used for: Based on the configuration constraint specifications defined in the target vehicle open system architecture, verify the validity of the value of each candidate configuration item in the target configuration information one by one; Verify one by one whether each candidate configuration item in the target configuration information is defined in the target vehicle open system architecture; Candidate configuration items whose values ​​are valid and defined in the target vehicle open system architecture are identified as target configuration items that pass the verification.

[0136] In some embodiments of this application, to improve the standardization of the configuration file, the verification module is further configured to: Before generating the configuration file based on the target configuration information through the third intelligent agent, the configuration items in the target configuration information are sorted by the fourth intelligent agent based on the configuration item dependency graph corresponding to the target vehicle open system architecture to obtain a sorting result; the configuration item dependency graph is used to indicate the dependency relationship between configuration items, and the dependency relationship includes one of the following: inclusion, reference, initialization; the sorting result meets the sorting conditions, and the sorting conditions include at least one of the following: the referenced item is located before the referencing item, the parent container is after its child element, and the sibling configuration items are arranged in the recommended order of the target vehicle open system architecture; Delivery module 604 is used for: A configuration file is generated based on the sorting results using a third intelligent agent.

[0137] In some embodiments of this application, to improve flexibility, the verification module is also used for: Before sorting the configuration items in the target configuration information based on the configuration item dependency graph, a second input from the user is received to select the configuration file generation mode; The configuration items in the target configuration information are sorted based on the configuration item dependency graph corresponding to the target vehicle open system architecture to obtain the sorting result, including: When the configuration file generation mode selected by the second input is the first mode, the configuration items in the target configuration information are sorted based on the configuration item dependency graph to obtain the sorting result. The first mode is used to indicate that the configuration items are sorted before generating the configuration file. The verification module is also used for: When the configuration file generation mode selected by the second input is the second mode, the configuration items in the target configuration information are grouped based on the configuration item dependency graph to obtain the grouping results, and the configuration file is generated based on the grouping results. The second mode is used to indicate that the configuration items are grouped before generating the configuration file.

[0138] Figure 7 A schematic diagram of the hardware structure of the electronic device provided in an embodiment of this application is shown.

[0139] The electronic device 700 may include a processor 701 and a memory 702 storing computer program instructions.

[0140] Specifically, the processor 701 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0141] Memory 702 may include mass storage for data or instructions. For example, and not limitingly, memory 702 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 702 may include removable or non-removable (or fixed) media. Where appropriate, memory 702 may be internal or external to electronic device 700. In a particular embodiment, memory 702 is a non-volatile solid-state memory.

[0142] In a specific embodiment, the memory 702 can be implemented as ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 702 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 702 and executed by the processor 701. The processor 701 reads and executes the computer program instructions stored in the memory 702 to implement any of the configuration information extraction methods in the above embodiments.

[0143] In one example, the electronic device 700 may also include a communication interface 703 and a bus 710. Wherein, as... Figure 7 As shown, the processor 701, memory 702, and communication interface 703 are connected through bus 710 and complete communication with each other.

[0144] The communication interface 703 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0145] Bus 710 includes hardware, software, or both, that couples components of electronic device 700 together. For example, and not limitingly, the bus may include Accelerated Graphics Port (AGP) or other graphics buses, Enhanced Industry Standard Architecture (EISA) buses, Front Side Bus (FSB), HyperTransport (HT) interconnects, Industry Standard Architecture (ISA) buses, Infinite Bandwidth Interconnects, Low Pin Count (LPC) buses, memory buses, Microchannel Architecture (MCA) buses, Peripheral Component Interconnect (PCI) buses, PCI-Express (PCI-X) buses, Serial Advanced Technology Attachment (SATA) buses, Video Electronics Standards Association Local (VLB) buses, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 710 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, any suitable bus or interconnect is contemplated herein.

[0146] The electronic device 700 can execute the configuration information extraction method in the embodiments of this application based on the first intelligent agent, the second intelligent agent, the third intelligent agent, the fourth intelligent agent, and the knowledge base, thereby achieving a combination of Figure 1 and Figure 5 The method described is for extracting configuration information.

[0147] Furthermore, in conjunction with the configuration information extraction methods described in the above embodiments, this application embodiment can provide a computer storage medium for implementation. This computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the configuration information extraction methods described in the above embodiments.

[0148] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0149] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute the configuration information extraction method as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0150] This application also provides a computer program product, including a computer program, which, when executed, implements any of the configuration information extraction methods described in the above embodiments.

[0151] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0152] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0153] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0154] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0155] The above description is merely a specific implementation of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A method for extracting configuration information, characterized in that, The method includes: The first intelligent agent obtains preprocessed information of the technical specification document input by the user and determines the target automotive open system architecture associated with the technical specification document. Through a second intelligent agent, a pre-trained large language model is used to extract target configuration information that matches the target vehicle open system architecture from the pre-processed information. The large language model is trained based on knowledge in the automotive electronics field. The target configuration information includes the value of at least one configuration item of the automotive electronic control unit. A configuration file is generated based on the target configuration information through a third intelligent agent.

2. The method according to claim 1, characterized in that, The preprocessing information for obtaining user-input technical specification documents includes: The document identifier for generating the aforementioned technical specification document; The document identifier is matched with the document identifiers stored in the knowledge base to obtain the matching result. The knowledge base is used to store preprocessing information and document identifiers of historical technical specification documents. If the matching result indicates that the knowledge base contains a target historical technical specification document whose document identifier matches the document identifier of the technical specification document, the preprocessing information of the target historical technical specification document is obtained as the preprocessing information of the technical specification document. If the matching result indicates that the knowledge base does not contain a target historical technical specification document whose document identifier matches the document identifier of the technical specification document, the technical specification document is preprocessed to obtain preprocessed information of the technical specification document, and the document identifier and preprocessed information of the technical specification document are stored in the knowledge base.

3. The method according to claim 1, characterized in that, The determination of the target automotive open system architecture associated with the technical specification document includes: Receive the first input from the user specifying the version of the automotive open system architecture; Upon receiving the first input, the automotive open system architecture corresponding to the automotive open system architecture version specified by the first input is determined as the target automotive open system architecture associated with the technical specification document; If the first input is not received, the latest version of the stored automotive open system architecture is identified as the target automotive open system architecture associated with the technical specification document.

4. The method according to claim 1, characterized in that, Before the step of extracting target configuration information matching the target vehicle open system architecture from the preprocessed information using a pre-trained large language model via a second intelligent agent, the method further includes: The second intelligent agent receives the target function description information input by the user. The user intent is determined based on the target function description information using the large language model, and the user intent is used to indicate the target function to be configured by the user. The step of extracting target configuration information matching the target vehicle open system architecture from the preprocessed information using a pre-trained large language model via a second intelligent agent includes: Based on the user intent, the target configuration information that matches the target vehicle open system architecture is extracted from the preprocessed information using the large language model.

5. The method according to claim 4, characterized in that, The step of extracting target configuration information matching the target vehicle open system architecture from the preprocessed information based on the user intent using the large language model includes: Based on the user intent and the configuration structure and naming rules in the target vehicle open system architecture, candidate configuration item information that conforms to the user intent is determined. The candidate configuration item information includes the name and semantics of the candidate configuration item related to the target function. The preprocessed information is retrieved based on the candidate configuration item information to obtain retrieval results, which include the context content of each candidate configuration item. For each candidate configuration item, the value of the candidate configuration item under the target function is determined based on the context content of the candidate configuration item using a large language model. The target configuration information is constructed based on the name and value of the candidate configuration items.

6. The method according to any one of claims 1-5, characterized in that, Before generating the configuration file based on the target configuration information through a third intelligent agent, the method further includes: The fourth intelligent agent verifies each candidate configuration item in the target configuration information to obtain the target configuration item that passes the verification. The verification includes at least one of the following: numerical validity verification and automotive open system architecture version compatibility verification. The step of generating a configuration file based on the target configuration information through a third intelligent agent includes: An ECU configuration file is generated by a third intelligent agent based on the information corresponding to the target configuration item in the target configuration information.

7. The method according to claim 6, characterized in that, The step of validating each candidate configuration item in the target configuration information to obtain the target configuration item that passes the validation includes: Based on the configuration constraint specifications defined in the target vehicle open system architecture, verify the validity of the value of each candidate configuration item in the target configuration information one by one; Verify one by one whether each candidate configuration item in the target configuration information is defined in the target vehicle open system architecture; Candidate configuration items whose values ​​are valid and defined in the target vehicle open system architecture are identified as target configuration items that pass the verification.

8. The method according to any one of claims 1-5, characterized in that, Before generating the configuration file based on the target configuration information through a third intelligent agent, the method further includes: The fourth intelligent agent sorts the configuration items in the target configuration information based on the configuration item dependency graph corresponding to the target vehicle open system architecture, and obtains the sorting result. The configuration item dependency graph is used to indicate the dependency relationship between configuration items. The dependency relationship includes one of the following: inclusion, reference, initialization. The sorting result meets the sorting conditions. The sorting conditions include at least one of the following: the referenced item is located before the referencing item, the parent container is after its child element, and the sibling configuration items are arranged in the recommended order of the target vehicle open system architecture. The step of generating a configuration file based on the target configuration information through a third intelligent agent includes: A configuration file is generated based on the sorting results using a third intelligent agent.

9. The method according to claim 8, characterized in that, Before sorting the configuration items in the target configuration information based on the configuration item dependency graph corresponding to the target vehicle open system architecture and obtaining the sorting result, the method further includes: Receive a second input from the user to select the configuration file generation mode; The configuration items in the target configuration information are sorted based on the configuration item dependency graph corresponding to the target vehicle open system architecture to obtain the sorting result, including: When the configuration file generation mode selected by the second input is the first mode, the configuration items in the target configuration information are sorted based on the configuration item dependency graph corresponding to the target vehicle open system architecture to obtain the sorting result. The first mode is used to indicate that the configuration items are sorted before generating the configuration file. The method further includes: When the configuration file generation mode selected by the second input is the second mode, the configuration items in the target configuration information are grouped based on the configuration item dependency graph to obtain the grouping results, and the configuration file is generated based on the grouping results. The second mode is used to indicate that the configuration items are grouped before generating the configuration file.

10. A configuration information extraction device, characterized in that, The device includes: The document parsing module is used to obtain preprocessed information of the technical specification document input by the user through the first intelligent agent, and to determine the target automotive open system architecture associated with the technical specification document; The configuration item mapping module is used to extract target configuration information that matches the target vehicle open system architecture from the preprocessed information by a second intelligent agent using a pre-trained large language model. The large language model is trained based on knowledge in the automotive electronics field. The target configuration information includes the value of at least one configuration item of the automotive electronic control unit. The delivery module is used to generate a configuration file based on the target configuration information through a third intelligent agent.

Images