Vehicle diagnostic software launch method, system, vehicle diagnostic device, and media

Abstract

The application relates to the technical field of vehicle diagnosis, and discloses a vehicle diagnosis software starting method and system, a vehicle diagnosis device and a medium. The vehicle diagnosis device comprises the following steps: acquiring a voice instruction for starting vehicle diagnosis software input by a user, and converting the voice instruction into voice text; performing entity recognition on the voice text to acquire target entities in the voice text; analyzing the association relationship between the target entities based on a knowledge graph, and determining target vehicle diagnosis software corresponding to the target entities in the knowledge graph, wherein the knowledge graph comprises the association relationship between brands, vehicle models and vehicle diagnosis software in different languages; and generating a structured instruction corresponding to the target vehicle diagnosis software to start the target vehicle diagnosis software through the structured instruction. The application realizes the conversion from a voice instruction to an executable diagnosis command, reduces the step of manually searching for diagnosis software, improves the interaction efficiency, and solves the problems of synonyms, multiple languages and multiple brand terms.

Description

Technical Field

[0001] This application relates to the field of vehicle diagnostic technology, and in particular to a vehicle diagnostic software startup method, system, vehicle diagnostic equipment, and medium. Background Technology

[0002] In the field of intelligent vehicles and intelligent devices, users often need to control vehicle diagnostic software of specific brands via voice commands. Existing technologies have the following drawbacks: most existing voice systems are primarily single-language and cannot recognize or adapt to brand and model names in other language environments; they cannot accurately identify brands and models, often leading to confusion, causing vehicle diagnostic software to call errors or fail to respond; the mapping relationship between software and brands / models is unclear, resulting in delayed voice command responses and a poor user experience. Summary of the Invention

[0003] In view of this, embodiments of this application provide a vehicle diagnostic software startup method, system, vehicle diagnostic equipment, and medium, which can effectively solve the problems of easy errors and slow response during vehicle diagnostic software startup.

[0004] In a first aspect, embodiments of this application provide a method for starting vehicle diagnostic software, including:

[0005] Acquire the user's voice command to start the vehicle diagnostic software, and convert the voice command into speech text;

[0006] Entity recognition is performed on the spoken text to obtain the target entities in the spoken text;

[0007] The association between the target entities is analyzed based on the knowledge graph, and the target vehicle diagnostic software corresponding to the target entity in the knowledge graph is determined. The knowledge graph includes the association between brands, models and vehicle diagnostic software in different languages.

[0008] Generate structured instructions corresponding to the target vehicle diagnostic software, and start the target vehicle diagnostic software through the structured instructions.

[0009] In a first possible embodiment of the first aspect, the target entity includes a target brand and a target vehicle model, and the step of performing entity recognition on the speech text to obtain the target entity in the speech text includes:

[0010] The semantics of the speech text are obtained by performing semantic parsing on the shared underlying layer of the pre-trained multilingual understanding model.

[0011] The entity recognition head of the multilingual understanding model performs entity recognition on the semantic vector to obtain the target brand and target vehicle model in the speech text.

[0012] In a second possible embodiment of the first aspect, the step of analyzing the relationships between the target entities based on the knowledge graph and determining the target vehicle diagnostic software corresponding to the target entities in the knowledge graph includes:

[0013] The graph self-attention network of the multilingual understanding model is used to analyze the relationships between the target entities in the knowledge graph and generate relationship tags between the target car model and the target brand; the relationship tags between the target entities represent the attribution relationship between the target car model and the target brand.

[0014] Given that the target vehicle model belongs to the target brand, the target vehicle diagnostic software corresponding to the target vehicle model and the target brand in the knowledge graph is determined by the graph self-attention network.

[0015] In a third possible embodiment of the first aspect, generating the structured instructions corresponding to the target vehicle diagnostic software includes:

[0016] The decoder of the multilingual understanding model uses the semantic vector of the speech text and the target vehicle diagnostic software to generate structured instructions corresponding to the target vehicle diagnostic software.

[0017] In a fourth possible embodiment of the first aspect, the pre-training process of the multilingual understanding model includes:

[0018] The collected multilingual data will undergo data augmentation processing;

[0019] Based on the multilingual data after data augmentation, a model training task is performed on the neural network model to obtain a multilingual model after the training task is performed; the model training task includes a masked language training task and a language prediction task.

[0020] Construct multiple independent task heads for the multilingual model after the training task is performed, and jointly optimize each of the independent task heads based on the multi-task loss function.

[0021] In a fifth possible embodiment of the first aspect, the process of constructing the knowledge graph includes:

[0022] Collect multilingual brand, model and vehicle diagnostic software, and perform data cleaning processing on the collected multilingual brand, model and vehicle diagnostic software;

[0023] Construct a multilingual mapping table for each of the aforementioned brands, vehicle models, and vehicle diagnostic software, and construct the knowledge graph based on the multilingual mapping table.

[0024] In a sixth possible embodiment of the first aspect, it further includes:

[0025] The acoustic model is used to identify whether the user's input voice text is a preset wake word;

[0026] The voiceprint recognition is activated when the preset wake word is detected, the user's voiceprint features are extracted, and the user's identity is verified based on the comparison results between the voiceprint features and the pre-stored voiceprint database.

[0027] If the user authentication is successful, the voice command to start the vehicle diagnostic software will be converted into the voice text.

[0028] Secondly, embodiments of this application provide a vehicle diagnostic software startup system, comprising:

[0029] A voice recognition module is used to acquire the user's voice command to start the vehicle diagnostic software and to convert the voice command into speech text.

[0030] An entity recognition module is used to perform entity recognition on the speech text in order to obtain the target entities in the speech text.

[0031] The knowledge graph analysis module is used to analyze the relationships between the target entities based on the knowledge graph, and to determine the target vehicle diagnostic software corresponding to the target entities in the knowledge graph. The knowledge graph includes the relationships between brands, vehicle models and vehicle diagnostic software in different languages.

[0032] The software startup module is used to generate structured instructions corresponding to the target vehicle diagnostic software, so as to start the target vehicle diagnostic software through the structured instructions.

[0033] Thirdly, embodiments of this application provide a vehicle diagnostic device, which includes a processor and a memory. The memory stores a computer program, and the processor executes the computer program to implement the vehicle diagnostic software startup method described above.

[0034] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed on a processor, implements the vehicle diagnostic software startup method described above.

[0035] The embodiments of this application have the following beneficial effects:

[0036] This embodiment of a vehicle diagnostic software startup method includes: acquiring a user-inputted voice command to start the vehicle diagnostic software, and converting the voice command into speech text; performing entity recognition on the speech text to obtain target entities in the speech text; analyzing the relationships between target entities based on a knowledge graph, and determining the target vehicle diagnostic software corresponding to the target entities in the knowledge graph, wherein the knowledge graph includes the relationships between brands, vehicle models, and vehicle diagnostic software in different languages; generating a structured instruction corresponding to the target vehicle diagnostic software, so as to start the target vehicle diagnostic software through the structured instruction. This application achieves unified semantic representation in multiple languages, and can recognize and unify the standard brand and vehicle model regardless of whether the user uses Chinese, English, Spanish, etc. This application achieves end-to-end automation from voice command to executable diagnostic command, reducing the steps of manually searching for diagnostic software, improving interaction efficiency, supporting multilingual input and entity recognition, and solving the problem of multilingual and multi-brand terminology. It significantly improves the accuracy of speech recognition in multilingual scenarios and reduces the command failure rate caused by language mismatch. Attached Figure Description

[0037] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This paper illustrates a first flowchart of a vehicle diagnostic software startup method according to an embodiment of this application.

[0039] Figure 2 This paper illustrates a second flowchart of the vehicle diagnostic software startup method according to an embodiment of this application.

[0040] Figure 3 This paper illustrates a third flowchart of the vehicle diagnostic software startup method according to an embodiment of this application.

[0041] Figure 4 A schematic diagram of a vehicle diagnostic software startup system according to an embodiment of this application is shown.

[0042] Explanation of key component symbols:

[0043] 200 - Vehicle diagnostic software startup system; 210 - Voice recognition module; 220 - Entity recognition module; 230 - Knowledge graph analysis module; 240 - Software startup module. Detailed Implementation

[0044] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0045] The components of the embodiments of this application described and illustrated in the accompanying drawings can be arranged and designed in a variety of different configurations. Therefore, the following detailed description of the embodiments of this application provided in the drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0046] In the following text, the terms "comprising," "having," and their cognates, which may be used in various embodiments of this application, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more combinations thereof. Furthermore, the terms "first," "second," "third," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance.

[0047] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of this application pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be construed as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of this application.

[0048] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0049] The following describes the method for starting the vehicle diagnostic equipment using specific embodiments.

[0050] Figure 1 A flowchart illustrating a vehicle diagnostic device startup method according to an embodiment of this application is shown. Exemplarily, the vehicle diagnostic device startup method includes the following steps:

[0051] S110 acquires the user's voice command to start the vehicle diagnostic software and converts the voice command into speech text.

[0052] In one embodiment, this application uses an acoustic model to identify whether the user-input speech text is a preset wake word; if the preset wake word is identified, voiceprint recognition is activated to extract the user's voiceprint features, and the user's identity is verified based on the comparison results between the voiceprint features and a pre-stored voiceprint database. In this embodiment, this application uses a CNN (Convolutional Neural Network) + GRU (Gated Recurrent Unit) acoustic model to identify whether the user-input wake word is a preset wake word, and supports custom preset wake words. The CNN is used to extract local speech features of the user-input speech using multiple convolutional kernels, and the GRU is used to receive the local speech features output by the CNN, model the temporal dependencies of the speech signal, and output whether it is a wake word (binary classification) or belongs to which wake word (multi-class classification).

[0053] This application uses traditional acoustic features or deep learning features to extract the voiceprint vector of the user's input speech. For example, i-vector (Identity Vector) is a low-dimensional speaker feature representation method that maps speech segments to a fixed-length voiceprint vector to represent the speaker's identity information. Upon first use, users need to record a speech segment to construct their pre-stored voiceprint template. This template is stored as an embedded vector in a pre-stored voiceprint library, supporting multiple user registrations, and each user can set a personalized wake-up word. The voiceprint vector extracted from the real-time speech stream is compared with the pre-stored voiceprint template using cosine similarity; if a match is successful and the wake-up word is correct, a personalized wake-up is triggered; otherwise, it is ignored or an insufficient permission message is displayed. Introducing voiceprint recognition to identify user identity improves system intelligence, security, and user experience.

[0054] In another embodiment, upon successful user authentication, the voice command to activate the vehicle diagnostic software is converted into speech-to-text. This application uses ASR (Automatic Speech Recognition) to convert the user-inputted voice command to wake up the vehicle diagnostic software into speech-to-text, supporting recognition of multiple languages ​​such as Chinese, English, and Spanish.

[0055] S120 performs entity recognition on the speech text to obtain the target entities in the speech text.

[0056] In one embodiment, this application uses a pre-trained multilingual understanding model to perform entity recognition, knowledge graph analysis, and structured instruction generation tasks. For example... Figure 2 As shown, the pre-training process of the multilingual understanding model includes:

[0057] S121, perform data augmentation processing on the collected multilingual data.

[0058] By way of example, this application uses multilingual Wikipedia data as multilingual data for model training. Multilingual Wikipedia data refers to language data from multiple different language versions on the Wikipedia platform, which is the data source for building a multilingual understanding model.

[0059] In one embodiment, the data augmentation process of this application includes: Synonym Replacement, constructing a cross-lingual thesaurus containing variations of expressions in different languages, and randomly replacing keywords in multilingual data during training; for example, the original sentence: "Open the diagnostic system of a certain brand and model," replaced with: "Start the diagnostic tool of a certain brand and model." The data augmentation process of this application also includes: randomly replacing vehicle model entities in multilingual data, limiting the replacement scope to the same brand, enhancing the model's understanding of brand-vehicle relationships, avoiding model dependence on a single vehicle model, and improving robustness. Semantic category replacement is performed on vehicle diagnostic software; the replacement rules are based on the functional similarity of the vehicle diagnostic software, improving the model's ability to recognize software intent categories, promoting intent generalization learning, and improving intent recognition accuracy.

[0060] S122, Based on the multilingual data after data augmentation, perform a model training task on the neural network model to obtain a multilingual model after the training task is performed; the model training task includes a masked language training task and a language prediction task.

[0061] In this embodiment, the neural network model can be a Transformer architecture. Based on the Transformer architecture, this application constructs a unified multilingual semantic space to achieve cross-language knowledge transfer and generalized understanding.

[0062] In one embodiment, this application trains the model to predict the masked words by randomly masking approximately 15% of the tokens in the input text using a masked language training task. The training objective of the masked language training task is to learn contextual semantic information, enabling unsupervised training and improving the model's ability to understand multilingual text. This application also trains the model to understand the logical relationships between contexts through a language prediction task (Next Sentence Prediction). The language prediction task involves selecting two consecutive sentences A and B from multilingual data, and randomly selecting another sentence C from multilingual data. The model uses a classifier to determine the relationship between sentence pairs, enabling the model to learn the contextual logical relationship between two consecutive sentences. The model concatenates A and B, and B and C, into input sequences, adding special markers before and after each sequence. For example, the concatenated sequence of A and B is: [CLS] + A + [SEP] + B + [SEP], where: [CLS] is the classification label, used by the classifier to extract the overall semantic representation of A and B; [SEP] is the separator label, distinguishing between the two sentences; the two concatenated sequences are passed through a fully connected layer (usually a linear classifier) ​​to output the probabilities of the two categories, including continuous sentences and random sentences. This application uses the cross-entropy loss function for training, and the optimization goal is to enable the model to correctly distinguish between continuous sentences and random sentences.

[0063] S123, construct multiple independent task heads for the multilingual model after the training task is performed, and jointly optimize each independent task head based on the multi-task loss function.

[0064] In this embodiment, multiple independent task heads refer to the simultaneous execution of multiple independent tasks within a multilingual model. This sharing of underlying representations enhances the overall generalization ability and learning efficiency of the model. These multiple independent task heads include an entity recognition task, a knowledge graph analysis task, and a structured instruction generation task. The entity recognition task identifies target entities in the speech text; the knowledge graph analysis task determines the relationships between target entities and identifies the target vehicle diagnostic software corresponding to each target entity in the knowledge graph; and the structured instruction generation task generates structured instructions corresponding to the target vehicle diagnostic software.

[0065] In one embodiment, this application employs a multi-task learning (MTL) architecture to simultaneously optimize multiple independent tasks, and to jointly optimize multiple independent tasks through a weighted loss function, wherein the expression for the weighted loss function is:

[0066]

[0067] in, The joint loss represents the total loss that the multilingual understanding model aims to minimize during training. The joint loss takes into account the losses of the three independent tasks. This represents the entity recognition loss, used to measure whether the entity recognition task accurately identifies the target entity; This represents the knowledge graph analysis loss, which measures the prediction error of the knowledge graph analysis task in determining the relationship between two target entities and identifying the target vehicle diagnostic software corresponding to the target entity. The structured instruction generation loss is used to measure the error in generating the structured instructions corresponding to the target vehicle diagnostic software. The entity recognition loss, knowledge graph analysis loss, and structured instruction generation loss can be calculated using the cross-entropy loss, which is not limited here. This indicates the importance of the entity recognition task in the total loss function. This indicates the importance of knowledge graph analysis loss in the total loss function. This indicates the importance of the structured instruction generation loss in the total loss function. For example, It can be set to 0.4. It can be set to 0.3. It can be set to 0.3.

[0068] In one embodiment, the target entities include the target brand and the target vehicle model. This application performs semantic parsing on the speech text through the shared underlying layer of a pre-trained multilingual understanding model to obtain the semantic vector of the speech text; and performs entity recognition on the semantic vector through the entity recognition head of the multilingual understanding model to obtain the target brand and the target vehicle model in the speech text.

[0069] In one implementation, the shared bottom layer is an encoding layer that acts as a feature extractor, providing shared semantic vectors for entity recognition and relation classification. This encoding layer uses a shared vocabulary to map similar semantics between speech texts from different languages ​​to similar vector spaces. A shared vocabulary refers to sharing a unified lexical representation space across multiple languages. Unlike methods where each language constructs its own vocabulary, a shared vocabulary encodes all lexical units from all languages ​​into a single vocabulary. This allows identical or similar semantic words in different languages ​​to be encoded using the same lexical unit vectors, facilitating transfer learning in unseen language tasks.

[0070] In one embodiment, the entity recognition head of the multilingual understanding model classifies each word and determines whether it belongs to a certain entity category (such as brand, car model, etc.). The entity recognition head maps semantic vectors to entity category labels through fully connected layers. Fully connected layers are one of the fundamental layers in neural networks. They linearly transform the input vector through a weight matrix and a bias vector, and the output label space represents the probability distribution of the word belonging to different entity category labels. The category with the highest probability is taken as the category of the word to identify target brand and car model entities. The entity category labels include brand and car model.

[0071] S130, based on knowledge graph analysis, analyzes the relationships between target entities and determines the target vehicle diagnostic software corresponding to the target entities in the knowledge graph. The knowledge graph includes the relationships between brands, models and vehicle diagnostic software in different languages.

[0072] In one embodiment, such as Figure 3 As shown, the knowledge graph construction process includes:

[0073] S131 collects multilingual brand, model, and vehicle diagnostic software, and performs data cleaning on the multilingual brand, model, and vehicle diagnostic software.

[0074] In one embodiment, this application can use a web crawling framework (such as Scrapy or Selenium) to crawl multilingual web pages, obtain multilingual brand, vehicle model, and vehicle diagnostic software from the multilingual web pages, and obtain vehicle diagnostic software usage records from automotive forums / communities, obtaining multilingual brand, vehicle model, and vehicle diagnostic software from the vehicle diagnostic software usage records. This application can also collect software call logs output by the vehicle diagnostic interface, obtaining multilingual brand, vehicle model, and vehicle diagnostic software from the software call logs.

[0075] In one embodiment, this application uses the SimHash algorithm (Similarity Hashing Algorithm) to detect duplicate text, which can be used to quickly identify duplicate or similar content, such as duplicate brand, vehicle model, and vehicle diagnostic software names in multiple forums, logs, and official websites. This application deletes the detected duplicate brand, vehicle model, and vehicle diagnostic software names, retaining only one. Furthermore, this application unifies the cleaned brand, vehicle model, and software information into a structured format to facilitate subsequent multilingual alignment, knowledge graph construction, and other operations.

[0076] S132, construct a multilingual mapping table for each brand, model, and vehicle diagnostic software, and build a knowledge graph based on the multilingual mapping table.

[0077] In one embodiment, the multilingual mapping table adopts a nested JSON (JavaScript Object Notation) structure, constructing entity mapping relationships around the language dimension. Its core structure consists of three nested layers: Brand, storing the multilingual mapping of car brands; Vehicle Model, recording the multilingual mapping of specific vehicle models; and Vehicle Diagnostic Software, containing the multilingual mapping of the in-vehicle software system. Each entity item contains a multilingual dictionary, supporting unlimited expansion of language types. Type-safe expansion is achieved through JSON Schema validation; adding a new language only requires appending key-value pairs to the corresponding entity dictionary, without modifying the overall structure.

[0078] In this embodiment, a knowledge graph is a structured semantic network used to represent entities (brands, vehicle models, and vehicle diagnostic software) and the relationships between them. Nodes in the knowledge graph represent entities and record various languages ​​associated with each entity; edges represent relationships between entities, such as production, adaptation, and belonging. Attributes of the knowledge graph describe the metadata of nodes or edges, such as the year the brand was established, the vehicle model's powertrain, and the software version.

[0079] In one embodiment, this application analyzes the relationships between target entities in a knowledge graph using a graph attention network (GAT) of a multilingual understanding model to generate relationship labels between target vehicle models and target brands; the relationship labels between target entities represent the attribution relationship between the target vehicle model and the target brand; and under the condition that the target vehicle model belongs to the target brand, the target vehicle diagnostic software corresponding to the target vehicle model and target brand in the knowledge graph is determined by the graph attention network.

[0080] In this embodiment, the graph-based attention network is used to dynamically aggregate node information in the knowledge graph through the attention mechanism. It can handle different types of nodes such as brand, vehicle model, and software, and can be used to determine whether the target vehicle model belongs to the target brand and to determine the target vehicle diagnostic software corresponding to the target vehicle model and target brand.

[0081] S140, Generate structured instructions corresponding to the target vehicle diagnostic software, so as to start the target vehicle diagnostic software through the structured instructions.

[0082] In one embodiment, this application uses a decoder in a multilingual understanding model to generate structured instructions corresponding to the target vehicle diagnostic software from the semantic vector of the speech text and the target vehicle diagnostic software. In this embodiment, the structured instruction generation task of the multilingual understanding model generates the structured instructions for the target vehicle diagnostic software through a decoder. This decoder can be a Transformer decoder, which fuses the semantic vector of the speech instructions with the target vehicle diagnostic software. Based on the fused semantic representation, it generates structured instructions corresponding to the target vehicle diagnostic software. These structured instructions are used to call the vehicle system interface, switch the software interface language, and perform access control, allowing the vehicle system to call and launch the target vehicle diagnostic software.

[0083] In one implementation, the semantic vector is used as the encoder output, and the target vehicle diagnostic software serves as the input context information for the decoder. The Transformer decoder uses the semantic vector as a global guide through a self-attention mechanism to ensure that the generated content aligns with the user's intent. Figure 1For example, if the semantic vector contains the instruction to activate vehicle diagnostic software, then {intent: activate_software} is generated. Furthermore, a self-attention mechanism is used to fuse the semantic vector with contextual information to generate structured fields. This supports JSON format output of structured instructions of arbitrary length. These structured instructions represent semantic intent information parsed from voice commands that can be used to perform operations. The structured instructions include an intent identifier, a target software identifier, and a language identifier. For example, the output format of this structured instruction could be: {intent: activate_software, software_id: UUID, lang: zh}, where intent represents the user's operational intent, activate_software indicates activating vehicle diagnostic software, software_id represents a unique software identifier, UUID represents a universally unique identifier for a specific vehicle diagnostic software, lang represents the language identifier used to identify the language version of the current voice command, and zh represents the Chinese language identifier.

[0084] Figure 4 A schematic diagram of a vehicle diagnostic software startup system 200 according to an embodiment of this application is shown. Exemplarily, the vehicle diagnostic software startup system 200 includes:

[0085] The voice recognition module 210 is used to acquire the user's voice command to start the vehicle diagnostic software and to convert the voice command into speech text.

[0086] The entity recognition module 220 is used to perform entity recognition on the speech text in order to obtain the target entity in the speech text.

[0087] The knowledge graph analysis module 230 is used to analyze the relationships between target entities based on the knowledge graph, and to determine the target vehicle diagnostic software corresponding to the target entities in the knowledge graph. The knowledge graph includes the relationships between brands, models and vehicle diagnostic software in different languages.

[0088] The software startup module 240 is used to generate structured instructions corresponding to the target vehicle diagnostic software, so as to start the target vehicle diagnostic software through the structured instructions.

[0089] It is understood that the system in this embodiment corresponds to the vehicle diagnostic software startup method in the above embodiment, and the options in the above embodiment are also applicable to this embodiment, so they will not be described again here.

[0090] This application also provides a vehicle diagnostic device, exemplary of which includes a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program to enable the vehicle diagnostic device to perform the functions of the various modules in the above-described vehicle diagnostic software startup method or the above-described vehicle diagnostic software startup system.

[0091] The processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including at least one of a Central Processing Unit (CPU), Graphics Processing Unit (GPU), Network Processor (NP), Digital Signal Processor (DSP), Application-Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the embodiments of this application.

[0092] Memory can be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), and Electrically Erasable Programmable Read-Only Memory (EEPROM). Memory is used to store computer programs, and the processor can execute these programs upon receiving execution instructions.

[0093] This application also provides a computer-readable storage medium for storing the computer program used in the aforementioned vehicle diagnostic equipment. For example, the computer-readable storage medium may include, but is not limited to, various media capable of storing program code, such as a USB flash drive, a portable hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0094] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that, as an alternative implementation, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0095] In addition, the functional modules or units in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0096] If a function is implemented as a software module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a smartphone, personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application.

[0097] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A vehicle diagnostic software startup method characterized by, The method comprises: acquiring a voice instruction for starting a vehicle diagnosis software input by a user, and converting the voice instruction into voice text; performing an entity recognition task, a knowledge graph analysis task and a structured instruction generation task through a pre-trained multi-language understanding model to start a target vehicle diagnosis software; wherein the entity recognition task comprises performing entity recognition on the voice text to acquire target entities in the voice text; the target entities comprise a target brand and a target vehicle model; the knowledge graph analysis task comprises analyzing the association relationship between the target entities based on a knowledge graph, and determining a target vehicle diagnosis software corresponding to the target entities in the knowledge graph; the knowledge graph comprises the association relationship between brands, vehicle models and vehicle diagnosis softwares in different languages; the structured instruction generation task comprises generating a structured instruction corresponding to the target vehicle diagnosis software to start the target vehicle diagnosis software through the structured instruction; the structured instruction is used to call a vehicle-mounted system interface and switch a software interface language, and is used to call the target vehicle diagnosis software by the vehicle-mounted system; wherein the knowledge graph analysis task comprises: analyzing the association relationship between the target entities in the knowledge graph through a graph self-attention network of the multi-language understanding model, and generating a relationship label of the target vehicle model and the target brand; the relationship label between the target entities is the ownership relationship between the target vehicle model and the target brand; under the condition that the target vehicle model belongs to the target brand, determining the target vehicle diagnosis software corresponding to the target vehicle model and the target brand in the knowledge graph through the graph self-attention network.

2. The vehicle diagnostic software launch method of claim 1, wherein, The entity recognition on the voice text to acquire the target entities in the voice text comprises: performing semantic analysis on the voice text through a shared bottom layer of a pre-trained multi-language understanding model to acquire a semantic vector of the voice text; performing entity recognition on the semantic vector through an entity recognition head of the multi-language understanding model to acquire the target brand and the target vehicle model in the voice text.

3. The vehicle diagnostic software launch method of claim 2, wherein, The generation of the structured instruction corresponding to the target vehicle diagnosis software comprises: generating the structured instruction corresponding to the target vehicle diagnosis software through a decoder of the multi-language understanding model based on the semantic vector of the voice text and the target vehicle diagnosis software.

4. The vehicle diagnostic software launch method of claim 2, wherein, The pre-training process of the multi-language understanding model comprises: performing data enhancement processing on collected multi-language data; performing a model training task on a neural network model based on the multi-language data after the data enhancement processing to obtain a multi-language model after the training task; the model training task comprises a mask language training task and a language prediction task; constructing a plurality of independent task heads of the multi-language model after the training task, and jointly optimizing each independent task head based on a multi-task loss function.

5. The vehicle diagnostic software launch method of claim 1, wherein, The construction process of the knowledge graph comprises: Collecting multi-language brands, vehicle models and vehicle diagnostic software, and performing data cleaning on the collected multi-language brands, vehicle models and vehicle diagnostic software; Building a multi-language mapping table for each brand, vehicle model and vehicle diagnostic software, and building the knowledge graph based on the multi-language mapping table.

6. The vehicle diagnostic software launch method of claim 1, wherein, Further comprising: Identifying whether the voice text input by the user is a preset wake-up word through an acoustic model; Activating voiceprint recognition under the condition of identifying the preset wake-up word, extracting the voiceprint features of the user, and verifying the identity of the user according to the comparison result of the voiceprint features and the pre-stored voiceprint library; Under the condition that the user identity verification is successful, the voice instruction for starting the vehicle diagnostic software is converted into the voice text.

7. A vehicle diagnostic software launch system, characterized by, Comprise: A voice recognition module for obtaining a voice instruction input by a user for starting a vehicle diagnostic software, and converting the voice instruction into a voice text; Performing an entity recognition task, a knowledge graph analysis task and a structured instruction generation task through a pre-trained multi-language understanding model to start a target vehicle diagnostic software; The entity recognition task includes an entity recognition module for performing entity recognition on the voice text to obtain target entities in the voice text; the target entities include target brands and target vehicle models; The knowledge graph analysis task includes a knowledge graph analysis module for analyzing the association between the target entities based on a knowledge graph, and determining a target vehicle diagnostic software corresponding to the target entities in the knowledge graph, the knowledge graph including the association between brands, vehicle models and vehicle diagnostic software in different languages; The structured instruction generation task includes a software starting module for generating a structured instruction corresponding to the target vehicle diagnostic software to start the target vehicle diagnostic software through the structured instruction; the structured instruction is used to call a vehicle-mounted system interface and switch software interface language for the vehicle-mounted system to call and start the target vehicle diagnostic software; The knowledge graph analysis module for analyzing the association between the target entities based on a knowledge graph, and determining a target vehicle diagnostic software corresponding to the target entities in the knowledge graph, includes: Analyzing the association between the target entities in the knowledge graph through a graph self-attention network of the multi-language understanding model, and generating a relationship label of the target vehicle model and the target brand; the relationship label between the target entities is the ownership relationship between the target vehicle model and the target brand; Under the condition that the target vehicle model belongs to the target brand, the graph self-attention network determines the target vehicle diagnostic software corresponding to the target vehicle model and the target brand in the knowledge graph.

8. A vehicle diagnosis device characterized by comprising: The vehicle diagnostic device comprises a processor and a memory, the memory stores a computer program, and the processor is configured to execute the computer program to implement the vehicle diagnostic software starting method of any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, It stores a computer program, which, when executed on a processor, implements the vehicle diagnostic software starting method according to any one of claims 1-6.

Images