Vehicle-mounted human-computer interaction method and system based on vector retrieval, computer equipment and storage medium

Abstract

The invention discloses a vehicle-mounted human-computer interaction method and system based on vector retrieval, computer equipment and a storage medium, efficient semantic retrieval is achieved by converting application function guide document content into word embedding vectors and storing the word embedding vectors into a vector database, when a user initiates an inquiry, a cloud side quickly finds out related documents through vector retrieval, and the user experience is improved. And the application end has an execution link framework, and can identify the current interface display content to execute a related action link. According to the method, the retrieval accuracy and efficiency are improved, and the man-machine interaction system can better understand and respond to the query intention of the user, so that the user experience is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle networks, and in particular relates to a vehicle-mounted human-computer interaction method, system, computer equipment and storage medium based on vector retrieval. Background Art

[0002] As the functions of the Internet of Vehicles become more and more complex, the number of functions that can be controlled by the car computer will undoubtedly increase the user's learning cost. When users want to find certain settings, it may take a long time. Although most of the current function settings can be opened directly by voice, users cannot observe the specific execution steps. The next time they want to find it manually, they may still not be able to accurately locate it. Traditional novice guidance is often a preset fixed process, while the vector retrieval-based system can provide personalized guidance based on the user's specific situation and needs. For example, the system can dynamically adjust the guidance content based on the user's query history, behavior patterns, etc. to ensure the relevance and practicality of the information.

[0003] In the prior art, the storage and retrieval of document content mainly rely on traditional database systems, such as relational databases and search engines. These systems perform well in processing structured data, but have limitations when processing unstructured or semi-structured text data. For example, keyword search cannot understand the semantics of the query well, resulting in a large number of irrelevant results or results that omit relevant information. In addition, as the amount of data increases, the retrieval efficiency and accuracy of traditional databases often cannot meet real-time requirements. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, the present invention provides a vehicle-mounted human-computer interaction method, system, computer device and storage medium based on vector retrieval, which realizes efficient semantic retrieval by converting the document content guided by the application function into a word embedding vector and storing it in a vector database. When a user initiates an inquiry, the cloud uses vector retrieval to quickly find relevant documents and returns the results to the application end to execute the corresponding action. The application end has an execution link framework that can identify the current interface display content to execute related action links. The present invention not only improves the accuracy and efficiency of retrieval, but also enables the human-computer interaction system to better understand and respond to the user's query intention, thereby improving the user experience.

[0005] The objective of the present invention is achieved through the following technical solutions:

[0006] In a first aspect, the present invention provides a vehicle-mounted human-computer interaction method based on vector retrieval, comprising the following steps:

[0007] S1. Create a vector database, convert the "User Guide Document Manual" into a word embedding vector and store it in the vector database;

[0008] S2. The vehicle application performs semantic analysis on the user's voice input and sends the analyzed voice input text content to the cloud server;

[0009] S3. The cloud server converts the voice input content into vector form, searches in the vector database, and returns the search results to the vehicle application end:

[0010] S4. The vehicle application executes the corresponding action link according to the search results returned by the cloud server:

[0011] S41. The vehicle application terminal is pre-loaded with a large number of action resources. After the vehicle application terminal receives the search results, it matches and loads a list of action sets that match the search results in the action resources according to the search results;

[0012] S42. The vehicle application loops through all action instructions in the action set list matched in step S41:

[0013] S421. Take out an action instruction from the action set list each time to execute;

[0014] S422. When each action is executed, the result is verified according to the expected effect of the execution interface of the action. If the final interface result does not match the expected effect of the action, the action is considered to have failed. In the action set list, all action chains following this action will not be executed further. If the final interface result matches the expected effect of the action, return to step S421 to obtain the next action instruction for execution.

[0015] S423. Finally, if all the action instructions in the action set list are executed successfully, it means that the function inquired by the user is executed successfully and the guidance ends.

[0016] Furthermore, the step S1 comprises:

[0017] S11. Create a vector database designed for high-dimensional data in a cloud server;

[0018] S12. Use natural language processing technology to convert the document content of the "User Guide Document Manual" into word embedding vectors, each vector representing the semantic information of a word or phrase;

[0019] S13. Store the word embedding vector converted in step S12 in the created vector database.

[0020] Furthermore, in the step S1, the bge-large-zh semantic vector model, which is open source from a third party, is used to perform word embedding vector conversion; and the vector database uses the Faiss database, which is open source from a third party.

[0021] Furthermore, the step S3 comprises:

[0022] S31. After receiving the request, the server converts the user's voice input content into a vector form to generate a query vector;

[0023] S32. Perform similarity search in the vector database according to the query vector to obtain the document with the highest similarity as the query result;

[0024] S33. The cloud server returns the search results to the vehicle application via a network request.

[0025] Furthermore, in step S32, for a given query vector, the Faiss database calculates the similarity between the query vector and all data vectors through the index structure, retrieves the most similar vector, and returns the most similar document as the query result.

[0026] Furthermore, the matching rule of step S41 is: each action instruction has a key header named action, and if its value is consistent with the value returned in the final search result of the cloud server, it is added to the action set list, and finally an action chain is formed.

[0027] In a second aspect, the present invention provides a vehicle-mounted human-computer interaction system based on vector retrieval, which is used to implement the vehicle-mounted human-computer interaction method, including a cloud server and a vehicle-mounted computer application end, wherein the cloud server and the vehicle-mounted computer application end are connected via network communication;

[0028] The cloud server includes:

[0029] A vector database, which is used to store word embedding vectors, each of which represents the semantic information of a word or phrase;

[0030] The vector conversion module is used to convert text content into word embedding vectors through the bge-large-zh semantic vector model;

[0031] A vector retrieval module is used to search for the document result with the highest similarity in the vector database according to the retrieval vector;

[0032] The vehicle machine application end includes:

[0033] The semantic analysis module is used to parse the user's voice input command into voice input content text and upload it to the cloud server;

[0034] The action link execution module is used to match the action set according to the search results sent by the cloud server, and traverse and execute all action instructions in the action set.

[0035] Furthermore, the action link execution module includes:

[0036] An action resource database, which stores a large number of action instructions;

[0037] An action set matching unit, used to match action instructions in an action resource database according to the search results and generate an action set;

[0038] An action traversal unit, used for traversing and executing all action instructions in the generated action set;

[0039] An action execution setting unit, used to set a delay for the execution of each action or an expected interface effect;

[0040] The action execution verification unit is used to verify the result according to the expected effect of the set action execution interface.

[0041] In a third aspect, the present invention further provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a vehicle-mounted human-computer interaction method based on vector retrieval as described in the present invention is implemented.

[0042] In a fourth aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a vehicle-mounted human-computer interaction method based on vector retrieval as described in the present invention.

[0043] The present invention has the following advantages:

[0044] The present invention provides a vehicle-mounted human-computer interaction method and system, whereby a user can ask questions and obtain instant feedback. Such interactive learning can arouse the user's interest and sense of participation better than one-way information transmission, thereby improving learning efficiency.

[0045] The present invention uses natural language processing technology to convert the content of the application function guidance document into word embedding vectors, each of which represents the semantic information of a word or phrase. Therefore, users can interact with the system using natural language without having to adapt to the system's instructions or keywords. This approach reduces the user's learning cost and makes the guidance process more intuitive and friendly.

[0046] The present invention constructs a vector database designed specifically for high-dimensional data, and stores the converted word embedding vectors in the vector database for efficient similarity calculation and retrieval.

[0047] The present invention adopts a cloud retrieval mechanism. When a user initiates a query, the query content is converted into a vector form, and the most similar vector set, that is, the relevant document content, is searched in the vector database.

[0048] The application end of the present invention executes corresponding action links according to the vector search results, such as opening a certain interface, playing related videos, etc. Users do not need to search for answers through complex menus or help documents, but only need to ask questions to get guidance, which reduces the user's operating burden and improves the user experience.

[0049] The present invention has scalability, and as the business develops, the knowledge base and functions of the interactive system can be easily expanded without the need to reconstruct the entire guidance system. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings used in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the contents of the embodiments of the present invention and these drawings without creative work.

[0051] Figure 1 This is a flow chart of a vehicle-mounted human-computer interaction method based on vector retrieval as described in Example 1 of the present invention;

[0052] Figure 2 This is a schematic diagram of the composition of a vehicle-mounted human-computer interaction system based on vector retrieval as described in Example 2 of the present invention;

[0053] Figure 3 This is a schematic diagram of the structure of a computer device described in Example 3 of the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] Example 1

[0056] like Figure 1 As shown, this embodiment is a vehicle-mounted human-computer interaction method based on vector retrieval, comprising the following steps:

[0057] S1. Create a vector database, convert the "User Guide Document Manual" into a word embedding vector and store it in the vector database:

[0058] S11. Create a vector database designed for high-dimensional data in a cloud server.

[0059] S12. Use natural language processing technology to convert the document content of the "User Guide Document Manual" into word embedding vectors, each vector represents the semantic information of a word or phrase. In this embodiment, the bge-large-zh semantic vector model open sourced by a third party is used for word embedding vector conversion. The principle of the bge-large-zh semantic vector model mainly involves converting text into high-dimensional vector representations. These vectors capture the semantic information in the text, making semantically similar texts closer in the vector space. This representation method facilitates subsequent similarity searches and clustering.

[0060] S13. The word embedding vector converted in step S12 is stored in the created vector database. In this embodiment, the third-party open source Faiss database is used to perform efficient similarity calculation and retrieval.

[0061] S2. The vehicle computer application performs semantic analysis on the user's voice input and sends the analyzed content to the cloud server:

[0062] S21. The vehicle application obtains the user's voice input, performs semantic analysis on the voice input, and obtains the content for the voice input;

[0063] S22. The vehicle computer application sends the acquired voice input content to the cloud server via an http network request.

[0064] S3. The cloud server converts the voice input content into vector form, searches in the vector database, and returns the search results to the vehicle application end:

[0065] S31. After receiving the request, the server converts the user's voice input content into a vector form to generate a query vector;

[0066] S32. Perform similarity search in the vector database according to the query vector, and obtain the document with the highest similarity as the query result. In this embodiment, for a given query vector, the Faiss database quickly retrieves the most similar vector through the index structure. It can perform nearest neighbor search (NN) or range search, support precise search and approximate search algorithms, and the precise search calculates the similarity between the query vector and all data vectors, and returns the most similar result;

[0067] S33. The cloud server returns the search results to the vehicle application via a network request.

[0068] S4. The vehicle application executes the corresponding action link according to the search results returned by the cloud server:

[0069] S41. The vehicle application end is preloaded with a large number of action resources. After receiving the search results, the vehicle application end matches and loads the action instructions matching the search results in the action resources according to the search results, and generates an action set list. The matching rule is: each action has a key header named action. If its value is consistent with the value returned in the final search result of the server, it is added to the action set, and finally forms an action chain.

[0070] S42. The vehicle application loops through all action instructions in the action set list matched in step S41:

[0071] S421. Take out one action instruction from the action set list each time for execution.

[0072] The actions include pulling up the application interface, clicking or sliding, etc.; pulling up the interface can call the Android native startActivity() method, and the clicking or sliding actions can be implemented using the Android accessibility service, which can automatically realize page switching changes without the user manually clicking.

[0073] Before each action is executed, you can specify the delay in milliseconds before the action starts, as well as the expected interface effect when the action is completed.

[0074] During the execution of each action, the accessibility service function of the Android system is used to constantly observe changes in the vehicle computer interface.

[0075] S422. When each action is completed, the result can be checked according to the final expected effect of the action. If the final interface result does not match the expected effect of the action, the action is considered to have failed. In the action set list, all action chains following this action will not be executed. At the same time, voice TTS broadcast will be used to prompt the user of the execution failure. If the final interface result matches the expected effect of the action, return to step S421 to obtain the next action for execution.

[0076] S423. Finally, if all actions in the action set are executed successfully, it means that the function inquired by the user is executed successfully and the guidance ends.

[0077] The following is a specific example to illustrate the method provided by the present invention. Assuming that the user asks "How to buy traffic?", the system first converts the query into a vector representation, and then retrieves the most similar document vector in the vector database. Assuming that the retrieved document contains the steps and techniques for buying traffic, the system organizes this information into an answer and returns it to the application. The application side executes a series of action links based on the content, including but not limited to TTS broadcast and interface jump. The specific execution steps are:

[0078] User voice question: "How to purchase traffic?"

[0079] Step 1: The server converts the "User Guide Document Manual" into word embedding vectors through the bge-large-zh semantic vector model, which is open sourced by three parties. Each vector represents the semantic information of a word or phrase. The principle of the bge-large-zh semantic vector model mainly involves converting text into high-dimensional vector representations. These vectors capture the semantic information in the text, making semantically similar texts closer in the vector space. This representation method facilitates subsequent similarity search and clustering.

[0080] Step 2: Store the converted vectors in a vector database designed for high-dimensional data. This solution uses the third-party open source Faiss database to perform efficient similarity calculation and retrieval.

[0081] Step 3: The vehicle computer application obtains the content of the user's voice input, such as "how to purchase traffic".

[0082] Step 4: The vehicle application sends the acquired content to the server via an http network request.

[0083] Step 5: After receiving the request, the server searches the vector database for the user's voice input content.

[0084] Step 6: For a given query vector, the Faiss database quickly retrieves the most similar vectors through the index structure. It can perform nearest neighbor search (NN) or range search, and supports exact search and approximate search algorithms. Exact search calculates the similarity between the query vector and all data vectors and returns the most similar results.

[0085] Step 7: The server returns the search results to the vehicle application through a network request.

[0086] Step 8: The vehicle application loads and stores a large number of action resources.

[0087] Step 9: After receiving the search results, the vehicle application matches and loads the action set that matches the search results. Matching rules: Each action has a key header called action. If its value is consistent with the value returned in the final search result of the server, it is added to the action set, eventually forming an action chain.

[0088] Step 10: The car machine application loops through the action set, and takes out an action from the action set list each time to execute. (This action may be to pull up the application interface, or it may be a click or slide, etc.) Pulling up the interface can call the Android native startActivity() method, and the click or slide action can be implemented using the Android accessibility service. The page switching change can be automatically realized without the user manually clicking.

[0089] Step 11: Before each action is executed, you can specify the delay in milliseconds before the action is executed, as well as the expected interface effect when the action is completed.

[0090] Step 12: During the execution of each action, use the accessibility service function of the Android system to always observe the changes in the interface of the car computer.

[0091] Step 13: When each action is completed, the result can be verified according to the final expected effect of the action. If the final interface result does not match the expected effect of the action, the action is considered to have failed. Then, in the action set list, all action chains following this action will not be executed further, and the user will be prompted with voice TTS to indicate the failure. Otherwise, repeat step 10 to obtain the next action for execution.

[0092] Step 14: Finally, if all actions in the action set are executed successfully, it means that the function inquired by the user is executed successfully and the guidance ends.

[0093] Example 2

[0094] like Figure 2 As shown, this embodiment is a vehicle-mounted human-computer interaction system based on vector retrieval, including a cloud server and a vehicle-machine application end, and the cloud server and the vehicle-machine application end are connected through network communication;

[0095] The cloud server includes:

[0096] A vector database, which is used to store word embedding vectors, each of which represents the semantic information of a word or phrase;

[0097] The vector conversion module is used to convert text content into word embedding vectors through the bge-large-zh semantic vector model open sourced by a third party;

[0098] A vector retrieval module is used to search for the document result with the highest similarity in the vector database according to the retrieval vector;

[0099] The vehicle machine application end includes:

[0100] The semantic analysis module is used to parse the user's voice input command into voice input content text and upload it to the cloud server;

[0101] The action link execution module is used to match the action set according to the search results sent by the cloud server, and traverse and execute all action instructions in the action set.

[0102] The action link execution module includes:

[0103] An action resource database, which stores a large number of action instructions;

[0104] An action set matching unit, used to match action instructions in an action resource database according to the search results and generate an action set;

[0105] An action traversal unit, used for traversing and executing all action instructions in the generated action set;

[0106] An action execution setting unit, used to set a delay for the execution of each action or an expected interface effect;

[0107] The action execution verification unit is used to verify the result according to the expected effect of the set action execution interface.

[0108] Example 3

[0109] Figure 3 This is a schematic diagram of the structure of a computer device in Example 3 of the present invention. Figure 3 A block diagram of an exemplary computer device 12 suitable for use in implementing embodiments of the present invention is shown. Figure 3 The computer device 12 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0110] like Figure 3 As shown, the computer device 12 is in the form of a general-purpose computing device. The components of the computer device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and a bus 18 that connects various system components (including the system memory 28 and the processing unit 16).

[0111] Bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus architectures. By way of example, these architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MAC) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0112] The computer device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computer device 12, including volatile and non-volatile media, removable and non-removable media.

[0113] The system memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. The computer device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 34 may be used to read and write non-removable, non-volatile magnetic media ( Figure 3 not shown, usually called a "hard drive"). Although Figure 3 Not shown in the figure, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 18 via one or more data medium interfaces. The memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present invention.

[0114] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in the memory 28. Such program modules 42 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. The program modules 42 generally perform the functions and / or methods of the embodiments described herein.

[0115] The computer device 12 may also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), may also communicate with one or more devices that enable a user to interact with the computer device 12, and / or communicate with any device that enables the computer device 12 to communicate with one or more other computing devices (e.g., network card, modem, etc.). Such communication may be performed through an input / output (I / O) interface 22. In addition, the computer device 12 in this embodiment, the display 24 does not exist as an independent individual, but is embedded in the mirror surface, and when the display surface of the display 24 is not displayed, the display surface of the display 24 and the mirror surface are visually integrated. In addition, the computer device 12 may also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN) and / or public network, such as the Internet) through a network adapter 20. As shown in the figure, the network adapter 20 communicates with other modules of the computer device 12 through a bus 18. It should be understood that although not shown in the figure, other hardware and / or software modules can be used in conjunction with the computer device 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0116] The processing unit 16 executes various functional applications and data processing by running programs stored in the system memory 28, for example, implementing a vehicle-mounted human-computer interaction method based on vector retrieval provided in an embodiment of the present invention.

[0117] Example 4

[0118] Embodiment 4 of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, a vehicle-mounted human-computer interaction method based on vector retrieval is implemented as provided in all the embodiments of the invention of this application.

[0119] Any combination of one or more computer-readable media may be used. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, device, or device.

[0120] Computer-readable signal media may include a data signal propagated in baseband or as part of a carrier wave, which carries a computer-readable program code. Such propagated data signals may take a variety of forms, including, but not limited to, electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0121] Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0122] Computer program code for performing the operations of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0123] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A vehicle-mounted human-computer interaction method based on vector retrieval, characterized in that: The following steps are involved: S1. Create a vector database, convert the "User Guide Document Manual" into a word embedding vector and store it in the vector database; S2. The vehicle application performs semantic analysis on the user's voice input and sends the analyzed voice input text content to the cloud server; S3. The cloud server converts the voice input content into vector form, searches in the vector database, and returns the search results to the vehicle application end: S4. The vehicle application executes the corresponding action link according to the search results returned by the cloud server: S41. The vehicle application terminal is pre-loaded with a large number of action resources. After the vehicle application terminal receives the search results, it matches and loads a list of action sets that match the search results in the action resources according to the search results; S42. The vehicle application loops through all action instructions in the action set list matched in step S41: S421. Take out an action instruction from the action set list each time to execute; S422. When each action is executed, the result is verified according to the expected effect of the execution interface of the action. If the final interface result does not match the expected effect of the action, the action is considered to have failed. In the action set list, all action chains following this action will not be executed further. If the final interface result matches the expected effect of the action, return to step S421 to obtain the next action instruction for execution. S423. Finally, if all the action instructions in the action set list are executed successfully, it means that the function inquired by the user is executed successfully and the guidance ends.

2. The vehicle-mounted human-computer interaction method based on vector retrieval as claimed in claim 1, characterized in that: The step S1 comprises: S11. Create a vector database designed for high-dimensional data in a cloud server; S12. Use natural language processing technology to convert the document content of the "User Guide Document Manual" into word embedding vectors, each vector representing the semantic information of a word or phrase; S13. Store the word embedding vector converted in step S12 in the created vector database.

3. The vehicle-mounted human-computer interaction method based on vector retrieval as claimed in claim 2, characterized in that: In the step S1, the bge-large-zh semantic vector model provided by a third party is used to perform word embedding vector conversion; the vector database uses the Faiss database provided by a third party.

4. The vehicle-mounted human-computer interaction method based on vector retrieval as claimed in claim 1, characterized in that: The step S3 comprises: S31. After receiving the request, the server converts the user's voice input content into a vector form to generate a query vector; S32. Perform similarity search in the vector database according to the query vector to obtain the document with the highest similarity as the query result; S33. The cloud server returns the search results to the vehicle application via a network request.

5. The vehicle-mounted human-computer interaction method based on vector retrieval as claimed in claim 4, characterized in that: In step S32, for a given query vector, the Faiss database calculates the similarity between the query vector and all data vectors through the index structure, retrieves the most similar vector, and returns the most similar document as the query result.

6. The vehicle-mounted human-computer interaction method based on vector retrieval as claimed in claim 1, characterized in that: The matching rule of step S41 is: each action instruction has a key header named action, and if its value is consistent with the value returned in the final search result of the cloud server, it is added to the action set list, and finally an action chain is formed.

7. A vehicle-mounted human-computer interaction system based on vector retrieval, which is used to implement the vehicle-mounted human-computer interaction method according to any one of claims 1 to 6, characterized in that: It includes a cloud server and a vehicle computer application end, and the cloud server and the vehicle computer application end are connected via network communication; The cloud server includes: A vector database, which is used to store word embedding vectors, each of which represents the semantic information of a word or phrase; The vector conversion module is used to convert text content into word embedding vectors through the bge-large-zh semantic vector model; A vector retrieval module is used to search for the document result with the highest similarity in the vector database according to the retrieval vector; The vehicle machine application end includes: The semantic analysis module is used to parse the user's voice input command into voice input content text and upload it to the cloud server; The action link execution module is used to match the action set according to the search results sent by the cloud server, and traverse and execute all action instructions in the action set.

8. The vehicle-mounted human-computer interaction system based on vector retrieval as claimed in claim 7, characterized in that: The action link execution module includes: An action resource database, which stores a large number of action instructions; An action set matching unit, used to match action instructions in an action resource database according to the search results and generate an action set; An action traversal unit, used for traversing and executing all action instructions in the generated action set; An action execution setting unit, used to set a delay for the execution of each action or an expected interface effect; The action execution verification unit is used to verify the result according to the expected effect of the set action execution interface.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the in-vehicle human-computer interaction method as described in any one of claims 1 to 6 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the vehicle-mounted human-computer interaction method as described in any one of claims 1 to 6 is implemented.

Images