Vehicle-mounted space transformation method, device, equipment and medium

Abstract

The invention provides a vehicle-mounted space transformation method and device, equipment and a medium. The method comprises the following steps: the cloud server obtains a user input instruction through the vehicle machine, and obtains a user intention according to the user input instruction; the AI large model service of the cloud server obtains an SOA atomic item recommendation result according to the user intention; wherein the SOA atomic item recommendation result comprises recommendation working states of at least part of SOA atomic items in a cabin scene; and controlling a cabin execution module based on the SOA atomic item recommendation result. According to the method, the resources in the intelligent automobile operation system can be reasonably allocated, and the sensory requirements of the user can be accurately covered.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent automobile operating systems, and in particular to a vehicle-mounted space transformation method, device, equipment and medium. Background Art

[0002] The car computer refers to the in-vehicle information system installed in the intelligent car operating system. The car computer can functionally realize information communication between people and cars, and cars and the outside world (cars and cars); through the in-vehicle information system, the driver can control the realization of different functions in the vehicle cabin, including the scene layout of seats, windows, air conditioning, music and other functions.

[0003] In traditional scene layout, the in-vehicle information system uses the car computer voice and user content to understand and learn user intentions, and finally hands it over to the downstream service for scene layout recommendation based on traditional service algorithms, for example, scene layout recommendation based on Natural Language Understanding (NLU) algorithm; however, traditional scene layout has the problem of improper resource allocation in the intelligent car operating system, narrow coverage of recommended scenes, and failure to accurately cover user sensory needs.

[0004] Therefore, how to reasonably allocate resources in the smart car operating system and accurately cover the user's sensory needs when making scene layout recommendations is an urgent problem to be solved. Summary of the invention

[0005] The present application provides a vehicle-mounted space transformation method, device, equipment and medium to solve the problems in the prior art of improper resource allocation in the intelligent vehicle operating system, narrow recommended scene coverage, and inability to accurately cover user sensory needs.

[0006] In a first aspect, the present application provides a vehicle-mounted space transformation method, which is applied to a cloud server connected to an intelligent vehicle operating system and a vehicle computer of the intelligent vehicle operating system, and the method includes:

[0007] The cloud server obtains the user input command through the vehicle computer, and obtains the user intention according to the user input command;

[0008] The AI ​​big model service of the cloud server obtains the SOA atomic item recommendation result according to the user intention; wherein the SOA atomic item recommendation result includes the recommended working status of at least some SOA atomic items in the cockpit scene; the SOA atomic item includes the cockpit vehicle control service, or includes the cockpit vehicle control service and multimedia, the multimedia includes audio and images, and the images include static images and dynamic images;

[0009] Based on the SOA atomic item recommendation result, the cockpit execution module is controlled.

[0010] In a possible implementation, the cloud server obtains the user input instruction through the vehicle computer, and obtains the user intention according to the user input instruction, including:

[0011] The cloud server obtains the user input command through the vehicle computer and identifies the command type of the user input command; wherein the command type includes a voice input command and a text input command, the voice input command is obtained through the voice recognition program of the cloud server, and the text input command is obtained through the text query program of the cloud server;

[0012] The cloud server obtains the user intention through the instruction type and user historical input.

[0013] In a possible implementation, the AI ​​big model service of the cloud server obtains the SOA atomic item recommendation result according to the user intention, including:

[0014] The cloud server constructs a prompt project for the AI ​​big model service based on fine-tuning of the user's historical usage of the SOA atomic item;

[0015] The prompt project generates an intention text describing the user intention according to the user intention;

[0016] The cloud server generates the SOA atomic item recommendation result according to the intention text.

[0017] In a possible implementation, the cloud server generates the SOA atomic item recommendation result according to the intention text, including:

[0018] The cloud server performs pre-training, parameter optimization and posture adjustment on the intended text to obtain the image generation result in the SOA atomic item recommendation result; wherein the cloud server includes a program for performing the pre-training, the parameter optimization and the posture adjustment, and the parameter optimization is used to adjust the content in the intended text.

[0019] In a possible implementation, the controlling the cockpit execution module based on the SOA atomic item recommendation result includes:

[0020] If the SOA atomic item recommendation result includes an image result, the cloud server adjusts the size of the image result according to the size of the vehicle computer, and controls the vehicle computer to display the image result after the size adjustment through the cockpit execution module;

[0021] If the SOA atomic item recommendation result includes a non-image result, the cockpit execution module is controlled to achieve the non-image result; wherein the image result and the non-image result are recommended working states corresponding to different SOA atomic items.

[0022] In a possible implementation, the cloud server adjusts the size of the image result according to the size of the vehicle computer, including:

[0023] The prompt project generates an image describing the user's intention according to the user's intention, enlarges the image, and trims the size and quality of the enlarged image to obtain the image result after the size adjustment.

[0024] In one possible implementation, the AI ​​big model service is deployed in the cloud server configured with concurrency function.

[0025] In a second aspect, the present application provides a vehicle-mounted space transformation device, comprising:

[0026] An acquisition module, used for the cloud server to acquire a user input instruction through the vehicle computer, and to acquire a user intention according to the user input instruction;

[0027] A processing module, used for the AI ​​big model service of the cloud server to obtain the SOA atomic item recommendation result according to the user intention; wherein the SOA atomic item recommendation result includes the recommended working status of at least some SOA atomic items in the cockpit scene; the SOA atomic item includes the cockpit vehicle control service, or includes the cockpit vehicle control service and multimedia, the multimedia includes audio and image, and the image includes static image and dynamic image;

[0028] The output module is used to control the cockpit execution module based on the SOA atomic item recommendation result.

[0029] In a third aspect, the present application provides a vehicle-mounted space transformation device, including: at least one processor and a memory;

[0030] The memory stores computer-executable instructions;

[0031] The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the in-vehicle space transformation method as described above.

[0032] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the vehicle-mounted space transformation method as described above.

[0033] The present application provides a vehicle-mounted space transformation method, device, equipment and medium, which are applied to a cloud server connected to an intelligent vehicle operating system and a vehicle computer of the intelligent vehicle operating system. The method includes:

[0034] The cloud server obtains user input instructions through the vehicle computer, and obtains user intentions according to the user input instructions; the AI ​​big model service of the cloud server obtains SOA atomic item recommendation results according to the user intentions; wherein the SOA atomic item recommendation results include recommended working states of at least some SOA atomic items in the cockpit scene; the SOA atomic items include cockpit vehicle control services, or include cockpit vehicle control services and multimedia, the multimedia includes audio and images, and the images include static images and dynamic images; based on the SOA atomic item recommendation results, the cockpit execution module is controlled.

[0035] In the above method, SOA atomic items are various atomic capabilities in the cockpit scene, including air conditioning, etc.; the AI ​​large model service that can recommend SOA atomic items is deployed in the cloud server, and the vehicle computer is responsible for receiving user input instructions and transmitting the user input instructions to the cloud server, allowing the cloud server to infer the user's intention based on the user input instructions, and give SOA atomic item recommendations, so that each SOA atomic item has a recommended working state that satisfies the user's perception; the cloud server feeds back the inferred SOA atomic item recommendation results to the vehicle computer, allowing the vehicle computer to control each SOA atomic item in the cockpit scene through the cockpit execution module to reach the recommended working state, accurately covering the user's sensory needs, and reasonably utilizing the resources in the intelligent car operating system in the task allocation between the vehicle computer and the cloud server. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0037] Figure 1 A schematic diagram of a scene of vehicle-mounted space transformation provided in an embodiment of the present application;

[0038] Figure 2 A schematic diagram of a vehicle space transformation method provided in an embodiment of the present application Figure 1 ;

[0039] Figure 3 A module implementation diagram of a vehicle-mounted space transformation provided in an embodiment of the present application;

[0040] Figure 4A schematic diagram of a vehicle space transformation method provided in an embodiment of the present application Figure 2 ;

[0041] Figure 5 A schematic diagram of a vehicle space transformation method provided in an embodiment of the present application Figure 3 ;

[0042] Figure 6 A schematic diagram of size adjustment for vehicle-mounted space transformation provided in an embodiment of the present application;

[0043] Figure 7 A basic dependency diagram of a vehicle-mounted space transformation provided in an embodiment of the present application;

[0044] Figure 8 A diagram of a vehicle-mounted space transformation device provided by an embodiment of the present invention;

[0045] Fig. 9 A hardware schematic diagram of a vehicle-mounted space transformation device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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.

[0047] The car computer refers to the in-vehicle information system installed in the intelligent car operating system. The car computer can realize the information communication between people and cars, cars and the outside world (cars and cars). Through the in-vehicle information system, the driver can control the realization of different functions in the vehicle cabin, including the scene layout of seats, windows, air conditioning, music and other functions.

[0048] The different functions in the vehicle cockpit are atomic item functions of the service-oriented architecture (SOA). SOA atomic items are various atomic capabilities in the cockpit scene, namely, the air conditioning, seats, windows, ambient lights, massage and other functions of the smart cockpit, as well as the entertainment services of the car computer such as music, audio radio, video and other online service capabilities are all SOA atomic items.

[0049] In traditional scene layout, the in-vehicle information system uses the car computer voice and user content to understand and learn user intentions, and finally hands it over to the downstream service for scene layout recommendation based on traditional service algorithms, for example, scene layout recommendation based on Natural Language Understanding (NLU) algorithm; however, traditional scene layout has the problem of improper resource allocation in the intelligent car operating system, narrow coverage of recommended scenes, and failure to accurately cover user sensory needs.

[0050] Therefore, how to reasonably allocate resources in the smart car operating system and accurately cover the user's sensory needs when making scene layout recommendations is an urgent problem to be solved.

[0051] The following describes the implementation process of a vehicle-mounted space transformation method proposed in the present application in combination with the accompanying drawings and specific embodiments.

[0052] Figure 1 A schematic diagram of a vehicle-mounted space transformation scenario provided in an embodiment of the present application. Figure 1 As shown, the system includes: a cloud server connected to a smart car operating system, and a vehicle computer of the smart car operating system;

[0053] The user can interact with the vehicle computer, and the user can convey the intention to the vehicle computer by giving user input instructions; wherein the user input instructions include instruction types, for example, the instruction type is "I want to sleep for a while in the bamboo forest in the rain."

[0054] The car computer can interact with the cloud server. After obtaining the user input command, the car computer can pass the user input command to the cloud server, allowing the cloud server to understand the user's intention. The cloud server is equipped with an AI big model service, which can be used for SOA scheduling, that is, to infer the recommended working status of the SOA atomic item according to the user's intention, obtain the SOA atomic item recommendation result, and feed back the SOA atomic item recommendation result to the car computer, so that the car computer and the cloud server can interact to figure out how to schedule the cabin.

[0055] For example, the AI ​​big model service infers the SOA atomic items that users need to adjust based on keywords such as "rain", "bamboo forest", and "sleep", including air-conditioning temperature, audio, seat height, window switch status, image display on the car display screen, etc. The image display can be a bamboo forest with rain, and the audio can include slight rain sounds.

[0056] The vehicle computer receives the SOA atomic item recommendation results from the cloud server and can control each SOA atomic item in the cockpit scene to reach the recommended working state through the cockpit execution module, wherein the cockpit execution module may include a computing program that adjusts and changes the functions corresponding to the SOA atomic items according to the instructions.

[0057] Figure 2 A schematic diagram of a vehicle space transformation method provided in an embodiment of the present application Figure 1 .like Figure 2 As shown, the method includes:

[0058] S201. The cloud server obtains a user input instruction through the vehicle computer, and obtains a user intention according to the user input instruction.

[0059] This embodiment is applied to a cloud server connected to a smart car operating system and a car computer of a smart car operating system; a person can interact with the car computer of the smart car operating system to realize operations such as entertainment, communication and car control, and the interaction methods include buttons, touch screens, voice, etc.; the button-type interaction is generally set to relatively simple operations, such as clicks, switches, etc.; the touch screen-type interaction can add gesture recognition, text input and other operations on the basis of the button-type operation; the voice-type interaction allows users to control the car computer by chatting, and users can express their needs by voice, and the microphone service in the car computer can collect the needs in the user's voice;

[0060] User input commands are demand commands transmitted by users to the vehicle computer. For example, the vehicle computer needs to give a corresponding response to the information input by the user after understanding the user input information. The service responsible for understanding the user input information is set on the cloud server. After obtaining the user input command, the vehicle computer can send the user input command to the cloud server. The cloud server parses the user's intention according to the user input command, identifies the content expressed by the user, and can also combine the user's historical expression to organize the user's overall expression.

[0061] S202. The AI ​​big model service of the cloud server obtains the SOA atomic item recommendation result according to the user intention; wherein the SOA atomic item recommendation result includes the recommended working status of at least some SOA atomic items in the cockpit scene; the SOA atomic item includes the cockpit vehicle control service, or includes the cockpit vehicle control service and multimedia, the multimedia includes audio and images, and the images include static images and dynamic images.

[0062] Artificial Intelligence (AI) is a human-like intelligence that is realized by using technologies such as machine learning, natural language processing, computer vision, and deep learning. The realization of artificial intelligence requires the support of model algorithms, such as AI big model services. AI big model services include artificial intelligence generated content (AIGC) technology.

[0063] The AI ​​big model service is set up on the cloud server. After the cloud server understands the user's intention, it can infer the user's intention and generate corresponding content through the AI ​​big model service. The purpose of inferring the user's intention is to accurately understand the user's intention. For example, if the user needs a bamboo forest with rain, then the bamboo forest is more critical. Based on the key points understood by inferring the user's intention, corresponding content is generated. For example, an image with a bamboo forest is generated, and a rainy atmosphere can be created based on the bamboo forest.

[0064] There are various SOA atomic item capabilities in the cockpit scene, and the image display of the car computer is one of them; the SOA atomic item recommendation results include recommendation results for at least cockpit vehicle control services, and more recommendation results for multimedia, and the recommendation results include recommended working status; cockpit vehicle control services include lighting control, air conditioning control, seat adjustment, etc., and multimedia includes audio and images. Among them, the recommendation results for images, such as recommending that the working status of the car computer is to display the generated bamboo forest picture, if the user still needs to take a nap, then the SOA atomic item recommendation results also include adjusting the seat to a comfortable angle, and the recommended working status of the seat is this angle; SOA atomic items based on cockpit vehicle control services, such as adjusting the air conditioning mode, seat position, etc., combined with cockpit vehicle control services and multimedia can create a more immersive user experience. If necessary, multimedia can also include video display.

[0065] S203: Control the cockpit execution module based on the SOA atomic item recommendation result.

[0066] The cloud server is responsible for analyzing and recommending how to create the cockpit scene required by users, which corresponds to the SOA atomic item recommendation results of each SOA atomic item. The vehicle computer is responsible for controlling the vehicle to implement the SOA atomic item recommendation results made by the cloud server; the module in the vehicle computer responsible for controlling the work of each SOA atomic item in the cockpit scene is the cockpit execution module, according to which the corresponding SOA atomic item can be controlled to reach the recommended working state; for example, the seat is controlled to reach the angle recommended by the cloud server.

[0067] In the embodiment of the present application, SOA atomic items are various atomic capabilities in the cockpit scene, including air conditioning, etc.; the AI ​​large model service that can recommend SOA atomic items is deployed in the cloud server, and the vehicle computer is responsible for receiving user input instructions and transmitting the user input instructions to the cloud server, allowing the cloud server to infer the user's intention based on the user input instructions, and give SOA atomic item recommendations, so that each SOA atomic item has a recommended working state that satisfies the user's perception; the cloud server feeds back the inferred SOA atomic item recommendation results to the vehicle computer, allowing the vehicle computer to control each SOA atomic item in the cockpit scene through the cockpit execution module to reach the recommended working state, accurately covering the user's sensory needs, and reasonably utilizing the resources in the intelligent car operating system in the task allocation between the vehicle computer and the cloud server.

[0068] Figure 3 A module implementation diagram of a vehicle-mounted space transformation provided in an embodiment of the present application. Figure 4 A schematic diagram of a vehicle space transformation method provided in an embodiment of the present application Figure 2 .like Figure 4 As shown, the method includes:

[0069] S401, the cloud server obtains the user input instruction through the vehicle computer and identifies the instruction type of the user input instruction; wherein the instruction type includes voice input instruction and text input instruction, the voice input instruction is obtained through the voice recognition program of the cloud server, and the text input instruction is obtained through the text query program of the cloud server.

[0070] like Figure 3 As shown, in the AI ​​big model service of the cloud server, multiple work procedures are included, such as pre-interaction, model reasoning and content generation; this embodiment mainly describes the pre-interaction process, and the pre-interaction includes voice ASR, context and Query Text;

[0071] ASR stands for Automatic Speech Recognition, which is responsible for speech recognition. Voice ASR can convert audio into text. Context includes the user's current input and historical input. Query Text is a way to query text in the AI ​​big model service, which can query text input. The user's voice input command is obtained through the cloud server's speech recognition program, and the user's text input command is obtained through the cloud server's text query program.

[0072] The cloud server obtains the user input command through the vehicle computer, and the user input command includes the command type; the command type includes voice input command and text input command. Generally, voice input and text input are not performed at the same time; when the cloud server receives the voice input, the voice can be converted into text through voice ASR. If the user performs voice input multiple times within a certain period of time, all of them can be converted into text. There is a contextual relationship between the multiple voice inputs. For example, before entering "bamboo forest in the rain" (below), the user may also have said "it's so hot today" (above); the specific content of the text can be queried by querying the text, such as querying for text containing "bamboo forest in the rain".

[0073] S402: The cloud server obtains the user intention through the instruction type and the user's historical input.

[0074] After the cloud server completes the recognition of the instruction type, it needs to combine the user's historical input, that is, the context, to confirm the user's overall intention that the user wants to express, and the user's intention exists in the form of text information.

[0075] In the embodiment of the present application, the user input instructions are parsed through the pre-interaction in the AI ​​big model service of the cloud server to clarify the user's intention.

[0076] Figure 5 A schematic diagram of a vehicle space transformation method provided in an embodiment of the present application Figure 3 .like Figure 5 As shown, the method includes:

[0077] S501. The cloud server constructs a prompt project for the AI ​​big model service based on fine-tuning of the user's historical usage of the SOA atomic items.

[0078] This embodiment mainly describes Figure 3 The model reasoning process in , which includes prompt engineering and fine-tuning;

[0079] Fine-tuning is a technical processing method that makes partial adjustments to the existing model according to the current model requirements. For example, the training set is appropriately adjusted while retaining the weights of the original model, and the model is retrained. At this time, the adjusted training set is more similar to the original training set.

[0080] The AI ​​big model service of the cloud server can be fine-tuned based on the user's historical usage of SOA atomic items to ensure that the data makes reasonable recommendations; for example, when obtaining the user's history for vehicle control, the user's historical usage of SOA atomic items is obtained, and the user's historical usage is constructed as a knowledge graph, and the user's usage tendency for SOA atomic items under different conditions is obtained from the knowledge graph; based on the SOA atomic item usage data corresponding to the usage tendency, the AI ​​big model service is fine-tuned to allow the AI ​​big model service to learn the SOA atomic item usage data with usage tendency; the AI ​​big model service can compare the user's intention and the user's historical usage, where the user's historical usage can be stored in the knowledge base recall scenario;

[0081] The AI ​​big model service can also establish prompt projects, which include keywords (prompt) and reverse keywords (negative_prompt); keywords are related words that meet user intentions, and reverse keywords are related words that are inconsistent with user intentions. For example, if the user wants to generate a dog but does not want the dog to sit, the keyword is dog, and the reverse keyword is not sit.

[0082] S502, the prompt project generates an intention text describing the user's intention according to the user's intention.

[0083] User intent is the content expressed by the user. The prompt project is responsible for extracting keywords from the user intent, describing the user intent with appropriate keywords, and forming an intent text that describes the user intent. Based on the model reasoning capability in the AI ​​big model service, the prompt project's ability to generate keywords can be optimized, allowing the prompt project to generate sufficiently accurate keywords.

[0084] S503: The cloud server generates the SOA atomic item recommendation result according to the intention text.

[0085] Through the AI ​​big model service of the cloud server, model inference is performed to obtain the intent text, and the SOA atomic item recommendation results related to the keywords in the intent text can be generated based on the intent text; the prompt project in the AI ​​big model service can realize the processes of text-to-image (txt2img), text-to-audio (txt2music) and text-to-video (txt2video), which correspond to different SOA atomic item recommendation results respectively;

[0086] In the process of generating the content of the text map:

[0087] For example, the cloud server performs pre-training, parameter optimization and posture adjustment on the intended text to obtain the image generation result in the SOA atomic item recommendation result; wherein the cloud server includes a program for performing the pre-training, the parameter optimization and the posture adjustment, and the parameter optimization is used to adjust the content in the intended text.

[0088] In the text generation process, based on the pre-training, parameter optimization and posture adjustment process, the image generation results in the SOA atomic item recommendation results can be obtained; Figure 3 The content generation process in the process includes Lora pre-training, parameter optimization and posture adjustment; the generated image content can be controlled by the cockpit execution module of the vehicle computer to be displayed on the display screen of the vehicle computer;

[0089] Lora pre-training automatically adjusts the weight of the current layer of neural network by learning the correlation between the previous layer of neural network and the next layer of neural network, thereby improving the performance of the model; parameter optimization can optimize the parameters of the intent text. For example, the image generated according to the intent text has omissions. The dog generated according to the keyword "dog" has no mouth. In this case, the dog with a mouth is generated again after parameter optimization. The keyword "lawn" can also be added autonomously to put the dog on the lawn. ControlNet posture adjustment is used to adjust the posture of the image generated according to the intent text. For example, the dog's posture is limited to running, and the final image generated is a dog running on the lawn.

[0090] In the embodiment of the present application, through model reasoning and content generation of the AI ​​big model service on the cloud server, appropriate SOA atomic item recommendation results are inferred to ensure accurate visual sensory experience for users.

[0091] The AI ​​big model service of the cloud server is responsible for inferring and generating the SOA atomic item recommendation results, and the vehicle computer is responsible for controlling the cockpit scene to execute the recommendation; for SOA atomic items that include images and SOA atomic items that do not include images:

[0092] For example, if the SOA atomic item recommendation result includes an image result, the cloud server adjusts the size of the image result according to the size of the vehicle computer, and controls the vehicle computer to display the image result after the size adjustment through the cockpit execution module;

[0093] If the SOA atomic item recommendation result includes a non-image result, the cockpit execution module is controlled to achieve the non-image result; wherein the image result and the non-image result are recommended working states corresponding to different SOA atomic items.

[0094] SOA atomic items such as background images and videos include image results; SOA atomic items such as air conditioning and seats do not include image results; for SOA atomic item recommendation results that include image results, it is necessary to adjust the image size according to the size of the vehicle computer's display screen, so that the image is fed back to the vehicle computer in an appropriate size, and the vehicle computer's cockpit execution module controls the vehicle computer's display screen to display the resized image results; for SOA atomic item recommendation results that do not include image results, the SOA atomic item recommendation results are fed back to the vehicle computer, so that the vehicle computer's cockpit execution module controls the corresponding SOA atomic items in the cockpit scene to reach the corresponding working state.

[0095] Figure 6 A schematic diagram of the size adjustment of a vehicle-mounted space transformation provided in an embodiment of the present application. Figure 6 As shown, the resizing process includes: source file input (Source Txt Input), image generator (Img Generator), image amplifier (Upscaler), image repairer (Resize & Repair Img), image output (Target Img Output); the image generator can use an end-to-end Wensheng graph model, such as StableDiffusionPipeline models; the image repairer includes image quality repair, and the image quality repair method is such as SwinInIR;

[0096] The source file input in the source file input process is text containing the user's intention; the image generator can convert the text into an image, and the image size is the original size; the image amplifier can enlarge the image to a preset size, which is larger than the original size; the image trimmer can further trim the enlarged image to ensure that it meets the needs of the vehicle computer; the trimmed image can be output.

[0097] The resizing process can be performed under the prompt project:

[0098] For example, the prompt project generates an image describing the user's intention according to the user's intention, enlarges the image, and trims the size and quality of the enlarged image to obtain the image result after the size adjustment.

[0099] The resulting resized image is Figure 6 The prompt project obtains the text with the user's intention, and processes the text into an image through the image generator to obtain an image (original size) that describes the user's intention; after enlarging the image that describes the user's intention, the image is trimmed, including size and quality trimming, and the enlarged image is adjusted (cut) to a size suitable for the vehicle computer, and the quality is reconstructed to ensure that the image is clear enough.

[0100] Figure 7 A basic dependency diagram of a vehicle-mounted space transformation provided by an embodiment of the present application. Figure 7 As shown, the basic dependencies include: algorithm base (AIGC algorithm base), service instance (web service instance), head unit;

[0101] In the cloud server, the algorithm foundation includes source text input, prompt engineering, text-to-img output and other algorithms; service instances include GPU Compute Unit, process queue, basic data feedback and other services;

[0102] In the head unit, it includes basic services involving AI large model services (AIGC based services).

[0103] Example: The AI ​​big model service is deployed in the cloud server configured with concurrency function.

[0104] Under the basic system construction based on AI big model service, the basic dependence of productization is the cornerstone of the overall service; the overall involves the connection between cloud server and vehicle computer, as well as the reasoning of AI big model service, and its concurrency based on the reasoning of AI big model service does not support tens of millions of levels of concurrency. The task construction of AI big model service is based on the pipeline queuing mechanism, that is, the cloud server configuration supports the concurrency function of AI big model service; through this pipeline, multiple commands can be sent to the cloud server at one time and the results can be returned to the vehicle computer at one time after execution.

[0105] In the embodiment of the present application, SOA atomic item recommendation for vehicle space transformation is performed under basic dependencies, and queue resources are reasonably allocated based on concurrent functions.

[0106] In an environment without a car computer, the SOA atomic item recommendation results of the car computer can be displayed through the experience station; if the user inputs the information "I want to sleep for a while in the bamboo forest in the rain", the experience station can display the content, including: image (bamboo forest), air conditioning volume (0), seat ventilation (off), ambient light color (empty), air conditioning (off), windows (all closed), air conditioning circulation mode (internal circulation), fragrance (off), seat massage intensity (moderate), air conditioning wind direction (blowing glass), seat heating (off), seat posture (unchanged).

[0107] Figure 8 A vehicle-mounted space conversion device is provided in an embodiment of the present invention, such as Figure 8 As shown, the device includes: an acquisition module 801, a processing module 802 and an output module 803;

[0108] The acquisition module 801 is used for the cloud server to acquire the user input instruction through the vehicle computer, and acquire the user intention according to the user input instruction.

[0109] The acquisition module 801 is also used for the cloud server to acquire the user input command through the vehicle computer and identify the command type of the user input command; wherein the command type includes a voice input command and a text input command, the voice input command is acquired through the voice recognition program of the cloud server, and the text input command is acquired through the text query program of the cloud server;

[0110] The cloud server obtains the user intention through the instruction type and user historical input.

[0111] Processing module 802, for the AI ​​big model service of the cloud server to obtain SOA atomic item recommendation results according to the user intention; wherein the SOA atomic item recommendation results include recommended working states of at least some SOA atomic items in the cockpit scene; the SOA atomic items include cockpit vehicle control services, or include cockpit vehicle control services and multimedia, the multimedia includes audio and images, and the images include static images and dynamic images.

[0112] The processing module 802 is also used for the cloud server to build a prompt project for the AI ​​big model service based on fine-tuning of the user's historical usage of the SOA atomic item;

[0113] The prompt project generates an intention text describing the user intention according to the user intention;

[0114] The cloud server generates the SOA atomic item recommendation result according to the intention text.

[0115] The processing module 802 is also used for the cloud server to pre-train, optimize parameters and adjust the posture of the intended text, and obtain the image generation result in the SOA atomic item recommendation result; wherein the cloud server includes a program for performing the pre-training, the parameter optimization and the posture adjustment, and the parameter optimization is used to adjust the content in the intended text.

[0116] The output module 803 is used to control the cockpit execution module based on the SOA atomic item recommendation result.

[0117] The output module 803 is further configured to, if the SOA atomic item recommendation result includes an image result, cause the cloud server to adjust the size of the image result according to the size of the vehicle computer, and control the vehicle computer to display the image result after the size adjustment through the cockpit execution module;

[0118] If the SOA atomic item recommendation result includes a non-image result, the cockpit execution module is controlled to achieve the non-image result; wherein the image result and the non-image result are recommended working states corresponding to different SOA atomic items.

[0119] The present application also provides a vehicle-mounted space transformation device, comprising: at least one processor and a memory;

[0120] The memory stores computer-executable instructions;

[0121] The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the in-vehicle space transformation method.

[0122] Fig. 9 The hardware diagram of the vehicle-mounted space conversion device provided by the embodiment of the present invention is shown in FIG. Fig. 9 As shown, the vehicle-mounted space transformation device 90 provided in this embodiment includes: at least one processor 901 and a memory 902. The device 90 also includes a communication component 903. The processor 901, the memory 902 and the communication component 903 are connected via a bus 904.

[0123] In the specific implementation process, at least one processor 901 executes the computer-executable instructions stored in the memory 902, so that at least one processor 901 executes the above-mentioned vehicle-mounted space transformation method.

[0124] The specific implementation process of the processor 901 can be found in the above method embodiment, and its implementation principle and technical effect are similar, so this embodiment will not be repeated here.

[0125] In the above Fig. 9 In the illustrated embodiment, it should be understood that the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the invention may be directly implemented as being executed by a hardware processor, or may be executed by a combination of hardware and software modules in the processor.

[0126] The memory may include a high-speed memory (Random Access Memory, RAM), and may also include a non-volatile memory (Non-volatile Memory, NVM), such as at least one disk memory.

[0127] The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the bus in the drawings of the present application is not limited to only one bus or one type of bus.

[0128] The present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the vehicle-mounted space transformation method as described above is implemented.

[0129] The computer-readable storage medium mentioned above can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general or special-purpose computer.

[0130] An exemplary readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (Application Specific Integrated Circuits, referred to as: ASIC). Of course, the processor and the readable storage medium can also exist in the device as discrete components.

[0131] The division of the units is only a logical function division, and there may be other divisions in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, which may be electrical, mechanical or other forms.

[0132] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0133] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

[0134] Finally, it should be noted that those skilled in the art will readily conceive of other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variation, use or adaptation of the present invention, which follows the general principles of the present invention and includes common knowledge or conventional techniques in the art not disclosed by the present invention, is not limited to the precise structure described above and shown in the drawings, and may be modified and changed in various ways without departing from the scope thereof.

Claims

1. A vehicle-mounted space transformation method, characterized in that: Applied to a cloud server connected to a smart car operating system, and a vehicle computer of the smart car operating system, the method includes: The cloud server obtains the user input command through the vehicle computer, and obtains the user intention according to the user input command; The AI ​​big model service of the cloud server obtains the SOA atomic item recommendation result according to the user intention; wherein the SOA atomic item recommendation result includes the recommended working status of at least some SOA atomic items in the cockpit scene; the SOA atomic item includes the cockpit vehicle control service, or includes the cockpit vehicle control service and multimedia, the multimedia includes audio and images, and the images include static images and dynamic images; Based on the SOA atomic item recommendation result, the cockpit execution module is controlled.

2. The method according to claim 1, characterized in that The cloud server obtains the user input command through the vehicle computer, and obtains the user intention according to the user input command, including: The cloud server obtains the user input command through the vehicle computer and identifies the command type of the user input command; wherein the command type includes a voice input command and a text input command, the voice input command is obtained through the voice recognition program of the cloud server, and the text input command is obtained through the text query program of the cloud server; The cloud server obtains the user intention through the instruction type and user historical input.

3. The method according to claim 1, characterized in that The AI ​​big model service of the cloud server obtains the SOA atomic item recommendation result according to the user intention, including: The cloud server constructs a prompt project for the AI ​​big model service based on fine-tuning of the user's historical usage of the SOA atomic item; The prompt project generates an intention text describing the user intention according to the user intention; The cloud server generates the SOA atomic item recommendation result according to the intention text.

4. The method according to claim 3, characterized in that The cloud server generates the SOA atomic item recommendation result according to the intention text, including: The cloud server performs pre-training, parameter optimization and posture adjustment on the intended text to obtain the image generation result in the SOA atomic item recommendation result; wherein the cloud server includes a program for performing the pre-training, the parameter optimization and the posture adjustment, and the parameter optimization is used to adjust the content in the intended text.

5. The method according to claim 3, characterized in that: The controlling cockpit execution module based on the SOA atomic item recommendation result includes: If the SOA atomic item recommendation result includes an image result, the cloud server adjusts the size of the image result according to the size of the vehicle computer, and controls the vehicle computer to display the image result after the size adjustment through the cockpit execution module; If the SOA atomic item recommendation result includes a non-image result, the cockpit execution module is controlled to achieve the non-image result; wherein the image result and the non-image result are recommended working states corresponding to different SOA atomic items.

6. The method according to claim 5, characterized in that The cloud server adjusts the size of the image result according to the size of the vehicle computer, including: The prompt project generates an image describing the user's intention according to the user's intention, enlarges the image, and trims the size and quality of the enlarged image to obtain the image result after the size adjustment.

7. The method according to claim 1, characterized in that The AI ​​big model service is deployed in the cloud server configured with concurrency function.

8. A vehicle-mounted space conversion device, characterized in that: Applicable to a cloud server connected to a smart car operating system, and a vehicle computer of the smart car operating system, including: An acquisition module, used for the cloud server to acquire a user input instruction through the vehicle computer, and to acquire a user intention according to the user input instruction; A processing module, used for the AI ​​big model service of the cloud server to obtain the SOA atomic item recommendation result according to the user intention; wherein the SOA atomic item recommendation result includes the recommended working status of at least some SOA atomic items in the cockpit scene; the SOA atomic item includes the cockpit vehicle control service, or includes the cockpit vehicle control service and multimedia, the multimedia includes audio and image, and the image includes static image and dynamic image; The output module is used to control the cockpit execution module based on the SOA atomic item recommendation result.

9. A vehicle-mounted space conversion device, characterized in that: include: at least one processor and memory; The memory stores computer-executable instructions; The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the vehicle-mounted space transformation method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the vehicle-mounted space transformation method as described in any one of claims 1 to 7 are implemented.

Images