Vehicle control method, server and computer readable storage medium

Abstract

The invention discloses a vehicle control method, a server and a computer readable storage medium. The method comprises the steps of receiving a voice request forwarded by a vehicle; then, according to the voice request, a task list is determined, and the task list comprises at least one subtask; then, according to the sub-tasks, target sub-agents corresponding to the sub-tasks are determined; and then, according to a processing result of the target sub-agent on the corresponding sub-task, a vehicle control instruction is determined. And finally, sending a vehicle control instruction to the vehicle to control the vehicle to execute the vehicle control instruction. Therefore, the voice request is disassembled into the sub-tasks, and the sub-tasks are processed by the special sub-agents, so that the task processing efficiency and accuracy can be improved. Moreover, according to different task requirements, different sub-agents are flexibly selected, complex tasks are completed through cooperation of multiple agents, finally, control over the vehicle is achieved, the adaptability to diversified tasks and the overall performance are improved, and convenient and comfortable driving experience is provided for users.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle control, and particularly to a vehicle control method, a server, and a computer-readable storage medium. Background Art

[0002] In the related art, the vehicle cockpit system interacts with users through an intelligent agent to improve driving convenience. However, since the intelligent agent is usually optimized only for specific application scenarios, it has insufficient understanding of complex voice requests when processing complex voice requests and cannot accurately process the complex voice requests, which affects the user experience. Summary of the Invention

[0003] This application provides a vehicle control method, a server, and a computer-readable storage medium.

[0004] An embodiment of this application provides a vehicle control method, and the method includes:

[0005] Receiving a voice request forwarded by a vehicle;

[0006] Determining a task list according to the voice request, where the task list includes at least one subtask;

[0007] Determining a target sub-intelligent agent corresponding to the subtask according to the subtask;

[0008] Determining a vehicle control instruction according to the processing result of the target sub-intelligent agent for the corresponding subtask;

[0009] Sending the vehicle control instruction to the vehicle to control the vehicle to execute the vehicle control instruction.

[0010] In this way, the server receives the voice request forwarded by the vehicle. Then, the server determines a task list according to the voice request, and the task list includes at least one subtask. Subsequently, the server determines a target sub-intelligent agent corresponding to the subtask according to the subtask. Then, the server determines a vehicle control instruction according to the processing result of the target sub-intelligent agent for the corresponding subtask. Finally, the server sends the vehicle control instruction to the vehicle to control the vehicle to execute the vehicle control instruction. In this way, by decomposing the voice request into subtasks and processing them by dedicated sub-intelligent agents, the efficiency and accuracy of task processing can be improved. Moreover, according to different task requirements, different sub-intelligent agents can be flexibly selected, and complex tasks can be completed through multi-intelligent agent collaboration, ultimately realizing the control of the vehicle, improving the adaptability and overall performance of the system to diverse tasks, and providing a convenient and comfortable driving experience for users.

[0011] In some embodiments, the determining a task list according to the voice request includes:

[0012] Perform a splitting process on the voice request to determine at least one of the subtasks;

[0013] Determine the task list according to the subtasks.

[0014] In this way, the server performs a splitting process on the voice request to determine at least one subtask. Then, the server determines the task list according to the subtasks. In this way, through the splitting process of the voice request, the system can more accurately understand the user's intention, accurately determine the task list, realize the decomposition and execution of complex tasks, improve the efficiency and flexibility of the system, provide intelligent and convenient services for users, and enhance the user experience.

[0015] In some embodiments, the determining, according to the subtasks, a target sub-agent corresponding to the subtasks includes:

[0016] Determine the task type of the subtask according to the subtask;

[0017] Determine a target sub-agent corresponding to the subtask according to the task type.

[0018] In this way, the server determines the task type of the subtask according to the subtasks. Then, the server determines the target sub-agent corresponding to the subtask according to the task type. In this way, by assigning different types of tasks to specific sub-agents, professional division of labor can be achieved, the execution efficiency can be improved, and the user experience can be enhanced.

[0019] In some embodiments, the determining, according to the processing result of the target sub-agent on the corresponding subtask, a vehicle control instruction includes:

[0020] Determine the current planning information according to the current subtask, the current target sub-agent corresponding to the current subtask, and the current display content information of the vehicle display components;

[0021] Determine the vehicle control instruction corresponding to the current target subtask according to the current planning information.

[0022] In this way, the server determines the current planning information according to the current subtask, the current target sub-agent corresponding to the current subtask, and the current display content information of the vehicle display components. Then, the server determines the vehicle control instruction corresponding to the current target subtask according to the current planning information. In this way, by analyzing the current subtask, the current target sub-agent, and the current display content information, the system can dynamically adjust according to the current state of the vehicle and the external environment, so as to generate accurate vehicle control instructions.

[0023] In some embodiments, the method further includes:

[0024] When the vehicle control instruction is executed, determine the status information of the current subtask according to the current subtask, the current target sub-agent, the current display content information, the current planning information, and the vehicle control instruction.

[0025] In this way, when the vehicle control instruction is executed, the server determines the status information of the current subtask according to the current subtask, the current target sub-agent, the current display content information, the current planning information, and the vehicle control instruction. Thus, after the vehicle control instruction that can realize the user requirements represented by the current subtask is executed, determine the status information of the current subtask according to the current subtask, the current target sub-agent, the current display content information, the current planning information, and the vehicle control instruction, so as to determine the realization of the user requirements according to the status information subsequently.

[0026] In some embodiments, the method further includes:

[0027] Determine whether the user requirements corresponding to the voice request are met according to the voice request and the status information corresponding to the last subtask in the status information set, so as to verify the rationality of the task list, where the status information set includes the status information of each subtask.

[0028] In this way, the server determines whether the user requirements corresponding to the voice request are met according to the voice request and the status information corresponding to the last subtask in the status information set, so as to verify the rationality of the task list, where the status information set includes the status information of each subtask. Thus, by verifying the rationality of the task list, it can be ensured that the original requirements of the user are met, thereby improving the user satisfaction.

[0029] In some embodiments, the method further includes:

[0030] Generate feedback information according to the status information set when it is confirmed that the task list is reasonable;

[0031] Send the feedback information to the vehicle to control the vehicle to broadcast the feedback information.

[0032] In this way, when it is confirmed that the task list is reasonable, the server generates feedback information according to the status information set. Then, the server sends the feedback information to the vehicle to control the vehicle to broadcast the feedback information. Thus, by generating and broadcasting the feedback information to the user, the user can be informed of the task execution result in a timely manner, enabling the user to understand the progress and final result of the task, thereby increasing the interaction between the vehicle and the user and enhancing the user experience.

[0033] In some embodiments, the method further includes:

[0034] When it is confirmed that the task list is unreasonable, task splitting auxiliary information is generated according to the task list and a preset experience library.

[0035] An optimized task list is determined according to the task splitting auxiliary information and the voice request, and the optimized task list includes at least one optimized subtask.

[0036] A target sub-agent corresponding to the optimized subtask is determined according to the optimized subtask.

[0037] An optimized vehicle control instruction is determined according to the processing result of the target sub-agent for the corresponding optimized subtask.

[0038] The optimized vehicle control instruction is sent to the vehicle to control the vehicle to execute the optimized vehicle control instruction.

[0039] In this way, when it is confirmed that the task list is unreasonable, the server generates task splitting auxiliary information according to the task list and a preset experience library. Then, the server determines an optimized task list according to the task splitting auxiliary information and the voice request, and the optimized task list includes at least one optimized subtask. Next, the server determines a target sub-agent corresponding to the optimized subtask according to the optimized subtask. Subsequently, the server determines an optimized vehicle control instruction according to the processing result of the target sub-agent for the corresponding optimized subtask. Finally, the server sends the optimized vehicle control instruction to the vehicle to control the vehicle to execute the optimized vehicle control instruction. In this way, by generating an optimized task list, unnecessary tasks can be avoided, ensuring that all subtasks can be successfully completed and user requirements are met, enhancing the user experience.

[0040] An embodiment of the present application provides a server, which includes a processor and a memory. A computer program is stored on the memory, and when the computer program is executed by the processor, the above vehicle control method is implemented.

[0041] An embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the vehicle control method as described above are implemented.

[0042] Additional aspects and advantages of the embodiments of the present application will be given in part in the following description, become apparent in part from the following description, or be understood through the practice of the embodiments of the present application. Description of the Drawings

[0043] The above and / or additional aspects and advantages of the present application will become apparent and be easily understood from the description of the embodiments in conjunction with the following drawings, where:

[0044] Figure 1 is one of the flow diagrams of the vehicle control method according to some embodiments of the present application;

[0045] Figure 2 is the second of the flow diagrams of the vehicle control method according to some embodiments of the present application;

[0046] Figure 3 is the third of the flow diagrams of the vehicle control method according to some embodiments of the present application;

[0047] Figure 4 is the fourth of the flow diagrams of the vehicle control method according to some embodiments of the present application;

[0048] Figure 5 is the fifth of the flow diagrams of the vehicle control method according to some embodiments of the present application;

[0049] Figure 6 is the sixth of the flow diagrams of the vehicle control method according to some embodiments of the present application;

[0050] Figure 7 is the seventh of the flow diagrams of the vehicle control method according to some embodiments of the present application;

[0051] Figure 8 is the eighth of the flow diagrams of the vehicle control method according to some embodiments of the present application. Detailed Embodiments

[0052] The following describes in detail the embodiments of the present application. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary only for explaining the embodiments of the present application and should not be construed as limiting the embodiments of the present application.

[0053] In the current technical environment, the vehicle cockpit system interacts with users through agents to improve driving convenience. However, current vehicle cockpit agents are usually optimized for specific application scenarios, such as navigation, music playback, phone calls, etc.

[0054] This results in that when dealing with complex voice requests involving multiple steps or subtasks, the agent may not be able to effectively decompose and assign these tasks to different agents for processing. Moreover, it is difficult for a single agent to cooperate with other applications or services, which limits its performance in scenarios where multiple applications need to cooperate to complete tasks.

[0055] Specifically, agents are usually optimized only for specific application scenarios and lack the ability to handle cross-application tasks. For example, an agent optimized for navigation may not be able to handle complex requests involving navigation and music playback.

[0056] Based on the above problems, please refer to Figure 1 , an embodiment of the present application provides a vehicle control method, the method comprising:

[0057] 01: Receiving a voice request forwarded by a vehicle;

[0058] 02: Determining a task list according to the voice request;

[0059] 03: Determining a target sub-agent corresponding to the subtask according to the subtask;

[0060] 04: Determining a vehicle control instruction according to the processing result of the target sub-agent for the corresponding subtask;

[0061] 05: Sending a vehicle control instruction to the vehicle to control the vehicle to execute the vehicle control instruction.

[0062] An embodiment of the present application further provides a server, comprising a memory and a processor. The vehicle control method of the embodiment of the present application can be implemented by the server of the embodiment of the present application. Specifically, a computer program is stored in the memory, and the processor is configured to receive a voice request forwarded by the vehicle. And determine a task list according to the voice request. And determine a target sub-agent corresponding to the subtask according to the subtask. The processor is configured to determine a vehicle control instruction according to the processing result of the target sub-agent for the corresponding subtask. And send a vehicle control instruction to the vehicle to control the vehicle to execute the vehicle control instruction.

[0063] An embodiment of the present application further provides a vehicle control device. The vehicle control method of the embodiment of the present application can be implemented by the vehicle control device of the embodiment of the present application. Specifically, the vehicle control device comprises a receiving module, a determining module and a sending module. The obtaining module is configured to receive a voice request forwarded by the vehicle. The determining module is configured to determine a task list according to the voice request. The determining module is further configured to determine a target sub-agent corresponding to the subtask according to the subtask. The determining module is further configured to determine a vehicle control instruction according to the processing result of the target sub-agent for the corresponding subtask. The sending module is configured to send a vehicle control instruction to the vehicle to control the vehicle to execute the vehicle control instruction.

[0064] Specifically, a task list refers to a set that includes at least one subtask. Each subtask is a part of the user's request, which is decomposed from the user's voice request. Each subtask corresponds to the generation of a vehicle control instruction, such as information search, navigation route, temperature adjustment, and driving mode switching. For example, if the user's voice request is "Take me to the nearest gas station", then the task list may include two subtasks: "Search for the nearest gas station" and "Plan the route to the gas station".

[0065] A sub-agent refers to a small agent optimized for a specific task type. For example, sub-agent A is good at searching for the strategy of application A, sub-agent B is good at querying and booking air tickets or train tickets, sub-agent C is good at using application B, sub-agent D is good at understanding general pictures, and sub-agent E is good at switching between applications and handling abnormal situations. Each sub-agent focuses on a specific field and has corresponding knowledge and skills to efficiently complete the tasks assigned to it. In some embodiments, the training method of the sub-agent may be to use a large pre-trained model as the base model and fine-tune the same base model according to the prompt words or training data corresponding to the specific task type. For example, a large language model can be used for pre-training and then fine-tuned according to the task of searching for the strategy of application A.

[0066] A target sub-agent refers to an agent that is selected and assigned to execute a specific subtask. Each target sub-agent is determined according to the requirements and characteristics of the subtask to ensure that the task can be processed most effectively. For example, a task list includes three subtasks: task 1, task 2, and task 3. Among them, the sub-agent corresponding to task 1 is sub-agent A, and the sub-agents corresponding to task 2 and task 3 are both sub-agent D. Then the target sub-agents are sub-agent A and sub-agent D.

[0067] The processing result of the target sub-agent for the corresponding subtask refers to the output or result of the task completed when a certain sub-agent in the system is assigned to execute a specific subtask. This result can be any form of information, depending on the specific content of the subtask.

[0068] The user inputs an instruction by voice. For example, the user's voice request is "Help me compare which air tickets from address E to address F on application C and application D are cheaper".

[0069] Next, the system analyzes and decomposes the voice request into multiple subtasks according to the voice request, and determines a task list. Continuing with the above example, the user's voice request is "Help me compare which flight tickets from address E to address F on application C and application D are cheaper", and the determined task list is ["Search for flight tickets from address E to address F on application C", "Search for flight tickets from address E to address F on application D", "Compare which of the two flight tickets is cheaper"].

[0070] Then, the system determines the most suitable sub-agent, that is, the target sub-agent, according to the characteristics of each subtask in the above task list. Continuing with the above example, according to the task list ["Search for flight tickets from address E to address F on application C", "Search for flight tickets from address E to address F on application D", "Compare which of the two flight tickets is cheaper"], the determined target sub-agents are "agent B, agent B, and agent D" respectively.

[0071] Subsequently, the selected target sub-agent processes the corresponding subtask according to its proficient field, obtains a processing result, and generates a corresponding vehicle control instruction.

[0072] Finally, after the system integrates the vehicle control instructions generated by all sub-agents, it sends the vehicle control instructions to the vehicle to control the vehicle to execute the vehicle control instructions.

[0073] In summary, in the vehicle control method and server provided by the embodiment of the present application, the server receives the voice request forwarded by the vehicle. Next, the server determines a task list according to the voice request, and the task list includes at least one subtask. Subsequently, the server determines the target sub-agent corresponding to the subtask according to the subtask. Then, the server determines the vehicle control instruction according to the processing result of the target sub-agent for the corresponding subtask. Finally, the server sends the vehicle control instruction to the vehicle to control the vehicle to execute the vehicle control instruction. In this way, by disassembling the voice request into subtasks and processing them by dedicated sub-agents, the efficiency and accuracy of task processing can be improved. Moreover, according to different task requirements, different sub-agents can be flexibly selected, and complex tasks can be completed through multi-agent collaboration, ultimately realizing the control of the vehicle, enhancing the adaptability and overall performance of the system to diverse tasks, and providing users with a convenient and comfortable driving experience.

[0074] Please refer to Figure 2 , in some embodiments, step 02 (determine a task list according to the voice request) includes:

[0075] 021: Perform a splitting process on the voice request to determine at least one subtask;

[0076] 022: Determine the task list according to the subtask.

[0077] In some embodiments, the determination module is configured to split a voice request, determine at least one subtask, and determine a task list according to the subtasks.

[0078] In some embodiments, the processor is further configured to split a voice request, determine at least one subtask, and determine a task list according to the subtasks.

[0079] Specifically, the splitting process refers to the process of decomposing the user's original voice request into smaller and more specific subtasks, enabling the system to more accurately understand and execute the user's request. In some embodiments, the Prompt for the model to perform the splitting process is:

[0080] Suppose you are an agent planning assistant. Please split the user request into the operation intensity of a single app according to the following information.

[0081] The task categories of the user's request are as follows:

[0082] 1. Search for a strategy in Application A

[0083] 2. Book a train ticket

[0084] 3. Order takeout

[0085] 4. Send a message in Application B

[0086] Please output the appropriate task category for the user request.

[0087] For example, the user query: "Help me send a reminder to Xiaohong in Application B to bring something."

[0088] Output subtasks: ["Send a message in Application B"].

[0089] The server splits the received voice request, decomposing it into smaller and executable subtasks. Continuing the above example, when the user says "Help me compare which flight tickets from Address E to Address F on Application C and Application D are cheaper", the main Agent will split it into three subtasks: "Search for flight tickets from Address E to Address F on Application C", "Search for flight tickets from Address E to Address F on Application D", and "Compare which of the two flight tickets is cheaper".

[0090] Then, according to the split subtasks, determine the final task list. Continuing the above example, the task list includes the three subtasks mentioned above.

[0091] In this way, through the splitting process of voice requests, the system can more accurately understand the user's intention, accurately determine the task list, realize the decomposition and execution of complex tasks, improve the efficiency and flexibility of the system, provide users with intelligent and convenient services, and enhance the user experience.

[0092] Please refer to Figure 3 , in some embodiments, step 03 (determining the target sub-agent corresponding to the subtask according to the subtask) includes:

[0093] 031: Determine the task type of the subtask according to the subtask;

[0094] 032: Determine the target sub-agent corresponding to the subtask according to the task type.

[0095] In some embodiments, the determination module is further configured to determine the task type of the subtask according to the subtask, and determine the target sub-agent corresponding to the subtask according to the task type.

[0096] In some embodiments, the processor is further configured to determine the task type of the subtask according to the subtask, and determine the target sub-agent corresponding to the subtask according to the task type.

[0097] Specifically, the task type refers to the category obtained by classifying the subtask according to the nature and requirements of the subtask, including navigation tasks, information search, information sending, and application switching, etc. Different task types may require different processing methods and expertise. By identifying and classifying the task types, the system can more effectively allocate tasks to the appropriate sub-agents and ensure that the tasks can be executed correctly and efficiently.

[0098] The server determines the task type to which it belongs according to the content and characteristics of the subtask. For example, the subtask "search for air tickets from address E to address F on application C" belongs to the "air ticket query" type, the subtask "search for air tickets from address E to address F on application D" belongs to the "air ticket query" type, and the subtask "compare which of the two air tickets is cheaper" belongs to the "picture understanding" type.

[0099] Next, select the target sub-agent. The server selects the target sub-agent corresponding to each subtask according to the determined task type. For example, for subtasks of the "air ticket query" type, "agent B" can be selected to execute; for subtasks of the "picture understanding" type, "agent D" can be selected to execute.

[0100] In this way, by assigning different types of tasks to specific sub-agents, professional division of labor can be achieved, the execution efficiency can be improved, and the user experience can be enhanced.

[0101] Please refer toFigure 4 , in some embodiments, step 04 (determining a vehicle control instruction according to the processing result of the corresponding subtask by the target sub-agent) includes:

[0102] 041: determining current planning information according to the current subtask, the current target sub-agent corresponding to the current subtask, and the current display content information of the vehicle display component;

[0103] 042: determining a vehicle control instruction corresponding to the current target subtask according to the current planning information.

[0104] In some embodiments, the determining module is further configured to determine current planning information according to the current subtask, the current target sub-agent corresponding to the current subtask, and the current display content information of the vehicle display component. And determine a vehicle control instruction corresponding to the current target subtask according to the current planning information.

[0105] In some embodiments, the processor is further configured to determine current planning information according to the current subtask, the current target sub-agent corresponding to the current subtask, and the current display content information of the vehicle display component. And determine a vehicle control instruction corresponding to the current target subtask according to the current planning information.

[0106] Specifically, the current display content information of the vehicle display component refers to the information content currently presented on the in-vehicle display screen or other display devices. For example, navigation-related content such as the current driving route, destination, estimated arrival time, and traffic conditions, or entertainment system-related content such as a music playlist, track information, and radio frequency, and can also be communication-related content such as incoming call information, text message content, and contact list. The display content information is an important part of the vehicle intelligent control system. The display content information can provide key driving information and entertainment services for the driver, and is also an important interface for system-user interaction. It should be noted that in the embodiments of the present application, the in-vehicle display screen is used as the vehicle display component to illustrate the vehicle control method, that is, the current display content information is a screenshot of the in-vehicle display screen.

[0107] It should also be noted that the current display content information is obtained through optical character recognition technology (Optical Character Recognition, OCR) technology and XML file parsing technology. First, compare the text set extracted by OCR and the control set parsed by XML to find the overlapping elements, that is, the OCR text and the XML control point to the same screen position. Then, perform a similarity calculation, use the area ratio formula to calculate the similarity between the OCR text and the XML control, and judge whether the two are associated. The area ratio formula calculates the ratio of the overlapping area of the OCR text and the XML control to the union area, and the calculation formula is Finally, perform fusion. If the similarity is greater than the threshold (e.g., 0.9), it is considered that the OCR text is associated with the XML control, and the information of the two is combined to form a new information set, including the control name, attributes and types, text content, and location information, that is, generate T_merge = {(t_p, n_p, attr_p, (x1_p, y1_p, x2_p, y2_p)) | p = 1, 2,..., n}.

[0108] OCR technology is used to recognize the text content on the screen and record its location information. Moreover, OCR technology can extract the text content into a text set, and each element contains the text content and location information. The form of the text set is: T_ocr = {(t_i, (x1_j, y1_j, x2_j, y2_j)) | i = 1, 2,..., n}. Among them, t_i represents the i-th string, and (x1_i, y1_i, x2_i, y2_i) represents the location of t_i on the screen.

[0109] XML file parsing is used to extract the structured information of the screen, and the structured information is extracted into a set, and each element contains the control name, attributes and types, and location information. The form of the set is: T_xml = {(n_j, attr_j, (x1_j, y1_j, x2_j, y2_j)) | j = 1, 2,..., n}. Among them, n_j represents the j-th control name, attr_j represents the j-th attributes and types, and (x1_j, y1_j, x2_j, y2_j) represents the location of n_j on the screen.

[0110] In some embodiments, the Retrieval Augmented Generation (RAG) technology can also be used to associate the page elements on the screen with an external knowledge base, so as to obtain richer information. That is, by analyzing the screen screenshot and the XML file, the page elements and structured information are extracted, and RAG information is generated. The RAG information includes the similarity between the page elements and the corresponding external knowledge base entries, so as to judge the specific meaning represented by the page elements.

[0111] The planning information refers to the steps and actions to complete the subtasks, such as "1. Open application A; 2. Click on the search box", etc.

[0112] Vehicle control instructions refer to the current planning information, which determines and executes specific operations, including click operations, swipe operations, typing operations, etc. For example, "1. Click on location A (coordinate position) in the in-vehicle screen; 2. Click on the search box and type in the search information". A click operation refers to an operation such as clicking on an icon or button to activate a specific function or service. For example, when the user says "Open music", the system may need to perform an operation of clicking on the music application icon in the in-vehicle entertainment system. A swipe operation refers to an operation where when the current page information list is folded or too long, the screen needs to be swiped or paged to view more information. For example, when the route information displayed by the navigation system is too long, the system may need to perform a screen swipe operation to display the complete route. A typing operation refers to an operation of clicking on the search box and typing in content to find specific information or services in the system. For example, when the user says "Search for nearby restaurants", the system may need to perform an operation of clicking on the search box and entering "nearby restaurants". A typing operation can also indicate entering and sending text messages in the in-vehicle system, such as sending text messages or making comments on social media. For example, when the user says "Send a message to Zhang San", the system may need to perform operations of opening the messaging application, selecting contact Zhang San, and opening the chat window.

[0113] After the server determines the current subtask and the current target sub-intelligent agent corresponding to the current subtask, it obtains the current display content information of the vehicle display components, that is, at the start time when the target sub-intelligent agent is needed to process the current subtask, it obtains the current display content information of the vehicle display components. And, the server generates the current planning information based on the above information.

[0114] Next, the server determines the vehicle control instructions corresponding to the current target subtask according to the current planning information.

[0115] Continuing with the above example, the current subtask is "Search for flight tickets from address E to address F on application C", and the current in-vehicle display screen is on the main interface. Based on the received "Current subtask: Search for flight tickets from address E to address F on application C; Current display content information: Main interface;", the current planning information is generated as "1. Open application A; 2. Click on the search box and type in address E to address F; 3. Click on Search; 4. Determine the ticket price". Next, the server determines the vehicle control instructions corresponding to the current target subtask according to the current planning information.

[0116] In this way, by analyzing the current subtask, the current target sub-intelligent agent, and the current display content information, the system can dynamically adjust according to the current state of the vehicle and the external environment, so as to generate accurate vehicle control instructions.

[0117] Please refer to Figure 5 , in some embodiments, the method further includes:

[0118] 043: When the vehicle control instruction is executed, determine the status information of the current subtask according to the current subtask, the current target sub-agent, the current display content information, the current planning information, and the vehicle control instruction.

[0119] In some embodiments, the determination module is further configured to determine the status information of the current subtask according to the current subtask, the current target sub-agent, the current display content information, the current planning information, and the vehicle control instruction when the vehicle control instruction is executed.

[0120] In some embodiments, the processor is further configured to determine the status information of the current subtask according to the current subtask, the current target sub-agent, the current display content information, the current planning information, and the vehicle control instruction when the vehicle control instruction is executed.

[0121] Specifically, the status information refers to the information set of the current subtask completed, the current target sub-agent used, the display content information at the time of subtask completion, the current planning information, and the vehicle control instruction recorded when a certain subtask is completed, which can be used for subsequent result fusion to generate the final execution result.

[0122] It should be noted that after determining the subtask and the corresponding target sub-agent, an information set including the current subtask, the current target sub-agent, the current display content information, the current planning information, and the vehicle control instruction will be generated, such as S_agent=(T_subtask, N_agent, T_merge, A, P). Where T_subtask is the currently executed task, N_agent is the selected target sub-agent, T_merge is the current screen information. A is the action list executed by the target sub-agent, with a default value of [], which is cleared each time the target sub-agent is switched. P is the planning list of the target sub-agent, with a default value of [], which is cleared each time the target sub-agent is switched.

[0123] After a vehicle control instruction in the information set is executed, it will be updated until all vehicle control instructions are completed, and then the information set will be recorded as status information. For example, on agent B, the task is "Search for flight tickets from address E to address F on application C". S1 = (Current subtask: "Search for flight tickets from address E to address F on application C", Current target sub-agent: agent B, Current display content information: T1, Current planning information: [a1, a2,..., a_n], Vehicle control instructions: [p1, p2,..., p_n]). When vehicle control instruction p1 is completed, S2 = (Current subtask: "Search for flight tickets from address E to address F on application C", Current target sub-agent: agent B, Current display content information: T1, Current planning information: [a1(finish), a2,..., a_n], Vehicle control instructions: [p1(finish), p2,..., p_n]). The finally recorded status information is S3 = (Current subtask: "Search for flight tickets from address E to address F on application C", Current target sub-agent: agent B, Current display content information: T1, Current planning information: [a1, a2,..., a_n(finish)], Vehicle control instructions: [p1, p2,..., p_n(finish)]).

[0124] Obtain the information of the current subtask, including the task content of the subtask, the target sub-agent, the current display content information, the current planning information, and the vehicle control instructions. Analyze the above information to determine the completion status of the vehicle control instructions. When the vehicle control instructions are executed, record the current subtask, the current target sub-agent, the current display content information, the current planning information, and the vehicle control instructions at this time as the status information of the current subtask.

[0125] In this way, after the vehicle control instructions that can achieve the user requirements represented by the current subtask are executed, determine the status information of the current subtask according to the current subtask, the current target sub-agent, the current display content information, the current planning information, and the vehicle control instructions, so as to determine the implementation status of the user requirements according to the status information later.

[0126] Please refer to Figure 6 , in some embodiments, the method further includes:

[0127] 044: Determine whether the user requirements corresponding to the voice request are met according to the voice request and the status information corresponding to the last subtask in the status information set, so as to verify the rationality of the task list.

[0128] In some embodiments, the determination module is further configured to determine whether the user requirements corresponding to the voice request are met according to the voice request and the status information corresponding to the last subtask in the status information set, so as to verify the rationality of the task list.

[0129] In some embodiments, the processor is further configured to determine whether the user requirements corresponding to the voice request are met according to the voice request and the status information corresponding to the last subtask in the status information set, so as to verify the rationality of the task list.

[0130] Specifically, the server determines whether the user requirements corresponding to the voice request are met according to the voice request and the status information corresponding to the last subtask in the status information set, so as to verify the rationality of the task list, where the status information set includes the status information of each subtask.

[0131] Continuing with the above example, the status information set is obtained, including the status information of each subtask, that is, the completion status of "searching for flight tickets from address E to address F on application C", the completion status of "searching for flight tickets from address E to address F on application D", and the completion status of "comparing which of the two flight tickets is cheaper". In some embodiments, the status information set is "S3 = (\"searching for flight tickets from address E to address F on application C\", agentB, T_2, [a1, a2,..., a_n,..., a_m(finish)], [p1, p2,..., p_n,..., p_m(finish)]); S5 = (\"searching for flight tickets from address E to address F on application D\", agentB, T_4, [a1, a2,..., a_n,..., a_m(finish)], [p1, p2,..., p_n,..., p_m(finish)]); S7 = (\"comparing which of the two flight tickets is cheaper\", agentD, T_6, [a1, a2,..., a_n,..., a_m(finish)], [p1, p2,..., p_n,..., p_m(finish))".

[0132] Next, use natural language processing technology to analyze the voice request and extract the user requirements. For example, the user needs to compare the prices of flight tickets from address E to address F on application C and application D. Compare the user requirements with the status information corresponding to the last subtask in the status information set to determine whether the user requirements are met. If the status information set shows that "comparing which of the two flight tickets is cheaper" has been completed and the result is that "the flight ticket on application C is cheaper", it means that the user requirements have been met.

[0133] It should be noted that if the user's requirements have been met, it indicates that the task list is reasonable and subsequent tasks can be continued or the task can be ended. If the user's requirements are not met, it indicates that there may be something unreasonable in the task list and the task decomposition and planning need to be redone.

[0134] Continuing with the above example, if "Compare which air ticket is cheaper" in the status information set is not completed, it is necessary to check whether the task decomposition is reasonable and whether some subtasks are omitted.

[0135] In this way, by verifying the reasonableness of the task list, it can be ensured that the user's original requirements are met, thereby improving user satisfaction.

[0136] Please refer to Figure 7 , in some embodiments, the method further includes:

[0137] 045: When it is confirmed that the task list is reasonable, generate feedback information according to the status information set;

[0138] 046: Send the feedback information to the vehicle to control the vehicle to broadcast the feedback information.

[0139] In some embodiments, the vehicle control device further includes a generation module, and the generation module is further configured to generate feedback information according to the status information set when it is confirmed that the task list is reasonable. The sending module is further configured to send the feedback information to the vehicle to control the vehicle to broadcast the feedback information.

[0140] In some embodiments, the processor is further configured to generate feedback information according to the status information set when it is confirmed that the task list is reasonable. And send the feedback information to the vehicle to control the vehicle to broadcast the feedback information.

[0141] Specifically, judge whether the task list is reasonable according to the user's requirements and the status information set to ensure that all subtasks have been successfully completed and the user's requirements are met. Continuing with the above example, if the user requests "Help me compare which air ticket from address E to address F on application C and application D is cheaper", and the status information set shows that all subtasks have been successfully completed, then it can be confirmed that the task list is reasonable.

[0142] Then, generate feedback information according to the status information set, including the task execution result and the satisfaction of the user's requirements.

[0143] Continuing with the above example, if the status information set shows that "searching for flight tickets from address E to address F on application C" is completed, "searching for flight tickets from address E to address F on application D" is completed, "comparing which of the two flight tickets is cheaper" is completed, and the result is that "the flight ticket on application C is cheaper", then feedback information can be generated: "The cheapest flight ticket from address E to address F on application C is 1111 yuan, and the cheapest flight ticket from address E to address F on application D is 1111 yuan. After comparison, the flight ticket on application C is cheaper."

[0144] Finally, the generated feedback information is sent to the vehicle to control the vehicle to broadcast the feedback information.

[0145] Continuing with the above example, the vehicle can voice broadcast: "The cheapest flight ticket from address E to address F on application C is 1111 yuan, and the cheapest flight ticket from address E to address F on application D is 1111 yuan. After comparison, the flight ticket on application C is cheaper."

[0146] In this way, by generating feedback information and broadcasting it to the user, the user can be informed of the task execution result in a timely manner, enabling the user to understand the progress and final result of the task, thereby increasing the interaction between the vehicle and the user and enhancing the user experience.

[0147] Please refer to Figure 8 , in some embodiments, the method further includes:

[0148] 047: In the case of confirming that the task list is unreasonable, generate task splitting assistance information according to the task list and a preset experience library;

[0149] 048: Determine an optimized task list according to the task splitting assistance information and the voice request;

[0150] 049: Determine a target sub-intelligent agent corresponding to the optimized subtask according to the optimized subtask;

[0151] 050: Determine an optimized vehicle control instruction according to the processing result of the target sub-intelligent agent for the corresponding optimized subtask;

[0152] 051: Send the optimized vehicle control instruction to the vehicle to control the vehicle to execute the optimized vehicle control instruction.

[0153] In some embodiments, the generation module is further configured to generate task splitting auxiliary information according to the task list and a preset experience library when it is confirmed that the task list is unreasonable. The determination module is further configured to determine an optimized task list according to the task splitting auxiliary information and the voice request. And determine a target sub-agent corresponding to the optimized subtask according to the optimized subtask. And determine an optimized vehicle control instruction according to the processing result of the target sub-agent on the corresponding optimized subtask. The sending module is configured to send the optimized vehicle control instruction to the vehicle to control the vehicle to execute the optimized vehicle control instruction.

[0154] In some embodiments, the processor is further configured to generate task splitting auxiliary information according to the task list and a preset experience library when it is confirmed that the task list is unreasonable. And determine an optimized task list according to the task splitting auxiliary information and the voice request. And determine a target sub-agent corresponding to the optimized subtask according to the optimized subtask. The processor is further configured to determine an optimized vehicle control instruction according to the processing result of the target sub-agent on the corresponding optimized subtask. And send the optimized vehicle control instruction to the vehicle to control the vehicle to execute the optimized vehicle control instruction.

[0155] Specifically, the preset experience library refers to a data set that stores the decision-making process, successful execution cases, and failure lessons of the in-vehicle intelligent assistant, enabling the in-vehicle intelligent assistant to learn from past experiences and improve the quality and efficiency of its decision-making. The decision-making process record details the decision-making process of the intelligent assistant when processing user requests, including decision-making inputs (historical operation descriptions of the user and the vehicle's perception state information at that time), execution results after executing operations, and timestamps (the time when the decision occurs), etc. The successful execution cases record the situations where the intelligent assistant successfully executes tasks, including decision-making inputs, executed operations, and final successful results. The failure lessons record the failure situations encountered by the intelligent assistant when executing tasks, including the reasons for failure, errors that occurred, and possible improvement measures. The above decision-making process, successful execution cases, and failure lessons are collectively referred to as preset reflection information, that is, the preset experience library includes multiple pieces of preset reflection information, and each piece of preset reflection information includes preset historical operation description information, preset state information, preset execution results, and preset execution operation information.

[0156] Judge whether the task list is reasonable according to the user requirements and the state information set. If there are subtasks not completed or the user requirements are not met, it means the task list is unreasonable.

[0157] In the case where the task list is confirmed to be unreasonable, analyze the reasons for the unreasonableness of the task list based on the task list and the preset experience library, and generate task splitting auxiliary information, such as which subtasks are recommended to be added or which subtasks are to be modified. Continuing with the above example, if after detection, it is found that the subtask of comparing ticket prices is missing from the task list, then task splitting auxiliary information can be generated: "It is recommended to add the subtask of comparing which of the two tickets is cheaper".

[0158] Determine the optimized task list based on the task splitting auxiliary information and the voice request, including at least one optimized subtask. Continuing with the above example, the optimized task list can be modified to "Search for tickets from address E to address F on application C", "Search for tickets from address E to address F on application D", and "Compare which of the two tickets is cheaper".

[0159] Next, based on the optimized task list, determine the target sub-intelligent agents corresponding to each optimized subtask. Continuing with the above example, for the subtask "Search for tickets from address E to address F on application C", the target sub-intelligent agent is "agent B". For the subtask "Search for tickets from address E to address F on application D", the target sub-intelligent agent is "agent B". For the subtask "Compare which of the two tickets is cheaper", the target sub-intelligent agent is "agent D".

[0160] Subsequently, determine the optimized vehicle control instruction based on the processing results of the target sub-intelligent agents for the corresponding optimized subtasks.

[0161] Finally, send the optimized vehicle control instruction to the vehicle to control the vehicle to execute the optimized vehicle control instruction.

[0162] It should be noted that all subtasks need to be processed through the above steps, which will not be elaborated here.

[0163] In this way, by generating an optimized task list, unnecessary tasks can be avoided, ensuring that all subtasks can be successfully completed and the user's needs are met, enhancing the user experience.

[0164] This application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the vehicle control method as described above are implemented.

[0165] It can be understood that a computer program includes computer program code. The computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable storage medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution medium, etc.

[0166] In the description of this specification, the descriptions referring to terms such as "specifically", "further", "specially", "understandably", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiments or examples are included in at least one embodiment or example of the present application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0167] Any process or method description shown in the flowchart or described in other ways herein can be understood as representing a module, segment, or part of executable request code including one or more steps for implementing a specific logical function or process, and the scope of the preferred embodiments of the present application includes additional implementations, where the functions can be executed in a substantially simultaneous manner or in the reverse order according to the functions involved, rather than in the order shown or discussed, which should be understood by those skilled in the art of the embodiments of the present application.

[0168] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present application. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A vehicle control method, characterized in that: The method comprises: receiving a voice request; Determine a task list according to the voice request, wherein the task list includes at least one subtask; According to the subtask, determining a target sub-agent corresponding to the subtask; Determining a vehicle control instruction according to a processing result of the target sub-agent on the corresponding sub-task; The vehicle control instruction is sent to a vehicle to control the vehicle to execute the vehicle control instruction.

2. The method according to claim 1, characterized in that The step of determining a task list according to the voice request includes: Splitting the voice request to determine at least one of the subtasks; The task list is determined according to the subtasks.

3. The method according to claim 1, characterized in that The step of determining a target sub-agent corresponding to the sub-task according to the sub-task includes: According to the subtask, determining the task type of the subtask; According to the task type, a target sub-agent corresponding to the sub-task is determined.

4. The method according to claim 1, characterized in that: The step of determining the vehicle control instruction according to the processing result of the target sub-agent on the corresponding sub-task comprises: Determine current planning information according to the current subtask, the current target sub-agent corresponding to the current subtask, and the current display content information of the vehicle display component; The vehicle control instruction corresponding to the current target subtask is determined according to the current planning information.

5. The method according to claim 4, characterized in that The method further comprises: When the vehicle control instruction is executed, the status information of the current subtask is determined according to the current subtask, the current target sub-agent, the current display content information, the current planning information and the vehicle control instruction.

6. The method according to claim 5, characterized in that The method further comprises: According to the voice request and the status information corresponding to the last subtask in the status information set, determine whether the user demand corresponding to the voice request is met to verify the rationality of the task list, wherein the status information set includes the status information of each subtask.

7. The method according to claim 6, characterized in that The method further comprises: When confirming that the task list is reasonable, generating feedback information according to the state information set; The feedback information is sent to the vehicle to control the vehicle to broadcast the feedback information.

8. The method according to claim 6, characterized in that The method further comprises: When it is confirmed that the task list is unreasonable, generating task splitting auxiliary information according to the task list and a preset experience library; Determine an optimization task list according to the task splitting auxiliary information and the voice request, wherein the optimization task list includes at least one optimization subtask; According to the optimization subtask, determining a target sub-agent corresponding to the optimization subtask; Determining an optimized vehicle control instruction according to the processing result of the target sub-agent on the corresponding optimization sub-task; The optimized vehicle control instruction is sent to a vehicle to control the vehicle to execute the optimized vehicle control instruction.

9. A server, characterized in that: The server includes a processor and a memory, wherein a computer program is stored in the memory. When the computer program is executed by the processor, the method according to any one of claims 1 to 8 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 steps of the method according to any one of claims 1 to 8 are implemented.

Images