Vehicle-mounted user intention recognition and task generation system based on large model

Abstract

The invention relates to a vehicle-mounted user intention recognition and task generation system based on a large model, and relates to the technical field of automobile interaction. The user intention analysis module is used for obtaining voice input of a user in a vehicle and analyzing the voice input to obtain a plurality of user intentions; the situation awareness module is used for collecting current environment data of the vehicle and obtaining a current situation based on the current environment data through a large model; the task generation module is used for searching a task capability interface of a vehicle according to the user intention and the current situation to generate a task combination; the task combination is to call different agents; the reflection module is used for evaluating whether the generated task combination is consistent with the user intention and the current situation and adjusting the inconsistent task combination; and the task sequence arrangement module is used for confirming the control sequence of different series-parallel tasks according to the task combination and calling the corresponding agents.

Description

Technical Field

[0001] The present application relates to the technical field of automobile interaction, and in particular to an in-vehicle user intention recognition and task generation system based on a large model. Background Art

[0002] At present, the semantics of vehicle cockpit dialogue systems are all in the form of instructions. When users give clear instructions, the in-vehicle dialogue system performs semantic recognition and takes the next step based on the instructions. Current solutions generally have insufficient generalization capabilities. For example, they can recognize the instruction to turn on the air conditioner, but cannot recognize the generalized instruction to adjust the air conditioner to a suitable temperature.

[0003] To solve the generalization problem, large model capabilities are widely used in dialogue systems. With the help of large model reasoning and context understanding capabilities, traditional predictive agents can be further improved to identify the specific intentions of users, which has a very good effect. The introduction of large model capabilities has improved the generalization ability of speech, making the dialogue between people and cars more and more natural. However, it also brings new problems. The introduction of large model technology has made users have higher requirements for dialogue systems, and they expect the system to better understand the intentions in the dialogue. Since there are naturally multi-intention scenarios in human dialogue speech, the current command-based accurate classification method in the in-vehicle dialogue system is no longer applicable to the broad generalized intentions after the support of large models, affecting the user experience.

[0004] Therefore, how to achieve smarter human-vehicle interaction that better meets user needs has become a problem that needs to be solved urgently. Summary of the invention

[0005] In view of the problems in the prior art such as insufficient generalization ability, inability to accurately identify multiple user intentions, and insufficient intelligence in task generation logic, the present invention provides an in-vehicle user intention recognition and task generation system based on a large model.

[0006] The large model-based in-vehicle user intention recognition and task generation system provided in this application adopts the following technical solution.

[0007] The in-vehicle user intention recognition and task generation system based on a large model includes:

[0008] User intention recognition module; used to: obtain the voice input of the user in the car, and analyze the voice input to obtain multiple user intentions;

[0009] A situational awareness module is used to: collect the current environmental data of the vehicle and obtain the current situation based on the current environmental data through a large model;

[0010] A task generation module is used to: generate a task combination according to the user intention and the task capability interface of the current context vehicle search; the task combination is to call different intelligent agents;

[0011] A reflection module, used to: evaluate whether the generated task combination is consistent with the user intention and the current situation and adjust the inconsistent task combination;

[0012] The task sequence scheduling module is used to: confirm the control order of different series and parallel tasks according to the task combination and call the corresponding intelligent agent.

[0013] Optionally, the user intention recognition module includes:

[0014] Input data processing unit; receiving and processing user's voice input;

[0015] Large model understanding unit: Uses a large model to analyze the user's voice input and identify the user's subjective intention;

[0016] Intent classification unit: converts user intentions into demands of different dimensions; the demands of different dimensions include: rest and relaxation, entertainment and leisure, soothing and comfort, environmental control, navigation and driving, communication and social networking, safety reminders and work production.

[0017] Optionally, the context awareness module includes:

[0018] Environmental data collection unit; collects the current environmental data of the vehicle, the current environmental data includes: vehicle operation status, vehicle interior environment, vehicle exterior environment, safety status and user status;

[0019] Large model situation analysis unit; analyzes current environmental data based on the large model to obtain the current situation.

[0020] Optionally, the system also includes: a self-learning module; the self-learning module is used to: collect user feedback data and perform intent correlation analysis; generate a standard positive database and a negative database based on the feedback data and adjust the large model.

[0021] Optionally, a standard positive database and negative database are generated based on the feedback data, including:

[0022] After each task is executed, the user's feedback data will be recorded; the feedback data includes: task completion status, user satisfaction evaluation, behavior patterns and task execution results; when the user manually adjusts the tasks generated by the system, the relevant task generation data and task arrangement order will be used as negative feedback.

[0023] Optionally, the task generation module includes:

[0024] Task search unit: searches for the vehicle's task capability interface based on the user's intent and current context;

[0025] Large model task generation unit: Generates task combinations that meet user intentions and situations based on the pre-trained large model.

[0026] Optionally, the self-learning module is further used to: during the execution of the task combination, monitor the completion status of the task and the changes in the current environmental data in real time, and re-evaluate the rationality of the task combination according to the current situation;

[0027] The self-learning module is also used to record and analyze the user's long-term behavior data and extract the user's habits and preferences.

[0028] Optionally, the reflection module includes:

[0029] Reflection evaluation unit: evaluates whether the generated task combination is consistent with the user's intention and the current situation;

[0030] Safety assessment unit; ensure the rationality and safety of task combination;

[0031] Task correction unit: adjusts task combinations that do not meet expectations.

[0032] The in-vehicle user intention recognition and task generation system based on a large model provided by the present invention significantly improves the accuracy of user intention recognition by introducing the semantic reasoning and context understanding capabilities of the large model, and can parse multiple intentions in complex contexts and dynamically generate task combinations that meet actual needs in combination with the current situation of the vehicle. The system optimizes the execution order of tasks through the task sequence arrangement module to ensure the efficient coordination of serial and parallel tasks, and at the same time evaluates and adjusts the rationality and safety of tasks through the reflection module to improve the safety and reliability of the system. The self-learning module can continuously optimize the weights and task generation logic of the large model according to user behavior and feedback, realize personalized customization, and gradually meet the unique needs of users. The situational awareness module combines the internal and external environmental data of the vehicle, and accurately adapts to diversified driving scenarios through large model reasoning and analysis, providing a more accurate reference basis for task generation. The system effectively solves the problems of insufficient generalization ability, inability to accurately identify multiple intentions and unintelligent task generation in the prior art, greatly improves the intelligence level of the in-vehicle system and the user interaction experience, and provides users with safer, more comfortable and efficient driving services. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a system block diagram of the in-vehicle user intention recognition and task generation system based on a large model in an embodiment of the present application. DETAILED DESCRIPTION

[0034] Large model capabilities are widely used in dialogue systems. With the help of the reasoning and context understanding capabilities of large models, traditional predictive agents can be further improved to identify the specific intentions of users, which has a very good effect. The introduction of large model capabilities has improved the generalization ability of speech, making the dialogue between people and cars more and more natural. However, it also brings new problems. The introduction of large model technology has made users have higher requirements for dialogue systems, and they expect the system to better understand the intentions in the dialogue. Since there are naturally multiple intention scenarios in human dialogue speech, the current command-based accurate classification method in the in-vehicle dialogue system can no longer be applied to the broad generalized intentions after the blessing of large models, affecting the user experience. For example, when a user in the dialogue system requests: What is the weather like in Beijing during this period? Following up on actual user surveys, the following possible intentions exist:

[0035] Today's weather forecast for Beijing, the agent corresponds to the weather report;

[0036] The overall change of climate over a long period of time recently, which corresponds to the overall weather query of the agent;

[0037] Based on the geographical location and historical situation, the overall weather conditions in Beijing during this period, the agent corresponds to encyclopedic knowledge;

[0038] Since the current in-vehicle dialogue system understands human language through a single classification, a single agent and slot are selected to achieve a closed loop of functions. However, the user's specific intention is related to the vehicle status, geographic location, and mood. A single agent classification cannot meet the user's situational needs.

[0039] In order to solve the above problems, this application proposes an in-vehicle user intention recognition and task generation system based on a large model.

[0040] The following is combined with Figure 1 The present application is further described with specific embodiments:

[0041] First of all, it should be noted that in the description of this application, if the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside" and other directional words appear, the orientation or position relationship indicated is based on the orientation or position relationship shown in the drawings, which is only for the convenience of description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on this application; in addition, if the terms "first", "second", "third" and other numerical quantifiers appear, they are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in this application, unless otherwise clearly specified and limited, if the terms "installed", "connected", and "connected" appear, they should be understood in a broad sense, for example, it can be a fixed connection, or a detachable connection, a limited connection such as an interference fit, a transition fit, or an integral connection; it can be directly connected or indirectly connected through an intermediate medium; therefore, for ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to the specific circumstances.

[0042] The embodiment of the present application discloses a vehicle-mounted user intention recognition and task generation system based on a large model. Figure 1 As an implementation of the vehicle-mounted user intention recognition and task generation system based on a large model, the vehicle-mounted user intention recognition and task generation system based on a large model includes:

[0043] User intention recognition module 101; used to: obtain voice input from a user in the car, and analyze the voice input to obtain multiple user intentions;

[0044] The context perception module 102 is used to collect the current environment data of the vehicle and obtain the current context based on the current environment data through a large model;

[0045] The task generation module 103 is used to: generate a task combination according to the user intention and the task capability interface of the current context vehicle search; the task combination is to call different agents;

[0046] A reflection module 104 is used to: evaluate whether the generated task combination is consistent with the user intention and the current situation and adjust the inconsistent task combination;

[0047] The task sequence arrangement module 105 is used to: confirm the control order of different serial and parallel tasks according to the task combination and call the corresponding intelligent agent.

[0048] Specifically, the user intention recognition module 101 analyzes the user's voice input to understand the multiple intention expressions in a complex context. The context perception module 102 collects data on the internal and external environment of the vehicle and uses a large model to analyze these data to obtain the current context information. The task generation module 103 combines the user's intention and the current context to determine the most appropriate task combination, and completes these tasks by calling different agents, ensuring the effectiveness and efficiency of task execution. For example, if the user simultaneously asks to play music and set a destination, the system will start the navigation first and then play the music to prevent the navigation voice prompt from being covered by the music. The reflection module 104 is responsible for evaluating whether the generated task combination matches the user's real needs. If not, it will automatically adjust the task combination to ensure that the final execution result can better meet the user's needs. The task sequence arrangement module 105 is responsible for determining the execution order between tasks and how to coordinate the simultaneous operation of multiple tasks. The in-vehicle user intention recognition and task generation system based on the large model has significantly improved the intelligence level of the in-vehicle system through the application of deep learning and artificial intelligence technology, providing users with a safer, more convenient and comfortable driving environment.

[0049] The following is a detailed description based on a scenario:

[0050] In a car scenario, the user inputs "I'm a little tired" by voice. The system first recognizes the user's demand as "rest and relax" through the user intention recognition module. Then, the context perception module collects the current environmental data of the vehicle, such as the speed is 0, the gear is in the parking gear, the air conditioner is turned on, and the user is alone in the car. The large model task generation module will generate task combinations based on such contexts and user intentions, such as adjusting the seat to a comfortable position, playing light music, turning on the nap mode, etc. The reflection mechanism module will check whether these tasks are consistent with the current context and whether there are safety hazards. If it is found that the task does not meet expectations, such as adjusting the window and the air conditioning state conflict, the task will be corrected and regenerated. The task sequence arrangement module will arrange the tasks in parallel or serially to ensure that the execution of the tasks is logical. After the task is executed, the user switches the music to another light music, then the self-learning module will collect the user data and use it as a bad case for fine-tuning the task generation large model. After a certain amount of data training, the model will meet the personalized response to the user. Through this method, the present invention can intelligently identify user intentions and generate appropriate task combinations according to specific situations, greatly improving the user's driving experience and the intelligence level of the system.

[0051] As a specific implementation of the in-vehicle user intention recognition and task generation system based on a large model, a voice input of a user in the vehicle is obtained, and the voice input is parsed to obtain multiple user intentions, including:

[0052] Parsing the voice input of the user in the car to obtain a user input x;

[0053] The user input x is inferred using a large model to obtain the output probability corresponding to each agent; wherein the large model is provided with an agent set and an agent-agent weight; each element in the agent set corresponds to an intention; the agent-agent weight includes a preset weight for each agent; the weight is the importance of the agent in the inference process;

[0054] For each agent, determine whether its output probability is greater than the preset probability threshold; if so, it is determined to meet the conditions;

[0055] Output all agents that meet the conditions as multiple user intentions obtained by parsing.

[0056] Specifically, the voice input of the user in the car is parsed to obtain the user input x, and the original voice data is accurately converted into a text form that can be used for subsequent processing; the user input x is inferred using a large model to obtain the output probability corresponding to each agent, and the relevance of each agent to the user's intention is evaluated, thereby improving the accuracy of intent recognition. Each agent has a specific weight, which helps the model better understand which agents are more important for a specific task, thereby giving higher priority when identifying user intentions. For each agent, it is determined whether its output probability is greater than a preset probability threshold, so that only agents with high probability matching will be selected, thereby reducing the situation of mismatching and improving the system's response quality and user experience. The system outputs all agents that meet the conditions as multiple user intentions obtained by parsing, and screens out a group of agents that are most likely to meet the user's intention, which enables the system to respond to user requests more accurately.

[0057] As a specific implementation of the vehicle-mounted user intention recognition and task generation system based on a large model, the control sequence of different serial and parallel tasks is determined according to the task combination and the corresponding intelligent agent is called, including:

[0058] Determine the agent's response delay weight: ; Among them, each agent S i The response delay is t i , the lower the latency, the higher the weight;

[0059] Determine agent priority weights ; Each agent Corresponding to a static priority weight ;

[0060] Determine the weight of user personalized preferences ; Used to represent users’ preferences for different agents;

[0061] Determine vehicle restriction status weights ;

[0062] The arbitration function A is determined based on the response delay weight, the agent priority weight, the user personalized preference weight and the vehicle restriction state weight; wherein the formula of the arbitration function A is: ; It is an intelligent agent The output probability distribution of represents the probability that the user intends to belong to the agent, where i=1,2,…,n, where n is the total number of agents;

[0063] Select the agent with the largest arbitration function A The first agent to respond responds to other agents in the order of decreasing values ​​of arbitration function A.

[0064] Specifically, by considering the response time of the agent, those agents with fast response speed can be given priority, thereby speeding up the response speed of the entire system; each agent has a static priority weight, and some key or commonly used functions will be given priority to ensure that important functions are processed first and improve user experience. Considering the personalized needs of users, the system can adjust the selection of agents according to the user's historical behavior and preferences, so that the service is more in line with the user's habits and enhances user satisfaction. The state of the vehicle (such as driving status, safety requirements, etc.) will affect the available agents and services. By considering these restrictions, the system can make more reasonable and safe service decisions. By comprehensively considering the above factors, the arbitration function A can comprehensively evaluate the applicability of each agent, so as to select the agent that best suits the current situation. The agent with the largest value of arbitration function A is selected as the output result of the agent, ensuring that the best agent is selected in the current situation, maximizing the satisfaction of user needs while also taking into account factors such as the system's response speed and safety.

[0065] As a specific implementation of the large-model-based in-vehicle user intention recognition and task generation system, the system also includes: a self-learning module; the self-learning module is used to: collect user feedback data and perform intention correlation analysis; generate a standard positive database and a negative database based on the feedback data and adjust the large model.

[0066] Specifically, after each task is executed, the system will record user feedback data, such as task completion status, user satisfaction rating, behavior patterns, and task execution results. If users frequently manually adjust system-generated tasks (such as canceling a task or changing the order of tasks), the system will treat this data as negative feedback and believe that the current task generation logic may be insufficient.

[0067] The system gradually learns the user's preferences and habits by analyzing these feedback data. For example, if a user is used to playing soft music and turning off the lights in the car while taking a break, the system will automatically generate a combination of these two tasks when it recognizes the user's "rest" intention in the future, so that task generation is more in line with the user's expectations.

[0068] As a specific implementation of the vehicle-mounted user intention recognition and task generation system based on a large model, a standard positive database and a negative database are generated according to feedback data, including:

[0069] After each task is executed, the user's feedback data will be recorded; the feedback data includes: task completion status, user satisfaction evaluation, behavior patterns and task execution results; when the user manually adjusts the tasks generated by the system, the relevant task generation data and task arrangement order will be used as negative feedback.

[0070] As a specific implementation of the large model-based in-vehicle user intention recognition and task generation system, the self-learning module is also used to:

[0071] Dynamically update the agent’s weights based on user feedback and preferences;

[0072] Based on user feedback and preferences, the weights of the agent are dynamically updated, including:

[0073] Set the weight vector w and update function U; the weight vector w contains each agent The update function U is used to dynamically adjust the weight;

[0074] Set the feedback vector r; the feedback vector r contains the user's feedback information on each agent's output;

[0075] The update function U takes the feedback vector r and the current weight vector w as input to output a new weight vector ;in ;in is the learning rate, used to control the update speed, ;

[0076] Updated weight vector It will be used in the next agent arbitration process. During the arbitration process, each agent The output probability distribution of Will be weighted according to the corresponding weight; ; n is the total number of agents.

[0077] Specifically, by collecting user feedback and preferences, the system can continuously adjust the weights of the agents to make them more in line with the user's personal preferences. The weight vector w contains the weights of each agent, and the update function U is used to dynamically adjust these weights based on user feedback. The feedback vector r contains the user's feedback information on the output of each agent, which is used to guide the update of the weights. By utilizing user feedback, the system can better understand the user's needs and make more accurate service decisions based on this. By using the learning rate α to control the speed of weight updates, it can ensure that the system can quickly adapt to user needs while avoiding instability caused by over-adjustment. During the arbitration process, the output probability distribution of each agent is weighted according to the corresponding weight. The system will be more inclined to select those agents that have been verified by users to perform better, thereby providing higher quality services.

[0078] As one of the implementation methods of the in-vehicle user intention recognition and task generation system based on a large model, the self-learning module is also used to: during the execution of the task combination, monitor the completion status of the task and the change of the current environmental data in real time, and re-evaluate the rationality of the task combination according to the current situation;

[0079] The self-learning module is also used to record and analyze the user's long-term behavior data and extract the user's habits and preferences.

[0080] Specifically, during the execution of a task, the system can monitor the completion of the task and changes in the environment in real time, and re-evaluate the rationality of the task based on the current situation. For example, if the system generates a task to play music after recognizing the user's "rest" intention, but detects a sudden increase in noise outside the car (such as traffic noise), the system will adjust the music volume or switch to noise-reducing music based on this situational change. In addition, if a task conflict occurs during task execution (such as opening the window and starting the air conditioner at the same time), the reflection mechanism will intervene and adjust the task generation strategy to ensure the consistency and rationality of task execution in the new situation.

[0081] The system will continuously record and analyze the user's long-term behavior data to extract the user's habits and preferences. For example, if a user often chooses to play a specific type of music or adjust the color of the interior lights when driving at night, the system will learn this pattern and prioritize the generation of tasks that meet the user's habits in similar situations.

[0082] The system will not only predict possible future needs based on the user's current needs, but also combine their past behavior patterns to generate relevant task combinations in advance. For example, if a user is accustomed to taking a break every hour during a long-distance drive at night, the system will automatically suggest that the user take regular breaks or arrange appropriate rest points when recognizing the long-distance driving situation.

[0083] For example, during many long-distance driving trips, a user often takes a break after driving for 2 hours. During the break, he likes to play light music and lower the seat back angle. The system gradually learns the user's rest preferences by recording these behaviors. Next, when the user inputs "I'm a little tired" by voice during a long-distance drive, the system will generate a task combination, suggesting to stop and rest, and automatically play light music and lower the seat angle. If the user also manually adjusts the brightness of the lights in the car, the system will also record it and add the light adjustment task to the next rest task.

[0084] When performing the same task next time, the system will directly generate such a personalized task combination, thereby reducing the user's operation steps and providing services that better meet personal needs.

[0085] As one implementation of the in-vehicle user intention recognition and task generation system based on a large model, the current environment data of the vehicle is collected and the current context is obtained based on the current environment data through the large model. The specific steps include:

[0086] Performing a preliminary analysis on the vehicle status information to identify at least one key feature for current situation judgment;

[0087] Set a starting reasoning path and feature weights based on the branching logic of the decision tree;

[0088] Perform reasoning based on the current reasoning path and adjust the path based on output feedback and adjust feature weights through backpropagation during the path adjustment process;

[0089] After multiple reasoning paths, a convergent reasoning path is obtained, and the current situation is obtained based on the convergent reasoning path.

[0090] Specifically, by conducting a preliminary analysis of the key features of the vehicle status information, the system can quickly capture important clues to the current situation; through the branching logic of the decision tree, the system can select the appropriate reasoning path based on the current key features and set the corresponding weights according to the importance of the features; reasoning is performed based on the current reasoning path and the path is adjusted based on the output feedback and the feature weights are adjusted by backpropagation during the path adjustment process. The system dynamically adjusts the reasoning path based on the feedback of the reasoning results and adjusts the feature weights by backpropagation, continuously optimizing the process of situation recognition and improving accuracy. Through multiple iterative reasoning until the path converges, the system can ensure the stability of the situation recognition results and reduce the errors caused by a single reasoning path.

[0091] As one implementation of the in-vehicle user intention recognition and task generation system based on a large model, the user intention recognition module includes:

[0092] Input data processing unit; receiving and processing user's voice input;

[0093] Large model understanding unit: Uses a large model to analyze the user's voice input and identify the user's subjective intention;

[0094] Intention classification unit: converts user intention into demands of different dimensions; the demands of different dimensions include: rest and relaxation, entertainment and leisure, soothing and comforting, environmental control, navigation and driving, communication and social interaction, safety reminders and work production;

[0095] The situation awareness module includes:

[0096] Environmental data collection unit; collects the current environmental data of the vehicle, the current environmental data includes: vehicle operation status, vehicle interior environment, vehicle exterior environment, safety status and user status;

[0097] Large model situation analysis unit: analyzes the current environmental data based on the large model to obtain the current situation;

[0098] The task generation module includes:

[0099] Task search unit: searches for the vehicle's task capability interface based on the user's intent and current context;

[0100] Large model task generation unit: Generates task combinations that meet user intentions and situations based on the pre-trained large model;

[0101] The reflection module includes:

[0102] Reflection evaluation unit: evaluates whether the generated task combination is consistent with the user's intention and the current situation;

[0103] Safety assessment unit; ensure the rationality and safety of task combination;

[0104] Task correction unit: adjusts task combinations that do not meet expectations.

[0105] It should be noted that the above embodiments are only used to illustrate the present application and are not intended to limit the technical solutions described in the present application. Although the present application has been described in detail in this specification with reference to the above embodiments, a person of ordinary skill in the art should understand that a person of ordinary skill in the art can still modify or make equivalent substitutions to the present application, and all technical solutions and improvements thereof that do not depart from the spirit and scope of the present application should be included in the scope of the claims of the present application.

Claims

1. A large model-based vehicle-mounted user intention recognition and task generation system, characterized by: include: User intention recognition module; Used to: obtain voice input from a user in the car, and analyze the voice input to obtain multiple user intentions; A situational awareness module is used to: collect the current environmental data of the vehicle and obtain the current situation based on the current environmental data through a large model; A task generation module is used to: generate a task combination according to the user intention and the task capability interface of the current context vehicle search; the task combination is to call different intelligent agents; A reflection module, used to: evaluate whether the generated task combination is consistent with the user intention and the current situation and adjust the inconsistent task combination; The task sequence scheduling module is used to: confirm the control order of different series and parallel tasks according to the task combination and call the corresponding intelligent agent.

2. The large model-based vehicle-mounted user intention recognition and task generation system according to claim 1 is characterized in that: The user intention recognition module comprises: Input data processing unit; receiving and processing user's voice input; Large model understanding unit: Uses a large model to analyze the user's voice input and identify the user's subjective intention; Intent classification unit: converts user intentions into demands of different dimensions; the demands of different dimensions include: rest and relaxation, entertainment and leisure, soothing and comfort, environmental control, navigation and driving, communication and social networking, safety reminders and work production.

3. The large model-based vehicle-mounted user intention recognition and task generation system according to claim 2 is characterized in that: The context awareness module comprises: Environmental data collection unit; collects the current environmental data of the vehicle, the current environmental data includes: vehicle operation status, vehicle interior environment, vehicle exterior environment, safety status and user status; Large model situation analysis unit; analyzes current environmental data based on the large model to obtain the current situation.

4. The large model-based vehicle-mounted user intention recognition and task generation system according to claim 3 is characterized in that: The system also includes: a self-learning module; the self-learning module is used to: collect user feedback data and perform intention correlation analysis; generate a standard positive database and a negative database based on the feedback data and adjust the large model.

5. The large model-based vehicle-mounted user intention recognition and task generation system according to claim 4 is characterized in that: Generate standard positive and negative databases based on feedback data, including: After each task is executed, the user's feedback data will be recorded; the feedback data includes: task completion status, user satisfaction evaluation, behavior patterns and task execution results; when the user manually adjusts the tasks generated by the system, the relevant task generation data and task arrangement order will be used as negative feedback.

6. The large model-based vehicle-mounted user intention recognition and task generation system according to claim 5 is characterized in that: The task generation module includes: Task search unit: searches for the vehicle's task capability interface based on the user's intent and current context; Large model task generation unit; Generate task combinations that match user intentions and situations based on pre-trained large models.

7. The large model-based vehicle-mounted user intention recognition and task generation system according to claim 6 is characterized in that: The self-learning module is also used to: monitor the completion of the task and the change of the current environment data in real time during the execution of the task combination, and re-evaluate the rationality of the task combination according to the current situation; The self-learning module is also used to record and analyze the user's long-term behavior data and extract the user's habits and preferences.

8. The large model-based vehicle-mounted user intention recognition and task generation system according to claim 5, characterized in that: The reflection module includes: Reflection evaluation unit: evaluates whether the generated task combination is consistent with the user's intention and the current situation; Safety assessment unit; ensure the rationality and safety of task combination; Task correction unit: adjusts task combinations that do not meet expectations.

Images