Vehicle interaction method, vehicle interaction device and vehicle

Abstract

The invention provides a vehicle interaction method, a vehicle interaction device and a vehicle, and relates to the field of vehicles. The method comprises the following steps: acquiring vehicle data and user data of a vehicle; determining a user intention and a target function type of the vehicle based on the vehicle data and the user data; determining a target function based on the target function type; determining a target control instruction corresponding to the target function based on the user intention and the target function; and controlling the vehicle based on the target control instruction. The method can shorten the delay of system response so as to improve the user experience.

Description

Technical Field

[0001] The present application relates to the field of vehicles, and more specifically, to a vehicle interaction method, a vehicle interaction device, and a vehicle in the field of vehicles. Background Art

[0002] With the development of smart cars and connected car technologies, more and more vehicles are equipped with smart cockpits. Smart cockpit systems can combine user behavior and vehicle sensor data to provide users with intelligent decision-making.

[0003] After collecting user data, environmental data, and other relevant data, the vehicle inputs all of this data into the large model for processing. During the large model's data analysis and processing, if a large amount of data is input simultaneously, it may cause information overload. Furthermore, with the wide variety of in-vehicle functions, directly inputting all data into the large model increases the complexity of the model processing, resulting in lower data processing efficiency for the large model and, in turn, reduced responsiveness of the in-vehicle interactive system.

[0004] Therefore, how to shorten the system response delay to improve the user experience is an urgent problem to be solved. Summary of the Invention

[0005] The present application provides a vehicle interaction method, a vehicle interaction device and a vehicle, which can shorten the delay of system response to improve the user experience.

[0006] In a first aspect, a vehicle interaction method is provided, the method comprising:

[0007] Obtain vehicle data and user data of the vehicle;

[0008] Determine the user's intention and the target function type of the vehicle based on vehicle data and user data;

[0009] Determine the target function based on the target function type;

[0010] Based on user intention and target function, determine the target control instructions corresponding to the target function;

[0011] The vehicle is controlled based on the target control instructions.

[0012] In the above technical solution, the user intent and possible target function types of the vehicle user are determined by acquiring vehicle data and user data. Based on the user intent and target function types, the specific target function and the target control instructions corresponding to the target function are determined, and the vehicle is controlled according to the target control instructions. Compared to the prior art method of directly inputting vehicle data, user data, and all vehicle functions that the vehicle can provide into the model for processing to determine interaction information, the present application uses a two-step decision-making process to first determine the user intent and target function types corresponding to the vehicle data and user data, then select the target function from the target function types, and determine the corresponding target control instructions. This eliminates the need to select the function that matches the current scenario from all vehicle functions, narrowing the scope of data processing and avoiding the large amount of data to be processed simultaneously and the high number of functions involved in the decision-making process, which can lead to low processing efficiency. The present application uses data analysis to filter the types of vehicle functions and makes a second decision based only on the function types that match the current scenario, eliminating the need to process the decision based on all data. This reduces the amount of data required to be transmitted between the two decisions, shortening the time required for data processing, thereby reducing system response delays, outputting query information in a timely manner, and improving the user's vehicle experience.

[0013] In conjunction with the first aspect, in certain possible implementations, determining the user intent and the target function type of the vehicle based on the vehicle data and the user data includes:

[0014] Determine user intent based on vehicle data and user data;

[0015] Based on the user's intention and the vehicle's preset function types, a target function type is determined.

[0016] In the above technical solution, the user's intention is analyzed by combining multi-dimensional data including vehicle data and user data, which can improve the accuracy of user intention recognition; then, based on the user's intention and the preset function type, the target function type that matches the user's intention is determined from the preset function types, which can ensure the matching degree between the target function type and the user's intention, and thus improve the user's satisfaction with the interactive system; in the subsequent processing process, the function type that matches the user's intention can be processed in a targeted manner to reduce the amount of data processed by the system; in addition, the system only needs to search among the preset function types, which can reduce the time for the system to process and match functions, and further improve the response speed of the system.

[0017] In combination with the first aspect and the above implementations, in certain possible implementations, determining the user intent and the target function type of the vehicle based on the vehicle data and the user data includes:

[0018] The vehicle data, the user data and the first instruction are input into the interaction model to obtain the user intention and the target function type, wherein the first instruction is used to instruct the interaction model to output the user intention and the target function type.

[0019] In the above technical solution, after obtaining the vehicle data and user data, a corresponding first instruction can be determined based on the above data, and the output direction of the interaction model can be guided by the first instruction to ensure that the output content of the interaction model is the output content required for the interaction process; then the interaction model can analyze the input vehicle data and user data according to the first instruction to obtain the user intention and target function type. This process does not require manual input of the operations to be performed by the interaction model, and can avoid manual intervention or complex multi-step judgment processes, further shortening the response time of the system; with the help of the interaction model, by inputting vehicle data and user data into the model for comprehensive processing, it is possible to more comprehensively understand the user's needs and intentions, and ensure the accuracy of user intentions and target function types.

[0020] In combination with the first aspect and the above implementations, in some possible implementations, determining a target control instruction corresponding to the target function based on the user intention and the target function includes:

[0021] Based on the target function, determine the target configuration parameters corresponding to the target function;

[0022] Based on user intent and target configuration parameters, target control instructions are determined.

[0023] In the above technical solution, the specific target function of the vehicle is further determined based on the target function type. This avoids blind searches within the entire vehicle function library, narrows the search scope, and improves system efficiency. Based on the target function, the corresponding target configuration parameters are then determined. The system can then determine the target control instructions to be executed based on the user's intent and the target configuration parameters, ensuring that actual vehicle control meets the user's personalized settings. The system can also automatically analyze the user's current needs and provide decision recommendations, reducing user thinking and operation costs, improving interaction efficiency, and ultimately enhancing the user's vehicle experience.

[0024] In combination with the first aspect and the above implementations, in some possible implementations, determining target configuration parameters corresponding to the target function based on the target function includes:

[0025] Based on the target function, determining a target function interface corresponding to the target function from a candidate function interface library of the vehicle;

[0026] Obtain the target configuration parameters corresponding to the target function through the target function interface.

[0027] In this technical solution, the target function interface corresponding to the target function is determined from the vehicle's candidate function interface library. This eliminates the need for the system to blindly search within the complex vehicle functional architecture and directly locates the corresponding interface within the library based on the target function. This significantly reduces the time required to find the function interface, thereby improving the efficiency of the system invoking the target function. Obtaining the target configuration parameters corresponding to the target function through the target function interface ensures the accuracy and completeness of the data during the configuration parameter acquisition process. Furthermore, the target query information is determined based on the target configuration parameters, ensuring the accuracy of the target query information.

[0028] In combination with the first aspect and the above-mentioned implementation methods, in some possible implementation methods, the target control instruction is determined based on the user intention and the target configuration parameters, including: inputting the user intention, target configuration parameters, vehicle data, user data and a second instruction into the interaction model to obtain the target control instruction, wherein the second instruction is used to instruct the interaction model to output the target control instruction.

[0029] In the above technical solution, after analyzing the user intention and target configuration parameters, a corresponding second instruction can be determined based on the user intention, target configuration parameters, vehicle data and user data, and the second instruction is used to guide the interaction model to output the corresponding control instruction, avoiding the problem of diversity in the output content of the interaction model and ensuring the accuracy of the model output content; then the user intention, target configuration parameters, vehicle data, user data and the second instruction are input into the interaction model, and the interaction model can perform in-depth analysis based on the second instruction and the above information to output the final target control instruction; in addition, through the combination of user intention and target configuration parameters, the interaction model can more comprehensively and accurately understand the user's actual needs and the user's personalized configuration, thereby generating more accurate target control instructions, improving the success rate of user-system interaction, making the interaction process smoother, and thus improving the user's car experience.

[0030] In combination with the first aspect and the above implementation manner, in some possible implementation manners, the method further includes:

[0031] Obtaining data to be processed collected by vehicles;

[0032] The data to be processed is preprocessed to obtain vehicle data and user data, wherein the storage space occupied by the vehicle data and the user data is smaller than the storage space occupied by the data to be processed.

[0033] In the above technical solution, by pre-processing the data to be processed collected by the vehicle, the storage space occupied by the obtained vehicle data and user data is smaller than the data to be processed, redundant and unnecessary information is removed, and only key vehicle data and user data are retained, which greatly reduces the storage requirements of the system. When the vehicle data and user data are subsequently analyzed and processed (such as determining user intentions, matching functions, etc.), the amount of calculation is also reduced accordingly, allowing the system to complete data processing tasks in a shorter time, improving data processing efficiency and shortening the system response time.

[0034] In combination with the first aspect and the above implementation manner, in some possible implementation manners, the method further includes:

[0035] determining target query information corresponding to the target function based on the user intention and the target configuration parameters, and controlling the vehicle to output the target query information;

[0036] Based on the target control instructions, the vehicle is controlled, including:

[0037] When confirmation information of the target inquiry information is received, the vehicle is controlled based on the target control instruction.

[0038] In the above technical solution, based on the currently determined user intent and the target configuration parameters corresponding to the target function, the vehicle is first controlled to output the target query information corresponding to the target function, and the user's feedback is monitored. Upon receiving the user's confirmation information, the vehicle is further controlled accordingly based on the target control instructions. By outputting the target query information to determine the user's true intention, it is possible to avoid incorrect control of vehicle functions, improve the accuracy of vehicle function control and its matching with user needs. Only upon receiving the confirmation information, the vehicle computer is controlled, eliminating the need for manual operation by the user, which improves the convenience of vehicle computer control and enhances the user's driving experience.

[0039] In a second aspect, a vehicle interaction device is provided, the device comprising:

[0040] An acquisition module, used to acquire vehicle data and user data of a vehicle;

[0041] A first determination module is configured to determine a user intention and a target function type of the vehicle based on the vehicle data and the user data;

[0042] A second determining module is used to determine the target function based on the target function type;

[0043] A third determination module is used to determine a target control instruction corresponding to the target function based on the user intention and the target function;

[0044] The control module is used to control the vehicle based on the target control instructions.

[0045] In a third aspect, a vehicle is provided, comprising a memory and a processor. The memory is configured to store executable program code, and the processor is configured to retrieve and execute the executable program code from the memory, so that the vehicle executes the method of the first aspect or any possible implementation of the first aspect.

[0046] In a fourth aspect, a computer program product is provided, comprising: a computer program code, which, when executed on a computer, enables the computer to execute the method in the first aspect or any possible implementation of the first aspect.

[0047] In a fifth aspect, a computer-readable storage medium is provided, which stores a computer program code. When the computer program code runs on a computer, the computer executes the method in the above-mentioned first aspect or any possible implementation of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a data processing interaction diagram of a traditional in-vehicle interactive system provided by an embodiment of the present application;

[0049] Figure 2 is a schematic flow chart of a vehicle interaction method provided in an embodiment of the present application;

[0050] Figure 3 is an interaction diagram of a vehicle interaction method provided in an embodiment of the present application;

[0051] Figure 4 is a structural diagram of a vehicle interaction device provided in an embodiment of the present application;

[0052] Figure 5 It is a structural schematic diagram of a vehicle provided in an embodiment of the present application. DETAILED DESCRIPTION

[0053] The following will clearly and thoroughly describe the technical solutions in this application in conjunction with the accompanying drawings. In the description of the embodiments of this application, unless otherwise specified, " / " means or, for example, A / B can mean A or B: "and / or" in the text is only a description of the association relationship of associated objects, indicating that there can be three relationships, for example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more than two.

[0054] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be understood to imply or suggest relative importance or implicitly indicate the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features.

[0055] Today, more and more vehicles are equipped with intelligent cockpit in-vehicle interactive systems. With the development of artificial intelligence technology, in-vehicle interactive systems will become more intelligent, better able to understand user intentions and needs, and provide more personalized services. In-vehicle interactive systems can analyze relevant vehicle data and user behavior to meet user needs for vehicle functions.

[0056] Figure 1 This is a data processing interaction diagram of a traditional vehicle-mounted interactive system provided by an embodiment of the present application. Figure 1 As shown, during the data processing process of a traditional in-vehicle interactive system, the full amount of status data collected by the vehicle terminal 100 and the user instructions of the user 101 are directly sent to the traditional in-vehicle interactive system 102. The traditional in-vehicle interactive system 102 then processes all the data at once to obtain the operation instructions; the operation instructions are then returned to the vehicle terminal 100. This process involves a large amount of data to be processed. Directly transmitting such a large amount of data to the traditional in-vehicle interactive system 102 may cause the amount of data to be processed to exceed the data processing context window, for example, exceeding the 16k or 64k limit, resulting in a delay in the system's response. At the same time, due to the wide variety of in-vehicle functions, if the complete function description and status data are input into the traditional in-vehicle interactive system 102 for processing at once, it will not only increase the processing complexity but also reduce the robustness of the system, affecting the decision-making results.

[0057] In view of the above-mentioned problems in the existing technology, the embodiments of the present application provide a vehicle interaction method, a vehicle interaction device, and a vehicle. The method first obtains vehicle data and user data; then, based on the vehicle data and user data, determines the corresponding user intent and the target function type of the vehicle; then, based on the target function type, determines the target function under that type, and determines the corresponding target control instructions based on the target function and user intent; then, controls the vehicle according to the target control instructions. This method can shorten the response time of the in-vehicle interaction system and improve the user's vehicle experience.

[0058] The following combination Figures 2 to 3 The vehicle interaction method provided in the embodiment of the present application is described in detail.

[0059] Figure 2This is a schematic flow chart of a vehicle interaction method provided by an embodiment of the present application. It should be understood that the method can be applied to a vehicle; or, to a processor in a vehicle; or, to a chip mounted in a processor in a vehicle. Alternatively, the method can be applied to a server; or, to a processor in a server; or, to a chip mounted in a processor in a server.

[0060] For example, Figure 2 As shown, the method 200 includes:

[0061] S201, obtaining vehicle data and user data of the vehicle.

[0062] For example, during vehicle operation, sensors, cameras, and other monitoring devices installed on the vehicle continuously collect and monitor relevant vehicle data, thereby obtaining vehicle data and user data. Vehicle data may include current time data, vehicle location data, vehicle speed, and various sensor data; user data may include user behavior data, such as voice input or actions.

[0063] For example, the acquired vehicle data may include that the vehicle's current location is the user's home, the seat pressure sensor detects that the user has entered the vehicle, and the temperature sensor detects that the current temperature inside the vehicle is 25 degrees; the user data may include that no input voice is detected, that the user has entered the vehicle, etc.

[0064] In one implementation, data to be processed collected by a vehicle is obtained; the data to be processed is preprocessed to obtain vehicle data and user data, wherein the storage space occupied by the vehicle data and the user data is smaller than the storage space occupied by the data to be processed.

[0065] For example, the amount of data collected by the vehicle is large. In order to avoid response delays caused by the large amount of calculation in subsequent processing, the collected data to be processed can be preprocessed to obtain vehicle data and user data that occupy less storage space.

[0066] For example, the collected data to be processed may contain structured data, which occupies a large amount of storage space. Therefore, the structured data can be simplified to remove the data structure and reduce the memory usage. For example, if the data to be processed is "timestamp": "07:30" and "location": "home", the simplified data will be "07:30" and "home". The collected data to be processed may also contain noise and other influencing factors. During preprocessing, the data to be processed can also be denoised to improve data accuracy.

[0067] It should be understood that when the method of the embodiment of the present application is applied locally to the vehicle, the collected data to be processed can be processed directly locally in the vehicle to obtain vehicle data and user data; when the method is applied to the server, the vehicle can upload the collected data to be processed to the server in real time. After the server processes the data to be processed, it can perform subsequent processing on the server, or send the obtained vehicle data and user data to the vehicle for subsequent processing.

[0068] In an embodiment of the present application, by pre-processing the data to be processed collected from the vehicle, the storage space occupied by the obtained vehicle data and user data is smaller than the data to be processed, redundant and unnecessary information is removed, and only key vehicle data and user data are retained, which greatly reduces the storage requirements of the system. When the vehicle data and user data are subsequently analyzed and processed (such as determining user intentions, matching functions, etc.), the amount of calculation is also reduced accordingly, allowing the system to complete data processing tasks in a shorter time, improving the efficiency of data processing, and shortening the response time of the system.

[0069] S202 : Determine the user intention and the target function type of the vehicle based on the vehicle data and the user data.

[0070] For example, after obtaining the vehicle data and the user data, the user's current user intention may be analyzed based on the vehicle data and the user data, and the target function type of the vehicle that may currently need to be controlled may be determined.

[0071] The vehicle's functional type is obtained by classifying all functions that the vehicle can provide. For example, the vehicle's functional type may include an air conditioning control function type, a navigation control function type, and a media entertainment function type.

[0072] For example, if the vehicle data and user data indicate that the user leaves home and drives during working hours in the morning, it can be predicted that the user is currently in a commuting scenario, and the user intention indicates that the user is in the preparation stage for commuting to work; during the commuting process, it can be determined that the target function types that may be used include navigation control functions and entertainment media functions.

[0073] For example, if the current vehicle data indicates that the vehicle has turned on navigation or the user has turned on navigation, the navigation control function may not be used as the target function type to be controlled currently, so as to avoid the deviation between the predicted user intention and the user's actual intention, resulting in errors in vehicle function control and affecting the user's normal use.

[0074] In one implementation, a large amount of data can be analyzed in advance to classify different vehicle data and user data to determine the correspondence between them and user intentions and function types; and then, based on the above correspondence, the user intentions and target function types corresponding to the vehicle data and user data can be determined.

[0075] In one implementation, the process of determining the user intention and the target function type of the vehicle based on the vehicle data and the user data may specifically include:

[0076] Determine user intent based on vehicle data and user data;

[0077] Based on the user's intention and the vehicle's preset function types, a target function type is determined.

[0078] Among them, the vehicle's preset function types are used to indicate the function types that the current vehicle can provide to users. For example, the preset function types may include navigation control function, entertainment media function, air conditioning control function, and seat adjustment function.

[0079] For example, vehicle data and user data can reflect the user's current vehicle usage scenario. In this scenario, the user may not have yet input voice commands or control operations to control vehicle functions, or the user may not have given clear operating instructions. In this case, the user's intention can be predicted. Accordingly, after obtaining vehicle data and user data, the user's intention is first analyzed by combining these data. Then, from the preset function types available in the vehicle, the target function type that meets the user's intention is selected. This enables early prediction of user operations and facilitates the control of functions that meet the user's needs.

[0080] For example, if the current user intention indicates that the user needs to rest based on the analysis of vehicle data and user data, a target function type that can meet the user's rest needs can be selected from the vehicle's preset function types, such as a seat adjustment function, an ambient light adjustment function, or a cabin mode adjustment function.

[0081] In one implementation, a set of intent labels corresponding to various preset function types is preconfigured. The correspondence between these sets of intent labels can be stored locally in the vehicle or on a cloud platform. After determining the user's intent, the intent labels in the user's intent are analyzed to determine the degree of match between the intent labels and the sets of intent labels corresponding to the various preset function types. The function type with the highest degree of match is then selected as the target function type.

[0082] For example, the intent tag set corresponding to the navigation control function includes "going home," "going to work," and "traveling"; the intent tag set corresponding to the entertainment media function includes "commuting," "resting," and "entertainment"; the intent tag set corresponding to the air conditioning control function includes "high temperature" and "low temperature"; and the intent tag set corresponding to the seat adjustment function includes "resting" and "user change." If the user intent analyzed based on the currently acquired vehicle and user data indicates that the user is preparing for a commute to work, and the intent tags are determined to include "going to work" and "commuting," then the target function types matching the intent tags can be determined from the preset correspondence to be navigation control and entertainment media. If the user intent analyzed based on the currently acquired vehicle and user data indicates that the user is taking a nap and the current temperature is high, then the intent tags can be determined to include "resting" and "high temperature," and the corresponding target function types can be further determined to include air conditioning control and seat adjustment.

[0083] In an embodiment of the present application, the user's intention is analyzed in combination with multi-dimensional data including vehicle data and user data, which can improve the accuracy of user intention recognition; then, based on the user's intention and the preset function type, the target function type that matches the user's intention is determined from the preset function types, which can ensure the matching degree between the target function type and the user's intention, and thus improve the user's satisfaction with the interactive system; in the subsequent processing process, the function type that matches the user's intention can be targeted and processed to reduce the amount of data processed by the system; in addition, the system only needs to search among the preset function types, which can reduce the time for the system to process and match functions, and further improve the response speed of the system.

[0084] In one implementation, an interaction model may be used for data processing. Specifically, based on vehicle data and user data, the process of determining the user's intention and the target function type of the vehicle may include:

[0085] The vehicle data, the user data and the first instruction are input into the interaction model to obtain the user intention and the target function type, wherein the first instruction is used to instruct the interaction model to output the user intention and the target function type.

[0086] For example, when using an interactive model for data processing, only the data to be processed is input into the model, and the output of the model may not conform to the output content required by the current scenario. Therefore, it is necessary to combine vehicle data and user data to determine the corresponding instruction information - the first instruction, which is used to instruct the interactive model to output the user intention and target function type after analyzing the input user data and vehicle data; thereafter, the vehicle data, user data and the first instruction can be integrated into the prompt word prompt of the input interactive model, and input into the interactive model for processing, so that the user intention and target function type corresponding to the vehicle data and user data output by the interactive model can be obtained.

[0087] Exemplarily, after receiving the following data: "timestamp":"07:30", "location":"home" and "user_action":"get in the car", the system determines the data name of the above data, determines that the above data is vehicle data and user data to be processed, and then determines that the operation that needs to be performed by the interaction model is to output the user intention and target function type based on the above data, and generates a first instruction "Please output the user intention and function type based on the above information"; thereafter, the data such as "timestamp":"07:30", "location":"home" and "user_action":"get in the car" and the first instruction "Please output the user intention and vehicle function type based on the above information" are input into the interaction model as prompt words. The interaction model can analyze and process the input data according to the operation indicated by the first instruction, and output the user intention and target function type that match the input data.

[0088] In one implementation, vehicle data, user data, and preset function types of the vehicle are simultaneously input into the interaction model. The interaction model can determine the target function type that matches the vehicle data and user data from the preset function types based on the above data.

[0089] In an embodiment of the present application, after obtaining vehicle data and user data, a corresponding first instruction can be determined based on the above data, and the output direction of the interaction model can be guided by the first instruction to ensure that the output content of the interaction model is the output content required for the interaction process; the interaction model can then analyze the input vehicle data and user data according to the first instruction to obtain the user intention and target function type. This process does not require manual input of the operations to be performed by the interaction model, and can avoid manual intervention or complex multi-step judgment processes, further shortening the response time of the system; with the help of the interaction model, by inputting vehicle data and user data into the model for comprehensive processing, it is possible to more comprehensively understand the user's needs and intentions, and ensure the accuracy of the user's intentions and target function types.

[0090] S203: Determine the target function based on the target function type.

[0091] For example, each function type may include multiple different functions. For example, the air conditioning control function type may include air conditioning wind speed control, air flow control, and temperature control functions; the navigation control function type may include navigation, traffic query, and route optimization functions; and the media entertainment function type may include news playback, music playback, and volume adjustment functions. After determining the target function type of the vehicle based on the collected vehicle data and user data, the target functions included in the target function type may be further determined.

[0092] For example, if it is determined that the target function type is a navigation control function, it can be determined that the target functions include a navigation function, a traffic query function, a route optimization function, and the like.

[0093] S204: Determine a target control instruction corresponding to the target function based on the user intention and the target function.

[0094] For example, after determining the corresponding user intention and target function type based on vehicle data and user data, the target function included in the target function type can be further determined, and then the target control instructions corresponding to the target function can be determined based on the analyzed user intention and target function to achieve control of the target function.

[0095] In one implementation, based on the user intention and the target function, the process of determining the target control instruction corresponding to the target function may specifically include:

[0096] Based on the target function, determine the target configuration parameters corresponding to the target function;

[0097] Based on user intent and target configuration parameters, target control instructions are determined.

[0098] For example, after obtaining the target function type, the specific target function that the vehicle can provide in the target function type can be determined; then, the target configuration parameters under each target function can be determined; then, the corresponding target control instructions can be determined in combination with the user intention and the target configuration parameters.

[0099] Configuration parameters are used to indicate parameter settings related to each target function. For example, for the wind speed control function, air outlet mode control function, temperature control function, etc. in the air conditioning control function type, the corresponding configuration parameters may include the user's usual parameter settings in different vehicle usage scenarios, such as a usual temperature of 25°C; for the navigation function in the navigation control function type, the corresponding configuration parameters may include the user's destination information, such as the location of home or company, and the user's usual route preferences for navigation, such as shortest travel time, priority for highways, or no highways; for the music playback function in the media entertainment function type, the corresponding configuration parameters may include style preferences, language preferences, etc.

[0100] For example, taking the navigation control function as the target function type, the target functions included under the navigation control function type include navigation, real-time traffic query, and route optimization. After determining that the vehicle can provide the above target functions, it is possible to further obtain specific information under each function and target configuration parameters such as user preferences. For example, under the navigation function, it is necessary to obtain the user's target address and the user's route preferences when navigating, such as shortest travel time, priority for highway sections, or no highway sections. Then, combined with the current user intent and target configuration parameters, a target control instruction that matches the current scenario and user preferences can be determined. For example, when it is determined that the user intent is to commute to work and the route preference is determined to be the shortest travel time, the corresponding target control instruction may be "the navigation application navigates to the company using the shortest route preference."

[0101] In this embodiment of the present application, the specific target function of the vehicle is further determined based on the target function type. This avoids blindly searching the entire vehicle function library, narrows the search scope, and improves the system's operational efficiency. Based on the target function, the corresponding target configuration parameters are then determined. The system can then determine the target control instructions to be executed based on the user's intent and the target configuration parameters, ensuring that actual vehicle control meets the user's personalized settings. The system can also automatically analyze the user's current needs and output decision recommendations, reducing the user's thinking and operating costs, improving interaction efficiency, and ultimately enhancing the user's vehicle experience.

[0102] In one implementation, based on the target function, a specific process of determining the target configuration parameters corresponding to the target function may include:

[0103] Based on the target function, determining a target function interface corresponding to the target function from a candidate function interface library of the vehicle;

[0104] Obtain the target configuration parameters corresponding to the target function through the target function interface.

[0105] For example, to standardize the data transmission format, parameter types, and calling rules in a vehicle, different functional interfaces are typically configured for different functions in the vehicle. These interfaces serve as standardized interaction channels connecting upper-layer applications (such as functions triggered by user intent) with underlying hardware / software resources. For all functions a vehicle can provide, corresponding functional interfaces exist, and the correspondence between multiple functional interfaces and functions is pre-stored in a candidate functional interface library in the vehicle. For example, the candidate functional interface library includes the functional interface corresponding to the navigation function, called navigate_function, and the target interface corresponding to the seat adjustment function, called seat_adjust_function.

[0106] For example, different vehicle functions and related parameters correspond to different functional interfaces. A candidate functional interface library may be pre-stored in the vehicle, storing the functional interfaces of the vehicle's available in-vehicle functions. After determining the vehicle's target function, the target functional interface corresponding to the target function can be retrieved from the candidate functional interface library. The target configuration parameters corresponding to the target function can then be obtained from the target functional interface.

[0107] Exemplarily, after determining the target function interface that matches the target function, the metadata corresponding to the interface can be read from the vehicle's system configuration file or database through the name of the target function interface. For example, for the charging pile recommendation function, the current power parameters required to use the charging pile recommendation function and the user's preference setting parameters for the function (for example, distance priority) can be obtained from the vehicle data and user data corresponding to the function interface through the function interface corresponding to the charging pile recommendation function.

[0108] In an embodiment of the present application, the target function interface corresponding to the target function is determined from the vehicle's candidate function interface library. The system does not need to blindly search within the complex vehicle functional architecture, but can directly locate the corresponding interface in the library based on the target function. This significantly reduces the time required to find the function interface, thereby improving the efficiency of the system invoking the target function. By obtaining the target configuration parameters corresponding to the target function through the target function interface, the accuracy and completeness of the data during the configuration parameter acquisition process can be ensured. Furthermore, the target query information can be determined based on the target configuration parameters, thereby ensuring the accuracy of the target query information.

[0109] In one implementation, when processing using an interaction model, the process of determining a target control instruction based on user intent and target configuration parameters may specifically include:

[0110] The user intention, target configuration parameters, vehicle data, user data and a second instruction are input into the interaction model to obtain a target control instruction, wherein the second instruction is used to instruct the interaction model to output the target control instruction.

[0111] Exemplarily, when the above data is processed using the interaction model, the vehicle data and user data are first input into the interaction model, and the interaction model outputs the user intention and target function type accordingly; then, the target configuration parameters corresponding to the target function are determined according to the target function type; and then, the corresponding second instruction can be determined based on the acquired original data - vehicle data, user data and the user intention and target configuration parameters determined based on the above data, to instruct the interaction model to output the corresponding target control instruction based on the above information; after determining the second instruction, the vehicle data, user data, user intention, target configuration parameters and the second instruction are integrated into the prompt word prompt input into the interaction model, and input into the interaction model again, and the interaction model processes the input content according to the second instruction and outputs the corresponding target control instruction.

[0112] Exemplarily, after determining the user intention and target configuration parameters, it is detected that the current data to be processed includes vehicle data and user data "timestamp":"07:30", "location":"home" and "user_action":"get in the car", etc., the user intention is "commuting preparation", and the target configuration parameters include {"navigateTo":{"target":"company address","routePreference":"fastest"},"queryRoadStatus":{}} and {"playMedia":{"content":"news channel"},"adjustVolume":{"level":"automatic"}}. The system analyzes the data name of the above data to be processed and determines that it includes vehicle data, user data, user intention and configuration parameters corresponding to different functions. Then, it can be determined that the operation to be executed by the interaction model is to output control instructions according to the above data, and generate a second instruction "Please output the final vehicle control instruction based on the above information"; thereafter, the above output to be processed and the second instruction are simultaneously input into the interaction model for analysis and processing, and the target control instruction output by the interaction model can be obtained.

[0113] In an embodiment of the present application, after analyzing the user intention and target configuration parameters, a corresponding second instruction can be determined based on the user intention, target configuration parameters, vehicle data and user data, and the second instruction is used to guide the interaction model to output the corresponding control instruction, thereby avoiding the problem of diversity in the output content of the interaction model and ensuring the accuracy of the model output content; then the user intention, target configuration parameters, vehicle data, user data and the second instruction are input into the interaction model, and the interaction model can perform in-depth analysis based on the second instruction and the above information, and output the final target control instruction; in addition, through the combination of user intention and target configuration parameters, the interaction model can more comprehensively and accurately understand the user's actual needs and the user's personalized configuration, thereby generating more accurate target control instructions, improving the success rate of user-system interaction, making the interaction process smoother, and thus improving the user's car experience.

[0114] Compared with the existing technology in which all vehicle-related data and data on all vehicle functions are input into the model as keywords for processing, this application uses the interactive model to process vehicle data and user data to obtain user intentions and target function types to achieve the first decision; and uses the interactive model to process user intentions and target function types to obtain target control instructions to achieve the second decision; there is no need to retrain the existing model on a large scale. By simplifying the keyword information input into the interactive model each time, the response time of the vehicle system can be shortened, the user experience can be improved, and thus low-cost optimization of the vehicle system can be achieved.

[0115] S205: Control the vehicle based on the target control instruction.

[0116] For example, after a final target control instruction is determined based on current data of the vehicle, a target function in the vehicle may be controlled based on the target control instruction.

[0117] For example, when it is determined that the target control instruction that meets the current scenario is "the navigation application navigates to the company with the shortest route preference", the navigation application in the vehicle or the navigation application downloaded by the user can be automatically controlled according to the target control instruction, so that the navigation application can use the company address pre-stored by the user as the destination and plan the route with the shortest time.

[0118] In one implementation, based on the user intention and the target configuration parameters, target query information corresponding to the target function is determined, and the vehicle is controlled to output the target query information;

[0119] Based on the target control instructions, the vehicle is controlled, including:

[0120] When confirmation information of the target inquiry information is received, the vehicle is controlled based on the target control instruction.

[0121] For example, after analyzing and obtaining the target function to be controlled, a target control instruction corresponding to the target function is determined. To further ensure that vehicle control meets the user's actual needs, before controlling the vehicle according to the target control instruction, corresponding target query information may be determined based on the user's intent and the target configuration parameters corresponding to the target function. The target query information is output via the vehicle's audio system or display screen to facilitate interaction between the vehicle and the user to determine whether to control the vehicle's target function according to the target control instruction. For example, if the user's intent is determined to be commuting, the target functions include navigation and route planning, the corresponding target configuration parameters include the company address, and the route preference is shortest time, the corresponding target query information may be "whether to navigate to the company via the shortest route." After the vehicle is controlled to output the target query information, user feedback is monitored in real time. When a confirmation message is received from the user regarding the target query information, indicating that the function control requirement indicated by the target query information exists, the relevant vehicle function may be controlled according to the target control instruction.

[0122] For example, based on the user's intention and the target function type, the target control instruction is determined to be navigating to the company by the shortest route. The vehicle outputs the target inquiry information "Do you want to navigate to the company by the shortest route?" If the user's confirmation information is detected, such as the user inputs the voice "OK", the vehicle's navigation system can be controlled to plan the route with the company as the destination and the preference setting of the shortest time.

[0123] In one implementation, the process of determining the target query information can also be combined with the interaction model, that is, user intention, target configuration parameters, vehicle data, user data and a third instruction are input into the interaction model, and the third instruction is used to instruct the interaction model to output the target query information; after the above prompt words are input into the interaction model at the same time, the interaction model can analyze the input data according to the operation indicated by the third instruction, and then obtain the corresponding target query information.

[0124] Optionally, if no confirmation information from the user is detected within a preset time period, the target inquiry information may be output again for inquiry.

[0125] Optionally, if it is detected that the user rejects the target query information, the vehicle may not be controlled; or, if it is detected that the user inputs new instruction information, the vehicle may be controlled according to the new instruction information.

[0126] For example, after the target query information is output, if a user voice input of "Let's take the highway today" is detected, the vehicle's navigation system can be controlled to plan the route with a preference setting of highway priority.

[0127] It should be understood that when the method of the embodiment of the present application is applied locally to the vehicle, the vehicle can be directly controlled to output the target query information; when the method is applied to the server, after determining the target query information, the server can send the target query information to the vehicle, and then control the vehicle to output the target query information.

[0128] Optionally, the target inquiry information may also be referred to as interactive speech, inquiry speech, etc., which is used for interaction with the user.

[0129] In the above technical solution, based on the currently determined user intent and target function type, the vehicle is first controlled to output target query information corresponding to the target function, and the user's feedback is monitored. Upon receiving the user's confirmation, the vehicle is further controlled accordingly based on the target control instructions. By outputting the target query information to determine the user's true intention, incorrect control of vehicle functions can be avoided, improving the accuracy of vehicle function control and its matching with user needs. Only upon receiving confirmation, vehicle control is performed, eliminating the need for manual user operation, which improves the convenience of vehicle control and enhances the user's driving experience.

[0130] In one implementation, after the target control instruction is determined, current scene information of the vehicle is acquired; and when it is determined that there is no conflict between the target control instruction and the current scene information, the vehicle is controlled to output target query information.

[0131] For example, during the process of determining a target control instruction based on vehicle data and user data, the vehicle data and user data may be constantly changing. If the target control instruction is modified in real time based on the changing vehicle data and user data, it may be impossible to determine the final control instruction, thereby causing a significant delay in the interaction between the vehicle and the user. To shorten the delay in human-vehicle interaction, after determining the target control instruction based on vehicle data and user data within a historical period (e.g., 2 seconds or 5 seconds), the vehicle's current scene information can be obtained; then, a determination can be made as to whether there is a conflict between the determined target control instruction and the current scene information; if there is no conflict, the vehicle is controlled to output target query information.

[0132] Optionally, if there is a conflict between the target control instruction and the current scene information, the vehicle may not be controlled; or the target control instruction may be updated and the vehicle may be controlled based on the updated target control instruction; or the target query information may be updated according to the current scene information of the vehicle and the updated target query information may be output.

[0133] For example, the conflict between the target control command and the current scene information may include a safety conflict. Based on relevant data from a historical period, it is determined that there is a current need to open the vehicle windows for ventilation. The corresponding target control command is determined to open all vehicle windows, and the target query information is "Should all vehicle windows be opened?"; if, before controlling the vehicle windows, the vehicle's current scene information indicates that a child is boarding the vehicle, and opening all windows could pose a safety risk to the child, it may be determined that there is a conflict between the target control command and the current scene information. In this case, the vehicle may not be controlled according to the current target control command. Alternatively, based on the child's position, the target window closest to the child's position may be determined, and the updated target control command may be determined to open all windows except the target window. Alternatively, the query information "Should all windows be opened except the window on the child's side" may be output.

[0134] For example, the conflict between the target control command and the current scene information may also include a hardware conflict. If the target control command is determined to open all vehicle windows, and if it is detected that the user is currently closing the right rear window, controlling the windows directly according to the command to open all vehicle windows may cause the window motor to malfunction. Therefore, the target control command may be updated to open all windows except the right rear window; or the target query information may be updated from "whether to open all vehicle windows" to "whether to open the front and left rear windows."

[0135] It should be noted that in the process of updating and adjusting the target control instructions or target query information, it is necessary to combine the actual scenario of the current vehicle, and the updated control instructions and query information need to avoid safety risks and failure risks.

[0136] In this embodiment of the present application, after determining the target control command, a determination is made as to whether there is a conflict between the target control command and the current scene information. If no conflict exists between the target control command and the current scene information, the vehicle is controlled to output the target query information, thereby avoiding vehicle anomalies caused by control conflicts during the vehicle control process, or outputting inappropriate query information that affects the user's vehicle experience and satisfaction. Furthermore, conflict determination can include safety conflicts, enabling detection of user intent while ensuring vehicle safety, thereby achieving more flexible and intelligent vehicle control.

[0137] In summary, in the embodiments of the present application, the user intent and the target function types that may be involved are determined by acquiring vehicle data and user data. Based on the user intent and target function types, the specific target function and the target control instructions corresponding to the target function are determined, and the vehicle is controlled according to the target control instructions. Compared to the prior art method of directly inputting vehicle data, user data, and all vehicle functions that the vehicle can provide into the model for processing to determine interaction information, the present application uses a two-step decision-making process to first determine the user intent and target function types corresponding to the vehicle data and user data, then select the target function from the target function types, and determine the corresponding target control instructions. This eliminates the need to select the function that matches the current scenario from all vehicle functions, narrowing the scope of data processing and avoiding the large amount of data to be processed simultaneously and the high number of functions involved in the decision-making process, which can lead to low processing efficiency. The present application uses data analysis to filter the types of vehicle functions and makes a second decision based only on the function types that match the current scenario, eliminating the need to process the decision based on all data. This reduces the amount of data required to be transmitted between the two decisions, shortening the time required for data processing, thereby reducing system response delays, outputting query information in a timely manner, and improving the user's vehicle experience.

[0138] Figure 3 It is an interactive diagram of a vehicle interaction method provided by an embodiment of the present application. It should be understood that the interactive process of the method involves the vehicle terminal 100, the user 101, the server 103, the large language model 104 and the functional interface library 105. Among them, the large language model 104 is equivalent to Figure 2 The interactive model in is used to indicate a model that has been pre-trained with large-scale data and has strong generation and understanding capabilities.

[0139] For example, Figure 3 As shown, the method 300 includes four stages, wherein the first stage is the intention recognition stage, the second stage is the functional interface retrieval stage, the third stage is the fine-grained decision stage, and the fourth stage is the execution feedback stage.

[0140] The first stage includes the following steps S301 to S303.

[0141] S301, sending multimodal status data.

[0142] Multimodal state data, or the vehicle's pending data, can include pending vehicle data, user data, and other relevant data that may impact human-vehicle interaction. For example, vehicle data can include in-vehicle and out-of-vehicle environment data, vehicle speed data, location data, and various sensor data; user data can include user behavior data and identity data; and multimodal state data can also include time data.

[0143] For example, during the driving of the vehicle, the vehicle terminal 100 will collect relevant data of the vehicle in real time; and the vehicle terminal 100 will send the multimodal status data to the server 103 for processing.

[0144] For example, the multimodal status data collected in real time by the vehicle-mounted terminal 100 may include time information, location data, sensor data, and user behavior data. For example, the time information indicates 7:30 AM. The location data indicates that the GPS positioning shows the current location as "home." The sensor data indicates that the seat pressure sensor triggered, apparently detecting that the user has entered the vehicle; the ambient temperature in the vehicle is 22°C, and the humidity in the vehicle is 50°C; the driving assistance camera detected the road conditions ahead, which is used for safety warnings. The user behavior data indicates that there was no voice input, indicating that the user did not actively issue a clear command; the vehicle-mounted system automatically detected the user's entry into the vehicle.

[0145] For example, after the above-mentioned relevant data is collected, the multimodal state data is sent from the vehicle terminal 100 to the server terminal 103 in a structured format. The structured format may be JSON format or the like.

[0146] The above data is represented in JSON format as follows:

[0147]

[0148]

[0149] Among them, the above data is used to indicate that the current time is "7:30"; the location is "home"; the seat pressure sensor detects the presence of a user; the temperature is 22; the humidity is 50; the road condition is "smooth"; and the user's boarding behavior is detected.

[0150] S302, input simplified state data and preset tool set.

[0151] Among them, the simplified state data is used to indicate the vehicle data and user data obtained after preprocessing the multimodal state data. The storage space occupied by the simplified state data is smaller than the storage space occupied by the multimodal state data; the preset tool set is also a collection of preset function types of the vehicle, which is used to indicate the function types that the current vehicle can provide to the user. For example, the preset tool set may include navigation control function, entertainment media function, air conditioning control function, and seat adjustment function, etc.

[0152] For example, after receiving the multimodal status data from the vehicle-mounted device 100, the server 103 can simplify the multimodal status data to extract key information, as the data is in a structured format and consumes a lot of memory and computation. For example, if the structured data is "timestamp": "07:30", the simplified data is "07:30", removing the structured format and reducing the storage space.

[0153] After the multimodal state data is simplified by the server 103, simplified state data can be obtained; then, the server 103 can form the simplified state data and the preset tool set into the input text (Prompt) of the first-level request, and input it into the large language model 104 for processing.

[0154] In one implementation, the content sent to the large language model 104 further includes a first instruction, which is used to instruct the large language model to output high-level intent and function classification based on the above information.

[0155] S303, returning high-level intent and function classification.

[0156] For example, after receiving the simplified state data and preset tool set sent by the server 103, the large language model 104 will analyze the above information to obtain the high-level intent corresponding to the simplified state data, that is, the user intent, and the recommended function classification, that is, the target function type.

[0157] For example, according to the example data in S301 , it can be determined that the high-level intent “high_level_intent” is that the user is in the stage of preparing for commuting to work; and further, it can be determined that the suggested function categories “suggested_function_categories” include navigation control functions and media entertainment functions.

[0158] Illustratively, after obtaining the high-level intent and function classification, the large language model 104 returns the above information to the server 103 .

[0159] In one implementation, if the information received by the large language model 104 does not contain instruction information, the corresponding first instruction can be determined based on the type of data input into the large language model; and the received simplified state data and preset tool set are processed according to the first instruction to obtain the corresponding high-level intent and functional classification.

[0160] The second stage includes the following steps S304 and S305.

[0161] S304, querying the functional interface according to the functional classification.

[0162] For example, after obtaining the function classification output by the large language model 104, the server 103 can query the function interface library 105 for a specific function interface based on the function classification. The function interface library 105 stores the specific function interface units and related parameters that the current vehicle can provide.

[0163] For example, after determining that the functional classification includes navigation control function and media entertainment function, the functional interface library 105 can be queried to obtain the functional interfaces such as "navigation", "real-time traffic query", "route optimization" and corresponding configuration parameters in the navigation control module; the media entertainment module contains functional interfaces such as "play news", "play music", "adjust volume" and corresponding configuration parameters.

[0164] S305: Return a specific functional interface list.

[0165] Exemplarily, the function interface library 105 may return a specific function interface list to the server 103 .

[0166] For example, the functional interface list includes the following content:

[0167] {

[0168] "Navigation Control":[

[0169] {"interface":"navigateTo","params":{"target":"address","routePreference":"fastest"}},

[0170] {"interface":"queryRoadStatus","params":{}}

[0171] ],

[0172] "Media Entertainment":[

[0173] {"interface":"playMedia","params":{"content":"News Channel"}},

[0174] {"interface":"adjustVolume","params":{"level":"automatic"}} ]

[0176] }.

[0177] The interface and parameters in the navigation control module indicate navigating to the aforementioned "address," with the route preference set to "fastest," and performing a traffic query. The interface and parameters in the media entertainment module indicate playing the "news channel" media content, with the volume adjustment mode set to automatic.

[0178] The third stage includes the following steps S306 to S309.

[0179] S306 , input high-level intent, functional interface list and simplified state data.

[0180] Exemplarily, after the functional interface library 105 returns the functional interface list to the server 103, the server 103 integrates the high-level intent returned by the large language model 104 in the first stage, the functional interface list returned by the functional interface library 105, and the original simplified state data into a new Prompt, and re-inputs the new Prompt into the large language model 104 for processing by the large language model 104.

[0181] For example, the information input by the server 103 into the large language model 104 may include the following: the scenario description includes the user getting in the car in the morning, the time is 07:30, the location is home, the sensor data in the car is normal (temperature 22°C, humidity 50%, smooth road conditions); the user did not explicitly issue a voice command. The high-level intention includes the user's intention of "commuting preparation" after preliminary analysis. The list of available functional interfaces includes the navigation control module: {"navigateTo":{"target":"Company Address","routePreference":"Fastest"},"queryRoadStatus":{}} Media entertainment module: {"playMedia":{"content":"News Channel"},"adjustVolume":{"level":"Automatic"}}.

[0182] In one implementation, the content sent to the large language model 104 also includes a second instruction, which is used to instruct the large language model to output final operation instructions and interactive dialogue based on the above information.

[0183] S307, returning the operation instruction and inquiry information.

[0184] Among them, the operation instruction is equivalent to Figure 2 The target control instruction, the query information is equivalent to Figure 2 Target query information in.

[0185] Exemplarily, the large language model 104 processes the high-level intent, the functional interface list, and the status data to obtain corresponding operation instructions and query information; thereafter, the operation instructions and query information may be returned to the server 103 .

[0186] For example, the operation instructions and query information returned by the large language model 104 may include:

[0187] {

[0188] "actions":[

[0189] {"type":"navigation","interface":"navigateTo","parameters":{"target":"Company Address","routePreference":"Fastest"}},

[0190] {"type":"media","interface":"playMedia","parameters":{"content":"News Channel"}}

[0191] ],

[0192] "message":"Good morning, are you ready for work? Do you want me to guide you to the office and play the morning news?"

[0193] }.

[0194] The operation instruction is used to instruct the user to navigate to the user's "company address" with the "fastest" navigation preference and to control the playback of the "news channel." The query information corresponds to the operation instruction and asks the user whether to plan the navigation and play the news.

[0195] In one implementation, if the information received by the large language model 104 does not contain instruction information, a corresponding second instruction can be determined based on the type of data input into the large language model; and the received high-level intent, functional interface list, and status data are processed according to the second instruction to obtain corresponding operation instructions and query information.

[0196] S308: Perform security verification on the operation instruction.

[0197] For example, after receiving the operation instructions generated by the large language model 104, the server 103 can perform security verification on the operation instructions and monitor the hardware status of the vehicle to avoid affecting the safety of the vehicle and the user due to safety hazards in the operation instructions or hardware abnormalities of the vehicle during the process of the vehicle performing function control according to the operation instructions.

[0198] For example, if the current vehicle speed is high and the operation instruction instructs to control the door to be unlocked or opened, it can be determined that there is a safety problem with the current operation instruction.

[0199] S309, returning the operation instruction and inquiry information.

[0200] For example, if it is detected that the current operation instruction has a security problem, the operation instruction will not be sent to the vehicle terminal 100; when it is determined that the operation instruction does not have a security problem, the service terminal 103 can return the operation instruction and query information to the vehicle terminal 100.

[0201] The fourth stage includes the following steps S310 and S311.

[0202] S310, execution result feedback.

[0203] For example, after receiving the query information and the operation instruction, the vehicle terminal 100 may output the query information to obtain the user's response. After the vehicle terminal 100 completes the execution, the execution result may be fed back to the server 103.

[0204] For example, if confirmation of the inquiry information is received from user 101, instructing the vehicle to navigate to the company and play a news channel, the vehicle can be directly controlled according to the operation instruction, and the execution result can indicate that the operation is successful or the matching degree is high. If no confirmation of the inquiry information is received from user 101, or other instructions are received from user 101, such as the user inputting "help me play music", it indicates that user 101 is not satisfied with the current operation instruction and inquiry information, then the vehicle can be controlled according to the user's actual needs, and the execution result can be fed back to the server 103.

[0205] S311, update according to the result feedback.

[0206] For example, after receiving the execution result, if the execution result indicates that the model output does not meet the actual needs of the user, the server 103 may update the large language model 104 according to the result feedback.

[0207] For example, after the user inputs "help me play music", the updated data combination is input into the large language model 104 for learning, so that when a similar scene is detected later, music is played for the user instead of the news channel.

[0208] In summary, in the embodiment of the present application, the huge amount of data to be processed is divided into multiple levels. Through the two-level processing process of the second and third stages, it can be ensured that only the necessary key information is transmitted between each layer, the context information is compressed, and the problem of unreliable reasoning results caused by context explosion is avoided; at the same time, the robustness of the reasoning results can be reduced through hierarchical decision-making. The function classification interface is used to search and retrieve the target functions that can actually be executed, without the need to screen among a large number of functions, which can improve the response speed and accuracy of the vehicle. At the same time, the voice scripts and control instructions for inquiry information are generated, which can not only realize interaction with the user, but also realize timely control of the vehicle, thereby improving the user's experience of using the in-vehicle interactive system. In addition, during the decision-making process, the vehicle's hardware is verified and the reasoning results are securely verified to ensure the safe execution of the control instructions; finally, the large language model is updated using the feedback mechanism to achieve dynamic optimization of the decision model, which can further improve the accuracy of the decision.

[0209] Combined with the above Figures 1 to 3 The vehicle interaction method provided by the embodiment of the present application is described in detail; Figure 4 and Figure 5 The device embodiments of the present application are described in detail. It should be understood that the devices in the embodiments of the present application can execute the various methods of the aforementioned embodiments of the present application, that is, the specific working processes of the following various products can refer to the corresponding processes in the aforementioned method embodiments.

[0210] Figure 4 It is a structural diagram of a vehicle interaction device provided in an embodiment of the present application.

[0211] For example, Figure 4 As shown, the vehicle interaction device 400 includes:

[0212] An acquisition module 401 is used to acquire vehicle data and user data of a vehicle;

[0213] A first determination module 402 is configured to determine user intention and a target function type of the vehicle based on vehicle data and user data;

[0214] A second determining module 403 is configured to determine a target function based on the target function type;

[0215] The third determining module 404 is configured to determine a target control instruction corresponding to the target function based on the user intention and the target function;

[0216] The control module 405 is used to control the vehicle based on the target control instruction.

[0217] In a possible implementation, the first determination module 402 is further configured to determine user intention based on vehicle data and user data; and determine a target function type based on the user intention and a preset function type of the vehicle.

[0218] In one possible implementation, the first determination module 402 is further used to input vehicle data, user data, and a first instruction into the interaction model to obtain user intent and target function type, wherein the first instruction is used to instruct the interaction model to output user intent and target function type.

[0219] In a possible implementation, the third determining module 404 is further configured to determine, based on the target function, target configuration parameters corresponding to the target function; and determine a target control instruction based on the user intention and the target configuration parameters.

[0220] In a possible implementation, the third determination module 404 is further configured to determine a target function interface corresponding to the target function from a candidate function interface library of the vehicle based on the target function; and obtain target configuration parameters corresponding to the target function through the target function interface.

[0221] In one possible implementation, the third determination module 404 is further used to input user intention, target configuration parameters, vehicle data, user data and a second instruction into the interaction model to obtain a target control instruction, wherein the second instruction is used to instruct the interaction model to output the target control instruction.

[0222] In one possible implementation, the vehicle interaction device 400 is also used to obtain data to be processed collected by the vehicle; preprocess the data to be processed to obtain vehicle data and user data, wherein the storage space occupied by the vehicle data and user data is smaller than the storage space occupied by the data to be processed.

[0223] In one possible implementation, the vehicle interaction device 400 is also used to determine the target query information corresponding to the target function based on the user intention and target configuration parameters, and to control the vehicle to output the target query information; the control module 404 is also used to control the vehicle based on the target control instruction when confirmation information of the target query information is received.

[0224] It should be noted that the above-mentioned vehicle interaction device is embodied in the form of a functional unit. The term "module" here can be implemented in the form of software and / or hardware, and is not specifically limited to this.

[0225] For example, a "module" may be a software program, a hardware circuit, or a combination of the two that implements the aforementioned functionality. The hardware circuit may include an application specific integrated circuit (ASIC), an electronic circuit, a processor (e.g., a shared processor, a dedicated processor, or a group of processors) and memory for executing one or more software or firmware programs, combined logic circuits, and / or other suitable components that support the described functionality.

[0226] Therefore, the units of each example described in the embodiments of this application can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0227] Figure 5 It is a structural schematic diagram of a vehicle provided in an embodiment of the present application.

[0228] For example, Figure 5 As shown, the vehicle 500 includes: a memory 501 and a processor 502, wherein the memory 501 stores an executable program code 503, and the processor 502 is used to call and execute the executable program code 503 to perform a vehicle interaction method.

[0229] In addition, an embodiment of the present application also protects a device, which may include a memory and a processor, wherein the memory stores executable program code, and the processor is used to call and execute the executable program code to perform a vehicle interaction method provided by an embodiment of the present application.

[0230] In this embodiment, the device can be divided into functional modules based on the above-described method examples. For example, each functional module can be mapped to a specific functional module, or two or more functions can be integrated into a single processing module. The integrated module can be implemented in hardware. It should be noted that the module division in this embodiment is illustrative and represents only a logical functional division. In actual implementation, other division methods may be used.

[0231] In the case of dividing the functional modules into corresponding functional modules, the device may further include an acquisition module, a first determination module, a second determination module, a third determination module, and a control module. It should be noted that all relevant contents of the various steps involved in the above method embodiment can be referred to the functional description of the corresponding functional modules and will not be repeated here.

[0232] It should be understood that the device provided in this embodiment is used to execute the above-mentioned vehicle interaction method, and thus can achieve the same effect as the above-mentioned implementation method.

[0233] In the case of an integrated unit, the device may include a processing module and a storage module. When the device is used in a vehicle, the processing module may be used to control and manage the vehicle's movements, while the storage module may be used to support the vehicle's execution of relevant program codes.

[0234] The processing module may be a processor or controller that implements or executes the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. The processor may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a digital signal processing (DSP) and a microprocessor, and the storage module may be a memory.

[0235] In addition, the device provided in the embodiments of the present application can specifically be a chip, component or module, and the chip may include a connected processor and memory; wherein the memory is used to store instructions, and when the processor calls and executes the instructions, the chip can execute a vehicle interaction method provided in the above embodiment.

[0236] This embodiment also provides a computer-readable storage medium, which stores computer program code. When the computer program code runs on a computer, the computer executes the above-mentioned related method steps to implement a vehicle interaction method provided by the above embodiment.

[0237] Among them, computer-readable storage media may include, but are not limited to, any type of disk, including floppy disks, optical disks, Digital Video Discs (DVDs), Compact Disc Read-Only Memory (CD-ROMs), microdrives and magneto-optical disks, Read-Only Memory (ROMs), Random Access Memory (RAMs), Erasable Programmable Read-Only Memory (EPROMs), Electrically Erasable Programmable Read-Only Memory (EEPROMs), Dynamic Random Access Memory (DRAMs), Video Random Access Memory (VRAMs), flash memory devices, magnetic or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0238] This embodiment also provides a computer program product. When the computer program product is run on a computer, it enables the computer to execute the above-mentioned related steps to implement a vehicle interaction method provided by the above embodiment.

[0239] Among them, the device, computer-readable storage medium, computer program product or chip provided in this embodiment are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0240] Through the description of the above implementation methods, technical personnel in the relevant field can understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0241] In the embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of modules or units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0242] The above content is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A vehicle interaction method, characterized in that: The method comprises: Obtain vehicle data and user data of the vehicle; determining a user intention and a target function type of the vehicle based on the vehicle data and the user data; determining a target function based on the target function type; Determining a target control instruction corresponding to the target function based on the user intention and the target function; The vehicle is controlled based on the target control instruction.

2. The method according to claim 1, characterized in that The determining, based on the vehicle data and the user data, the user intention and the target function type of the vehicle includes: determining the user intention based on the vehicle data and the user data; The target function type is determined based on the user intention and a preset function type of the vehicle.

3. The method according to claim 1, characterized in that The determining, based on the vehicle data and the user data, the user intention and the target function type of the vehicle includes: The vehicle data, the user data and a first instruction are input into the interaction model to obtain the user intention and the target function type, wherein the first instruction is used to instruct the interaction model to output the user intention and the target function type.

4. The method according to any one of claims 1 to 3, characterized in that The determining, based on the user intention and the target function, a target control instruction corresponding to the target function includes: Based on the target function, determining target configuration parameters corresponding to the target function; The target control instruction is determined based on the user intention and the target configuration parameter.

5. The method according to claim 4, characterized in that The determining, based on the target function, target configuration parameters corresponding to the target function includes: Based on the target function, determining a target function interface corresponding to the target function from a candidate function interface library of the vehicle; The target configuration parameters corresponding to the target function are obtained through the target function interface.

6. The method according to claim 4, characterized in that The determining the target control instruction based on the user intention and the target configuration parameter includes: The user intention, the target configuration parameter, the vehicle data, the user data and a second instruction are input into the interaction model to obtain the target control instruction, wherein the second instruction is used to instruct the interaction model to output the target control instruction.

7. The method according to any one of claims 1 to 3, characterized in that The method further comprises: Acquiring data to be processed collected by the vehicle; The data to be processed is preprocessed to obtain the vehicle data and the user data, wherein the storage space occupied by the vehicle data and the user data is smaller than the storage space occupied by the data to be processed.

8. The method according to any one of claims 1 to 3, characterized in that The method further comprises: determining target query information corresponding to the target function based on the user intention and the target configuration parameter, and controlling the vehicle to output the target query information; The controlling the vehicle based on the target control instruction includes: When confirmation information of the target inquiry information is received, the vehicle is controlled based on the target control instruction.

9. A vehicle interaction device, characterized in that: The device comprises: An acquisition module, used to acquire vehicle data and user data of a vehicle; a first determining module, configured to determine a user intention and a target function type of the vehicle based on the vehicle data and the user data; A second determining module is configured to determine a target function based on the target function type; a third determining module, configured to determine a target control instruction corresponding to the target function based on the user intention and the target function; A control module is used to control the vehicle based on the target control instruction.

10. A vehicle, characterized in that: The vehicle comprises: a memory for storing executable program code; A processor is configured to call and run the executable program code from the memory, so that the vehicle executes the method according to any one of claims 1 to 8.

Images