Human-vehicle interaction methods for two-wheeled electric vehicles

Abstract

This application provides a human-vehicle interaction method for a two-wheeled electric vehicle, relating to the field of human-vehicle interaction. The method includes: receiving first interaction information input by a user through an interaction hub of a first device and parsing it into a scheduling command. The interaction hub is a processing unit for receiving interaction information, parsing the interaction information, generating scheduling commands, and scheduling function modules. The first device is an in-vehicle terminal device or a mobile terminal device communicatively connected to an in-vehicle terminal device. The interaction hub invokes at least one function module to execute the scheduling command according to the scheduling command, obtaining corresponding first result information. The function module is an independent business service unit connected to the interaction hub through a standard interface. The interaction hub outputs the first result information. The method provided in this application improves the efficiency and reduces the complexity of human-vehicle interaction.

Description

Technical Field

[0001] This application relates to the field of human-vehicle interaction in two-wheeled electric vehicles, and more particularly to a human-vehicle interaction method for two-wheeled electric vehicles. Background Technology

[0002] As electric two-wheeled vehicles are a core mode of transportation for short-distance travel, users are increasingly demanding higher levels of intelligence from these vehicles.

[0003] However, in the existing technology, electric two-wheeled vehicles need to frequently switch between multiple independent applications or functional modules to meet various complex needs, which makes the operation cumbersome for users and inefficient in handling user needs.

[0004] Therefore, there is an urgent need for a simple and efficient human-vehicle interaction solution. Summary of the Invention

[0005] This application provides a human-vehicle interaction method for two-wheeled electric vehicles to solve the technical problems of cumbersome operation and low efficiency in human-vehicle interaction.

[0006] In a first aspect, this application provides a human-vehicle interaction method, comprising: receiving first interactive information input by a user through an interaction hub of a first device and parsing it into a scheduling instruction; the interaction hub being a processing unit for receiving interactive information, parsing interactive information to generate scheduling instructions, and scheduling function modules; the first device being an in-vehicle terminal device or a mobile terminal device communicatively connected to an in-vehicle terminal device; executing the scheduling instruction by calling at least one function module of the first device through the interaction hub according to the scheduling instruction, and obtaining corresponding first result information; the function module being an independent business service unit connected to the interaction hub through a standard interface; and outputting the first result information through the interaction hub.

[0007] In one possible implementation of the first aspect, the interaction hub calls at least one functional module of the first device to execute the scheduling instruction according to the scheduling instruction, and obtains the corresponding first result information. This includes: calling multiple functional modules to coordinately execute the scheduling instruction and obtain the corresponding first result information. The coordinated execution includes at least one of serial execution, parallel execution and interactive execution.

[0008] In one possible implementation of the first aspect, the interaction center calls at least one functional module of the first device to execute the scheduling instruction according to the scheduling instruction, and obtains the corresponding first result information, including: calling multiple functional modules to execute the scheduling instruction in a coordinated manner, and obtaining multiple second result information output by multiple functional modules; and the interaction center performs fusion processing on the second result information to obtain the first result information.

[0009] In one possible implementation of the first aspect, the interaction hub calls at least one functional module of the first device to execute the scheduling instruction according to the scheduling instruction, and obtains the corresponding first result information, including: calling at least one functional module according to the scheduling instruction and executing the scheduling instruction based on the first configuration information; wherein, the first configuration information is the configuration information corresponding to at least one functional module in a preset configuration information database, and the configuration information includes at least one of user preference configuration, interaction mode configuration, and functional module enable configuration.

[0010] In one possible implementation of the first aspect, the method further includes: receiving second interactive information input by the user through the interaction center of the first device and parsing it into a configuration adjustment instruction, wherein the second interactive information is used to indicate the adjustment of configuration information; and adjusting the configuration information based on the configuration adjustment instruction.

[0011] In one possible implementation of the first aspect, the interaction center is an agent, and the method further includes: determining a corresponding first task reward based on the first action performed by the agent in a training task round and the environmental change information corresponding to the first action; determining gradient information corresponding to a policy loss function based on multiple task rewards, wherein the multiple task rewards include the first task reward, and the policy loss function is used to determine the difference between the predicted environmental value and the actual environmental value of the agent in the training task round; and updating the parameters in the agent's policy neural network according to the gradient information.

[0012] In one possible implementation of the first aspect, the first device is an in-vehicle terminal device, and the method further includes: acquiring vehicle operation data; generating corresponding prompt information when the vehicle operation data meets a first condition, wherein the first condition is determined based on prompt configuration information.

[0013] In one possible implementation of the first aspect, the first device is an in-vehicle terminal device, and the method further includes: acquiring first operating data of the vehicle; performing data format conversion and / or data content conversion on the first operating data to obtain converted first operating data; extracting key operating data from the converted first operating data; encapsulating the key operating data and outputting the encapsulated key operating data.

[0014] In one possible implementation of the first aspect, the first device is an in-vehicle terminal device, and the method further includes: when the first interactive information is used to instruct vehicle control, performing security verification on the user, the security verification including at least one of voice verification, password verification and biometric verification; if the user passes the security verification, determining whether the current operating data of the vehicle meets the second condition; if the second condition is met, parsing the first interactive information into a scheduling instruction through the interaction center.

[0015] In one possible implementation of the first aspect, the method further includes: prohibiting the user from controlling the vehicle during a first time period if N consecutive interaction messages from the same user fail security verification; and / or generating a corresponding warning signal to indicate that N consecutive interaction messages have failed security verification.

[0016] In one possible implementation of the first aspect, the first interactive information is used to indicate the recommendation-related service. The first result information obtained by calling at least one functional module of the first device to execute the scheduling instruction through the interaction center according to the scheduling instruction includes: determining the corresponding recommendation service information based on user preference configuration and third-party data.

[0017] Secondly, this application provides a human-vehicle interaction device, comprising:

[0018] The receiving module is used to receive the first interactive information input by the user through the interaction center of the first device and parse it into a scheduling instruction. The interaction center is a processing unit used to receive interactive information, parse interactive information to generate scheduling instructions and scheduling function modules. The first device is an in-vehicle terminal device or a mobile terminal device that is communicatively connected to the in-vehicle terminal device.

[0019] The calling module is used to call at least one functional module of the first device to execute the scheduling instruction according to the scheduling instruction through the interaction center, and obtain the corresponding first result information. The functional module is an independent business service unit connected to the interaction center through a standard interface.

[0020] The output module is used to output the first result information through the interaction center.

[0021] The human-vehicle interaction method provided in this application receives and parses the user's first interaction information through the interaction center, and realizes the integration of vehicle ecosystem scenarios by combining the access of functional modules. This enables the linkage and integration between multiple scenarios, and users do not need to switch between multiple applications or modules. They can complete all operations by interacting with the interaction center, which simplifies the user's operation steps and improves the efficiency of handling user needs. Attached Figure Description

[0022] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0023] Figure 1 This application provides a schematic diagram of the architecture of a human-vehicle interaction system.

[0024] Figure 2 This is a schematic diagram of the architecture of another human-vehicle interaction system provided in an embodiment of this application;

[0025] Figure 3 This is a schematic diagram of the architecture of another human-vehicle interaction system provided in an embodiment of this application;

[0026] Figure 4 A flowchart illustrating a human-vehicle interaction method provided in an embodiment of this application;

[0027] Figure 5 A schematic diagram of a human-vehicle interaction method provided in an embodiment of this application;

[0028] Figure 6 A flowchart illustrating a human-vehicle interaction method provided in an embodiment of this application;

[0029] Figure 7 This is a schematic diagram of a human-vehicle interaction method provided in an embodiment of this application.

[0030] The accompanying drawings have illustrated specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to specific embodiments. Detailed Implementation

[0031] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0032] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of the relevant data all comply with the relevant laws, regulations, and standards of the relevant countries and regions, have taken necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation access points for users to choose to authorize or refuse.

[0033] The human-vehicle interaction method provided in this application embodiment can be applied to the vehicle's in-vehicle terminal or other terminal devices, such as smartphones, smart tablets, smartwatches, etc. Other terminal devices are communicatively connected to the vehicle's in-vehicle terminal and can realize the human-vehicle interaction method provided in this application embodiment through communication with the vehicle's in-vehicle terminal.

[0034] like Figure 1As shown, the human-vehicle interaction system 100 in this embodiment may include a user interaction layer 11, an interaction hub layer 12, a function module skill access layer 13, and a data storage layer 14. The layers interact with each other through data interfaces, as detailed below:

[0035] 1. User Interaction Layer 11: As the interaction entry point between the user and the human-vehicle interaction system, it can be used to receive interactive information such as voice information, text information, personalized configuration instructions, and operation instructions from the user, as well as to provide feedback to the user with various information (vehicle status information, reminder information, service recommendation information, etc.).

[0036] like Figure 1 As shown, this user interaction layer supports multiple interaction methods, including natural language interaction (voice interaction, text interaction) and manual operation interaction (terminal application interface, vehicle terminal interface), and can be applied to various terminal devices such as mobile terminal 111, vehicle terminal 112, and smartwatch terminal 113.

[0037] II. Interaction Hub Layer 12: Responsible for receiving interaction information from the user interaction layer, parsing user needs to generate corresponding scheduling instructions, and scheduling various skills in the skill access layer according to these instructions for data processing and forwarding. It can also store personalized user configuration information (such as skill activation status, preference settings, reminder rules, etc.) and dynamically adjust service response strategies based on user preferences and other configuration information, calling different functional modules to execute instructions.

[0038] For example, the interaction hub layer may include trained agents to implement the above functions. For example, the agent may be an AI operation assistant such as Openclaw personal assistant, CoPaw, and Kimi Claw. In addition, it may be other artificial intelligence (AI) agents, or other AI intelligent control systems that implement the same functions, such as edge operating system-level agents, vertical applications built based on large language models (LLM), etc. The specifics are not limited here.

[0039] like Figure 2 As shown, the interaction hub layer 12 may include the following core modules: instruction parsing module 121, skill scheduling module 122, personalized configuration module 123, and data storage module 124, wherein:

[0040] (1) Command parsing module 121: used to parse the interactive information (natural language, manual operation commands) input by the user, identify user needs (such as vehicle status query, nearby restaurant recommendations, personalized configuration adjustments, etc.), and convert the parsed needs into executable commands and send them to the skill scheduling module. For example, if the user sends the voice command "Check my car's current battery level", the command parsing module can parse it into the command "Vehicle status query - battery level query" and send it to the corresponding skill module.

[0041] For example Figure 2 The instruction parsing module shown may include a voice parsing unit 1211, a text parsing unit 1212, and an operation instruction parsing unit 1213, which are used to implement the functions of the instruction parsing module.

[0042] (2) Skill scheduling module 122: It is used to manage various access skills. According to the requirement command sent by the command parsing module, it schedules the corresponding skill to perform related operations. It can also realize the linkage scheduling between multiple skills at the same time to complete the fusion service of complex scenarios. For example, if the user command is "I want to go to a nearby scenic spot and check if the battery is enough", the skill scheduling module can simultaneously schedule the "vehicle status query skill" and the "scenic spot recommendation skill". First, it checks the vehicle battery. If the battery is sufficient, it pushes nearby scenic spots. If the battery is insufficient, it simultaneously schedules the "charging station query skill" to push charging stations around the scenic spot, realizing the linkage of scenarios.

[0043] For example Figure 2 As shown, the skill scheduling module may include a skill management unit 1221, a linkage scheduling unit 1222, and an instruction execution unit 1223, which are used to implement the functions of the skill scheduling module.

[0044] (3) Personalized configuration module 123: It is used to support users to set configuration information according to their personalized needs. It can include skill activation / disabling, reminder rule settings (such as reminder when the battery is below 2% or reminder when the tire pressure is abnormal), service preference settings (such as food and beverage taste preferences or attraction type preferences), interaction method settings (such as voice response speed and reminder method). Users can adjust the configuration information at any time, and the configuration information can be synchronized to the data storage module in real time.

[0045] For example Figure 2 As shown, the personalized configuration module may include a skill configuration unit 1231, a reminder rule configuration unit 1232, a preference configuration unit 1233, and an interaction method configuration unit 1234, which are used to implement the functions of the personalized configuration module.

[0046] (4) Data storage module 124: Used to store user's personalized configuration information, vehicle-related data (vehicle condition data, cycling trajectory data), skill operation data, service recommendation data, etc. It can adopt a dual storage mode of local + cloud. The device can store core privacy data (such as personalized configuration) locally, and non-privacy data (such as scenic spot information, catering information) in the cloud, ensuring data security and quick access, while supporting real-time data updates and backups.

[0047] For example Figure 2 As shown, the data storage module may include a local storage unit 1241, a cloud storage unit 1242, a data encryption unit 1243, and a data update unit 1244, which are used to implement the functions of the above data storage module.

[0048] III. Skill Access Layer 13: Responsible for integrating various skills related to the electric two-wheeled vehicle ecosystem, including vehicle control skills, vehicle condition monitoring skills, safety reminder skills, and travel-related service skills. These skills connect to the interaction hub layer through standardized interfaces, supporting hot-swapping and scalability. Users can add, delete, and update skills according to their needs without modifying the core system code.

[0049] like Figure 3 As shown, it can specifically include the following skills:

[0050] (1) Vehicle control skill131: Used to realize the remote control function of electric two-wheeled vehicles, including remote locking, remote unlocking, vehicle start, vehicle speed adjustment, light control and other control functions. It can be connected to the vehicle control system of electric two-wheeled vehicles, obtain vehicle control permissions through data interface, receive control instructions from the interaction center layer, execute the corresponding vehicle operation, and provide feedback on the operation results.

[0051] (2) Vehicle Condition Monitoring Skill 132: This skill is used to monitor the real-time operating status of electric two-wheelers, including data such as battery level, range, tire pressure, motor operating status, brake status, and battery health. It collects vehicle operating data through onboard sensors and transmits it to the interaction hub layer in real time, supporting user-initiated queries. It also provides data support for safety reminder skills and travel-related service skills. For example, this skill can collect battery power data in real time, and trigger a safety reminder skill when the battery level is lower than a user-set threshold.

[0052] (3) Safety Reminder Skill 133: Based on the data provided by the vehicle condition monitoring skill and combined with the reminder rules set by the user, various safety reminder functions can be implemented, including low battery reminder, abnormal tire pressure reminder, speeding reminder, battery over-discharge reminder, deviation from the preset route reminder, and abnormal vehicle movement reminder. The reminder methods can include voice reminder, text reminder, and pop-up reminder. Users can set the reminder threshold and reminder method themselves. For example, if the user sets a reminder when the battery is below 2%, when the vehicle condition monitoring skill collects that the battery is below 2%, the safety reminder skill will be triggered immediately and send a reminder message in the way preset by the user.

[0053] (4) Service-related skill 134: This skill is used to access travel-related services, including attraction recommendations, restaurant recommendations, charging station queries, parking lot queries, public transport transfer queries, and weather queries. It obtains relevant service data by connecting to third-party service platforms (such as map platforms and weather platforms), and combines the user's personalized preferences (such as restaurant tastes and attraction types) and vehicle status (such as battery level and range) to push accurate service recommendations to the user. For example, if a user sends the command "recommend nearby home-style restaurants" while riding, this skill can combine the user's set taste preferences and the vehicle's current location to push nearby home-style restaurants that meet the criteria, and mark information such as distance, rating, and average cost per person. At the same time, it can combine the vehicle's battery level to determine whether to recommend nearby charging stations simultaneously.

[0054] IV. Data Storage Layer 14: This layer stores foundational data for various ecosystem scenarios, including vehicle data (vehicle condition data and cycling trajectory data collected by onboard sensors), service data (scenic spot information, restaurant information, charging station information, etc.), and user data (personalized configuration information and usage habit data), providing data support for the entire system's operation. This layer employs data encryption technology to protect user privacy while ensuring real-time data updates to guarantee service accuracy and timeliness. For example, charging station information is synchronized in real-time with the latest data from third-party platforms, ensuring that the charging station status (availability, charging price) queried by users is accurate.

[0055] For example, such as Figure 1 As shown, the data storage layer may include a vehicle data storage module 141, a service data storage module 142, and a user data storage module 143, which are used to implement the functions of the aforementioned data storage layer.

[0056] As electric two-wheeled vehicles are a core mode of transportation for short-distance travel, users are increasingly demanding higher levels of intelligence from these vehicles.

[0057] However, in the existing technology, electric two-wheeled vehicles need to frequently switch between multiple independent applications or functional modules to meet various complex needs, which makes the operation cumbersome for users and inefficient in handling user needs.

[0058] The human-vehicle interaction method for two-wheeled electric vehicles provided in this application aims to solve the above-mentioned technical problems in the prior art.

[0059] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will be described below with reference to the accompanying drawings.

[0060] Figure 4 This is a flowchart illustrating a human-vehicle interaction method provided in an embodiment of this application, as shown below. Figure 4 As shown, the method includes:

[0061] S401. The first device receives first interactive information input by the user through the interaction center of the first device and parses it into a scheduling instruction. The interaction center is a processing unit for receiving interactive information, parsing interactive information to generate scheduling instructions and scheduling function modules. The first device can be an in-vehicle terminal device or a mobile terminal device that has established a communication connection with the in-vehicle terminal device.

[0062] The first device can interact with the user. For the first interactive information input by the user, the first device can parse it through the interaction center and generate corresponding scheduling instructions based on the parsed information.

[0063] For example, the first interactive information input by the user (such as voice, text, etc.) is first received by the interaction center. The parsing module of the interaction center (e.g., based on a large language model or rule engine) performs semantic understanding and intent recognition on the unstructured first interactive information, extracts key parameters, and finally transforms it into structured scheduling instructions.

[0064] The interaction hub refers to the core processing unit that receives and parses the first interactive information input by the user, thereby generating corresponding scheduling instructions, and scheduling the corresponding functional modules according to the scheduling instructions. For example, it could be an Openclaw personal assistant.

[0065] For example, if a user sends the text "Check battery level and recommend attractions", the interaction center can parse the text and generate corresponding dispatch instructions to call the vehicle condition monitoring skill and the attraction recommendation skill. The vehicle condition monitoring skill can be used to obtain the vehicle's battery level.

[0066] The first interactive information refers to the raw information input by the user to trigger system services. This can be natural language text, voice commands, or data requests in a specific format. For example, the text "Will it rain in Beijing tomorrow?" would be the first interactive information.

[0067] This scheduling instruction refers to at least one structured operation instruction generated after parsing user interaction information, used to trigger the execution of a functional module. For example, the intent to "check the weather" can be transformed into a JSON object containing the action type and parameters.

[0068] For example, "Battery Query + Attraction Recommendation" is parsed as a scheduling instruction that calls the vehicle condition monitoring skill and the attraction recommendation skill.

[0069] It is understandable that the scheduling instruction obtained by the interaction center from parsing the first interaction information can be a single instruction or multiple instructions; it can be multiple instructions that call the same functional module or multiple instructions that call multiple functional modules, and the specifics are not limited here.

[0070] S402. The first device calls at least one functional module of the first device to execute the scheduling instruction according to the scheduling instruction through the interaction center, and obtains the corresponding first result information. The functional module is an independent business service unit connected to the interaction center through a standard interface.

[0071] For example, the scheduling instruction specifies the type of task to be executed and the required parameters. The interaction hub locates the corresponding functional module in its internal functional module registry based on the scheduling instruction. Then, the interaction hub initiates a call request to the functional module through a standard interface, passing in the parameters from the scheduling instruction. As an independent business service unit, the functional module, upon receiving the request, can execute its specific internal business logic (such as querying the database or calling a third-party API), and encapsulate the execution result as first result information, returning it to the interaction hub through the standard interface.

[0072] This functional module refers to an independent business service unit that connects to the interaction center through a standardized interface. It can be used to execute specific business logic, such as vehicle control skills, vehicle condition monitoring skills, and safety alert skills. Each functional module can implement a corresponding function. For example, after receiving the corresponding angle command through the standardized interface, the vehicle condition monitoring skill can obtain the vehicle's battery level data.

[0073] This standard interface refers to the common data communication protocol or connection method between functional modules and the interaction hub, such as the representational state transfer (REST) ​​application programming interface (API) and plug-in interfaces. For example, the scenic spot recommendation skill exchanges data with the interaction hub through a plug-in interface.

[0074] The first result information refers to the processing result generated by the functional module after executing the scheduling command, which is used to provide feedback to the user. For example, the vehicle status data retrieved, feedback information on the completion of vehicle control, or error information on operation failure.

[0075] S403, The first device outputs the first result information through the interaction center.

[0076] After the functional module executes the scheduling instruction and returns the corresponding first result information to the interaction center, the first device can output the first result information to the user through the interaction center.

[0077] For example, after the interaction center performs necessary processing (such as formatting and aggregation) on the first result information, it feeds it back to the user through the original interaction channel, thereby completing the entire interaction process.

[0078] In this possible implementation, the interaction center receives and parses the user's first interaction information, and combines it with the access of functional modules to achieve the integration of vehicle ecosystem scenarios, realizing the linkage and integration between multiple scenarios. Users do not need to switch between multiple applications or modules; they can complete all operations by interacting with the interaction center, simplifying the user's operation steps and improving the efficiency of handling user needs.

[0079] In some embodiments, when a scheduling instruction corresponds to multiple service processing modules, the first device calls at least one functional module of the first device to execute the scheduling instruction according to the scheduling instruction, and obtains corresponding first result information, including:

[0080] The first device invokes multiple functional modules to collaboratively execute scheduling instructions and obtain corresponding first result information. Collaborative execution includes at least one of serial execution, parallel execution, and interactive execution. In addition, collaborative execution may include other execution methods, which are not limited here.

[0081] For example, the first device can parse the scheduling command through the interaction center to identify the complex intent contained therein and the functional modules that need to be invoked. Then, the interaction center selects the corresponding collaborative execution mode based on the preset business logic and the dependencies between modules.

[0082] This collaborative execution refers to an operational process in which at least two functional modules work together, such as the collaborative execution of battery level query and charging station recommendation. For example, after the vehicle status query skill checks the battery level, if the battery is low, it triggers a linked operation of the charging station query skill.

[0083] Specifically, the first device can analyze the relationships between scheduling instructions. If dependencies exist, multiple first instructions with dependencies are processed sequentially. Serial execution means that multiple modules execute multiple instructions in a sequential order, with the output of the previous module serving as the input for the next module.

[0084] If the instructions are independent of each other, then multiple independent first instructions are processed in parallel. Parallel execution means that multiple modules start and execute their corresponding instructions simultaneously within the same time period. There is no direct dependency between functional modules, and each module obtains the information required to execute the instruction.

[0085] If dynamic adjustments are required, interactive execution is used to process multiple first instructions. Interactive execution refers to bidirectional or multidirectional data exchange and feedback between multiple modules, where the execution process of one module is adjusted based on the real-time status of another module.

[0086] For example, when a user's battery is low during a ride, they can send a voice message: "Am I running out of power? Could you also find me a nearby charging station?" The interaction center can parse this message to understand the user's needs for "checking the current battery status" and "finding nearby charging stations," and generate corresponding dispatch commands to invoke the "vehicle status query skill" and the "charging station search skill." These two skills execute in parallel. The vehicle status query skill reads real-time data from the battery management system to obtain information such as the current battery percentage and estimated remaining range. The charging station search skill obtains the vehicle's GPS location and searches for charging station locations, availability, and charging prices within a 5-kilometer radius.

[0087] Then, when the current battery level is low, the "route planning skill" is invoked. Combining information such as the remaining driving range provided by the "vehicle status query skill" and the location of charging stations provided by the "charging station search skill", a route to the charging station can be planned before the battery runs out.

[0088] Understandably, in the example above, the "Route Planning skill" can only plan a route after the "Vehicle Status Query skill" provides the remaining driving range; therefore, these two skills are executed sequentially. When planning the route, information such as the location of charging stations provided by the "Charging Station Finding skill" is needed, and the route may be adjusted based on different charging station locations; therefore, this is an interactive execution.

[0089] Then, after the interaction center obtains the execution results of the "vehicle status query skill", "charging station search skill" and "route planning skill", it can integrate information such as battery level, remaining mileage, recommended charging station locations and distances, and estimated riding time into the first result information.

[0090] This possible implementation introduces a multi-module collaborative execution mechanism (serial, parallel, and interactive execution), changing the existing operational mode where functional modules are isolated and users need to frequently switch between different applications. It achieves deep integration and linkage of business services, enabling efficient response to complex user commands. For example, in scenarios such as cycling navigation, battery management, and nearby service recommendations, multiple modules such as vehicle status, location navigation, and third-party services can be automatically scheduled to work in parallel or interactively, shortening information acquisition and service response time and improving operational efficiency and user experience. Furthermore, through the interactive execution between modules, the primary device can make decisions based on real-time vehicle status and environmental data, making service recommendations more accurate and intelligent, solving the problems of fragmented ecosystem scenarios, poor functional scalability, and unsatisfactory user experience.

[0091] In some embodiments, the first device invokes at least one functional module of the first device through an interaction center to execute scheduling instructions according to scheduling instructions, and obtains corresponding first result information, including:

[0092] The first device calls multiple functional modules to collaboratively execute scheduling instructions and obtains multiple second result information output by the multiple functional modules;

[0093] The first device integrates the second result information through the interaction center to obtain the first result information.

[0094] For example, the system can first perform distributed data acquisition through the interaction center, that is, the interaction center simultaneously calls multiple corresponding functional modules. After the multiple functional modules complete their tasks, each generates its own "second result information". For example, module A outputs "battery data", module B outputs "charging pile list", and module C outputs "route planning data".

[0095] All "secondary result information" generated by the functional modules are uniformly returned to the interaction center. The interaction center can perform fusion processing on this secondary result information to obtain the final primary result information output to the user. For example, this fusion processing may include at least one of the following methods:

[0096] Data cleaning process: Remove invalid or erroneous data (such as sensor data with unstable signals);

[0097] Data association processing: Establish logical connections between the data returned by different functional modules (e.g., match the "charging pile list" with the "route planning").

[0098] Data integration and processing: combining different types of data, such as numerical values, text, and location data, into a complete data structure.

[0099] Priority sorting: Sort the data according to business logic (e.g., put the nearest charging station at the top of the list). Other processing methods may also be included, but are not limited here.

[0100] The interactive hub then encapsulates the processed first result information into a standard format "first result information".

[0101] The second result information refers to the raw, partial execution results or data fragments output by each independent functional module after executing the instructions during the execution of the first instruction. This information is usually heterogeneous and unorganized.

[0102] The first result information refers to the complete interactive result that can ultimately be presented to the user. It is the information after integrating and processing all the "second result information".

[0103] In this possible implementation, the "second result information" output by multiple modules is integrated through an interaction hub, solving the problem of the difficulty in uniformly presenting multi-source heterogeneous data and realizing the transformation from "data patching" to "information understanding." As a unified data processing layer, the interaction hub can effectively shield the differences between functional modules, deeply correlate, clean, and integrate scattered raw data such as vehicle status, environmental data, and service information, eliminating information fragmentation and logical conflicts caused by independent module outputs. It also improves the logic, coherence, and readability of the final interaction result, enhancing the naturalness and intelligence of human-vehicle interaction.

[0104] In some embodiments, the first device invokes at least one functional module of the first device through an interaction center to execute scheduling instructions according to scheduling instructions, and obtains corresponding first result information, including:

[0105] The first device invokes at least one functional module according to the scheduling instruction and executes the scheduling instruction based on the first configuration information;

[0106] The first configuration information is the configuration information corresponding to at least one functional module in the preset configuration information database. The configuration information includes at least one of user preference configuration, interaction method configuration, and functional module enable configuration.

[0107] For example, the first device can first determine, through the interaction center, at least one functional module that the scheduling command needs to invoke. Then, based on the at least one functional module, it can search in a preset configuration information database to obtain the corresponding "first configuration information". For example, it can match the corresponding configuration information in the preset configuration information database based on at least one of the following: module identifier, user identifier, or scene tag.

[0108] Then, the first configuration information acquired by the first device is loaded into the current task execution environment as parameters or constraints when the functional module executes instructions. At least one functional module can read and apply the loaded configuration information during the execution of scheduling instructions, adjust its own execution actions, and generate the final execution result.

[0109] The first configuration information is a set of parameters stored in a preset configuration information database and associated with the corresponding functional module. This first configuration information can be used to adjust the behavior mode, output content, or interaction form of the functional module when performing tasks.

[0110] A configuration information database is a data warehouse used to centrally store and manage all user and system configuration information. It can be a local database or a cloud database; the specific type is not limited here.

[0111] User preference configuration refers to the settings recorded by the first device regarding a user's personalized habits and preferences. Examples include: navigation preferences (avoiding congestion, shortest distance), entertainment preferences (frequently listened-to playlists, volume levels), and service preferences (types of places frequently visited).

[0112] Interaction mode configuration refers to the way functional modules exchange information with users. For example: the speech rate and tone of voice broadcasts; the information density of screen display (simple mode or detailed mode); whether to enable vibration feedback, etc.

[0113] Function module enable configuration refers to the configuration information of whether a function module is enabled or disabled, and the switch configuration information under what conditions the function module is allowed to be called. For example: automatically disable the multimedia module during a specific time period (such as at night), and automatically enable the mute module in a specific area (such as the company).

[0114] In this possible implementation, by introducing first configuration information based on a preset database when executing scheduling instructions, dynamic personalization and scenario-based customization of the functional module execution process are achieved. This breaks away from the traditional fixed execution mode and enables the first device to automatically adjust the behavior of functional modules according to different user habits (such as navigation preferences and interaction habits) and the needs of different scenarios (such as commuting mode and night mode). This improves the convenience of user operation and the comfort of the interactive experience, avoids the tediousness of frequent manual settings by users, and also optimizes the utilization efficiency of system resources.

[0115] In some embodiments, the method further includes:

[0116] S404. The first device receives the second interactive information input by the user through the interaction center of the first device and parses it into a configuration adjustment command. The second interactive information is used to indicate the adjustment of configuration information.

[0117] S405. The first device adjusts the configuration information based on the configuration adjustment command.

[0118] For example, the interaction center of the first device receives the second interaction information from the user. The interaction center can call a natural language processing or instruction parsing engine to parse the second interaction information, identify the user's adjustment intention (such as "turn up the volume") and target object (such as "navigation sound"), and convert it into a structured configuration adjustment instruction.

[0119] Then, the interaction hub adjusts and modifies the configuration information accordingly based on the configuration adjustment commands. For example, if the interaction hub has read and write permissions to the configuration information database, it can directly execute the adjustment operation. Alternatively, the interaction hub can call the underlying configuration management service (such as a database interface) to complete the actual adjustment operation.

[0120] After the configuration information is adjusted, the interaction center can notify the relevant functional modules that the configuration has been changed (or provide the updated value when the functional module requests configuration again), and provide feedback to the user to indicate whether the configuration information has been adjusted.

[0121] The second interactive information refers to the interactive information input by the user, indicating an intention to modify the configuration information.

[0122] Configuration adjustment instructions are instructions generated by the interaction center after parsing the second interaction information, and are used to perform specific configuration information modification actions.

[0123] For example, if a user sends a second interactive message via voice: "Switch the navigation broadcast to simplified mode," the interaction center of the first device receives the voice text and recognizes that the user intends to adjust the navigation broadcast mode from "detailed mode" to "simplified mode." The interaction center can then generate a corresponding configuration adjustment command to instruct the navigation broadcast mode to be adjusted to "simplified mode."

[0124] Then, the interaction center updates the "Voice Mode" field of the "Navigation Module" in the configuration information database to "Simple Mode". Before the next broadcast, the navigation skill queries the interaction center or the configuration information database for the latest configuration information. After confirming that the navigation broadcast mode is "Simple Mode", the navigation skill adjusts its broadcast logic to the broadcast logic corresponding to Simple Mode.

[0125] In this possible implementation, the configuration adjustment function is integrated into the interaction hub, achieving centralized and intelligent configuration management. This allows users to directly modify configuration information through the most natural interaction methods (such as voice), without having to navigate through complex menu levels. This improves the convenience of operation and safety during riding, and enhances the intelligence of the primary device. Furthermore, by uniformly processing configuration adjustment commands through the interaction hub, the accuracy and consistency of configuration adjustments are ensured, avoiding potential conflicts when multiple modules modify configurations concurrently.

[0126] For example, such as Figure 5 As shown, the human-vehicle interaction method for this two-wheeled electric vehicle may include the following steps:

[0127] S501, the user can pre-set the configuration information, that is, the first device can receive the configuration information setting operation from the user and initialize the configuration information.

[0128] S502, Then the user can input interactive information, and the first device can receive the interactive information input by the user through the interactive center.

[0129] S503. The first device can parse the interaction information through the interaction center to obtain the intent of the interaction information. In some embodiments, a corresponding scheduling instruction can also be generated accordingly.

[0130] S504. Then the first device can determine whether the intent of the interaction information indicates that the configuration information should be adjusted; if it indicates that the configuration information should be adjusted, then the above step 501 is executed to set the configuration information.

[0131] If the intent of the interactive information is not to instruct the adjustment of configuration information, then proceed to step S505.

[0132] S505. The first device determines whether the interactive information is a complex requirement, i.e., whether a functional module is needed to implement the requirement, or whether a functional module is needed to implement the requirement.

[0133] If one functional module is required to fulfill this requirement, proceed to step S506; if multiple functional modules are required to fulfill this requirement, proceed to step S507.

[0134] S506. If a functional module is required to fulfill this requirement, the first device calls the corresponding functional module to execute the instruction corresponding to the interactive information.

[0135] S507. If multiple functional modules are required to fulfill this requirement, the first device shall call the corresponding multiple functional modules to execute the instructions corresponding to the interactive information.

[0136] S508. After the functional module completes the execution of the instruction corresponding to the interactive information, it can return the result information to the interaction center.

[0137] S509. After completing the configuration information settings, the first device can obtain the vehicle's operating data.

[0138] S510. Then, based on the configuration information, determine whether the vehicle's operating data triggers the generation of a prompt message. If not triggered, proceed to step S512; if triggered, proceed to step S511.

[0139] S511. When the vehicle's operating data triggers the generation of prompt information, the first device generates the corresponding prompt information.

[0140] S512. After the scheduling function module executes the instruction corresponding to the interaction information and returns the result to the interaction center, and after completing the detection of whether the vehicle's operation data triggers the generation of prompt information, the first device records the data used, thereby optimizing the service strategy.

[0141] The execution details of the corresponding steps in this embodiment are similar to those in the above embodiments, and will not be repeated here.

[0142] The following is a specific example of a user adjusting configuration information:

[0143] User C wants the human-vehicle interaction system of the electric two-wheeler to meet their personalized functional needs, such as recording riding routes and statistical analysis of riding health data (e.g., riding time, calorie consumption). Based on this, User C can activate the AI ​​intelligent control system via a mobile app, such as the Openclaw personal assistant, and then enable the "vehicle control skill" and "vehicle condition monitoring skill." Simultaneously, User C can independently activate the "riding route recording skill" and "riding health statistics skill" through the Openclaw personal assistant, completing the installation and activation of these skills.

[0144] Then, user C can set the following configuration information through the personalization module: the recording frequency of the "Cycling Track Recording skill" (record the location once every 10 seconds), the track storage time (retain for 30 days); set the statistical dimensions of the "Cycling Health Statistics skill" (cycling time, distance, calorie consumption, average speed), and set a daily cycling goal (such as cycling 5 kilometers), and trigger a reminder when the goal is reached.

[0145] Then, during user C's ride, the "Cycling Track Recording Skill" can collect the vehicle's GPS location data in real time, generate the cycling track, and store it in the data storage module; the "Cycling Health Statistics Skill" can combine the cycling track data and vehicle operation data to calculate cycling time, distance, calorie consumption, and other data in real time, and display them synchronously on the user's mobile application and the vehicle terminal display screen.

[0146] After the ride, user C can send a voice command to the Openclaw personal assistant: "View today's cycling data." The Openclaw personal assistant can then use the "Cycling Health Statistics skill" to provide data such as today's cycling time, distance, and calorie consumption, and compare it with the daily cycling goal, prompting: "You cycled 6 kilometers today, achieving your goal. Keep going!"

[0147] In some embodiments, the interaction hub is an intelligent agent, which can be pre-trained by the user, such as... Figure 6 As shown, the method also includes:

[0148] S601. The first device determines the corresponding first task reward based on the first action performed by the agent in the training task round and the environmental change information corresponding to the first action.

[0149] S602. The first device determines the gradient information corresponding to the policy loss function based on multiple task rewards, including the first task reward. The policy loss function is used to determine the difference between the predicted environment value and the actual environment value when the agent performs training task rounds.

[0150] S603, The first device updates the parameters in the agent's policy neural network based on the gradient information.

[0151] For example, the first device can train the agent. Specifically, the agent of the first device can perform a "first action" in a training task round and monitor environmental change information in real time. Then, the first device determines the corresponding "first task reward" based on the action and environmental change information (such as whether the user adopts the suggestion or whether the vehicle responds smoothly).

[0152] After collecting "task rewards" from multiple rounds (including the first task reward), the first device can calculate the value of the policy loss function based on these task rewards and solve for the gradient information corresponding to the function. This gradient information indicates the direction of the derivative of the policy loss function under the current parameters, revealing the "weaknesses" of the current policy neural network.

[0153] Then, the first device uses gradient information to update the weight parameters in the agent's policy neural network through the backpropagation algorithm. This makes the agent more likely to perform actions that yield high rewards when encountering similar environments in the future, thus continuously approaching the optimal processing strategy.

[0154] An intelligent agent is a software entity within an interaction hub that can perceive environmental states (such as user commands or vehicle status) and decide what action to take based on its internal policy neural network.

[0155] The first action refers to the specific operation performed by the agent in a training task round based on its understanding of the environmental state (e.g., selecting a response or adjusting vehicle parameters).

[0156] Environmental change information refers to the feedback data generated by the environment (including user feedback and changes in vehicle status) after the intelligent agent performs the "first action".

[0157] The first task reward refers to the quantitative evaluation given to the intelligent agent by the first device based on the "first action" and the resulting "environmental change information". The reward value reflects the quality of the action (e.g., a high reward for user satisfaction and a low reward for user interruption).

[0158] The policy loss function is a function used to measure the difference between an agent's predictions and the actual results. The policy loss function calculates the difference between the agent's predicted environmental value and the actual environmental value; the smaller the difference, the more accurate the policy chosen by the agent.

[0159] Gradient information refers to the direction of the derivative of the policy loss function with the current parameters. It can indicate how to adjust the parameters of the policy neural network to minimize the loss as quickly as possible.

[0160] For example, an intelligent agent can learn how to automatically adjust the priority of power output and voice prompts in complex road conditions.

[0161] When the intelligent agent observes that the vehicle is "climbing a hill" (environmental state), it can take the first action: maintain high power output while broadcasting road condition information.

[0162] At this point, the first device detects that the user manually cut off the voice, and the vehicle speed is not affected, meaning the first device has acquired information about the environmental change. Based on this, the first device can determine that the action was ineffective (interfering with the user) and assign a lower reward value for the first task.

[0163] In multiple similar training rounds, the agent collected a large amount of data showing that "silence is rewarded more when climbing hills" and "broadcasting is rewarded less when climbing hills." By calculating the gradient of the policy loss function and updating the parameters, the agent determined a new policy: in scenarios requiring user focus, such as climbing hills, it automatically reduces the priority of voice broadcasting or delays the broadcast.

[0164] In this possible implementation, the agent is used as the interaction hub, and a deep reinforcement learning framework based on a reward mechanism is introduced to achieve the autonomous evolution of the agent's decision-making ability. During the agent's training, the task reward is determined using the "first action" and "environmental change information," enabling the agent to objectively assess the impact of its own behavior and establish a precise feedback loop. By calculating the gradient information of the policy loss function based on multiple task rewards and updating the policy neural network parameters accordingly, the agent can continuously learn and gradually optimize its behavioral strategy, minimizing the deviation between its predicted and actual environmental values. This allows the agent in the first device to adapt to the personalized driving habits of different users, making more reasonable and intelligent decisions in complex environments, thus improving human-vehicle interaction capabilities and the vehicle's intelligence level.

[0165] In some embodiments, the first device is an in-vehicle terminal device, and the method further includes:

[0166] S406. The first device acquires vehicle operation data of the two-wheeled electric vehicle;

[0167] S407. The first device generates a corresponding prompt message when the vehicle operation data meets the first condition, and the first condition is determined according to the prompt configuration information.

[0168] For example, the first device can collect various sensor data of the two-wheeled electric vehicle in real time according to a preset frequency or in response to scheduling instructions to obtain the current vehicle operation data. Then, the first device can compare the acquired vehicle operation data with a preset "first condition" in real time. This "first condition" is determined and generated by the first device based on the "prompt configuration information" set by the user. For example, if the user configures a "low battery warning", the system will generate the first condition "battery level < 30%".

[0169] Then, when the first device detects that the vehicle's operating data meets the first condition (such as the real-time battery level dropping below 30%), the first device can generate a corresponding prompt message. This prompt message can then be transmitted to the user through a suitable output channel (such as a voice module or a display screen), completing proactive information interaction.

[0170] Among them, vehicle operation data refers to various physical quantities or digital signals used to indicate the current state of two-wheeled electric vehicles, including but not limited to parameters such as vehicle speed, motor speed, battery charge, motor temperature, ambient light intensity, and vehicle tilt angle.

[0171] The first condition refers to a preset logical threshold or range used to trigger the prompt action. This first condition is not fixed but is dynamically generated based on the "prompt configuration information".

[0172] The prompt configuration information refers to the configuration parameters preset by the user or the first device, which can be used to define under what vehicle conditions a prompt should be issued. For example, the user can set "remind to charge when the battery is below 30%" or "turn on night riding mode".

[0173] Prompt information refers to information generated by the system to inform users of the vehicle status or suggest actions. It can be voice broadcast, dashboard light flashing, mobile APP push notification, or text display on the vehicle screen.

[0174] In this possible implementation, a high degree of personalization and intelligence of the vehicle prompt system is achieved by comparing and judging vehicle operation data with a first condition generated based on the prompt configuration information. Users can customize the prompt configuration information according to their own needs, and the first device generates corresponding judgment conditions in real time accordingly, making the prompt logic more in line with actual use scenarios.

[0175] The following is a specific example of querying vehicle operation data and providing prompts based on that data:

[0176] User A commutes daily by electric two-wheeler and has high requirements for vehicle battery level and tire pressure safety, and desires convenient operation via voice interaction. Based on this, User A can launch the Openclaw personal assistant via the mobile application, enabling the corresponding functional modules such as "Vehicle Condition Monitoring Skill," "Safety Reminder Skill," and "Voice Interaction Skill," and setting the following configuration information: trigger a voice reminder when the battery level is below 20%, trigger a dual text and voice reminder when the tire pressure is below 1.8 bar, set the voice response speed to fast, and set the default interaction method to voice.

[0177] Vehicle operation data query: After the configuration information is set, user A can send the command "Query my vehicle's battery level and tire pressure" to the Openclaw personal assistant via voice. The command parsing module of the interaction center parses the interaction information and generates the corresponding dispatch command. The skill dispatch module of the interaction center can call the "vehicle condition monitoring skill" based on the dispatch command to obtain the battery data (such as 35%) and tire pressure data (such as 2.0 bar) collected by the vehicle sensors, and provide feedback to user A via voice: "Current vehicle battery level is 35%, range is about 40 kilometers, tire pressure is normal, normal travel is possible."

[0178] Proactive reminder: During user A's ride, the vehicle condition monitoring skill collects real-time battery data. When the battery drops to 20%, the safety reminder skill is automatically triggered, reminding the user via voice: "The current battery level is below 20%, with a range of approximately 25 kilometers. Please charge in time." At the same time, the skill scheduling module in the interaction center calls the "travel surrounding service skill" to push information on charging stations within 3 kilometers of the user's current route, along with information such as distance, charging price, and availability.

[0179] Configuration Adjustment: User A believes the battery reminder threshold is too low and uses the voice command "Adjust the battery reminder threshold to 30%". After receiving the command, the Openclaw personal assistant, which is the interaction hub, updates the reminder rules in real time through the personalized configuration module. Subsequently, when the battery drops to 30%, a reminder will be triggered to adapt to the user's changing needs.

[0180] In some embodiments, the first device is an in-vehicle terminal device, and the method further includes:

[0181] S408, The first device acquires the first operating data of the two-wheeled electric vehicle;

[0182] S409. The first device performs data format conversion and / or data content conversion on the first operating data to obtain the converted first operating data;

[0183] S410, The first device extracts key operational data from the converted first operational data;

[0184] S411, The first device encapsulates the key operating data and outputs the encapsulated key operating data.

[0185] For example, the first device can respond to a dispatch command and read the original "first operating data" of the two-wheeled electric vehicle. Then, the acquired raw data is preprocessed. Based on user requirements, the first device performs "data format conversion" (such as parsing binary packets) and / or "data content conversion" (such as unit conversion and outlier filtering) to obtain standardized intermediate data, i.e., the converted first operating data.

[0186] Then, the first device extracts "key operational data" from the transformed data stream based on a preset rule base or priority list, thereby reducing the amount of data and retaining only the core information. Finally, the first device encapsulates the extracted key data according to the interface specification, generates a data packet with a complete communication header and checksum, and outputs it through a specified channel (such as Bluetooth or internal bus).

[0187] The first operating data refers to the raw data of the vehicle directly collected from the first device, which may include noise or be unprocessed binary code.

[0188] Data format conversion refers to changing the physical representation of data, such as converting a binary stream to JSON format or converting data encoding from ASCII to Unicode, to adapt to the reading needs of different functional modules.

[0189] Data content conversion refers to processing the logical meaning of data, such as converting the raw voltage value of a sensor into a specific physical quantity (e.g., converting a 0-5V signal into a 0-100% throttle opening), or normalizing the data.

[0190] Key operational data refers to core parameters that, after screening, are crucial for vehicle control, status monitoring, or user interaction, such as vehicle speed, remaining mileage, and fault codes.

[0191] Encapsulation refers to packaging key data into data frames or data packets according to specific communication protocols (such as TCP / IP, CAN FD, or custom API interfaces), adding metadata such as checksums, source addresses, and destination addresses, and forming independent data units that can be transmitted over a network.

[0192] This possible implementation improves the intelligence and standardization of vehicle information interaction by performing multi-level transformation, extraction, and encapsulation of vehicle operation data. Data format and content conversion ensures seamless compatibility between heterogeneous systems and reduces data parsing error rates. Extracting key operational data effectively filters redundant information, reducing communication bandwidth consumption and device computational load, resulting in faster response times. Encapsulation not only guarantees the integrity and security of data transmission but also provides standardized interface specifications for third-party application access. Users can view vehicle operation data more smoothly, and devices can perform data analysis more efficiently, thus providing data support for vehicle-user interaction.

[0193] The following is a specific example of preprocessing vehicle operation data, including conversion and encapsulation:

[0194] User D uses an electric two-wheeler for commuting and shopping daily, and has extremely high requirements for system response speed and data accuracy. Therefore, it is necessary to perform on-vehicle data preprocessing (interface conversion, data encapsulation) to reduce data transmission latency and remove invalid data. At the same time, the raw data is transformed into key results that are easy for users to understand, thereby improving service response speed, accuracy and user experience.

[0195] In this embodiment, the vehicle-side data that can be preprocessed can be vehicle operation data collected by vehicle sensors and vehicle control systems, specifically including vehicle status data, operating parameter data, environmental adaptation data, and anomaly monitoring data. The focus of preprocessing is interface conversion (data format / content can be directly converted through a dedicated interface without the need for a large model) and data encapsulation (extracting core results and pushing key information, rather than all the original data).

[0196] For example, it may include: (1) Vehicle status data: power data (voltage, current, remaining power percentage), driving range data (real-time driving range, remaining driving range), tire pressure data (front and rear tire pressure, tire pressure change rate).

[0197] (2) Operating parameter data: vehicle speed data (real-time vehicle speed, average vehicle speed, maximum vehicle speed), motor operating data (motor speed, power, temperature), braking data (number of braking, braking force, braking duration);

[0198] (3) Environmental adaptation data: vehicle ambient temperature and humidity data, vehicle tilt angle data (anti-tipping warning), GPS positioning data (latitude and longitude, positioning accuracy);

[0199] (4) Abnormal monitoring data: abnormal battery discharge data, abnormal motor vibration data, sudden tire pressure data, and abnormal vehicle movement data.

[0200] Based on this, user D can launch the Openclaw personal assistant via the mobile app, enable the "Vehicle Condition Monitoring Skill" and "Data Preprocessing Skill" (the newly added extended skill), and set the core preprocessing rules: First, complete the data conversion through a dedicated interface (which can avoid the delay caused by the participation of large models); second, encapsulate various types of data, push only the key results, and remove redundant raw data; at the same time, set invalid data removal, data noise reduction and calibration rules to ensure data accuracy.

[0201] Then, during vehicle operation, onboard sensors and control systems can collect various types of vehicle-side data in real time. The "data preprocessing skill" allows data preprocessing to be completed locally on the vehicle, without needing to be uploaded to a large cloud model. Specifically, this can include:

[0202] S1. Basic Preprocessing: Basic preprocessing such as data filtering, noise reduction, and calibration is performed first, removing invalid data, eliminating fluctuation noise, and adjusting calibration time and errors to provide accurate raw data support for interface conversion and data encapsulation. Specifically, this may include the following preprocessing steps:

[0203] (1) Data filtering: Remove invalid data (such as GPS positioning data with an accuracy of less than 10 meters, and abnormal values ​​of power collection values ​​that exceed the range of 0-100%).

[0204] (2) Data noise reduction: For continuously collected data such as vehicle speed, tire pressure, and battery level, the mean filtering algorithm is used to remove instantaneous fluctuation noise and ensure data stability (e.g., the mean is taken every 5 seconds to replace the fluctuating data collected in a single time).

[0205] (3) Data calibration: For delayed data with a timestamp deviation of more than 1 second, timestamp calibration is performed to ensure that the data is synchronized with the actual operating status; error calibration is performed on voltage and current data to improve the accuracy of power calculation;

[0206] (4) Data classification and labeling: The pre-processed valid data is divided into "normal data" and "abnormal data". Abnormal data (such as battery over-discharge, abnormal tire pressure) is marked and transmitted to the Openclaw personal assistant core layer first. Normal data is transmitted at a preset frequency (such as every 2 seconds) to reduce the amount of data transmitted.

[0207] S2. Interface Conversion (No large model required, direct conversion via dedicated interface): For raw data requiring format / content conversion, it can be directly converted through a preset dedicated interface, improving conversion speed and accuracy. Examples include:

[0208] GPS latitude and longitude data conversion: The collected vehicle GPS raw latitude and longitude data (such as xx°xx′ North latitude, xx°xx′ East longitude) is directly converted into specific geographical location names (such as "surrounding area of ​​XX Road XX District XX City") through the dedicated geocoding interface built into the data preprocessing skill. There is no need to call the large model for reverse geocoding conversion, which improves the response speed and avoids the errors of large model conversion.

[0209] Sensor data interface adaptation and conversion: Raw data collected by vehicle sensors (such as analog signals like voltage, current, and speed) can be directly converted into standardized digital signals via a dedicated interface. This is compatible with the reading format of Openclaw personal assistant and the skill, eliminating the need for additional data parsing steps and reducing data transmission and processing latency.

[0210] Third-party data interface linkage conversion: The vehicle ambient temperature data can be directly linked with the local weather interface through a dedicated interface to convert it into an intuitive conclusion such as "Current temperature in the area is XX℃, suitable for riding", without the need for secondary processing.

[0211] S3. Data Encapsulation: Encapsulate the preprocessed valid data, extracting only the core results that users care about and pushing them to the user interaction layer. Do not display all the original vehicle status data. For example, it may include:

[0212] Vehicle-side low battery reminder packaging: After collecting and preprocessing raw data such as battery voltage, current, and remaining battery percentage, the system does not push all parameter data to the user, but only packages it into a key reminder result, such as "Current battery level is 18% (range of about 20 kilometers), which is lower than the preset threshold. It is recommended to go to XX charging station within 3 kilometers to charge". The system also includes the location of the charging station (the specific name after interface conversion).

[0213] Vehicle-side thermal runaway warning packaging: Real-time collection of raw data such as battery temperature, voltage change rate, and motor temperature. After preprocessing, if a risk of thermal runaway is detected, not all monitoring data is pushed, but only a warning result is packaged, such as "Battery temperature rises abnormally (currently 48℃), there is a risk of thermal runaway, please stop immediately and stay away from the vehicle". At the same time, the location of the nearest maintenance station is pushed.

[0214] Standard vehicle condition packaging: During daily riding, pre-processed data such as speed, tire pressure, and range are packaged into concise conclusions, such as "Current speed 25km / h (not exceeding the speed limit), tire pressure is normal, range remaining is 50km, suitable for normal commuting", to avoid users being disturbed by redundant raw data.

[0215] In this possible implementation, the interface conversion does not require the participation of a large model, which improves the data conversion speed, reduces conversion errors, and avoids the delays and deviations caused by large models such as reverse geocoding. Furthermore, after data encapsulation, users only receive key results without the need to filter redundant information, thus improving the user experience. It also reduces the overall data transmission volume, lowers data transmission latency, and improves system response speed. Invalid data is effectively removed, improving data accuracy and avoiding service misjudgments caused by invalid data.

[0216] In some embodiments, the first device is an in-vehicle terminal device, and the method further includes:

[0217] S412, when the first interactive information is used to instruct vehicle control, the first device performs security verification on the user, the security verification including at least one of voice verification, password verification, and biometric verification:

[0218] S413. If the user passes the security verification, the first device determines whether the vehicle's current operating data meets the second condition.

[0219] S414. If the second condition is met, the first device parses the first interactive information into a scheduling instruction through the interactive center.

[0220] For example, after the interaction center receives the first interaction information input by the user (such as the voice command "start vehicle"), it can first parse the intent of the interaction information. If it is determined that the first interaction information is used to instruct vehicle control, or its corresponding dispatch command is used to instruct vehicle control, then the security verification process is initiated.

[0221] The first device can perform one or more of the following verification mechanisms based on the first interactive information, the first interactive information input by the user, or the user's voice verification, password verification, or biometric verification: only after successful verification is the user considered to have control authority.

[0222] After the user's identity is verified, the first device further queries the vehicle's current operating data to determine whether the preset "second condition" is met. For example, if the instruction is "autonomous driving mode", the second condition may require "vehicle speed is 0" and "road conditions permit".

[0223] Then, only when both identity verification and operational status verification pass will the interaction center formally parse the original "first interaction information" into an executable "scheduling instruction" and call the corresponding functional module to execute it. If any step fails, the first device will block the instruction and issue a prompt.

[0224] Among them, security verification refers to the process by which the first device verifies the legitimacy of the user's identity before executing vehicle control.

[0225] Voice verification refers to a verification method that confirms a user's identity by comparing the user's voiceprint characteristics or a specific wake word.

[0226] Password verification refers to a verification method in which the first device confirms the user's identity based on the user's input of preset numbers, characters, or gesture patterns.

[0227] Biometric verification refers to the technology of identifying individuals using inherent physiological characteristics of the human body (such as fingerprints and facial recognition).

[0228] The second condition refers to a safety threshold or logical constraint related to the vehicle's current operating data or vehicle environment. Examples include "vehicle stationary" and "parking engaged," which are used to ensure that the vehicle is in a safe operating condition when executing commands.

[0229] Current operating data refers to data collected in real time by the first device that reflects the current state of the vehicle, such as vehicle speed, gear, braking status, tilt angle, etc.

[0230] This possible implementation enhances the safety and reliability of two-wheeled electric vehicles during intelligent interaction by establishing dual authentication methods: identity verification and operational condition verification. Multiple authentication methods, including voice, password, and biometrics, ensure that only authenticated users can perform critical operations such as vehicle control, reducing the risk of theft or accidental activation. Simultaneously, by introducing a "second condition" to monitor the vehicle's current operating data in real time, it prevents commands from being executed under dangerous conditions (such as switching power modes while driving), eliminating safety accidents caused by misoperation. This protects the user's personal and property safety and improves the vehicle's level of intelligence.

[0231] For example, such as Figure 7 As shown, the human-vehicle interaction method for this two-wheeled electric vehicle may include the following steps:

[0232] S701, The first device receives interactive information from the user through the interactive center.

[0233] S702. When the interaction center of the first device determines that the function module to be called for the interaction information is the vehicle-side query skill, the interaction center can directly call the vehicle-side query skill to execute the corresponding instruction.

[0234] S703. After the vehicle-side query skill executes the command, the data can be preprocessed before being transmitted to the interaction center, or it can be output directly.

[0235] S704. If the interaction center of the first device determines that the function module to be called for the interaction information is the vehicle control class skill, the first device needs to perform security verification first. The specific method is similar to the security verification method mentioned above, and will not be repeated here.

[0236] 705. After the user passes the security verification, the interaction center of the first device can call the vehicle-side control class skill to execute the corresponding instruction and return the corresponding execution result to the interaction center.

[0237] The execution details of the corresponding steps in this embodiment are similar to those in the above embodiments, and will not be repeated here.

[0238] In some embodiments, the method further includes:

[0239] S415. If N consecutive interaction messages from the same user fail security verification, the first device prohibits the user from controlling the vehicle for a first time period, where N is a positive integer; and / or,

[0240] The first device generates a corresponding warning signal, which indicates that N consecutive interactive messages have failed security verification.

[0241] For example, the first device may have a built-in counter for monitoring continuous interaction requests from the same user (or the same access device). Whenever an interaction is determined to be a "vehicle control" intent and enters the security verification process, the first device may record the corresponding verification result. If verification fails, the counter is incremented by 1; if verification succeeds, the counter is reset to zero. The first device may continuously determine whether the counter value reaches a preset threshold N.

[0242] If N consecutive interactions fail security verification, the first device's defense mechanism can be triggered. Firstly, during the "first time period," the user is prohibited from performing any vehicle control operations (i.e., the input channel is locked). Secondly, the first device can generate a corresponding "warning signal." This warning signal can be sent to the vehicle's alarm module (e.g., triggering an alarm sound), sent to the owner's mobile application (pushing a notification that "someone attempted to illegally start the vehicle"), or uploaded to the cloud backend for recording and monitoring. After the first time period ends, the first device can lift the prohibition, allowing the user to try again, while the counter remains reset to zero.

[0243] Among them, N consecutive interaction messages refer to N consecutive requests for vehicle control initiated by the first device during the same user session (N is a positive integer, such as 3 or 5 times).

[0244] "Failed security verification" means that when the above-mentioned interactive information is used for identity verification (such as incorrect password, voiceprint mismatch, or facial recognition failure), the first device determines that the operator is an unauthorized user or the information is invalid.

[0245] The first time period is a preset period of time during which the first device is disabled (such as 30 seconds, 5 minutes, or permanently locked until the administrator unlocks it). During this period, the first device refuses to respond to any vehicle control commands from the user (or the input terminal).

[0246] A warning signal refers to information generated by the first device to indicate an abnormal state, which may include the time of the abnormality, the number of failures, and the possible risk level.

[0247] This possible implementation enhances the ability of two-wheeled electric vehicles to resist brute-force attacks and illegal intrusions by introducing a "continuous failure lockout" and "early warning generation" mechanism. This defense strategy, based on a counting threshold, effectively blocks attack paths that rely on brute-force guessing of passwords or voiceprints, improving the vehicle's anti-theft security level. The initial prohibition of operation not only protects the vehicle itself but also avoids wasted system resources or overheating of electronic components due to frequent erroneous commands. Furthermore, the generation and push of early warning signals enables real-time security linkage between the owner and the vehicle, allowing the owner to immediately grasp the vehicle's abnormal status and take corresponding measures, thus enhancing the user's sense of security and trust in the vehicle.

[0248] The following is a specific embodiment of a first device performing security verification:

[0249] User E is concerned that vehicle control commands may be illegally invoked or accidentally triggered, leading to loss of vehicle control. Therefore, a mandatory verification mechanism for vehicle control commands is needed to ensure vehicle control security. Based on this, User E can launch the Openclaw personal assistant through the in-vehicle terminal, enable the "Vehicle Control Skill" and "Security Verification Skill" (a new extended skill), and set mandatory verification rules for vehicle control commands: all vehicle control commands (remote locking, unlocking, starting, speed adjustment, light control, etc.) must undergo double verification. No third party or skill is allowed to directly invoke the in-vehicle control system; if verification fails, any control operation is prohibited.

[0250] For example, user E can set up a two-factor authentication method and choose according to their own needs. Specific authentication methods include:

[0251] First layer of verification: Identity verification, you can choose at least one of the following: "Voice verification" (preset exclusive voice command, such as "unlock vehicle + user name"), "Password verification" (preset 4-6 digit numeric password) or "Biometric verification" (fingerprint, face, compatible with in-vehicle terminals or mobile phones that support biometrics);

[0252] The second layer of verification: scenario verification. The system automatically detects whether the current scenario meets the control conditions (such as when remotely unlocking, it detects whether the current location of the vehicle is within the user's preset safe range (such as within 1 kilometer of home or company); when starting the vehicle, it detects whether the user is near the vehicle terminal (verified via Bluetooth connection status)).

[0253] Then, user E can send vehicle control commands (such as "remotely unlock the vehicle") to the Openclaw personal assistant via voice or manual operation; the Openclaw personal assistant can perform the first layer of authentication by calling the "security verification skill" to trigger authentication. The user completes the preset verification operation (such as entering a password or saying a personalized voice command). If the verification is successful, the second layer of scenario verification will be entered. If the verification fails, the command will be refused to be executed and a reminder will be sent to the user (such as "Authentication failed, unable to perform the unlock operation").

[0254] Then, a second layer of scenario verification is performed: the "security verification skill" detects the current scenario conditions. For example, when remotely unlocking, it checks whether the vehicle's current location is within the user's preset safe range and whether the vehicle is stationary. If the conditions are met, the verification passes; if the conditions are not met (such as the vehicle being in an unfamiliar area or in motion), the verification fails, the command is refused to be executed, and a scenario anomaly reminder is pushed (such as "The vehicle is currently in an unfamiliar area, remote unlocking is prohibited, please confirm vehicle safety").

[0255] After both verifications pass, the skill scheduling module of the Openclaw personal assistant sends an authorization command to the "vehicle control skill". The "vehicle control skill" sends control commands to the vehicle control system through a standardized interface, performs the corresponding operation (such as unlocking), and feeds back the operation result to the user. If either verification fails, it is prohibited to send any control commands to the vehicle control system, and direct invocation and intervention are strictly prohibited.

[0256] In some embodiments, the first device may also be configured with a verification failure protection mechanism. If authentication fails N times consecutively (e.g., 3 times), the vehicle control function will be automatically locked for an immediate period of time (e.g., 15 minutes), and a security warning message will be sent to the user's bound mobile phone. At the same time, the "security verification skill" will monitor the source of the control command in real time. If a call request not initiated by the Openclaw personal assistant (e.g., a third-party application or illegal instruction) is detected, it will be directly intercepted and access to the vehicle control system will be prohibited, thereby further ensuring vehicle control security and avoiding risks such as vehicle loss of control or theft.

[0257] In some embodiments, the first interactive information is used to indicate recommended related services. The first device, through an interactive hub, calls at least one functional module of the first device to execute the scheduling instruction according to the scheduling instruction, and obtains the corresponding first result information, including:

[0258] The first device determines the corresponding recommended service information based on user preference configurations and third-party data.

[0259] For example, the interaction center receives the user's first interaction information, and after parsing, confirms that the user's intention belongs to "recommendation related services" (such as the user asking "What are some fun things to do nearby?" or the system automatically triggering recommendations based on habits).

[0260] The first device can call corresponding functional modules (such as a service recommendation module) to read the "user preference configuration" stored locally to understand the user's long-term habits, and obtain real-time "third-party data" through a network interface to understand the current external environment. Then, the first device uses the functional modules to match and analyze the internal preferences with the external environment. Based on the matching results, it determines the "recommended service information" that best suits the current context. For example, it cross-filters the "user's preferred restaurant type" with the "list of restaurants near the current location" to determine recommended restaurants as the recommended service information.

[0261] Among them, recommended related services refer to information services that do not directly involve vehicle hardware control, but provide users with information services related to surrounding life, travel assistance, or entertainment value-added, such as nearby charging stations, repair shops, weather forecasts, and riding routes.

[0262] User preference configuration refers to the personalized options that users pre-set in the system, reflecting their habits and preferences. Examples include: frequently visited locations, preferred riding modes (sports / energy-saving), favorite music genres, and whether they pay attention to vehicle maintenance.

[0263] Third-party data refers to data from external vehicle systems or the internet, which typically includes real-time location information, weather forecast data, traffic condition data, charging station operator idle status data, and surrounding commercial (points of interest) data.

[0264] This possible implementation achieves precise and contextualized service recommendations for two-wheeled electric vehicles by deeply integrating "user preference configuration" and "third-party data." By incorporating personalized user preferences, the practical value of the recommended content and user satisfaction are enhanced. Utilizing real-time third-party data ensures the timeliness of the recommendations (e.g., the availability of charging stations), addressing the pain point of users searching for services and improving the intelligence level of the vehicle's infotainment system.

[0265] The following is a specific example of a service recommendation:

[0266] User B frequently rides an electric two-wheeler to nearby attractions and hopes the system can recommend suitable attractions and nearby restaurants, as well as provide convenient vehicle control. Based on this, User B can launch the Openclaw personal assistant through the vehicle terminal, enabling the "Attraction Recommendation Skill," "Restaurant Recommendation Skill," "Vehicle Control Skill," and "Travel Surrounding Services Skill," and setting the following configuration information, including service preferences: a preference for natural landscape attractions and a preference for light-tasting restaurants.

[0267] Then, when user B rides to area M, he sends an interactive message to the Openclaw personal assistant in the interaction center via text command: "Recommend nearby natural scenic spots and restaurants with light flavors." After the instruction parsing module of the Openclaw personal assistant parses the message, the skill scheduling module of the Openclaw personal assistant simultaneously calls the "scenic spot recommendation skill" and the "restaurant recommendation skill".

[0268] Then, the "Attraction Recommendation Skill" connects to a third-party map platform to obtain natural scenic spots within 5 kilometers of the user's current location. Based on the user's preference settings, it filters out three highly-rated attractions and pushes their names, distances, ticket information, and cycling routes. The "Restaurant Recommendation Skill," based on the user's preference for light-tasting food, recommends restaurants with a light diet within 2 kilometers of the attractions, noting their distance, ratings, and average cost per person. Simultaneously, the "Vehicle Condition Monitoring Skill" checks the vehicle's battery level; if the battery is sufficient (e.g., 60%), it directly pushes the above information.

[0269] If the battery is low (e.g., 15%), the "Charging Station Query Skill" will be invoked to push information on charging stations near attractions and restaurants, and recommend "charging + sightseeing + dining" routes.

[0270] In some embodiments, the first device can not only recommend relevant service information to the user, but also recommend appropriate skills, configuration information, configuration rules and other information based on the user's initial usage data (such as riding habits and demand commands). The user can directly confirm or fine-tune these options, which simplifies the user's configuration process and improves ease of use.

[0271] In this embodiment of the application, the integration of vehicle ecosystem scenarios is achieved by using an interaction hub as the core and combining it with skill access, which has the following beneficial effects:

[0272] 1. This invention addresses the issues of inaccurate personalized services and high privacy risks, improving the accuracy and security of personalized services. Existing technologies rely on user profiles, which are difficult to collect and update in a timely manner, resulting in low service accuracy and the risk of privacy leaks. This application's embodiment utilizes the personalized configuration function of the interaction hub, allowing users to independently set their needs and enable skills without forcibly collecting large amounts of privacy data. It only stores user-configured preference information and necessary vehicle data, significantly improving privacy protection. Simultaneously, users can adjust their configurations at any time to adapt to dynamic changes in needs. Combined with the long-term memory capability of the interaction hub, this improves the accuracy of personalized services. For example, after changing their commuting route, users can directly adjust the recommended scope and preferences for nearby services without waiting for user profile updates, resulting in more timely and accurate service responses.

[0273] 2. Deep integration of the electric two-wheeler ecosystem enhances the user experience. Existing technologies often present fragmented scenarios, requiring users to frequently switch between them, resulting in a cumbersome experience. This application's embodiment utilizes skill-based access to connect various scenarios such as vehicle control, vehicle status monitoring, safety alerts, and nearby travel services, achieving seamless integration between these scenarios. Users can complete all operations through the central interaction hub without switching between multiple applications or modules. Furthermore, the scenario-linking function automatically matches relevant services based on user needs; for example, it automatically recommends charging stations when the battery is low, eliminating the need for manual searching and further improving travel convenience.

[0274] 3. Strong functional scalability, adaptable to the personalized customization needs of different users: Existing technologies mostly have functions preset by manufacturers, which cannot be independently expanded and have poor flexibility. The embodiments of this application adopt an scalable skill access mode, which allows users to add, delete, and update skills according to their own needs, realizing personalized customization of functions and adapting to the differentiated needs of different users (such as commuters focusing on battery level and safety reminders, leisure cyclists focusing on scenic spot and restaurant recommendations, and tech enthusiasts focusing on cycling data statistics). At the same time, the hot-swappable nature of skills does not require modification of the core system code, reducing the difficulty of functional expansion. It can continuously expand into new ecological scenarios (such as cycling social networking and vehicle maintenance) according to industry development and user needs, improving the applicability and life cycle of the solution.

[0275] 4. Convenient interaction methods, proactive service, and enhanced intelligence: Existing technologies offer limited and cumbersome interaction methods and lack proactive service capabilities. This application's embodiments utilize a unified interaction hub, supporting multiple interaction methods such as voice, text, and manual operation. It is compatible with various devices including mobile phones and in-vehicle terminals. Users can conveniently operate the vehicle via voice during riding, freeing their hands and improving riding safety. Simultaneously, it can proactively sense user needs and automatically push reminders and related services based on vehicle condition data and usage habits, significantly enhancing the vehicle's intelligence level.

[0276] 5. Reduce manufacturers' R&D and maintenance costs and enhance product competitiveness: In existing technologies, manufacturers need to invest heavily in building user profiles, developing and maintaining multiple independent functional modules, and continuously optimizing profile algorithms, resulting in high costs. This application's embodiment adopts an interaction hub and skill access mode, eliminating the need for manufacturers to redevelop core interaction and scheduling systems. They only need to develop skills adapted to the electric two-wheeler scenario to achieve ecosystem integration, reducing R&D costs. Simultaneously, the modular design of the skills facilitates later maintenance and updates, reducing maintenance costs and enhancing the product's differentiated competitiveness.

[0277] 6. Enhance vehicle safety performance and reduce safety hazards: Existing safety reminder functions are relatively simple and mostly passive. This application's embodiment uses a vehicle condition monitoring skill to collect vehicle operating data in real time, combined with user-defined reminder rules, to achieve multi-dimensional proactive safety reminders (battery level, tire pressure, speeding, battery over-discharge, etc.), effectively preventing safety accidents caused by vehicle malfunctions. Simultaneously, the vehicle control skill employs a dual protection mechanism of "data isolation + access control," granting only necessary control permissions to eliminate the risk of unauthorized operation, further enhancing vehicle safety performance.

[0278] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of the claims of this application and their equivalents, this application also intends to include these modifications and variations.

[0279] In this application, the term "comprising" and its variations can refer to non-limiting inclusion; the term "or" and its variations can refer to "and / or". The terms "first", "second", etc., in this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. In this application, "multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0280] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0281] It should be further noted that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0282] It should be understood that the above-described device embodiments are merely illustrative, and the device of this application can also be implemented in other ways. For example, the division of units / modules in the above embodiments is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units, modules, or components may be combined, or integrated into another system, or some features may be ignored or not executed.

[0283] Furthermore, unless otherwise specified, the functional units / modules in the various embodiments of this application can be integrated into one unit / module, or each unit / module can exist physically separately, or two or more units / modules can be integrated together. The integrated units / modules described above can be implemented in hardware or as software program modules.

[0284] In the above embodiments, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0285] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0286] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A human-vehicle interaction method for a two-wheeled electric vehicle, characterized in that, The method includes: The interaction center of the first device receives the first interactive information input by the user and parses it into a scheduling instruction. The interaction center is a processing unit for receiving interactive information, parsing interactive information to generate scheduling instructions and scheduling function modules. The first device is an in-vehicle terminal device or a mobile terminal device that is communicatively connected to the in-vehicle terminal device. The interactive hub invokes at least one functional module of the first device to execute the scheduling instruction according to the scheduling instruction, and obtains the corresponding first result information. The functional module is an independent business service unit connected to the interactive hub through a standard interface. The first result information is output through the interaction center.

2. The method according to claim 1, characterized in that, The step of calling at least one functional module of the first device through the interaction center to execute the scheduling instruction according to the scheduling instruction, and obtaining the corresponding first result information, includes: The scheduling instruction is executed collaboratively by calling multiple functional modules to obtain the corresponding first result information. The collaborative execution includes at least one of serial execution, parallel execution, and interactive execution.

3. The method according to claim 2, characterized in that, The step of calling at least one functional module of the first device through the interaction center to execute the scheduling instruction according to the scheduling instruction, and obtaining the corresponding first result information, includes: The scheduling instruction is executed collaboratively by calling multiple functional modules to obtain multiple second result information output by the multiple functional modules; The second result information is fused and processed by the interaction center to obtain the first result information.

4. The method according to claim 1, characterized in that, The step of calling at least one functional module of the first device through the interaction center to execute the scheduling instruction according to the scheduling instruction, and obtaining the corresponding first result information, includes: At least one functional module is invoked according to the scheduling instruction, and the scheduling instruction is executed based on the first configuration information; The first configuration information is configuration information corresponding to the at least one functional module in a preset configuration information database. The configuration information includes at least one of user preference configuration, interaction method configuration, and functional module activation configuration.

5. The method according to claim 4, characterized in that, The method further includes: The first device receives second interactive information input by the user through its interactive center and parses it into a configuration adjustment command. The second interactive information is used to instruct the adjustment of the configuration information. The configuration information is adjusted based on the configuration adjustment instructions.

6. The method according to any one of claims 1-5, characterized in that, The interaction hub is an intelligent agent, and the method further includes: Based on the first action performed by the agent in the training task round and the environmental change information corresponding to the first action, the corresponding first task reward is determined. The gradient information corresponding to the policy loss function is determined based on multiple task rewards, wherein the multiple task rewards include the first task reward, and the policy loss function is used to determine the difference between the predicted environment value and the actual environment value when the agent performs training task rounds. The parameters in the agent's policy neural network are updated based on the gradient information.

7. The method according to claim 6, characterized in that, The first device is a vehicle-mounted terminal device, and the method further includes: Obtain vehicle operation data; If the vehicle operation data meets a first condition, a corresponding prompt message is generated, wherein the first condition is determined based on the prompt configuration information.

8. The method according to claim 1, characterized in that, The first device is a vehicle-mounted terminal device, and the method further includes: Obtain the vehicle's initial operating data; The first running data is converted in data format and / or in data content to obtain the converted first running data; Extract key operational data from the transformed first operational data; The key operational data is encapsulated and the encapsulated key operational data is output.

9. The method according to claim 1, characterized in that, The first device is a vehicle-mounted terminal device, and the method further includes: When the first interactive information is used to instruct vehicle control, the user undergoes security verification, which includes at least one of voice verification, password verification, and biometric verification. If the user passes security verification, determine whether the vehicle's current operating data meets the second condition; If the second condition is met, the first interactive information is parsed into the scheduling instruction through the interaction center.

10. The method according to claim 9, characterized in that, The method further includes: If N consecutive interaction messages from the same user fail security verification, the user is prohibited from controlling the vehicle during the first time period; and / or, A corresponding warning signal is generated, which is used to indicate that the N consecutive interactive messages have failed the security verification.

11. The method according to claim 6, characterized in that, The first interactive information is used to indicate the recommendation of related services. The step of calling at least one functional module of the first device through the interaction center to execute the scheduling instruction according to the scheduling instruction and obtain the corresponding first result information includes: The corresponding recommendation service information is determined based on user preference configurations and third-party data.

Images