Vehicle control method, server and computer readable storage medium

Abstract

The invention discloses a vehicle control method, a server and a computer readable storage medium. The method comprises the steps of receiving a first voice request forwarded by a vehicle and current state information of the vehicle; and based on a first preset model, according to the first voice request and the current state information, determining a target execution action, and issuing the target execution action to the vehicle to control the vehicle to execute the target execution action. And under the condition that the vehicle executes the target execution action abnormally, target reflection information is generated according to the first voice request and the current state information, and the vehicle is controlled according to the target reflection information. Therefore, by processing the first voice request and the current state information of the vehicle, the accuracy of analyzing the user request and determining the execution action is improved. Moreover, through a real-time state monitoring and exception handling mechanism, the server ensures the safety of vehicle operation, reduces potential errors and risks, and optimizes the performance and efficiency of the system, thereby improving the user experience.

Description

Technical Field

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

[0002] In-car smart assistants typically interact with users through cockpit screens. However, these assistants currently lack the ability to reflect on their actions during execution. If errors occur during operation, the voice assistants often fail to effectively adjust their strategies, resulting in unsuccessful task completion and a negative user experience. Summary of the Invention

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

[0004] An embodiment of the present application provides a vehicle control method, the method comprising:

[0005] receiving a first voice request forwarded by a vehicle and current status information of the vehicle;

[0006] Determining a target execution action based on a first preset model and the first voice request and the current state information, and sending the target execution action to the vehicle to control the vehicle to execute the target execution action;

[0007] In the event that an abnormality occurs when the vehicle executes the target execution action, target reflection information is generated according to the first voice request and the current state information, so as to control the vehicle according to the target reflection information.

[0008] In this way, the server receives the first voice request forwarded by the vehicle and the current status information of the vehicle. Then, based on the first preset model, the server determines the target execution action according to the first voice request and the current status information, and sends the target execution action to the vehicle to control the vehicle to execute the target execution action. Finally, in the event that an abnormality occurs when the vehicle executes the target execution action, the server generates target reflection information based on the first voice request and the current status information to control the vehicle according to the target reflection information. In this way, by processing the first voice request and the current status information of the vehicle, the accuracy of parsing user requests and determining execution actions is improved. In addition, through real-time status monitoring and exception handling mechanisms, the server ensures the safety of vehicle operation, enhances the system's reflection capabilities, reduces potential errors and risks, optimizes the system's performance and efficiency, and thus improves user experience.

[0009] In some embodiments, generating target reflection information according to the first voice request and the current state information, and controlling the vehicle according to the target reflection information, includes:

[0010] generating first output information according to the current state, the first voice request, and historical operation information associated with the first voice request, the historical operation information including historical operation description information and historical executed operation information;

[0011] Based on the second preset model, the target reflection information is generated according to the first output information and the preset experience library, wherein the preset experience library includes multiple preset reflection information, and the preset reflection information includes preset historical operation description information, preset status information, preset execution results and preset execution operation information.

[0012] In this way, the server generates first output information based on the current state, the first voice request, and the historical operation information associated with the first voice request, where the historical operation information includes historical operation description information and historical execution operation information. Next, based on the second preset model, the server generates target reflection information according to the first output information and the preset experience library, where the preset experience library includes multiple preset reflection information, and the preset reflection information includes preset historical operation description information, preset state information, preset execution results, and preset execution operation information. In this way, by integrating the current state, the first voice request, and the historical operation information associated with the first voice request, and referring to the preset experience library, the server can continuously generate target reflection information, improve the system's reflection ability and reflection speed, and thus improve the success rate of task execution and user experience.

[0013] In some embodiments, generating the target reflection information based on the second preset model according to the first output information and a preset experience library includes:

[0014] determining operation similarity based on the historical operation description information in the first output information and the preset historical operation description information;

[0015] determining a state similarity based on the current state in the first output information and the preset state information;

[0016] determining a target similarity value according to the operation similarity and the state similarity;

[0017] determining candidate preset reflection information from the preset experience library according to the target similarity value;

[0018] The target reflection information is generated according to the target similarity value, a preset threshold and the candidate preset reflection information.

[0019] In this way, the server determines the operation similarity based on the historical operation description information and the preset historical operation description information in the first output information. Next, the server determines the state similarity based on the current state and the preset state information in the first output information. Then, the server determines the target similarity value based on the operation similarity and the state similarity. Subsequently, the server determines the candidate preset reflection information from the preset experience library based on the target similarity value. Finally, the server generates the target reflection information based on the target similarity value, the preset threshold and the candidate preset reflection information. In this way, through similarity evaluation, the intelligent agent can accurately match the current problem with the historical cases in the experience library to generate target reflection information, thereby continuously optimizing its reflection ability and reflection speed.

[0020] In some embodiments, generating the target reflection information according to the target similarity value, a preset threshold, and the candidate preset reflection information includes:

[0021] In a case where the maximum target similarity value is greater than or equal to the preset threshold, the target reflection information is generated according to the candidate preset reflection information.

[0022] In this way, when the target similarity value is greater than or equal to the preset threshold, the server generates the target reflection information based on the candidate preset reflection information. In this way, by setting the threshold, it ensures that the candidate reflection information is only adopted when the similarity is high enough, thereby improving the accuracy of the decision.

[0023] In certain embodiments, the method further comprises:

[0024] When the preset execution result corresponding to the target reflection information is an execution exception, the preset execution operation information in the target reflection information is sent to the vehicle to complete the vehicle control.

[0025] In this way, if the target reflection information corresponds to a preset execution result that indicates an execution anomaly, the server sends the preset execution operation information in the target reflection information to the vehicle to complete vehicle control. This allows the system to quickly respond to execution anomalies and take corrective measures, improving the system's fault handling capabilities and enhancing the stability and reliability of vehicle control, thereby improving the user experience and overall system performance.

[0026] In certain embodiments, the method further comprises:

[0027] When the preset execution result corresponding to the target reflection information is successful execution, the target execution action is determined based on the first preset model, according to the first voice request and the current state information.

[0028] In this way, if the preset execution result corresponding to the target reflection information is successful, the server determines the target execution action based on the first preset model, the first voice request, and the current state information. This allows the system to use historical success cases to guide current decisions, thereby improving the accuracy and success rate of execution actions, enhancing the system's intelligent decision-making capabilities, and improving task execution efficiency and user experience.

[0029] In some embodiments, generating the target reflection information according to the target similarity value, a preset threshold, and the candidate preset reflection information includes:

[0030] When the target similarity value is less than the preset threshold, a template is generated based on preset reflection information, and the target reflection information is generated according to preset reflection knowledge, the first voice request and the current state information.

[0031] In this way, when the target similarity value is less than a preset threshold, the server generates target reflection information based on a preset reflection information generation template, taking into account the preset reflection knowledge, the first voice request, and the current state information. This allows the system to flexibly generate new reflection information when encountering unusual or novel situations, thereby improving the system's adaptability and intelligence, enhancing the system's ability to handle unknown situations, and promoting continuous learning and self-optimization.

[0032] In certain embodiments, the method further comprises:

[0033] When it is confirmed according to the current state information that the target detection object is in an abnormal state, it is confirmed that an abnormality occurs in the target execution action.

[0034] In this way, if the server confirms that the target detection object is in an abnormal state based on the current state information, it will also confirm that the target execution action has an abnormality. This allows the system to promptly detect and respond to vehicle status anomalies, thereby preventing potential failures or dangers, improving vehicle safety and reliability, and enhancing the system's intelligence and user experience.

[0035] An embodiment of the present application provides a server, which includes a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the above-mentioned vehicle control method is implemented.

[0036] An embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the vehicle control method as described above is implemented.

[0037] Additional aspects and advantages of the embodiments of the present application will be given in part in the description below, and in part will become obvious from the description below, or will be learned through practice of the embodiments of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0039] Figure 1 This is one of the flowcharts of the vehicle control method according to certain embodiments of the present application;

[0040] Figure 2 is a schematic diagram of a target reflection information generation process in certain embodiments of the present application;

[0041] Figure 3 This is a second flow chart of a vehicle control method according to certain embodiments of the present application;

[0042] Figure 4 This is a third flow chart of a vehicle control method according to certain embodiments of the present application;

[0043] Figure 5 This is a fourth flow chart of a vehicle control method according to certain embodiments of the present application;

[0044] Figure 6 This is a fifth flow chart of a vehicle control method according to certain embodiments of the present application;

[0045] Figure 7 This is the sixth flow chart of the vehicle control method according to certain embodiments of the present application;

[0046] Figure 8 This is the seventh flow chart of the vehicle control method according to certain embodiments of the present application;

[0047] Figure 9 This is the eighth flow chart of the vehicle control method of certain embodiments of the present application. DETAILED DESCRIPTION

[0048] The embodiments of the present application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the embodiments of the present application, and should not be understood as limiting the embodiments of the present application.

[0049] In-vehicle intelligent assistants typically use the cockpit screen as the primary interface, allowing users to interact with the assistant through touchscreen or voice commands. This interactive approach has become quite mature in providing functions such as navigation, entertainment, and vehicle diagnostics. However, while these in-vehicle intelligent assistants excel at performing routine tasks, their ability to reflect on complex situations or unexpected events is relatively weak.

[0050] Specifically, when an in-car intelligent assistant encounters errors or unexpected situations while performing a task, it often fails to effectively adjust its execution strategy. For example, if a user requests navigation to a specific location, but the navigation route never appears on the cockpit screen, a less intuitive in-car intelligent assistant might not automatically provide a solution, requiring the user to debug themselves, resulting in a poor user experience.

[0051] Based on the above questions, please refer to Figure 1 , an embodiment of the present application provides a vehicle control method, the method comprising:

[0052] 01: Receive the first voice request forwarded by the vehicle and the vehicle's current status information;

[0053] 02: Based on the first preset model, according to the first voice request and the current state information, determine the target execution action, and send the target execution action to the vehicle to control the vehicle to execute the target execution action;

[0054] 03: When an abnormality occurs when the vehicle executes the target execution action, target reflection information is generated according to the first voice request and the current state information to control the vehicle according to the target reflection information.

[0055] The embodiment of the present application also provides a server, including a memory and a processor. The vehicle control method of the embodiment of the present application can be implemented by the server of the embodiment of the present application. Specifically, a computer program is stored in the memory, and the processor is used to receive a first voice request forwarded by the vehicle and the current status information of the vehicle. And based on the first preset model, according to the first voice request and the current status information, the target execution action is determined, and the target execution action is sent to the vehicle to control the vehicle to perform the target execution action. And in the event that an abnormality occurs when the vehicle performs the target execution action, target reflection information is generated according to the first voice request and the current status information to control the vehicle according to the target reflection information.

[0056] The embodiment of the present application also provides a vehicle control device. The vehicle control method of the embodiment of the present application can be implemented by the vehicle control device of the embodiment of the present application. Specifically, the vehicle control device includes a receiving module, a determining module and a generating module. The receiving module is used to receive a first voice request forwarded by the vehicle and the current status information of the vehicle. The determining module is used to determine the target execution action based on the first preset model, according to the first voice request and the current status information, and send the target execution action to the vehicle to control the vehicle to perform the target execution action. The generating module is used to generate target reflection information according to the first voice request and the current status information when an abnormality occurs when the vehicle performs the target execution action, so as to control the vehicle according to the target reflection information.

[0057] Specifically, a first voice request refers to a voice command issued by a user through the in-vehicle voice assistant during a round of voice interaction, including vehicle function control, information query, or service request. For example, "Turn on the air conditioner," "What's the weather like today?", and "I need the nearest gas station."

[0058] The current status information refers to the current display content information of the vehicle display components and the current perception information of the vehicle.

[0059] The current display content information of a vehicle display component refers to the information content currently presented on the vehicle display screen or other display device. For example, navigation-related content such as the current driving route, destination, estimated arrival time, and traffic conditions, or entertainment system-related content such as music playlists, track information, and radio station frequencies, can also be communication-related content such as incoming call information, text message content, and contact lists. Display content information is an important component of the vehicle's intelligent control system. Display content information can provide the driver with key driving information and entertainment services, and is also an important interface for interaction between the system and the user. It should be noted that in the embodiment of the present application, the vehicle control method is described using the vehicle display screen as the vehicle display component, that is, the current display content information is a screenshot of the vehicle display screen.

[0060] Current perception information refers to the vehicle's current environmental status and the state of its onboard equipment functions, collected by its sensors. Environmental status information includes data about the vehicle's surroundings, such as traffic conditions, road conditions, weather conditions, and geographic location. Traffic conditions refer to road traffic conditions collected by sensors such as cameras, radar, or LiDAR, including the location and speed of other vehicles. Road conditions refer to information such as road slipperiness, potholes, and obstacles. Weather conditions refer to weather information such as temperature, rainfall, and wind speed, obtained through weather sensors or data exchange with external services. Geographic location refers to the vehicle's current location obtained through positioning systems such as GPS. Onboard equipment function status perception information includes status data of the vehicle's internal systems and devices, such as the status of the in-vehicle entertainment system, navigation system settings, air conditioning configuration, engine status, battery charge, fuel level, and tire pressure.

[0061] The target action refers to the specific action determined and performed by the system based on the first voice request, the currently displayed content, and the currently perceived information. These actions include tapping, swiping, and typing. A tap refers to actions such as tapping an icon or button to activate a specific function or service. For example, if a user says "open music," the system may tap the music app icon in the in-vehicle entertainment system. A swipe refers to swiping or turning pages to view more information when the current information list is collapsed or too long. For example, if a navigation system displays a long route, the system may swipe to display the full route. A typing action refers to tapping the search box and typing to find specific information or services within the system. For example, if a user says "search for nearby restaurants," the system may tap the search box and type "nearby restaurants." A typing action can also indicate entering and sending text messages within the in-vehicle system, such as sending a text message or commenting on social media. For example, if a user says "send a message to Zhang San," the system may open the messaging app, select Zhang San, and open a chat window.

[0062] A target action exception occurs when a problem or error occurs while the system attempts to perform a specific action determined based on the first voice request, the currently displayed content, and the current perception information, resulting in the action not being completed as expected. For example, if the user says "open music," the system should click the music app icon. However, if the icon doesn't respond or the app fails to launch, this constitutes a click operation exception. When the navigation system needs to display the complete route, the system should swipe the screen. If the screen doesn't scroll smoothly or the system doesn't respond to the swipe command, this constitutes a swipe operation exception. If the user says "search for nearby restaurants," the system should click the search box and enter "nearby restaurants." If the search box doesn't activate or the text entered is incorrect, this constitutes a typing operation exception. If the user says "send a message to Zhang San," the system should be able to open the messaging app and select a contact. If the messaging app doesn't open, a contact can't be selected, or the message fails to send, this constitutes a text message sending exception.

[0063] Target reflection information refers to the information generated by the system for adjusting or correcting the execution strategy based on the current situation, historical data and preset experience library when an abnormality occurs when the in-vehicle intelligent assistant performs a target execution action. This information reflects the system's analysis of the reasons for the failure of the current execution action and its thinking on how to better execute user instructions.

[0064] See also Figure 2 The server receives the first voice request forwarded by the vehicle and the vehicle's current status information, which may include but is not limited to the vehicle's temperature, humidity, speed, etc.

[0065] Next, the server analyzes the first voice request and the current state information based on a first preset model to determine a target action. For example, if the voice request is "raise the temperature" and the current temperature in the car is low, the server may determine that the target action is to turn on the heating system.

[0066] After determining the target action, the server sends the action to the vehicle. After receiving the command, the vehicle performs the corresponding operation, such as adjusting the temperature.

[0067] If an exception occurs when executing the target execution action (for example, the temperature fails to rise as expected), the server will generate target reflection information based on the first voice request and the current status information, including analyzing the cause of the exception and generating solutions, such as checking whether the heating system is working properly.

[0068] In summary, in the vehicle control method and server provided by the embodiments of the present application, the server receives the first voice request forwarded by the vehicle and the current status information of the vehicle. Then, based on the first preset model, the server determines the target execution action according to the first voice request and the current status information, and sends the target execution action to the vehicle to control the vehicle to execute the target execution action. Finally, in the event that an abnormality occurs when the vehicle executes the target execution action, the server generates target reflection information based on the first voice request and the current status information to control the vehicle according to the target reflection information. In this way, by processing the first voice request and the current status information of the vehicle, the accuracy of parsing user requests and determining execution actions is improved. In addition, through real-time status monitoring and exception handling mechanisms, the server ensures the safety of vehicle operation, reduces potential errors and risks, optimizes the performance and efficiency of the system, and thus improves user experience.

[0069] See also Figure 3 In some embodiments, step 03 (generating target reflection information according to the first voice request and the current state information, and controlling the vehicle according to the target reflection information) includes:

[0070] 031: Generate first output information according to the current state, the first voice request, and historical operation information associated with the first voice request;

[0071] 032: Based on the second preset model, target reflection information is generated according to the first output information and the preset experience library.

[0072] In some embodiments, the determination module is further configured to generate first output information based on the current state, the first voice request, and historical operation information associated with the first voice request, and to generate target reflection information based on the first output information and a preset experience library based on a second preset model.

[0073] In some embodiments, the processor is further configured to generate first output information based on the current state, the first voice request, and historical operation information associated with the first voice request, and to generate target reflection information based on the first output information and a preset experience library based on a second preset model.

[0074] Specifically, the first output information refers to data information generated based on the current state, the first voice request, and historical operation information associated with the first voice request, and is used for subsequent steps. In some embodiments, the template for generating the first output information is as follows:

[0075] Suppose you are an intelligent assistant. I will give you the following information. Based on this information, please output the operation steps and status of the agent:

[0076] User task: <user request query>

[0077] Current status: <current status>

[0078] Historical operation information: <historical plan:P>, <historical action:A>

[0079] Based on the above information, please provide a concise description of the agent's operation steps and status in 50 words or less:

[0080] describe:

[0081] A first output message generated based on the first output message generation template may be: In the last operation, the agent selected the "Go Here" button but failed to complete the task. The current screen in the car is an advertisement. The car's computer is in P gear, no music is playing, and the temperature inside the car is 22 degrees Celsius.

[0082] Historical operation information refers to the operation data collected and stored by the in-vehicle intelligent assistant in past interactions, reflecting the user's behavior patterns, preferences and past performance of the system. Historical operation information includes user voice requests, historical operation description information, historical execution operation information and screenshots of the in-vehicle display screen when the user's voice request is received. That is, for each operation state of a voice request, its i-th step can be defined as S_i(q, p_i, ai, pic_i), where q refers to the user voice request query, p_i refers to the plan of the current step, a_i refers to the action of the current step, and pic_i refers to the screenshot of the current step. Both plan and action can represent the operation that needs to be performed by the user's voice request. The difference is that plan refers to the description of the operation in text that ordinary users can understand, and action refers to the specific instructions of the operation that relevant technicians and machines can understand.

[0083] For an operation session of an in-vehicle intelligent assistant, we need to record its historical plan: P = [p_1, p_2, ..., p_n] and historical action: A = [a_1, a_2, ..., a_n].

[0084] The preset experience database refers to a data set that stores the decision-making process, successful execution cases, and failure lessons of the in-vehicle intelligent assistant. This enables the in-vehicle intelligent assistant to learn from past experience and improve the quality and efficiency of its decision-making. The decision-making process record details the decision-making process of the intelligent assistant when processing user requests, including decision input (the user's historical operation description and the vehicle's current perception status information), the execution result after the operation, and the timestamp (the time when the decision occurred). Successful execution cases record the successful execution of tasks by the intelligent assistant, including the decision input, the executed operation, and the final successful result. Failure lessons record the failures encountered by the intelligent assistant when performing tasks, including the reasons for the failure, the errors that occurred, and possible improvement measures. The above decision-making process, successful execution cases, and failure lessons are collectively referred to as preset reflection information. That is, the preset experience database includes multiple preset reflection information e_i, each of which includes preset historical operation description information, preset status information, preset execution results, and preset execution operation information. Where e_i = (p_i, s_i, r_i, a_i)

[0085] Where p_i refers to the preset historical operation description information of the recorded voice request, s_i refers to the preset status information of the recorded voice request, r_i refers to the preset execution result of the recorded voice request, and a_i refers to the preset execution operation information that can be output by the recorded voice request, that is, the next action.

[0086] The server generates first output information based on the current state, the first voice request, and historical operation information associated with the first voice request, where the historical operation information includes historical operation description information and historical execution operation information. Next, the server generates target reflection information based on the first output information and a preset experience library based on a second preset model. The preset experience library includes multiple preset reflection information pieces, each of which includes preset historical operation description information, preset state information, preset execution results, and preset execution operation information.

[0087] In this way, by integrating the current state, the first voice request and the historical operation information associated with the first voice request, and referring to the preset experience library, the server can continuously generate target reflection information, improve the system's reflection ability and reflection speed, thereby improving the success rate of task execution and user experience.

[0088] See also Figure 4 In some embodiments, step 032 (generating target reflection information based on the second preset model, the first output information and the preset experience library) includes:

[0089] 0321: Determine the operation similarity based on the historical operation description information and the preset historical operation description information in the first output information;

[0090] 0322: Determine state similarity based on the current state and preset state information in the first output information;

[0091] 0323: Determine the target similarity value based on the operation similarity and state similarity;

[0092] 0324: Determine candidate preset reflection information from the preset experience library based on the target similarity value;

[0093] 0325: Generate target reflection information according to the target similarity value, the preset threshold and the candidate preset reflection information.

[0094] In certain embodiments, the determination module is further configured to determine operation similarity based on historical operation description information and preset historical operation description information in the first output information; determine state similarity based on the current state and preset state information in the first output information; and determine a target similarity value based on the operation similarity and the state similarity. The determination module is further configured to determine candidate preset reflection information from a preset experience library based on the target similarity value; and generate target reflection information based on the target similarity value, a preset threshold, and the candidate preset reflection information.

[0095] In certain embodiments, the processor is further configured to determine operation similarity based on historical operation description information and preset historical operation description information in the first output information; determine state similarity based on the current state and preset state information in the first output information; and determine candidate preset reflection information from a preset experience library based on the target similarity value. The processor is further configured to determine candidate preset reflection information from a preset experience library based on the target similarity value; and generate target reflection information based on the target similarity value, a preset threshold, and the candidate preset reflection information.

[0096] Specifically, the preset execution operation information refers to a description of a series of standardized operation instructions that may be executed in the vehicle, for example, "the user requests navigation home at 8 pm" or "the user issues an instruction to turn on the wipers in the rain".

[0097] The preset status information describes the vehicle status and environment information when the vehicle performs a specific operation, for example, "the vehicle fuel level is less than 20%", "the outside temperature exceeds 28 degrees Celsius", etc.

[0098] The preset execution result records the execution result of the preset execution operation information and preset status information. For example, "the navigation operation is successful and the user has arrived at the destination", "the windshield wiper operation is successful and the wipers are turned on", etc.

[0099] Preset execution information includes the actions that the smart assistant should take under specific operations and conditions. For example, "remind the user to go to a gas station when the fuel level is low" or "automatically start the air conditioner when the temperature in the car is too high."

[0100] The server first compares the historical operation description information p in the first output information now and the preset historical operation description information p in the preset experience library i , to determine the operation similarity, that is, the operation similarity is sim(p now ,p i For example, if a user requests "navigate to the nearest gas station" multiple times and there are similar operation descriptions in the preset experience library, the operation similarity is high.

[0101] Next, the server generates the current state s according to the first output information. now and the preset state information s corresponding to the above preset historical operation description information item in the preset experience database i , determine the state similarity, that is, the state similarity is sim(s now ,s i For example, if the user's voice request is "Navigate to the nearest gas station" and the vehicle's fuel level is currently low, and there is a similar operation description in the preset experience library, and the description page of the status is consistent, then the status similarity is high.

[0102] Then, the server calculates the target similarity value based on the operation similarity and state similarity. That is, the similarity value between the first output information and each preset reflection information is calculated, and the calculation formula is sim now =sim(p now ,p i )+sim(s now ,s i ). Then compare these similarity values ​​and select the one with the largest similarity as the target similarity value, that is,

[0103] Subsequently, based on the target similarity value, the server selects candidate preset reflection information from the preset experience library, where the candidate preset reflection information is the historical case and its solution that is most similar to the current situation.

[0104] Finally, the server generates the target reflection information based on the target similarity value, a preset threshold (used to determine whether the similarity is high enough), and the candidate preset reflection information. For example, if the preset threshold is 0.85, if the target similarity value is greater than or equal to the preset threshold, the server may select the candidate preset reflection information as the target reflection information. If the target similarity value is less than the preset threshold, the server determines that there is no preset reflection information in the preset experience library that is similar to the first output information.

[0105] In this way, through similarity evaluation, the agent can accurately match the current problem with historical cases in the experience library to generate target reflection information, thereby continuously optimizing its reflection ability and reflection speed.

[0106] See also Figure 5 In some embodiments, step 0325 (generating target reflection information based on the target similarity value, the preset threshold, and the candidate preset reflection information) includes:

[0107] 03251: When the maximum target similarity value is greater than or equal to the preset threshold, target reflection information is generated according to the candidate preset reflection information.

[0108] In some embodiments, the determination module is further configured to generate target reflection information based on candidate preset reflection information when the maximum target similarity value is greater than or equal to a preset threshold.

[0109] In some embodiments, the processor is further configured to generate target reflection information based on candidate preset reflection information when the maximum target similarity value is greater than or equal to a preset threshold.

[0110] Specifically, the server compares the calculated target similarity value with a preset threshold. If the target similarity value is greater than or equal to the preset threshold, the server considers that the candidate preset reflection information is sufficiently related to the first output information, and the server can select the most similar candidate reflection information as the target reflection information.

[0111] In this way, by setting a threshold, it is ensured that the candidate preset reflection information is adopted only when the similarity is high enough, thereby improving the accuracy of decision making.

[0112] See also Figure 6 In certain embodiments, the method further comprises:

[0113] 03252: When the preset execution result corresponding to the target reflection information is an execution exception, the preset execution operation information in the target reflection information is sent to the vehicle to complete vehicle control.

[0114] In some embodiments, the determination module is further configured to send preset execution operation information in the target reflection information to the vehicle to complete vehicle control when the preset execution result corresponding to the target reflection information is an execution exception.

[0115] In some embodiments, the processor is further configured to send preset execution operation information in the target reflection information to the vehicle to complete vehicle control when the preset execution result corresponding to the target reflection information is an execution exception.

[0116] Specifically, preset execution operation information refers to a series of standardized operation instructions pre-set in the in-vehicle intelligent assistant system. These instructions are used to guide the vehicle's specific actions in specific situations. When the server detects that the preset execution result corresponding to the candidate preset reflection information is an execution exception, it will send this preset execution operation information to the vehicle to complete vehicle control. For example, if a program on the in-vehicle display screen becomes unresponsive, it will exit to the main interface and re-enter the program.

[0117] The server first selects the candidate preset reflection information that best matches the current situation from the preset experience library, namely the target reflection information. Next, the server checks the corresponding preset execution results in these candidate reflection information. If the preset execution result is marked as "execution exception", it means that under similar circumstances, the previous execution action failed to successfully complete the task. In this case, the server will send the preset execution operation information in the candidate preset reflection information to the vehicle. This operation information is a preset solution to the execution exception situation. The vehicle performs corresponding control actions based on the preset execution operation information sent to resolve or correct the execution exception.

[0118] In this way, the system can respond quickly when an execution anomaly is detected and take corrective measures, improving the system's fault handling capabilities and enhancing the stability and reliability of vehicle control, thereby improving the user experience and the overall performance of the system.

[0119] See also Figure 7 In certain embodiments, the method further comprises:

[0120] 03253: When the preset execution result corresponding to the target reflection information is successful execution, based on the first preset model, according to the first voice request and the current state information, determine the target execution action.

[0121] In some embodiments, the determination module is used to determine the target execution action based on the first preset model, the first voice request and the current state information when the preset execution result corresponding to the target reflection information is successful execution.

[0122] In some embodiments, the processor is further configured to determine the target execution action based on the first preset model, the first voice request, and the current state information when the preset execution result corresponding to the target reflection information is successful execution.

[0123] Specifically, the server first filters out candidate preset reflection information that best matches the current situation from the preset experience library. The server checks the corresponding preset execution results in these candidate reflection information. If the preset execution result is marked as "execution successful," this means that the previous execution action successfully completed the task under similar circumstances. In this case, the server will determine the target execution action based on the first voice request and the current state information. This is based on the experience of successful cases, and the server believes that similar execution actions are likely to be successful.

[0124] In this way, the system can use historical success cases to guide current decisions, thereby improving the accuracy and success rate of execution actions, enhancing the system's intelligent decision-making capabilities, and improving task execution efficiency and user experience.

[0125] See also Figure 8 In some embodiments, step 0325 (generating target reflection information based on the target similarity value, the preset threshold, and the candidate preset reflection information) includes:

[0126] 03254: When the target similarity value is less than a preset threshold, a template is generated based on the preset reflection information, and target reflection information is generated according to the preset reflection knowledge, the first voice request and the current state information.

[0127] In some embodiments, the determination module is used to generate a template based on preset reflection information when the target similarity value is less than a preset threshold, and generate target reflection information according to the preset reflection knowledge, the first voice request and the current state information.

[0128] In some embodiments, the processor is further configured to generate a template based on preset reflection information when the target similarity value is less than a preset threshold, and generate target reflection information according to the preset reflection knowledge, the first voice request, and the current state information.

[0129] Specifically, preset reflective knowledge refers to a series of rules and guidelines pre-set in the in-vehicle intelligent assistant system to help the intelligent assistant make correct decisions and operations in specific situations. For example, the preset reflective knowledge is "the agent can be operated when the vehicle is in P gear, and when the vehicle is in D gear, operations that may affect driving require secondary confirmation." The application scenario may be "When the user requests the intelligent assistant to perform an operation while driving, the system needs to check the current gear status of the vehicle. If the vehicle is in D gear (driving gear), the system will ask the user for secondary confirmation to ensure that the operation will not distract the driver's attention or affect driving safety." The preset reflective knowledge is "If the current vehicle screen is on an advertising page, you need to select the return key. If there is no cross or return button on the page, you can click on any blank space." The application scenario may be "When the user wants to perform an operation, but the vehicle screen displays an advertising page, the intelligent assistant needs to recognize this situation and guide the user to return to the main interface. If there is no obvious close button on the page, the intelligent assistant will instruct the user to click on a blank area on the screen to close the advertisement."

[0130] The server first calculates the target similarity value and then compares it with a preset threshold. The preset threshold is a standard set in the system to determine whether the similarity is high enough.

[0131] If the target similarity value is less than a preset threshold, the server deems the existing preset reflection information insufficiently compatible with the current situation. Therefore, based on a preset reflection information generation template, the server generates target reflection information based on the preset reflection knowledge, the first voice request, and the current state information. In some embodiments, the preset reflection information generation template is as follows:

[0132] Imagine you are an intelligent assistant and the agent is in some abnormal state. You need to reflect on the situation and output a reflective action that can help it continue to complete the original task. I will give you the following reference information. Please use this information to output a reflective plan and reflective action:

[0133] User task: <user voice request query>

[0134] Current status: <input information template>

[0135] Historical operation information: <historical plan:P>, <historical action:A>

[0136] Reference knowledge: <Pre-set reflective knowledge>

[0137] Based on the above information, please write your next reflection plan and reflection action in concise language, within 20 words:

[0138] Output:

[0139] A target reflection information generated based on the above-mentioned preset reflection information generation template may be: In the last operation, the agent selected the "Go here" button, but failed to complete the task. The current screen status in the car is an advertising page. The car computer status is P gear, no music is playing, and the temperature in the car is 22 degrees. Now turn off the advertising page on the car display. After the server generates the above-mentioned target reflection information, it processes the target reflection information according to the first preset model to determine whether the task can be carried out. If not, new target reflection information is regenerated until the preset target reflection information generation limit is reached or it is successfully executed. In this process, the target reflection information, voice request and current status generated each time are recorded so that the preset experience library can be updated later.

[0140] In this way, the system can flexibly generate new reflective information when encountering uncommon or novel situations, thereby improving the system's adaptability and intelligence, enhancing the system's ability to handle unknown situations, and promoting the system's continuous learning and self-optimization.

[0141] See also Figure 9 In certain embodiments, the method further comprises:

[0142] 04: When it is confirmed that the target detection object is in an abnormal state according to the current state information, it is confirmed that an abnormality occurs in the target execution action.

[0143] In some embodiments, the confirmation module is used to confirm that an abnormality has occurred in the target execution action when it is confirmed based on the current state information that the target detection object is in an abnormal state.

[0144] In some embodiments, the processor is further configured to confirm that an abnormality has occurred in the target execution action when it is confirmed based on the current state information that the target detection object is in an abnormal state.

[0145] Specifically, an abnormal state of a detected target object refers to a situation in which the detected target object (such as a vehicle component or external environmental factor) in the in-vehicle intelligent assistant system does not conform to normal or expected operation. Examples include: no operation on the in-vehicle display screen for more than n seconds; the vehicle computer status changes repeatedly within n seconds; the in-vehicle display screen status changes repeatedly within n seconds; the user does not respond within n seconds after a follow-up question; the in-vehicle display screen remains on an advertisement page; or the in-vehicle display screen is on an unclickable image page.

[0146] The server continuously receives and monitors the vehicle's current status information, which may include various vehicle performance indicators, sensor data, and user feedback. The server then analyzes this information based on a pre-set model to identify any abnormal conditions. If the analysis indicates an abnormal condition with a target detection object (such as a vehicle component or system), the server confirms that the target execution action has been abnormal. Once the abnormality is confirmed, the server generates target reflection information to ensure that the vehicle can respond safely and effectively to the current situation.

[0147] In this way, the system can promptly detect and respond to abnormalities in the vehicle status, thereby preventing potential failures or dangers, improving the safety and reliability of the vehicle, and enhancing the system's intelligence level and user experience.

[0148] The present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the vehicle control method described above.

[0149] It is understood that a computer program includes computer program code. The computer program code may be in source code form, object code form, executable file, or some intermediate form. Computer-readable storage media may include any entity or device capable of carrying computer program code, recording media, USB flash drives, removable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media.

[0150] In the description of this specification, the descriptions with reference to the terms "particularly", "further", "particularly", "understandably", etc. are intended to mean that the specific features, structures, materials or characteristics described in conjunction with the embodiments or examples are included in at least one embodiment or example of the present application. In this specification, the schematic expressions of the above terms are not intended to refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0151] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code that includes one or more executable requests for implementing a specific logical function or step of a process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may be performed in a different order than shown or discussed, including performing functions in a substantially simultaneous manner or in a reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present application belong.

[0152] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A vehicle control method, characterized in that: The method comprises: Receiving a first voice request forwarded by a vehicle and current status information of the vehicle; Based on a first preset model, determining a target execution action according to the first voice request and the current state information, and sending the target execution action to the vehicle to control the vehicle to execute the target execution action; In the event that an abnormality occurs when the vehicle executes the target execution action, target reflection information is generated according to the first voice request and the current state information, so as to control the vehicle according to the target reflection information.

2. The vehicle control method according to claim 1, characterized in that: The step of generating target reflection information according to the first voice request and the current state information, so as to control the vehicle according to the target reflection information, comprises: Generate first output information according to the current state, the first voice request, and historical operation information associated with the first voice request, wherein the historical operation information includes historical operation description information and historical execution operation information; Based on a second preset model, the target reflection information is generated according to the first output information and a preset experience library, wherein the preset experience library includes multiple preset reflection information, and the preset reflection information includes preset historical operation description information, preset status information, preset execution results and preset execution operation information.

3. The vehicle control method according to claim 2, characterized in that: The generating the target reflection information based on the second preset model according to the first output information and a preset experience library includes: Determining the operation similarity according to the historical operation description information in the first output information and the preset historical operation description information; Determining state similarity according to the current state in the first output information and the preset state information; Determining a target similarity value according to the operation similarity and the state similarity; Determining candidate preset reflection information from the preset experience library according to the target similarity value; The target reflection information is generated according to the target similarity value, a preset threshold and the candidate preset reflection information.

4. The vehicle control method according to claim 3, characterized in that: The step of generating the target reflection information according to the target similarity value, the preset threshold value and the candidate preset reflection information includes: In a case where the maximum value of the target similarity value is greater than or equal to the preset threshold, the target reflection information is generated according to the candidate preset reflection information.

5. The vehicle control method according to claim 4, characterized in that: The method further comprises: When the preset execution result corresponding to the target reflection information is an execution exception, the preset execution operation information in the target reflection information is sent to the vehicle to complete the vehicle control.

6. The vehicle control method according to claim 4, characterized in that: The method further comprises: When the preset execution result corresponding to the target reflection information is successful execution, the target execution action is determined based on the first preset model, according to the first voice request and the current state information.

7. The vehicle control method according to claim 3, characterized in that: The step of generating the target reflection information according to the target similarity value, the preset threshold value and the candidate preset reflection information includes: When the target similarity value is less than the preset threshold, a template is generated based on preset reflection information, and the target reflection information is generated according to preset reflection knowledge, the first voice request and the current state information.

8. The vehicle control method according to claim 1, characterized in that: The method further comprises: When it is confirmed according to the current state information that the target detection object is in an abnormal state, it is confirmed that an abnormality occurs in the target execution action.

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

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images