Voice instruction processing method and device, equipment, storage medium and program product

Abstract

The application provides a voice instruction processing method, device, equipment, storage medium and program product. It relates to the technical field of vehicle control. The method comprises the following steps: generating a semantic triple based on a voice wake-up instruction, wherein the semantic triple comprises a historical sound source direction, historical control content and a historical control state; receiving a current voice instruction within a preset time length and determining whether the current voice instruction is a simplified instruction; if it is determined that the current voice instruction is a simplified instruction, performing semantic rewriting on the current voice instruction based on the semantic triple to obtain a target instruction, and executing the target instruction. The method can quickly understand user instructions in a vehicle-mounted multi-user scenario by pre-constructing a semantic triple to perform semantic rewriting on a simplified instruction, greatly improves the response speed and recognition accuracy of voice interaction, simplifies the user operation process, and significantly improves driving comfort and user experience.

Description

Technical Field

[0001] This invention relates to the field of vehicle control technology, and in particular to a voice command processing method, apparatus, device, storage medium, and program product. Background Technology

[0002] With the rapid development of intelligent in-vehicle voice systems, voice interaction technology has become one of the important ways for humans and vehicles to interact. Users can conveniently control in-vehicle equipment and functions such as seats, air conditioning, windows, sunroof, lights, and multimedia through voice commands, without manual operation, thereby effectively improving driving safety and driving comfort.

[0003] In related technologies, when there are multiple users in a vehicle, existing in-vehicle voice systems can recognize the voice commands issued by each user and control the in-vehicle equipment to perform corresponding actions based on these commands. Each user's voice command must be standard and conform to a fixed pattern. However, when the first user issues a standardized command, if the second user has the same control intention, they will usually issue a simplified, colloquial command. In this case, the in-vehicle voice system cannot recognize the simplified command issued by the second user, leading to command parsing failure and a poor user experience. Summary of the Invention

[0004] One objective of this invention is to provide a voice command processing method to solve the technical problem that existing in-vehicle voice systems cannot recognize simplified spoken commands issued by users in multi-user scenarios, resulting in command parsing failure and poor user experience; a second objective is to provide a voice command processing device; a third objective is to provide a controller; a fourth objective is to provide a vehicle; a fifth objective is to provide a computer-readable storage medium; and a sixth objective is to provide a computer program product.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A voice command processing method, comprising:

[0007] A semantic triple is generated based on the voice wake-up command. The semantic triple includes the historical sound source location, historical control content, and historical control state.

[0008] Receive the current voice command within a preset time period and determine whether the current voice command is a simplified command;

[0009] If the current voice instruction is determined to be a simplified instruction, then the current voice instruction is semantically rewritten based on the semantic triple to obtain the target instruction, and the target instruction is executed.

[0010] Based on the above technical means, the accurate recognition and efficient execution of simplified commands in multi-user scenarios in vehicles can be achieved. After the user issues a standardized voice wake-up command, the user can directly issue a simplified command in spoken language without having to repeatedly send the standard command. This reduces the cost of user voice interaction, avoids the problem of simplified command parsing failure, and significantly improves the convenience and user experience of in-vehicle voice interaction.

[0011] Furthermore, determining whether the current voice command is a simplified command includes:

[0012] The text of the current voice command is matched with a preset simplified command library, which includes generalized expressions for repetition, following, and the same operation.

[0013] If a match is successful, the current voice command is determined to be a simplified command.

[0014] Based on the above technical means, a clear matching rule is established by pre-setting a simplified instruction library. The current voice instruction type can be quickly determined without complex semantic calculations. While reducing the computing burden on the controller, it can accurately distinguish between simplified and non-simplified instructions, providing a reliable basis for subsequent differentiated processing.

[0015] Furthermore, the step of semantically rewriting the current speech instruction based on the semantic triples to obtain the target instruction includes:

[0016] Obtain the current sound source location corresponding to the current voice command, wherein the current sound source location is the physical location of the user who issued the current voice command, or a specified location obtained from the semantic parsing of the current voice command;

[0017] Based on the semantic triples and combined with the current sound source location, the current voice instruction is semantically rewritten to obtain the target instruction.

[0018] Based on the above technical means, the entity issuing the multi-user command in the vehicle can be accurately located. By combining semantic triples to complete and simplify key information of the command, the confusion and parsing deviation of multi-user commands can be avoided, and the convenience of interaction and the accuracy of execution can be taken into account, thereby improving the in-vehicle voice interaction experience.

[0019] Furthermore, the step of semantically rewriting the current speech instruction based on the semantic triple and in conjunction with the current sound source location to obtain the target instruction includes:

[0020] If the current voice command does not include the current control content and / or does not include the current control state, then the corresponding historical item is extracted from the semantic triple according to the missing item, and the current voice command is completed in combination with the current sound source location to obtain the target command.

[0021] If the current voice command includes both the current control content and the current control state, then, referring to the structure of the semantic triple, the current voice command is semantically normalized by combining the current sound source location, the current control content, and the current control state to obtain the target command.

[0022] Based on the aforementioned technical methods, accurate completion is achieved for different missing scenarios in simplified instructions, while complete instructions are standardized and organized, and the location of the sound source is considered to ensure that the instructions match the user's intentions. Figure 1 This can further improve the accuracy of simplified command rewriting in in-vehicle scenarios and ensure the reliability of voice interaction.

[0023] Furthermore, the generation of semantic triples based on voice wake-up commands includes:

[0024] The location of the historical sound source corresponding to the voice wake-up command is obtained by sound source localization technology.

[0025] The voice wake-up command is semantically parsed to extract the corresponding historical control content and historical control state;

[0026] The historical sound source location, the historical control content, and the historical control state are combined to generate a semantic triple and temporarily stored.

[0027] Based on the aforementioned technical methods, the existing speech recognition and semantic parsing capabilities of the in-vehicle voice system can be fully utilized to obtain core interaction data without the need for additional dedicated hardware, effectively reducing the complexity and cost of system deployment. Simultaneously, by temporarily storing semantic triples, accurate historical context support can be provided for the subsequent intelligent rewriting of simplified commands, improving the convenience of voice control while ensuring the consistency and accuracy of semantic understanding.

[0028] Furthermore, the method also includes:

[0029] If it is determined that the current voice instruction is a non-simplified instruction, then the current voice instruction is parsed to generate the current semantic triple;

[0030] The current voice command is executed, and the current semantic triplet is used as the new historical semantic triplet. The original historical semantic triplet is then updated and replaced before being stored.

[0031] Based on the above technical means, the limitation of processing only simplified instructions can be overcome, and full-scenario coverage of in-vehicle voice commands can be achieved. By dynamically updating the historical semantic triplet, it can be ensured that the historical data on which subsequent simplified instructions are rewritten always match the user's latest control intentions, thereby enhancing the flexibility and adaptability of voice interaction.

[0032] Furthermore, the method also includes:

[0033] Record the execution result of the target instruction or the current voice instruction, wherein the execution result is successful or unsuccessful;

[0034] If the execution result is successful, then the user is notified of the execution completion information;

[0035] If the execution result is an execution failure, a compensation processing mechanism is triggered according to the reason for the failure. The compensation processing mechanism includes at least one of the following: retry mechanism, user operation prompt mechanism, sound source orientation calibration mechanism, and instruction integrity verification mechanism.

[0036] Based on the aforementioned technical means, feedback on execution results allows users to monitor the progress of the interaction in real time, enhancing the intuitiveness and user experience of voice interaction. At the same time, with the help of diverse compensation processing mechanisms, targeted remedies can be provided for scenarios where command execution fails, reducing interaction interruptions caused by factors such as semantic rewriting deviations, sound source localization errors, or insufficient command completeness. This improves the robustness and task completion rate of the voice control system, ensuring that users obtain a smoother and more reliable intelligent interaction experience in continuous dialogue scenarios.

[0037] A voice command processing device, comprising:

[0038] The first processing module is used to generate semantic triples based on voice wake-up commands. The semantic triples include historical sound source location, historical control content, and historical control status.

[0039] The second processing module is used to receive the current voice command within a preset time period and determine whether the current voice command is a simplified command.

[0040] The third processing module is used to, if it is determined that the current voice instruction is a simplified instruction, semantically rewrite the current voice instruction based on the semantic triplet to obtain the target instruction, and execute the target instruction.

[0041] A controller includes: a memory and a processor;

[0042] The memory stores computer-executed instructions;

[0043] The processor executes computer execution instructions stored in the memory, causing the processor to perform the voice command processing method as described in any of the above.

[0044] A vehicle includes: a vehicle body and the controller described above.

[0045] A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the voice command processing method as described in any of the preceding claims.

[0046] A computer program product includes a computer program that, when executed by a processor, is used to implement the voice command processing method as described in any of the preceding claims.

[0047] The beneficial effects of this invention are:

[0048] (1) It can adapt to in-vehicle multi-user interaction scenarios and break through the dependence of existing in-vehicle voice systems on standardized commands. By retaining historical control context through semantic triples, simplified commands can be accurately rewritten, thereby supporting multiple users to continuously issue spoken simplified commands. This effectively solves the problem of command parsing failure caused by the inability to recognize simplified commands in multi-user scenarios, significantly reducing the threshold of user voice interaction and improving user experience.

[0049] (2) It can make full use of the vehicle's existing sound source localization, speech recognition and semantic parsing capabilities. Through lightweight processing logic such as mapping matching and temporary storage, it can not only reduce the system deployment cost and hardware complexity, but also effectively reduce the computing burden of the controller, thereby ensuring the efficiency and real-time responsiveness of voice command processing.

[0050] (3) By using the execution result feedback and diversified compensation processing mechanism, the user can be notified in a timely manner when the instruction is executed successfully, and targeted remedies can be carried out when the execution fails, thereby reducing the occurrence of interaction interruption, improving the robustness of voice control, and allowing users to have a smoother and more reliable interactive experience. Attached Figure Description

[0051] Figure 1 A flowchart illustrating the voice command processing method provided in this application embodiment. Figure 1 ;

[0052] Figure 2 A flowchart illustrating a first example of the voice command processing method provided in this application.

[0053] Figure 3 A flowchart illustrating the voice command processing method provided in this application embodiment. Figure 2 ;

[0054] Figure 4 A flowchart illustrating a second example of the voice command processing method provided in this application.

[0055] Figure 5 A flowchart illustrating a third example of the voice command processing method provided in this application.

[0056] Figure 6 A flowchart illustrating a fourth example of the voice command processing method provided in this application.

[0057] Figure 7 A flowchart illustrating the voice command processing method provided in this application embodiment. Figure 3 ;

[0058] Figure 8 A flowchart illustrating the voice command processing method provided in this application embodiment. Figure 4 ;

[0059] Figure 9 This is a schematic diagram of the structure of the voice command processing device provided in the embodiments of this application;

[0060] Figure 10 This is a schematic diagram of the controller provided in an embodiment of this application. Detailed Implementation

[0061] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0062] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0063] First, the background technology involved in this application will be explained in detail.

[0064] With the rapid development of intelligent in-vehicle voice systems, voice interaction has become one of the core methods of human-vehicle interaction. Users can conveniently control various in-vehicle devices and related functions such as seats, air conditioning, windows, sunroof, lights, and multimedia with voice commands, without the need for manual operation, thereby effectively improving driving safety and vehicle comfort.

[0065] This interaction method is particularly common in multi-user car travel scenarios: when traveling with family, after the driver turns on the seat heating, the front passenger may need to issue a command with the same control intent; when traveling with multiple people, after the front passenger adjusts the multimedia volume, the rear passengers may want to adjust it to the same volume simultaneously; in addition, when the same user issues commands with the same control intent repeatedly within a short period of time, they will habitually use simplified expressions, without having to repeat the complete standardized command. However, the simplified commands in these scenarios are often colloquial and non-fixed paradigms, which are significantly different from the standardized commands required by existing in-vehicle voice systems.

[0066] In related technologies, existing in-vehicle voice systems can only recognize standardized voice commands that conform to fixed paradigms, and have not designed adaptation mechanisms for simplified commands in multi-user scenarios. However, the use of simplified commands is extremely common when multiple users are traveling together, which leads to frequent command parsing failures in the system. This not only prevents users from quickly controlling in-vehicle devices and reduces the convenience and efficiency of voice interaction, but also forces some users to abandon voice control and switch to manual operation of in-vehicle devices. Especially while the vehicle is in motion, manual operation can distract the driver and increase driving safety risks.

[0067] In summary, existing in-vehicle voice systems lack the ability to adapt to simplified commands in multi-user scenarios, resulting in numerous problems such as parsing failures, cumbersome operations, and security risks. This significantly reduces the practicality of voice interaction, preventing users from quickly fulfilling their needs through voice and ultimately leading to a poor user experience.

[0068] Based on the above problems, the technical concept of this application is as follows: A voice command processing method is provided, which can generate semantic triples based on voice wake-up commands. The semantic triples include historical sound source location, historical control content, and historical control state. The method receives the current voice command within a preset time period and determines whether the current voice command is a simplified command. If the current voice command is determined to be a simplified command, the semantics of the current voice command can be rewritten based on the semantic triples to obtain the target command, and the target command can be executed. This solves the shortcomings of existing in-vehicle voice systems, such as inability to recognize simplified commands in multi-user scenarios, frequent parsing failures, cumbersome operation, driving safety risks, and poor user experience. The method achieves accurate recognition and execution of simplified voice commands in multi-user scenarios.

[0069] The technical solution of this application and how it solves the above-mentioned technical problems will be 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.

[0070] Figure 1A flowchart illustrating the voice command processing method provided in this application embodiment. Figure 1 .like Figure 1 As shown, the method includes:

[0071] S101. Generate semantic triples based on voice wake-up commands.

[0072] The execution entity in this application embodiment can be a controller deployed in a vehicle, or a voice command processing device installed in the controller. The voice command processing device can be implemented through software or a combination of software and hardware. The voice command processing device can be a processor in the controller. For ease of understanding, the following description uses a controller as the execution entity.

[0073] Specifically, voice wake-up commands are complete voice commands issued by the user that conform to the system's recognition standards, used to trigger the interactive response of the in-vehicle voice system. For example, commands such as "turn on the driver's seat heater" issued by the driver, "adjust the passenger side air conditioning temperature to 24 degrees" issued by the passenger, and "open the rear window" issued by the rear passenger are all complete statements that clearly include the controlled object and control operation, and can be directly recognized by the in-vehicle voice system and trigger the interaction.

[0074] In one specific implementation, the location of the historical sound source corresponding to the voice wake-up command can be obtained through sound source localization technology. This sound source localization technology can employ an in-vehicle microphone array localization algorithm. Multiple microphones built into the vehicle collect the sound signal of the voice wake-up command, and by combining the time difference and phase difference of the signal arriving at each microphone, the physical location of the voice is accurately calculated, thereby determining the location of the historical sound source. This historical sound source location corresponds to a fixed seating area in the vehicle, such as the driver's seat, front passenger seat, rear left seat, and rear right seat, ensuring the accuracy and specificity of location recognition.

[0075] After acquiring the location of historical sound sources, the controller can perform semantic parsing on the voice wake-up command. It converts the voice signal into text information using a preset speech recognition model, and then uses a semantic extraction algorithm to accurately extract the corresponding historical control content and historical control status from the text information. The historical control content refers to the in-vehicle control object indicated by the voice wake-up command, such as seat heating or air conditioning temperature; the historical control status refers to the specific operational requirements for that control object, such as turn on, lower by one level, turn on, raise by one level, raise by two levels, etc.

[0076] Finally, the controller can associate and combine the acquired historical sound source locations, extracted historical control content, and historical control states to generate a complete semantic triple, and temporarily store the semantic triple in the controller's memory.

[0077] Optionally, the duration of temporary storage for semantic triples can be flexibly set according to the actual in-vehicle interaction scenario. For example, it can be set to any duration between 1 minute and 5 minutes to meet the needs of subsequent related command invocation during continuous multi-user interaction.

[0078] In one optional implementation, the determination of the location of historical sound sources can follow a clear priority rule, with priority from high to low as follows: the specified location obtained from the semantic parsing of the voice wake-up command is higher than the actual physical location of the voice wake-up command issuer. That is, the location of historical sound sources is not necessarily the actual physical location of the command issuer, but rather the semantic parsing result of the voice wake-up command is taken first.

[0079] If the voice wake-up command explicitly specifies the location of the controlled object, then the historical sound source location refers specifically to that specified location, not the actual physical location of the command issuer. If the voice wake-up command does not explicitly specify any location, then the historical sound source location is the actual physical location of the command issuer obtained through sound source localization technology. For example, when a rear-seat user issues the voice wake-up command "Turn on driver's seat heating," the controller performs semantic parsing on the command and can extract the specified location as the driver's seat. According to the priority rules, in the semantic triplet generated at this time, the historical sound source location is the specified location (driver's seat) obtained from semantic parsing, not the rear seat (the actual physical location of the command issuer), the historical control content is seat heating, and the historical control status is "on."

[0080] Furthermore, if a user issues a related command subsequently, the controller will follow the same priority rule when determining the current sound source location: it will first determine the specified location based on the semantic parsing result of the current voice command; if the current voice command does not specify a location, it will then use the physical location of the user who issued the current command as the current sound source location, and use the current sound source location or the specified location obtained from semantic parsing to semantically rewrite the current voice command, ensuring that the location of the command execution accurately matches the current user's intent.

[0081] In the above implementation, the core interactive data can be acquired by leveraging the existing speech recognition and semantic parsing functions of the in-vehicle voice system, without the need for additional hardware investment, effectively reducing the difficulty and cost of system deployment. Furthermore, the temporary storage design of semantic triples provides reliable historical context data for the processing of subsequent related commands, improving the convenience of voice control while ensuring the consistency and accuracy of semantic understanding.

[0082] For example, when the driver issues the voice wake-up command "lower the driver's air conditioning temperature", the controller can generate a semantic triple based on the voice wake-up command. The semantic triple includes the historical sound source location [driver's seat], the historical control content [air conditioning temperature] obtained through semantic parsing, and the historical control status [lower].

[0083] S102. Receive the current voice command within a preset time period and determine whether the current voice command is a simplified command.

[0084] Optionally, the preset duration can be consistent with the duration of temporary storage of the aforementioned semantic triples. After generating and temporarily storing the semantic triples, the controller can start timing synchronously and continuously listen to the user's voice signal collected by the vehicle microphone within the preset duration to receive the current voice command issued by the user. If the preset duration is exceeded, the controller can automatically clear the temporarily stored semantic triples to avoid invalid data occupying memory space and ensure the operating efficiency of the vehicle voice system.

[0085] In one specific implementation, it can be determined whether the current voice command is a simplified command by matching the text of the current voice command with a preset simplified command library. If the match is successful, the current voice command is determined to be a simplified command. The simplified command library includes generalized expressions for representing repetition, following, and identical operations, such as "same," "me too," "lower it a little," "open it too," "set it to the same," "add 1," etc.

[0086] In the above implementation, by matching the text of the current voice command with a preset simplified command library, simplified commands that represent repetition, following, or similar operations can be quickly and accurately identified without complex parsing of the voice commands. This helps to improve command recognition efficiency and ensure the accuracy and reliability of semantic understanding.

[0087] For example, when the passenger in the front seat issues the current voice command "same" within a preset time, the controller can match the text of the current voice command with a preset simplified command library. If the match is successful, the current voice command is determined to be a simplified command.

[0088] S103. If it is determined that the current voice command is a simplified command, then the current voice command is semantically rewritten based on the semantic triple to obtain the target command, and the target command is executed.

[0089] Specifically, when the controller determines that the current voice command is a simplified command, it can obtain the current sound source location corresponding to the current voice command. Based on the semantic triple and combined with the current sound source location, the controller can semantically rewrite the current voice command to obtain the target command. Here, the current sound source location is the physical location of the user who issued the current voice command, or a specified location obtained from the semantic parsing of the current voice command.

[0090] In one specific implementation, regardless of whether the current voice command contains complete control information, as long as the user issues the current voice command, there is a clear current sound source location. However, at least one of the current control content or the current control state may be missing, making it impossible to form a complete and valid control command. In this case, by calling a temporarily stored semantic triple, the simplified command can be completed using the historical control content and historical control state recorded in the semantic triple. This effectively compensates for the missing control information in the simplified command, ensuring the integrity and executability of the target command.

[0091] Figure 2 This is a flowchart illustrating a first example of the voice command processing method provided in this application. Figure 2 As shown, this example mainly focuses on processing repetitive simplified commands. First, the driver issues the voice wake-up command "lower the air conditioning temperature". After semantic parsing of the command, the controller can generate a semantic triple containing the historical sound source location [driver's seat], the historical control content [air conditioning temperature], and the historical control status [lower], and temporarily save it.

[0092] Subsequently, the passenger in the front seat issues a simplified command "I want to too" within a preset time. After the controller determines that the command is a repetitive simplified command, it can obtain the current sound source location [passenger seat] and rewrite the simplified command by combining it with the saved semantic triplet to generate the target command "[passenger seat][air conditioning temperature][lower]" and execute it, thus completing the passenger in the front seat's control intention to adjust the air conditioning temperature in accordance with the driver's command, and the process ends.

[0093] In this embodiment, the controller can generate a semantic triplet containing historical sound source location, historical control content, and historical control state based on the voice wake-up command; receive the current voice command within a preset time period and determine whether the current voice command is a simplified command; if the current voice command is determined to be a simplified command, the controller can semantically rewrite the current voice command based on the generated semantic triplet to obtain the target command and execute the target command. In the above process, through precise binding and flexible adaptation of sound source location, control content, and control state, multiple users can repeatedly trigger historical operations through simplified commands, effectively solving the problems of difficult recognition of simplified commands and semantic parsing failure in multi-user scenarios in vehicles, thus improving the reliability of voice interaction and user experience.

[0094] Figure 3 A flowchart illustrating the voice command processing method provided in this application embodiment. Figure 2 .like Figure 2 As shown, based on the above embodiment one, the specific implementation of the target instruction by semantically rewriting the current voice instruction based on semantic triples in step S103 may include the following steps:

[0095] S301. Obtain the current sound source location corresponding to the current voice command.

[0096] Specifically, after determining that the current voice command is a simplified command, the controller can determine the current sound source location by following the priority rules consistent with the historical sound source location: firstly, based on the semantic parsing results of the current voice command, the explicitly specified control location is extracted; if the current voice command does not explicitly specify any location, the physical location of the user who issued the current command can be obtained through the sound source localization technology of the vehicle microphone array, and this location is used as the current sound source location.

[0097] For example, when the passenger in the front seat issues the simplified command "I want it too", the command does not explicitly specify the control position. The controller can obtain the current physical location of the user as the passenger seat through sound source localization technology, and therefore determine the current sound source location as [passenger seat].

[0098] S302. Based on semantic triples and combined with the current sound source location, the current speech command is semantically rewritten to obtain the target command.

[0099] Specifically, the controller can perform differentiated semantic rewriting of commands based on semantic triples and the current sound source location, according to the control information contained in the current voice command, including:

[0100] First, if the current voice command does not include the current control content and / or the current control state, then the corresponding historical item is extracted from the semantic triple based on the missing item, and the current voice command is completed by combining the current sound source location to obtain the target command.

[0101] Figure 4 This is a flowchart illustrating a second example of the voice command processing method provided in this application. Figure 4 As shown, this example mainly focuses on processing repetitive simplified commands (without specifying control content and control state). First, the driver issues the voice wake-up command "lower the air conditioning temperature". After semantic parsing of the command, the controller can generate a semantic triple containing the historical sound source location [driver's seat], the historical control content [air conditioning temperature], and the historical control state [lower], and temporarily save it.

[0102] Subsequently, the rear-seat user issues a simplified command "rear seat also" within a preset time. After the controller determines that the command is a repetitive simplified command, it can obtain the current sound source location [rear seat] and rewrite the simplified command by combining it with the saved semantic triplet to generate the target command "[rear seat][air conditioning temperature][lower]" and execute it. After the rear-seat user follows the driver's intention to adjust the air conditioning temperature, the process ends.

[0103] Figure 5This is a flowchart illustrating a third example of the voice command processing method provided in this application. Figure 5 As shown, this example mainly focuses on processing simplified adjustment commands (without specifying the control content). First, any user (such as the driver) issues the voice wake-up command "Driver's seat heating up 2 levels". After semantic parsing of the command, the controller can generate a semantic triple containing the historical sound source location [driver's seat], the historical control content [seat heating], and the historical control status [up 2 levels], and temporarily save it.

[0104] Subsequently, any user issues a simplified command "Raise the passenger seat by 1 level" within a preset time. After the controller determines that the command is an adjustment-type simplified command, it can obtain the current sound source location [passenger seat], and rewrite the simplified command by combining it with the saved semantic triplet, generating the target command "[passenger seat][seat heating][Raise by 1 level]" and executing it to complete the adjustment control intention of the passenger seat heating, and the process ends.

[0105] Figure 6 This is a flowchart illustrating a fourth example of the voice command processing method provided in this application. Figure 6 As shown, this example mainly focuses on processing simplified adjustment commands (without specifying the control state). First, any user (such as the driver) issues a voice wake-up command "Driver's screen brightness decrease by 10". After semantic parsing of the command, the controller can generate a semantic triple containing the historical sound source location [driver's seat], the historical control content [screen brightness], and the historical control state [decrease by 10], and temporarily save it.

[0106] Subsequently, if any user issues a simplified command "passenger screen brightness is also" within a preset time period, the controller will determine that the command is an adjustment-type simplified command. It can obtain the current sound source location [passenger] and rewrite the simplified command by combining it with the saved semantic triplet to generate the target command "[passenger][screen brightness][lower by 10]" and execute it to complete the intention to adjust the brightness of the passenger screen, and the process ends.

[0107] Secondly, if the current voice command includes both the current control content and the current control state, then the current voice command is semantically normalized by referring to the structure of the semantic triple, combining the current sound source location, the current control content, and the current control state, to obtain the target command.

[0108] For example, if any user (such as the driver) issues a voice wake-up command "Adjust driver's seat heating to level 2", the controller generates a semantic triple [driver's seat][seat heating][adjust to level 2] and saves it temporarily.

[0109] Subsequently, any user issues the voice command "Raise the passenger seat heating by 1 level" within a preset time period. This command simultaneously includes the current control content [seat heating] and the current control status [raise by 1 level]. After determining that the command is a simplified command, the controller obtains the current sound source location [passenger seat], and, referring to the three-element structure of the semantic triple, performs semantic normalization by combining the current sound source location, current control content, and current control status to generate the target command [passenger seat][seat heating][raise by 1 level] and executes it, thus completing the adjustment control intention for the passenger seat heating.

[0110] In the above implementation method, by implementing differentiated processing strategies for simplified instructions composed of different information, accurate completion is achieved by relying on semantic triples when information is missing, and standardization and regularization are performed by referring to the triple structure when information is complete. Combined with the current sound source location, the directional matching of the instruction intent can be achieved, which can effectively improve the accuracy and adaptability of voice command rewriting in in-vehicle multi-user scenarios.

[0111] In this embodiment, the controller can obtain the current sound source location corresponding to the current voice command, and based on the semantic triplet, semantically rewrite the current voice command in combination with the current sound source location to obtain the target command. In this method, by accurately obtaining the current sound source location and semantically rewriting the current voice command in combination with the semantic triplet, the binding between the command and the user's location can be simplified, adapting to the control needs of different users in multi-user vehicle scenarios, avoiding command execution deviations caused by sound source location confusion, and ensuring the accuracy and relevance of the target command.

[0112] Figure 7 A flowchart illustrating the voice command processing method provided in this application embodiment. Figure 3 .like Figure 7 As shown, based on any of the above embodiments, the voice command processing method further includes:

[0113] S701. If it is determined that the current voice command is a non-simplified command, then parse the current voice command to generate the current semantic triplet.

[0114] Specifically, when the controller determines that the current voice command is a non-simplified command, that is, the current voice command contains a clear control location, control content and control state, and does not need to rely on semantic triples for semantic rewriting, it can directly perform semantic parsing on the current voice command, extract the sound source location, control content and control state, and generate a complete current semantic triple.

[0115] For example, if the driver issues a non-simplified command "increase the driver's air conditioning temperature by 3 degrees", the controller performs semantic parsing on the command and extracts the current sound source location as [driver's seat], the current control content as [air conditioning temperature], and the current control state as [increase by 3 degrees], thereby generating the current semantic triple [driver's seat][air conditioning temperature][increase by 3 degrees].

[0116] S702. Execute the current voice command, and use the current semantic triplet as the new historical semantic triplet. Update and replace the original stored historical semantic triplet before storing it.

[0117] Specifically, after generating the current semantic triplet, the controller can directly execute the current voice command and simultaneously use the current semantic triplet as the new historical semantic triplet, overwriting and replacing the original stored historical semantic triplet, thus completing the storage update.

[0118] For example, in the scenario of "raising the driver's side air conditioning temperature by 3 degrees", if the original stored historical semantic triplet is [driver's side][seat heating][raise by 2 levels], after the controller executes the "raise the driver's side air conditioning temperature by 3 degrees" instruction, it can use the newly generated current semantic triplet [driver's side][air conditioning temperature][raise by 3 degrees] as the new historical semantic triplet, overwriting and replacing the original stored [driver's side][seat heating][raise by 2 levels], thus completing the storage update. This allows subsequent simplified instructions to be rewritten based on the latest historical context.

[0119] In this embodiment, when the controller determines that the current voice command is a non-simplified command, it performs semantic parsing to generate a current semantic triplet, executes the current voice command, and uses the current semantic triplet as a new historical semantic triplet to update and replace the original stored historical semantic triplet before storing it. In this process, by directly parsing the complete non-simplified command and updating the historical context, the latest voice interaction information can be continuously maintained, ensuring that subsequently received simplified commands are always semantically rewritten based on the latest control scenario, thereby improving the coherence and accuracy of in-vehicle multi-user voice interaction.

[0120] Figure 8 A flowchart illustrating the voice command processing method provided in this application embodiment. Figure 4 .like Figure 8 As shown, based on any of the above embodiments, the voice command processing method further includes:

[0121] S801: Record the execution result of the target command or the current voice command.

[0122] Specifically, after the controller completes the execution of the target instruction or the current voice instruction, it can record the execution result, which includes two states: execution success and execution failure.

[0123] Optionally, while recording the execution results, auxiliary information related to the execution of the instructions can also be recorded, such as the instruction content, execution time, sound source location, and device status.

[0124] For example, after the controller executes the target command "[rear row][air conditioning temperature][lower]", it can record the execution status of the command as successful and simultaneously record information such as the location of the sound source and the execution time corresponding to the command.

[0125] S802. If the execution result is successful, then send a message to the user indicating that the execution is complete.

[0126] Specifically, when the controller determines that the instruction has been executed successfully, it can send the corresponding execution completion information to the user who issued the instruction, so that the user is aware that the instruction has been successfully executed.

[0127] The feedback methods can adopt common forms of in-vehicle systems, such as voice broadcast, on-screen text prompts, and icon status updates, to ensure that users can intuitively and quickly obtain the execution results; the feedback content can correspond to the instruction content and concisely and clearly display the execution effect to avoid repeated prompts or invalid prompts.

[0128] For example, after successfully executing "[Passenger seat][Screen brightness][Reduce by 10]", the controller can provide voice feedback to the user that "Passenger seat screen brightness has been adjusted", thus providing timely notification of the execution result.

[0129] S803. If the execution result is an execution failure, a compensation processing mechanism will be triggered based on the reason for the failure.

[0130] Specifically, when the controller determines that the command execution has failed, it can automatically trigger the corresponding compensation mechanism based on the identified reason for the failure, in order to restore the command execution as much as possible or to notify the user of the abnormal situation. The compensation mechanism includes at least one of the following: a retry mechanism, a user operation prompt mechanism, a sound source location calibration mechanism, and a command integrity verification mechanism.

[0131] The retry mechanism refers to a situation where instruction execution fails due to accidental events such as a brief system freeze or temporary communication interruption. Instead of immediately determining failure, the controller waits for a preset time before attempting to execute the instruction again. Automatic retries effectively prevent temporary anomalies from affecting user experience and improve the success rate of instruction execution in complex in-vehicle environments.

[0132] Regarding user operation prompts, this means that when equipment malfunctions or commands cannot be recognized—problems that cannot be automatically resolved—the controller can inform the user of the reason for the failure through voice or screen prompts. This allows the user to clearly understand why the command failed to execute, clarifying subsequent actions and making the entire voice interaction process clearer and more user-friendly.

[0133] The sound source location calibration mechanism can be interpreted as a failure to execute a command due to an error in the system's identification of the user's location. The controller can then re-acquire and analyze the user's voice signal to re-identify and calibrate the sound source location.

[0134] Optionally, if the user does not issue another voice command, the controller can recalculate and correct the orientation based on the sound data collected for this command. If higher accuracy is required, the controller can prompt the user to speak the command again, and complete the orientation calibration after acquiring new sound information. After calibration, the command is reprocessed and executed according to the correct orientation, avoiding command execution errors due to positioning deviations and improving the accuracy of voice interaction in multi-user scenarios.

[0135] The instruction integrity verification mechanism addresses situations where an instruction cannot be executed correctly due to incomplete information or incorrect formatting. In such cases, the controller can check the current instruction content, identify missing control information, and then, by combining this information with previously saved semantic triples, complete the missing content, restoring the instruction to an executable state. After completion, the controller can re-execute the corresponding operation according to the corrected, complete instruction, ensuring that the instruction can be correctly recognized and processed.

[0136] For example, when the controller processes simplified instructions, the generated target instructions may lack control content and therefore cannot be executed. The controller uses an instruction integrity verification mechanism to extract the corresponding control information from the historical semantic triples and complete the instructions. Then, it re-executes the complete instructions to ensure that the instructions take effect normally.

[0137] In this embodiment, the controller can record the execution result of the target command or the current voice command, provide feedback to the user on the execution completion information based on the execution result, or trigger a corresponding compensation mechanism based on the reason for the failure when execution fails. In the above process, by recording and providing feedback on the command execution result, the user can understand the command execution status in a timely manner, improving the transparency of voice interaction; by automatically triggering compensation mechanisms such as retry, prompts, orientation calibration, and command verification when execution fails, various abnormal scenarios can be effectively handled, improving the success rate and reliability of command execution, and further enhancing the user experience.

[0138] Figure 9 This is a schematic diagram of the structure of the voice command processing device provided in an embodiment of this application. Figure 9As shown, the voice command processing device 900 provided in this embodiment includes:

[0139] The first processing module 901 is used to generate semantic triples based on voice wake-up commands. The semantic triples include historical sound source location, historical control content, and historical control status.

[0140] The second processing module 902 is used to receive the current voice command within a preset time period and determine whether the current voice command is a simplified command.

[0141] The third processing module 903 is used to, if it is determined that the current voice instruction is a simplified instruction, semantically rewrite the current voice instruction based on the semantic triplet to obtain the target instruction, and then execute the target instruction.

[0142] Furthermore, the second processing module 902 is specifically used for:

[0143] Match the text of the current voice command with a preset simplified command library, which includes generalized expressions for repetition, following, and the same operation.

[0144] If a match is successful, the current voice command is determined to be a simplified command.

[0145] Furthermore, the third processing module 903 is specifically used for:

[0146] Get the current sound source location corresponding to the current voice command. The current sound source location is the physical location of the user who issued the current voice command, or a specified location obtained from the semantic parsing of the current voice command.

[0147] Based on semantic triples and combined with the current sound source location, the current speech command is semantically rewritten to obtain the target command.

[0148] Furthermore, the third processing module 903 is specifically used for:

[0149] If the current voice command does not include the current control content and / or does not include the current control state, then the corresponding historical item is extracted from the semantic triple based on the missing item, and the current voice command is completed by combining the current sound source location to obtain the target command.

[0150] If the current voice command includes both the current control content and the current control state, then the semantic triple structure is referenced, and the current sound source location, current control content, and current control state are combined to perform semantic normalization on the current voice command to obtain the target command.

[0151] Furthermore, the first processing module 901 is specifically used for:

[0152] The location of the historical sound source corresponding to the voice wake-up command is obtained through sound source localization technology.

[0153] Semantic parsing of voice wake-up commands is performed to extract the corresponding historical control content and historical control status;

[0154] The historical sound source location, historical control content, and historical control status are combined to generate semantic triples and temporarily stored.

[0155] Furthermore, the third processing module 903 is also used for:

[0156] If it is determined that the current voice command is a non-simplified command, then parse the current voice command to generate the current semantic triple;

[0157] Execute the current voice command, and use the current semantic triplet as the new historical semantic triplet. Update and replace the original stored historical semantic triplet before storing it.

[0158] Furthermore, the third processing module 903 is also used for:

[0159] Record the execution result of the target command or the current voice command, which is either successful or unsuccessful.

[0160] If the execution result is successful, then a completion message will be sent to the user.

[0161] If the execution result is an execution failure, a compensation mechanism will be triggered according to the reason for the failure. The compensation mechanism includes at least one of the following: retry mechanism, user operation prompt mechanism, sound source orientation calibration mechanism, and instruction integrity verification mechanism.

[0162] The voice command processing device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0163] Figure 10 This is a schematic diagram of the controller provided in an embodiment of this application. Figure 10 As shown, the controller 1000 provided in this embodiment can be an in-vehicle controller with control logic operation capabilities, such as a cockpit domain controller or a central domain controller in a vehicle. This application does not limit this.

[0164] The controller 1000 includes at least one processor 1001 and a memory 1002. Optionally, the controller 1000 also includes a communication component 1004. The processor 1001, memory 1002, and communication component 1004 are connected via a bus 1003.

[0165] In the specific implementation process, at least one processor 1001 executes computer execution instructions stored in memory 1002, causing at least one processor 1001 to execute the above-mentioned voice instruction processing method.

[0166] The specific implementation process of processor 1001 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0167] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0168] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0169] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0170] This application provides a vehicle, including a vehicle body and... Figure 10 The controller shown is used to implement the voice command processing method described in the above embodiments.

[0171] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0172] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0173] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0174] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0175] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0176] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0177] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0178] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0179] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0180] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

[0181] The above embodiments are merely preferred embodiments provided to fully illustrate the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are all within the scope of protection of the present invention.

Claims

1. A voice command processing method, characterized in that, include: A semantic triple is generated based on the voice wake-up command. The semantic triple includes the historical sound source location, historical control content, and historical control state. Receive the current voice command within a preset time period and determine whether the current voice command is a simplified command; If the current voice instruction is determined to be a simplified instruction, then the current voice instruction is semantically rewritten based on the semantic triple to obtain the target instruction, and the target instruction is executed.

2. The method according to claim 1, characterized in that, Determining whether the current voice command is a simplified command includes: The text of the current voice command is matched with a preset simplified command library, which includes generalized expressions for repetition, following, and the same operation. If a match is successful, the current voice command is determined to be a simplified command.

3. The method according to claim 1 or 2, characterized in that, The semantic rewriting of the current speech instruction based on the semantic triples to obtain the target instruction includes: Obtain the current sound source location corresponding to the current voice command, wherein the current sound source location is the physical location of the user who issued the current voice command, or a specified location obtained from the semantic parsing of the current voice command; Based on the semantic triples and combined with the current sound source location, the current voice instruction is semantically rewritten to obtain the target instruction.

4. The method according to claim 3, characterized in that, The process of semantically rewriting the current speech instruction based on the semantic triples and the current sound source location to obtain the target instruction includes: If the current voice command does not include the current control content and / or does not include the current control state, then the corresponding historical item is extracted from the semantic triple according to the missing item, and the current voice command is completed in combination with the current sound source location to obtain the target command. If the current voice command includes both the current control content and the current control state, then, referring to the structure of the semantic triple, the current voice command is semantically normalized by combining the current sound source location, the current control content, and the current control state to obtain the target command.

5. The method according to claim 1 or 2, characterized in that, The generation of semantic triples based on voice wake-up commands includes: The location of the historical sound source corresponding to the voice wake-up command is obtained by sound source localization technology. The voice wake-up command is semantically parsed to extract the corresponding historical control content and historical control state; The historical sound source location, the historical control content, and the historical control state are combined to generate a semantic triple and temporarily stored.

6. The method according to claim 1 or 2, characterized in that, The method further includes: If it is determined that the current voice instruction is a non-simplified instruction, then the current voice instruction is parsed to generate the current semantic triple; The current voice command is executed, and the current semantic triplet is used as the new historical semantic triplet. The original historical semantic triplet is then updated and replaced before being stored.

7. The method according to claim 1 or 2, characterized in that, The method further includes: Record the execution result of the target instruction or the current voice instruction, wherein the execution result is successful or unsuccessful; If the execution result is successful, then the user is notified of the execution completion information; If the execution result is an execution failure, a compensation processing mechanism is triggered according to the reason for the failure. The compensation processing mechanism includes at least one of the following: retry mechanism, user operation prompt mechanism, sound source orientation calibration mechanism, and instruction integrity verification mechanism.

8. A voice command processing device, characterized in that, include: The first processing module is used to generate semantic triples based on voice wake-up commands. The semantic triples include historical sound source location, historical control content, and historical control status. The second processing module is used to receive the current voice command within a preset time period and determine whether the current voice command is a simplified command. The third processing module is used to, if it is determined that the current voice instruction is a simplified instruction, semantically rewrite the current voice instruction based on the semantic triplet to obtain the target instruction, and execute the target instruction.

9. A controller, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-7.

10. A vehicle, characterized in that, include: The vehicle body and the controller as described in claim 9.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-7.

12. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method described in any one of claims 1-7.

Images