In-vehicle voice interaction method and apparatus, vehicle, medium, and program

Abstract

The present application relates to an in-vehicle voice interaction method and apparatus, a vehicle, a medium, and a program. The method comprises: recognizing voice data inputted by a user; detecting a network status of a vehicle, and switching a processing link of the voice data on the basis of the network status; and processing the voice data on the basis of the processing link to generate a voice recognition result, and on the basis of the voice recognition result, controlling the vehicle to perform a target operation. The present application solves the problems that voice data processing is not intelligent and flexible enough in end-cloud collaboration, and a service policy cannot be dynamically adjusted on the basis of a real-time network and user requirements, thereby affecting overall user experience, quality of service, and the like.

Description

In-vehicle voice interaction methods, devices, vehicles, media and programs

[0001] This application claims priority to Chinese patent application No. 202510004973.0, filed on January 2, 2025, entitled "In-vehicle Voice Interaction Method, Device, Vehicle, Medium and Program", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of vehicle technology, and in particular to an in-vehicle voice interaction method, device, vehicle, medium, and program. Background Technology

[0003] With the rapid development of intelligent vehicles, in-vehicle voice assistants, as an important component of human-computer interaction, are gradually becoming a key technology for improving the driving experience and ensuring driving safety. However, current in-vehicle voice assistants on the market still have many problems, especially in terms of network adaptability, command understanding, and execution. These problems not only affect the user experience but also limit the expansion and application of in-vehicle voice assistant functions to some extent.

[0004] In related technologies, in-vehicle voice assistant solutions are mainly divided into two categories: cloud-based processing and services, and processing models that rely entirely on the vehicle's edge. Cloud-based processing mainly utilizes powerful cloud computing resources to provide more complex and accurate speech recognition and natural language processing services. However, poor network conditions can lead to service delays or unavailability, severely impacting user experience. Edge-based processing, on the other hand, does not rely on the network and can operate without an internet connection. It offers faster response times, but due to the limited computing power and storage resources of in-vehicle devices, edge-based models are usually simpler, resulting in limited speech recognition and natural language understanding capabilities, making it unable to handle complex user requests. Summary of the Invention

[0005] This application provides an in-vehicle voice interaction method, device, vehicle, medium, and program to address the problems in related technologies where voice data processing is not intelligent and flexible enough in terms of end-to-cloud collaboration, and cannot dynamically adjust service strategies according to real-time network and user needs, thus affecting the overall user experience and service quality.

[0006] The first aspect of this application provides an in-vehicle voice interaction method, comprising the following steps: recognizing voice data input by a user; detecting the network status of the vehicle and switching the processing link of the voice data according to the network status; processing the voice data according to the processing link to generate a voice recognition result, and controlling the vehicle to perform a target operation according to the voice recognition result.

[0007] Optionally, the step of switching the voice data processing link according to the network status includes: if the network status is in a first target status, then switching the voice data processing link to a cloud link; if the network status is in a second target status, then switching the voice data processing link to a vehicle-side link.

[0008] Optionally, the step of processing the voice data according to the processing link to generate a voice recognition result includes: if the processing link is a cloud link, then processing the voice data using a voice recognition model and a collaborative model to generate a voice recognition result; if the processing link is a vehicle-side link, then processing the voice data using a voice recognition algorithm and a natural language processing model to generate a voice recognition result.

[0009] Optionally, the step of using a speech recognition model and a collaborative model to process the speech data and generate a speech recognition result includes: inputting the speech data into a speech recognition model, the speech recognition model outputting corresponding speech text, wherein the speech recognition model extracts time-series features from the speech data; matching the time-series features with a preset database to generate corresponding text content; inputting the text content into the collaborative model, the collaborative model outputting a speech recognition result, wherein the collaborative model performs intent recognition and semantic parsing on the text content to generate a corresponding user control intent, and generates a corresponding speech recognition result based on the user control intent.

[0010] Optionally, the step of using a speech recognition algorithm and a natural language processing model to process the speech data and generate a speech recognition result includes: using a speech recognition algorithm to extract key speech features of the speech data to generate corresponding text data; inputting the text data into the natural language processing model, and the natural language processing model outputting a speech recognition result, wherein the natural language processing model performs semantic parsing and intent inference on the text data to generate the corresponding speech recognition result.

[0011] Optionally, the voice recognition result includes an instruction result and a query result, wherein controlling the vehicle to perform the target operation based on the voice recognition result includes: if the voice recognition result is an instruction result, controlling the intelligent agent corresponding to the vehicle to execute the target instruction; if the voice recognition result is a query result, controlling the vehicle to generate corresponding text data or voice data to feed back to the user.

[0012] A second aspect of this application provides an in-vehicle voice interaction device, comprising: a recognition module for recognizing voice data input by a user; a detection module for detecting the network status of the vehicle and switching the processing link of the voice data according to the network status; and a processing module for processing the voice data according to the processing link to generate a voice recognition result and controlling the vehicle to perform a target operation according to the voice recognition result.

[0013] A third aspect of this application provides a vehicle, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to perform the in-vehicle voice interaction method as described in the above embodiments.

[0014] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to perform the in-vehicle voice interaction method as described in the above embodiments.

[0015] A fifth aspect of this application provides a computer program product, including a computer program or instructions, characterized in that, when the computer program or instructions are executed, they implement the in-vehicle voice interaction method as described in the above embodiments.

[0016] Therefore, this application has at least the following beneficial effects:

[0017] (1) The embodiments of this application can recognize the voice data input by the user; switch the voice data processing link according to the network status; process the voice data according to the processing link to generate voice recognition results, and control the vehicle to perform target operations according to the voice recognition results. By monitoring the network status in real time and flexibly switching the cloud-side or end-side link, it ensures that the user's instructions can be responded to accurately and quickly, greatly improving the user experience and enhancing the overall performance and adaptability of the in-vehicle voice assistant, so that it can play a good role in different usage scenarios.

[0018] (2) The embodiments of this application can use speech recognition models and collaborative models to process speech data and generate speech recognition results in the cloud, and use speech recognition algorithms and natural language processing models to process speech data and generate speech recognition results on the vehicle side. This enhances the ability of the in-vehicle voice assistant to cope with complex network environments, greatly improves the accuracy and generalization of command processing, realizes efficient and accurate voice interaction, and meets the diverse and natural command needs of users.

[0019] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0020] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0021] Figure 1 is a flowchart of an in-vehicle voice interaction method according to an embodiment of this application;

[0022] Figure 2 is a schematic diagram of a method for constructing an in-vehicle voice assistant based on a large model according to an embodiment of this application;

[0023] Figure 3 is a schematic diagram of the process for generating enhanced vehicle control agent retrieval according to an embodiment of this application;

[0024] Figure 4 is a block diagram of an example of an in-vehicle voice interaction device provided according to an embodiment of this application;

[0025] Figure 5 is a structural schematic diagram of a vehicle provided according to an embodiment of this application. Detailed Implementation

[0026] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote 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 intended to explain this application, and should not be construed as limiting this application.

[0027] The following description, with reference to the accompanying drawings, describes an in-vehicle voice interaction method, device, vehicle, storage medium, and program according to embodiments of this application.

[0028] Specifically, Figure 1 is a flowchart illustrating an in-vehicle voice interaction method provided in an embodiment of this application.

[0029] As shown in Figure 1, the in-vehicle voice interaction method includes the following steps:

[0030] In step S101, the voice data input by the user is recognized.

[0031] It is understood that the embodiments of this application can recognize the voice data input by the user in order to facilitate the subsequent parsing of the voice content.

[0032] In step S102, the network status of the vehicle is detected, and the voice data processing link is switched according to the network status.

[0033] It is understood that the embodiments of this application can detect the network status of the vehicle and switch the voice data processing link according to the network status. By monitoring the network status in real time and flexibly switching the cloud-side or end-side link, it ensures that user commands can be responded to accurately and quickly, greatly improving the user experience.

[0034] It should be noted that network condition monitoring requires accurate network signal measurement tools and algorithms that can acquire signal strength (in dBm) and data transmission rate (in Mbps) in real time.

[0035] In this embodiment of the application, switching the voice data processing link according to the network status includes: if the network status is in a first target status, then switching the voice data processing link to a cloud link; if the network status is in a second target status, then switching the voice data processing link to a vehicle-side link.

[0036] The first target state can be a good network state, and the second target state can be a poor network state or a state with no network connection.

[0037] It is understood that, in the embodiments of this application, when the network status is in the first target state, the voice data processing link is switched to the cloud link, and when the network status is in the second target state, the voice data processing link is switched to the vehicle-side link. By monitoring the network status in real time and flexibly switching between the cloud-side or vehicle-side links, it is ensured that user commands can be responded to accurately and quickly, which greatly improves the user experience.

[0038] It should be noted that the network monitoring module in the vehicle-mounted equipment continuously collects and analyzes parameters such as network connection signal strength and data transmission rate in real time. If the data transmission rate is higher than a set threshold, such as 5Mbps, the network is considered good; if it is lower than this threshold, the network is considered poor. The threshold for judging a good network can be adjusted according to actual application needs and network environment, and is generally set at a data transmission rate level that can ensure the stable and efficient operation of cloud services.

[0039] In step S103, voice data is processed according to the processing link to generate a voice recognition result, and the vehicle is controlled to perform the target operation based on the voice recognition result.

[0040] It is understood that the embodiments of this application can process voice data according to the processing link to generate voice recognition results, and control the vehicle to perform target operations according to the voice recognition results, thereby meeting the user's control needs, enhancing the overall performance and adaptability of vehicle voice recognition, enabling it to play a good role in different usage scenarios, and improving the user experience.

[0041] In this embodiment of the application, the process of processing voice data according to the processing link to generate a voice recognition result includes: if the processing link is a cloud link, then the voice data is processed using a voice recognition model and a collaborative model to generate a voice recognition result; if the processing link is a vehicle-side link, then the voice data is processed using a voice recognition algorithm and a natural language processing model to generate a voice recognition result.

[0042] It is understood that, in the embodiments of this application, if the processing link is a cloud link, then the speech recognition model and the collaborative model are used to process the speech data to generate the speech recognition result; if the processing link is a vehicle-side link, then the speech recognition algorithm and the natural language processing model are used to process the speech data to generate the speech recognition result, so as to accurately and flexibly process the speech data, meet the diverse and natural command needs of users, and improve the user experience.

[0043] In this embodiment, speech data is processed using a speech recognition model and a collaborative model to generate speech recognition results, including: inputting speech data into a speech recognition model, which outputs corresponding speech text, wherein the speech recognition model extracts time-series features from the speech data; matching the time-series features with a preset database to generate corresponding text content; inputting the text content into a collaborative model, which outputs speech recognition results, wherein the collaborative model performs intent recognition and semantic parsing on the text content to generate corresponding user control intents, and generates corresponding speech recognition results based on the user control intents.

[0044] The preset database can be selected based on actual needs, without specific limitations.

[0045] It is understood that in the embodiments of this application, speech data can be input into a speech recognition model, and the speech recognition model can output corresponding speech text. The speech recognition model extracts time series features from the speech data; generates corresponding text content by matching the time series features with a preset database; inputs the text content into a collaborative model, and the collaborative model outputs speech recognition results. The collaborative model performs intent recognition and semantic parsing on the text content to generate corresponding user control intents, and generates corresponding speech recognition results based on the user control intents, so as to improve the accuracy of speech recognition in the cloud link.

[0046] It should be noted that cloud-based speech recognition utilizes the powerful computing capabilities and advanced speech recognition algorithms of cloud servers to accurately convert user-input speech into text. Collaborative model processing, based on deep learning, combines a large amount of text data and contextual information to accurately determine user intent and, according to preset rules and algorithms, invoke other intelligent agents for input.

[0047] Specifically, as shown in Figure 2, cloud-side link processing:

[0048] (1) Cloud-based speech recognition: When a user issues a voice command, the in-vehicle device quickly transmits the voice data to the cloud server. The cloud server utilizes its powerful computing resources and advanced speech recognition models, such as deep learning-based neural network models, to extract features and match patterns from the speech, accurately converting it into clear text information.

[0049] (2) Collaborative Model Processing: After receiving the speech-to-text result, the collaborative model combines the contextual information obtained from the in-vehicle device, such as previous dialogue content and user preferences, to accurately determine the user's intent through deep analysis and semantic understanding. Then, according to preset rules and agent invocation strategies, the task is assigned to the corresponding agent for processing, such as the vehicle control agent.

[0050] In this embodiment of the application, speech data is processed using a speech recognition algorithm and a natural language processing model to generate speech recognition results, including: extracting key speech features of the speech data using a speech recognition algorithm to generate corresponding text data; inputting the text data into a natural language processing model, and the natural language processing model outputting speech recognition results, wherein the natural language processing model performs semantic parsing and intent inference on the text data to generate corresponding speech recognition results.

[0051] It is understood that the embodiments of this application can use speech recognition algorithms to extract key speech features of speech data to generate corresponding text data; input the text data into a natural language processing model, and the natural language processing model outputs speech recognition results. The natural language processing model performs semantic parsing and intent inference on the text data to generate corresponding speech recognition results, which can not only accurately understand the user's instructions, but also provide more intelligent and personalized services, greatly enhancing the user experience and system reliability.

[0052] It should be noted that (1) End-side speech recognition: With the help of the local speech recognition chip and optimization algorithm of the vehicle device, speech is recognized and converted into text. In the case of poor network conditions, the local speech recognition module of the vehicle device starts working. This module uses an optimized speech recognition algorithm to perform noise reduction, feature extraction and other processing on the collected speech, and convert the speech into text.

[0053] (2) Edge-side Natural Language Understanding: A locally pre-loaded natural language processing model is used to analyze and understand the text. The edge-side pre-loaded natural language understanding model performs semantic parsing and intent inference on the generated text. This model has been compressed and optimized to adapt to the limited computing resources of the in-vehicle device and is able to understand common instructions and user needs.

[0054] Specifically, as shown in Figure 3, vehicle control and ecosystem application commands are completed using the concept of intelligent agents. This abandons the traditional slot-based keyword matching method and adopts a large model to deeply understand user commands, generating voice commands that meet the requirements of the vehicle-mounted system interface. By using function calling technology, which is the process of the user's voice command calling a specific function or service, efficient and accurate interaction with the vehicle-mounted system interface is achieved, and the command is executed.

[0055] Step 1: The user makes an inquiry.

[0056] Step 2: Intent Understanding: Intent understanding refers to receiving a user query, understanding the user's intent in the query, breaking it down into several selection modules, and then using a traditional natural language classification model for judgment.

[0057] Step 3: Query rewriting: Retrieval fusion, similarity construction is performed based on the original query, each query is retrieved using different retrieval methods, and the retrieval content is sorted in reverse order and fused as model input.

[0058] Step 4: Multi-path retrieval, including keyword retrieval, semantic retrieval, and online query.

[0059] Keyword search: The query is segmented into words, and the segmented content is then searched in the knowledge base.

[0060] Semantic retrieval: The query is vectorized, and the query content is retrieved in a vectorized manner.

[0061] Online search: Search for related content via the internet.

[0062] Step 5: Reorder the results. Based on the results of the multi-way queries in Step 4, sort the results in reverse order and select the most relevant results.

[0063] Step 6: Create prompts or guides based on the rearranged content.

[0064] Step 7: Input the data into the large model, and the large model generates the content.

[0065] In this embodiment of the application, the voice recognition result includes an instruction result and an inquiry result. Controlling the vehicle to perform a target operation based on the voice recognition result includes: if the voice recognition result is an instruction result, controlling the intelligent agent corresponding to the vehicle to execute the target instruction; if the voice recognition result is an inquiry result, controlling the vehicle to generate corresponding text data or voice data to feed back to the user.

[0066] It is understood that in the embodiments of this application, if the voice recognition result is an instruction result, the intelligent agent corresponding to the vehicle is controlled to execute the target instruction; if the voice recognition result is a query result, the vehicle is controlled to generate corresponding text data or voice data to feed back to the user, so as to meet the diverse needs of the user and improve the user experience.

[0067] It should be noted that the voice recognition results of this application may include adjusting the in-vehicle temperature, changing radio channels, etc., without being specifically limited.

[0068] According to the in-vehicle voice interaction method proposed in this application, the user's voice input data is recognized; the voice data processing link is switched according to the network status; the voice data is processed according to the processing link to generate a voice recognition result; and the vehicle is controlled to perform the target operation according to the voice recognition result. By monitoring the network status in real time and flexibly switching the cloud-side or end-side link, the user's commands are ensured to receive accurate and rapid responses, greatly improving the user experience and enhancing the overall performance and adaptability of the in-vehicle voice assistant, enabling it to play a good role in different usage scenarios.

[0069] The in-vehicle voice interaction method of this application will be described in detail below with reference to specific embodiments. The in-vehicle device includes a microphone for acquiring voice, a processor for running local speech recognition and natural language understanding modules, a network module for communicating with a cloud server, and a storage unit for storing local models and data. The cloud server has a powerful computing cluster, large-capacity storage devices, and high-speed network connections to support cloud-side speech recognition and collaborative model processing, as detailed below:

[0070] Example 1:

[0071] Voice interaction processing based on cloud links:

[0072] (1) Voice data acquisition and preprocessing

[0073] Microphone capture: Users issue voice commands through the microphone on the in-vehicle device.

[0074] Preprocessing: The system performs preliminary processing on the acquired speech signals, such as noise reduction and gain adjustment, to improve the quality of subsequent processing.

[0075] (2) Upload to cloud server

[0076] Data transmission: The pre-processed voice data is quickly uploaded to the cloud server to ensure processing under good network conditions and with the help of powerful computing resources.

[0077] (3) Speech recognition model processing

[0078] Input speech data: Input the speech data into the pre-trained speech recognition model.

[0079] Time series feature extraction:

[0080] Feature extraction: The speech recognition model first extracts time-series features from the speech data. These features may include phonemes, Mel-frequency coefficients (MFCCs), etc., which can capture the temporal variation characteristics of speech.

[0081] Feature vectorization: Converting extracted time series features into a vector form suitable for model processing.

[0082] Matching preset database:

[0083] Text generation: Based on the extracted time series features, the speech recognition model searches for the most similar pattern in its internal pre-set speech-text mapping database and generates the corresponding text content.

[0084] Text correction: Utilizes language models to further optimize the generated text, correcting any grammatical or semantic errors and ensuring text accuracy.

[0085] (4) Collaborative model processing

[0086] Text input: The generated text content is fed into the collaborative model.

[0087] Intent recognition and semantic parsing:

[0088] Contextual understanding: The collaborative model combines contextual information (such as previous conversation records, user preferences, etc.) to accurately determine the user's intent through deep analysis and semantic understanding.

[0089] Intent classification: Based on the user's intent, tasks are assigned to the corresponding intelligent agents for processing. For example, the vehicle control intelligent agent is responsible for vehicle control-related commands.

[0090] User control intent generation: The collaborative model outputs the final user control intent, which may be a specific command (such as "set the air conditioning temperature to 22 degrees") or a query request (such as "location of nearby gas stations").

[0091] (5) Generate speech recognition results

[0092] Command results vs. query results:

[0093] If it is a result of an instruction, the collaborative model will generate a specific operation instruction and pass it to the corresponding agent for execution.

[0094] If it is a query result, then generate text or voice feedback to be returned to the user.

[0095] (6) Control the vehicle to perform the target operation

[0096] Invoking the intelligent agent: For the result of the instruction, the system invokes the vehicle control intelligent agent or other relevant modules to execute the specific command.

[0097] Feedback to users: Whether it is the result of an instruction or a query, the system will use speech synthesis technology to convert the result into voice feedback to the user, or display it directly on the in-vehicle screen.

[0098] (7) Exception handling

[0099] Error recovery mechanism: If a failure occurs during cloud processing, such as a network interruption, the system will immediately switch back to the vehicle-side link to continue processing the unfinished task, ensuring that the user experience is not affected.

[0100] User Tip: For commands that cannot be recognized or executed, the system should provide clear feedback to the user, explain the reason, and guide the next step.

[0101] For example, suppose a user says, "Set the air conditioner temperature to 22 degrees." The user speaks the command through a microphone, and the system performs noise reduction and gain adjustment on the voice signal. The pre-processed voice data is uploaded to a cloud server. The voice recognition model processes the data: based on time series features, it searches for the most similar pattern in a preset database and generates the text "Set the air conditioner temperature to 22 degrees." The collaborative model, combined with contextual information, understands that this is a control command, the intention of which is to set the air conditioner temperature. It then outputs the user's control intention, namely the instruction "Set the air conditioner temperature to 22 degrees." The system calls upon the vehicle control agent to execute the operation of setting the air conditioner temperature and informs the user through voice synthesis technology that "The air conditioner temperature has been set to 22 degrees."

[0102] Example 2:

[0103] Voice interaction processing based on vehicle-side link:

[0104] (1) Voice data acquisition and preprocessing

[0105] Microphone capture: Users issue voice commands through the microphone on the in-vehicle device.

[0106] Preprocessing: The system performs preliminary processing on the acquired speech signals, such as noise reduction and gain adjustment, to improve the quality of subsequent processing.

[0107] (2) Speech recognition algorithm processing

[0108] Extracting key speech features:

[0109] Feature extraction: Optimized speech recognition algorithms are used to extract key features from the speech data, such as phonemes, Mel-frequency coefficients (MFCCs), and prosodic features. These features capture the temporal variations in speech, helping to improve recognition accuracy.

[0110] Feature vectorization: Converting the extracted key features into a vector form suitable for algorithm processing, so as to facilitate subsequent analysis and matching.

[0111] Generate text data:

[0112] Pattern matching: Based on the extracted key speech features, the system searches for the most similar pattern in a local preset speech-text mapping database and generates the corresponding text content.

[0113] Text correction: The generated text is further optimized using a local language model to correct any grammatical or semantic errors and ensure the accuracy of the text.

[0114] (3) Natural Language Processing Model Processing

[0115] Text data input: The generated text data is fed into a natural language processing model pre-loaded on the in-vehicle device.

[0116] Semantic parsing and intent inference:

[0117] Contextual understanding: NLP models combine contextual information (such as previous conversation records, user preferences, etc.) to accurately determine the user's intent through deep analysis and semantic understanding.

[0118] Intent classification: Based on the user's intent, tasks are assigned to the corresponding intelligent agents for processing. For example, the vehicle control intelligent agent is responsible for vehicle control-related commands.

[0119] User control intent generation: The NLP model outputs the final user control intent, which may be a specific command (such as "set the air conditioning temperature to 22 degrees") or a query request (such as "location of nearby gas stations").

[0120] (4) Generate speech recognition results

[0121] Command results vs. query results:

[0122] If it is a result of an instruction, the NLP model will generate a specific operation instruction and pass it to the corresponding agent for execution.

[0123] If it is a query result, then generate text or voice feedback to be returned to the user.

[0124] (5) Control the vehicle to perform the target operation

[0125] Invoking the intelligent agent: For the result of the instruction, the system invokes the vehicle control intelligent agent or other relevant modules to execute the specific command.

[0126] Feedback to users: Whether it is the result of an instruction or a query, the system will use speech synthesis technology to convert the result into voice feedback to the user, or display it directly on the in-vehicle screen.

[0127] (6) Exception handling

[0128] Error recovery mechanism: If a problem is encountered during processing (such as an unrecognized command), the system will attempt to provide assistance or suggest that the user rephrase the command.

[0129] User Tip: For commands that cannot be recognized or executed, the system should provide clear feedback to the user, explain the reason, and guide the next step.

[0130] In summary, this application is based on speech recognition technology, natural language processing technology, and deep learning principles in artificial intelligence. Speech recognition converts speech into processable text by extracting features and matching patterns from speech signals. Natural language understanding utilizes deep learning models to perform semantic analysis and intent inference on text. Through learning from large amounts of text data, the model can automatically extract patterns and features in language to understand user needs. The neural network model in deep learning, through its multi-layered neuron structure and complex connections, can automatically learn hidden patterns and features in data, thereby achieving accurate speech recognition and natural language understanding. Function calling technology, based on predefined interface specifications and calling rules, enables efficient communication and collaboration between different modules.

[0131] Next, the in-vehicle voice interaction device proposed according to the embodiments of this application is described with reference to the accompanying drawings.

[0132] Figure 4 is a block diagram of an in-vehicle voice interaction device according to an embodiment of this application.

[0133] As shown in Figure 4, the in-vehicle voice interaction device 10 includes: a recognition module 100, a detection module 200, and a processing module 300.

[0134] The recognition module 100 is used to recognize the voice data input by the user; the detection module 200 is used to detect the network status of the vehicle and switch the voice data processing link according to the network status; the processing module 300 is used to process the voice data according to the processing link to generate voice recognition results, and control the vehicle to perform target operations according to the voice recognition results.

[0135] It should be noted that the foregoing explanation of the in-vehicle voice interaction method embodiment also applies to the in-vehicle voice interaction device of this embodiment, and will not be repeated here.

[0136] According to the vehicle-mounted voice interaction device proposed in the embodiments of this application, the device recognizes the voice data input by the user; switches the voice data processing link according to the network status; processes the voice data according to the processing link to generate a voice recognition result; and controls the vehicle to perform the target operation according to the voice recognition result. By monitoring the network status in real time and flexibly switching the cloud-side or end-side link, the device ensures that the user's commands can be responded to accurately and quickly, greatly improving the user experience and enhancing the overall performance and adaptability of the vehicle-mounted voice assistant, enabling it to play a good role in different usage scenarios.

[0137] Figure 5 is a structural schematic diagram of a vehicle provided in an embodiment of this application. The vehicle may include:

[0138] The memory 501, the processor 502, and the computer program stored on the memory 501 and capable of running on the processor 502.

[0139] When the processor 502 executes the program, it implements the in-vehicle voice interaction method provided in the above embodiments.

[0140] Furthermore, the vehicle also includes:

[0141] Communication interface 503 is used for communication between memory 501 and processor 502.

[0142] The memory 501 is used to store computer programs that can run on the processor 502.

[0143] The memory 501 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0144] If the memory 501, processor 502, and communication interface 503 are implemented independently, they can be interconnected via a bus to communicate with each other. 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 representation, only one thick line is used in Figure 5, but this does not indicate that there is only one bus or one type of bus.

[0145] Optionally, in a specific implementation, if the memory 501, processor 502, and communication interface 503 are integrated on a single chip, then the memory 501, processor 502, and communication interface 503 can communicate with each other through an internal interface.

[0146] Processor 502 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0147] This application also provides a computer-readable storage medium storing a computer program or instructions thereon, which, when executed by a processor, implements the above-described in-vehicle voice interaction method.

[0148] This application also provides a computer program product, including a computer program or instructions, characterized in that, when the computer program or instructions are executed, they implement the above-mentioned in-vehicle voice interaction method.

[0149] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0150] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0151] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0152] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or more of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0153] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

Claims

1. A vehicle-mounted voice interaction method, wherein, Includes the following steps: Recognize user-input voice data; Detect the vehicle's network status and switch the voice data processing link according to the network status; The voice data is processed according to the processing link to generate a voice recognition result, and the vehicle is controlled to perform the target operation based on the voice recognition result.

2. The in-vehicle voice interaction method according to claim 1, wherein, The step of switching the voice data processing link according to the network status includes: If the network status is in the first target state, then the voice data processing link is switched to the cloud link; If the network status is in the second target status, then the voice data processing link is switched to the vehicle-side link.

3. The in-vehicle voice interaction method according to claim 1, wherein, The step of processing the speech data according to the processing link to generate a speech recognition result includes: If the processing link is a cloud link, then the speech data is processed using a speech recognition model and a collaborative model to generate speech recognition results; If the processing link is a vehicle-side link, then the speech data is processed using a speech recognition algorithm and a natural language processing model to generate a speech recognition result.

4. The in-vehicle voice interaction method according to claim 3, wherein, The process of using a speech recognition model and a collaborative model to process the speech data and obtain speech recognition results includes: The speech data is input into a speech recognition model, and the speech recognition model outputs corresponding speech text, wherein the speech recognition model extracts time series features from the speech data; and generates corresponding text content by matching the time series features with a preset database. The text content is fed into the collaborative model, and the collaborative model outputs a speech recognition result. The collaborative model performs intent recognition and semantic parsing on the text content to generate a corresponding user control intent, and generates a corresponding speech recognition result based on the user control intent.

5. The in-vehicle voice interaction method according to claim 3, wherein, The process of using speech recognition algorithms and natural language processing models to process the speech data and generate speech recognition results includes: Key speech features of the speech data are extracted using a speech recognition algorithm to generate corresponding text data. The text data is input into the natural language processing model, and the natural language processing model outputs speech recognition results. The natural language processing model performs semantic parsing and intent inference on the text data to generate corresponding speech recognition results.

6. The in-vehicle voice interaction method according to claim 1, wherein, The voice recognition result includes command results and query results, wherein controlling the vehicle to perform the target operation based on the voice recognition result includes: If the voice recognition result is an instruction result, then control the intelligent agent corresponding to the vehicle to execute the target instruction; If the voice recognition result is a query result, the vehicle is controlled to generate corresponding text data or voice data to be fed back to the user.

7. An in-vehicle voice interaction device, wherein, include: The recognition module is used to recognize the voice data input by the user; The detection module is used to detect the network status of the vehicle and switch the processing link of the voice data according to the network status. The processing module is used to process the voice data according to the processing link to generate a voice recognition result, and to control the vehicle to perform the target operation according to the voice recognition result.

8. A vehicle, wherein, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the in-vehicle voice interaction method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program or instructions stored thereon, wherein, When the computer program or instructions are executed by the processor, they are used to implement the in-vehicle voice interaction method as described in any one of claims 1-6.

10. A computer program product comprising a computer program or instructions, wherein, When the computer program or instructions are executed, they implement the in-vehicle voice interaction method as described in any one of claims 1-6.

Images