Information interaction method, equipment, device and vehicle

Abstract

The embodiment of the invention discloses an information interaction method, information interaction equipment, an information interaction device and a vehicle. A target scene identifier is determined according to user input information, current voice style information and vehicle state information are obtained, a plurality of candidate corpora corresponding to the current voice style information are obtained by executing a preset script, and the target scene identifier is obtained based on a predetermined corresponding relation between the candidate corpora and the candidate scene identifier. And determining a target corpus from the plurality of candidate corpora according to the target scene identifier, so that the preset script determines reply information according to the vehicle state information, the user input information and the target corpus. Therefore, the user can experience the reply information of different voice styles on the premise of not relying on the overall upgrading of the vehicle-mounted terminal system, the operation is simple, and the user experience can be improved.

Description

Technical Field

[0001] The present invention relates to the field of voice technology, and in particular to an information interaction method, equipment, device and vehicle. Background Art

[0002] With the continuous improvement of automobile intelligence, the car system is equipped with many pre-set functions to improve the user experience. For example, the vehicle voice system can perform pre-set operations based on the user's voice information, and then generate voice information for reply based on the execution result. However, as the car system software becomes more and more complex, how to quickly update and upgrade the software has become an urgent problem to be solved.

[0003] In the prior art, OTA technology (Over The Air Technology) is usually used to update and upgrade the vehicle system (such as the voice system, etc.). Specifically, the vehicle downloads the upgrade package from the cloud server through the Internet of Vehicles terminal, processes and analyzes the upgrade package, and then updates and upgrades the vehicle system software.

[0004] On the one hand, the existing technology uses OTA technology to perform an overall upgrade of the vehicle system, which requires the development of a corresponding vehicle system upgrade package. Due to the long development cycle, the vehicle voice system update is not flexible enough and has great limitations. On the other hand, the interactive method of generating voice messages based on the execution results in the existing technology may make users feel rigid and inflexible, and cannot meet the differentiated user needs. Summary of the invention

[0005] In view of this, the purpose of the embodiments of the present invention is to provide an information interaction method, device, apparatus and vehicle, which can allow users to experience reply messages in different voice styles without relying on the overall upgrade of the vehicle system, and is simple to operate, thereby improving the user experience.

[0006] In a first aspect, an embodiment of the present invention provides an information interaction method, the method comprising:

[0007] Determine the target scene identifier based on the user input information, and obtain the current voice style information and vehicle status information;

[0008] Execute a preset script to obtain multiple candidate corpora corresponding to the current voice style information from a database;

[0009] Based on the predetermined correspondence between the candidate corpora and the candidate scene identifiers, determining a target corpus from the plurality of candidate corpora according to the target scene identifier;

[0010] Reply information is determined according to the vehicle state information, the user input information and the target corpus.

[0011] In some embodiments, the method further comprises:

[0012] Determine the information interaction interface;

[0013] The target scene identifier, the current voice style information, the vehicle status information and the user input information are imported into the preset script through the information interaction interface.

[0014] In some embodiments, the method further comprises:

[0015] Downloading a speech style information set and a corpus set, wherein the speech style information set includes a plurality of speech style information, and the corpus set includes a plurality of candidate corpora;

[0016] Determine candidate corpora corresponding to each piece of speech style information according to the speech style information set and the corpus set;

[0017] The candidate corpora corresponding to each of the tone style information and each of the voice style information are stored in a database respectively.

[0018] In some embodiments, the preset script is a Lua script, and the method further includes:

[0019] An update file is downloaded, and the correspondence between the candidate corpus and the candidate scene identifier of the Lua script is updated according to the update file.

[0020] In some embodiments, the method further comprises:

[0021] In response to receiving a voice style switching instruction, determining target voice style information from the plurality of voice style information according to the voice style switching instruction;

[0022] The current voice style information is switched to the target voice style information.

[0023] In some embodiments, determining the target scene identifier according to the user input information includes:

[0024] The user input information is semantically understood to determine the target scene identifier.

[0025] In a second aspect, an embodiment of the present invention provides an information interaction device, the device comprising:

[0026] An information acquisition unit, used to determine a target scene identifier according to user input information, and to acquire current voice style information and vehicle status information;

[0027] A candidate corpus determining unit, configured to execute a preset script to obtain a plurality of candidate corpora corresponding to the current voice style information from a database;

[0028] a target corpus determining unit, configured to determine a target corpus from a plurality of the candidate corpora according to the target scene identifier based on a predetermined correspondence between the candidate corpora and the candidate scene identifiers;

[0029] A reply information determination unit is used to determine reply information according to the vehicle state information, the user input information and the target corpus.

[0030] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium storing computer program instructions, wherein the computer program instructions implement the method described in the first aspect when executed by a processor.

[0031] In a fourth aspect, an embodiment of the present invention provides an information interaction device, the device comprising:

[0032] a microphone configured to obtain user input information;

[0033] A controller comprises a memory and a processor, wherein the memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method as described in the first aspect.

[0034] In a fifth aspect, an embodiment of the present invention provides a vehicle, the vehicle comprising:

[0035] Vehicle body;

[0036] A sensor configured to obtain vehicle status information;

[0037] The information interaction device as described in the fourth aspect.

[0038] The embodiment of the present invention determines the target scene identifier according to the user input information, obtains the current voice style information and the vehicle status information, executes the preset script to obtain multiple candidate corpora corresponding to the current voice style information, and determines the target corpora from the multiple candidate corpora according to the target scene identifier based on the predetermined correspondence between the candidate corpora and the candidate scene identifier, and then the preset script determines the reply information according to the vehicle status information, the user input information and the target corpora. Therefore, without relying on the overall upgrade of the vehicle system, the user can experience the reply information of different voice styles, and the operation is simple, which can improve the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The above and other objects, features and advantages of the present invention will become more apparent through the following description of the embodiments of the present invention with reference to the accompanying drawings, in which:

[0040] Figure 1is a flow chart of an information interaction method according to an embodiment of the present invention;

[0041] Figure 2 is a flow chart of an information interaction method according to an embodiment of the present invention;

[0042] Figure 3 is a flow chart of updating voice style information and preset scripts in an embodiment of the present invention;

[0043] Figure 4 is a schematic diagram of the corresponding relationship between speech style information and candidate corpora in an embodiment of the present invention;

[0044] Figure 5 is a data flow diagram in an embodiment of the present invention;

[0045] Figure 6 is a schematic diagram of an application scenario of the information interaction method according to an embodiment of the present invention;

[0046] Figure 7 Schematic diagram of an information interaction device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The present invention is described below based on embodiments, but the present invention is not limited to these embodiments. In the detailed description of the present invention below, some specific details are described in detail. It is possible for a person skilled in the art to fully understand the present invention without the description of these details. In order to avoid confusing the essence of the present invention, known methods, processes, flows, components and circuits are not described in detail.

[0048] In addition, persons of ordinary skill in the art will appreciate that the drawings provided herein are for illustration purposes and are not necessarily drawn to scale.

[0049] Unless the context clearly requires otherwise, words such as “include”, “including” and the like in the specification should be interpreted as including rather than being exclusive or exhaustive; that is, as meaning “including but not limited to”.

[0050] In the description of the present invention, it is to be understood that the terms "first", "second", etc. are only configured for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0051] In the following description, the information interaction method is applied to the interactive scenario of the vehicle voice system as an example for explanation. Among them, the vehicle voice system is such as VOS (Voice Operating System), VCS (Voice Control System), Cortana, Alexa Auto, etc. Among them, VOS and VCS can convert the driver's voice commands into control commands for vehicle functions, such as air conditioning control, music playback, navigation, phone calls, etc. through voice recognition technology, without the need to use manual control buttons or touch screens, and the interaction process is more convenient. Cortana and Alexa Auto belong to a vehicle voice assistant system, which is usually used for vehicle voice control. It can perform a variety of predetermined operations according to the user's predetermined voice, such as sending messages, searching for information, reminding schedules, etc. In the following description, the vehicle voice system is VCS, and the vehicle voice device (that is, the information interaction device) includes a microphone, a speaker and a controller, and the controller includes a vehicle voice system as an example for explanation. Among them, the controller can be an electronic device with functions such as data transmission, data processing and data storage, which is used to generate reply information according to user input information. It should be understood that the information interaction method of the embodiment of the present invention can also be applied to various other types of vehicle systems, such as tailgate control systems (such as voice control to manage vehicle tailgates), battery management systems (such as voice control to obtain information such as vehicle battery power), etc. At the same time, the information interaction method of the embodiment of the present invention can also be applied to various other types of scenarios that require information interaction, such as smart furniture management (such as voice-controlled desk lamps), etc.

[0052] Figure 1 is a flow chart of the information interaction method according to an embodiment of the present invention. Figure 1 As shown, the information interaction process of this embodiment includes the following steps:

[0053] Step S100: determine the target scene identifier according to the user input information, and obtain the current voice style information and vehicle status information.

[0054] In this embodiment, due to the interactive method of generating voice information for replying according to the execution result in the prior art, the user may feel that it is relatively rigid and inflexible, and cannot meet the differentiated user needs. In view of this situation, this embodiment performs voice recognition and semantic understanding on the user input information to determine the target scene identification, and obtains the current voice style information and vehicle status information, and then generates reply information based on the user input information, the target scene identification, the current voice style information and the vehicle status information, so that the user can experience reply information with different voice styles, thereby improving the user experience. Among them, the vehicle status information represents the status information of the vehicle itself, such as control status information (such as air conditioning up, air conditioning down, etc.), parameter status information (such as current vehicle speed, battery power), etc.

[0055] Step S200: Execute a preset script to obtain multiple candidate corpora corresponding to the current voice style information from a database.

[0056] Step S300: Based on the predetermined correspondence between the candidate corpora and the candidate scene identifiers, a target corpus is determined from the plurality of candidate corpora according to the target scene identifier.

[0057] In this embodiment, user input information, target scene identification, current voice style information and vehicle status information can be imported into a preset script, and the preset script determines the target corpus from multiple candidate corpora corresponding to the current voice style information according to the target scene identification. Since the OTA technology is used in the prior art to update and upgrade the vehicle system, a long development cycle is required to develop the upgrade package, resulting in the vehicle voice system update being not flexible enough and having large limitations. In view of this situation, the preset script used in this embodiment is a Lua script, which can be a script file written in C language according to the voice interaction function of the vehicle voice system (ie, VCS) to realize the user's personalized voice style reply requirements. The Lua script is a lightweight scripting language that can be embedded in an application or vehicle system, thereby providing flexible expansion and customization functions for the application or vehicle system. In addition, the Lua script has the characteristics of fast development speed and short online cycle, so the corresponding Lua script can be developed according to the user's personalized customization needs to be uploaded to the cloud server, and the Lua script can be downloaded by the vehicle voice system for update, so that the user can experience the reply information of different voice styles without upgrading the vehicle system as a whole, so as to improve the user experience.

[0058] Optionally, the preset script can also be implemented through Python scripts, JavaScript scripts and Ruby scripts. Among them, Python scripts are written in Python language, which has high text processing and data processing characteristics, so that voice information can be efficiently processed and responses can be generated. JavaScript scripts can be used in browsers or embedded in vehicle voice systems. It has a large number of Web development tools and frameworks, making development more convenient. Ruby scripts have the characteristics of easy maintenance, so that the vehicle voice system can be updated and maintained efficiently.

[0059] Step S400: Determine reply information according to the vehicle state information, the user input information and the target corpus.

[0060] In this embodiment, the Lua script performs combined processing based on the vehicle status information, the user input information and the target corpus to generate the reply information for playback.

[0061] The embodiment of the present invention determines the target scene identifier according to the user input information, obtains the current voice style information and the vehicle status information, executes the preset script to obtain multiple candidate corpora corresponding to the current voice style information, and determines the target corpus from the multiple candidate corpora according to the target scene identifier based on the predetermined correspondence between the candidate corpora and the candidate scene identifier, and then the preset script determines the reply information according to the vehicle status information, the user input information and the target corpus. In this way, without relying on the overall upgrade of the vehicle system, the user can experience the reply information of different voice styles, and the operation is simple, which can improve the user experience

[0062] Figure 2 is a flow chart of the information interaction method according to an embodiment of the present invention. Figure 2 As shown, the information interaction process of this embodiment includes the following steps:

[0063] Step S100': determine the target scene identifier according to the user input information, and obtain the current voice style information and vehicle status information.

[0064] In this embodiment, after receiving the user input information, the vehicle voice system can determine the target scene identifier according to the user input information.

[0065] In an optional implementation, the user input information may be audio input information. Specifically, the vehicle voice system may obtain the audio input information through a microphone disposed in the vehicle. Further, the vehicle voice system may convert the audio input information into a formatted digital signal, and then perform ASR processing (Automatic Speech Recognition) on the digital signal to obtain text input information corresponding to the audio input information.

[0066] In another optional embodiment, the user input information may be text input information. Specifically, the controller for setting the vehicle voice system may be an electronic device with a human-computer interaction function, which may receive instructions input by the user and may also display relevant information to the user. For example, the vehicle voice system may obtain the text input information through a display and an input device (such as a keyboard, etc.) provided in the vehicle. For another example, a touch screen may be provided in the vehicle, and the user may realize human-computer interaction by touching the screen, so that the vehicle voice system obtains the text input information through the touch screen. In the following description, the user input information is taken as an example of audio input information.

[0067] In this embodiment, the vehicle voice system may pre-process the text input information corresponding to the user input information to determine the target scene identification. The vehicle voice system may also import the text input information corresponding to the user input information into a Lua script, and determine the target scene identification according to the text input information through the Lua script.

[0068] In an optional embodiment, the vehicle voice system may perform NLU processing (Natural Language Understanding) on ​​the text input information corresponding to the user input information to determine the target scene information and the intent information, and then the vehicle voice system determines the target scene identifier corresponding to the target scene information from multiple candidate scene identifiers based on the correspondence between the preset candidate scene information (such as text information, etc.) and the candidate scene identifiers. For example, the vehicle voice system presets multiple candidate scene information "air conditioning" and "music", and the corresponding candidate scene identifiers are "N1" and "N2". The text input information corresponding to the user input information is "turn up the air conditioning by 2 degrees". After the vehicle voice system performs NLU processing on the text input information, the target scene information is obtained as "turn up the air conditioning", and the intent information is "turn up by 2 degrees". Then the vehicle voice system can determine that the target scene identifier is N1.

[0069] In another optional embodiment, the vehicle voice system can import the text input information corresponding to the user input information into the Lua script through the Lua engine, and the Lua script calls the semantic recognition library to determine the target scene information and intention information, and then the Lua script determines the target scene identification corresponding to the target scene information from multiple candidate scene identifications based on the correspondence between the preset candidate scene information and the candidate scene identification. Among them, the semantic recognition library includes CMUSphinx, DeepSpeech, etc. In the following description, the vehicle voice system is used to determine the target scene identification as an example.

[0070] In this embodiment, the vehicle voice system can obtain current voice style information. The voice style information includes a voice style identifier and a corresponding voice style configuration file. The voice style identifier can be implemented by a character string, etc. The voice style configuration file is used to define the sound characteristics of a specific voice style (such as a cute voice style, a humorous voice style, etc.), and may include parameter settings in terms of speech speed, pitch, volume, tone, etc.

[0071] In this embodiment, the vehicle voice system can obtain the corresponding vehicle status information at the current moment through sensors (such as temperature sensors, humidity sensors, speedometers, etc.) set in the vehicle according to the target scene information determined by the user input information. For example, if the target scene information is "air conditioning turned up", the vehicle voice system can obtain the corresponding vehicle status information at the current moment including air conditioning temperature, air conditioning humidity, etc. through temperature sensors and humidity sensors.

[0072] In the present embodiment, the Lua engine is used to realize the predetermined function, which can be realized by a software interface, such as a Lua virtual machine and a Lua interpreter, etc. Wherein, the Lua interpreter can read the Lua script code, and can execute the Lua script command to realize the information interaction with the vehicle voice system. The Lua virtual machine defines the execution rules and environment of the Lua script code, which can be embedded in the application and the vehicle voice system, and the Lua virtual machine can have an information interaction interface (such as API (Application Programming Interface, application programming interface), text information interface, etc.) for accessing the vehicle voice system resources. At the same time, the Lua virtual machine and the Lua interpreter, etc. have a custom extension interface, which can be expanded according to user needs, and have high flexibility. In the following description, the software program written by the Lua engine according to the information interaction process described below is used as an example to illustrate.

[0073] Optionally, by setting up a Lua engine in the vehicle voice system, the Lua engine can automatically detect whether there are updated Lua scripts and voice style information sets on the cloud server. If the Lua engine detects that there are updated Lua scripts (that is, update files), voice style information sets, and corpus sets on the cloud server, the Lua engine can automatically download the updated Lua scripts, voice style information sets, and corpus sets for storage. In this way, the vehicle voice system can automatically download updated Lua scripts, voice style information sets, and corpus sets through the Lua engine without the user actively downloading the upgrade package, thereby simplifying user operations. Specifically, before step S100', the information interaction method of this embodiment may also include steps S100"-step S400". For details, please refer to Figure 3 .

[0074] Figure 3 1 is a flow chart of updating voice style information and preset scripts in an embodiment of the present invention. Figure 3 As shown, the process of updating the voice style information and the preset script in this embodiment includes the following steps:

[0075] Step S100 : downloading a speech style information set and a corpus set, wherein the speech style information set includes a plurality of speech style information, and the corpus set includes a plurality of candidate corpora.

[0076] In this embodiment, a voice style information set with different voice style information (such as humorous voice style, cute voice style, etc.) and a corresponding corpus set can be developed and uploaded to the cloud server according to the voice control function of the vehicle voice system (such as air conditioning control, music control, etc.), or a voice style information set with different voice style information and a corresponding corpus set can be developed in an online editing manner on the cloud server system platform. Among them, each voice style information in the voice style information set has a corresponding voice style identifier and a voice style configuration file. The corpus set includes a plurality of candidate corpora, each candidate corpus has a corresponding candidate corpus identifier and a candidate corpus text, and each candidate corpus identifier corresponds to a predetermined candidate scene identifier, and the candidate scene identifier corresponds to the candidate scene information. For example, the voice style information identifier is "A1", corresponding to the "cute voice style configuration file", and the corresponding candidate corpus identifier is A11. The candidate corpus text corresponding to the candidate corpus identifier A11 is "Little brother, the air conditioner is turned on for you", the corresponding candidate scene identifier is A111, and the candidate scene information is "air conditioner is turned on".

[0077] Optionally, the voice style information in the voice style information set can be determined based on the audio information uploaded by the user. For example, the user can record his own audio information and upload it to the cloud server, and then a corresponding voice style information set and a corresponding corpus set can be generated based on the user's audio information. Then, the user can download the voice style information set and the corresponding corpus set corresponding to his own audio information from the cloud. In this way, it is possible to enable users to customize reply information with personalized voice styles, and the operation is simple, the online cycle is short, and the user experience can be improved.

[0078] In this embodiment, if the Lua engine detects that there are updated voice style information sets and corpus sets in the cloud server, the Lua engine can call the communication interface of the vehicle voice system to automatically download the updated voice style information sets and corpus sets for storage. Among them, the communication interface is such as HTTP interface (Hyper Text Transfer Protocol), FTP interface (File Transfer Protocol), WebSocket interface, etc. The communication between the vehicle voice system and the cloud server can be achieved through GSM (Global System for Mobile Communications), GPRS (General packet radio service), eMTC (LTE enhanced MTO, enhanced machine type communication), NB-IoT (NarrowBand Internet of Things) and other technologies.

[0079] In an optional embodiment, the Lua engine can automatically call the communication interface to send a polling signal to the cloud server at a preset time interval to determine whether there is an updated voice style information set and corpus set in the cloud server, and the polling signal can be implemented by RF signals (Radio Frequency, radio frequency signals) and the like. Correspondingly, the cloud server receives the polling signal and can feedback a polling response signal to the Lua engine to notify the Lua engine whether there is an updated voice style information set and corpus set. If the polling response signal indicates that there is an updated voice style information set and corpus set in the cloud server, the Lua script can call the communication interface to send a download request (such as an HTTP communication request, etc.) to the cloud server. The download request may include vehicle information and user information, such as a VIN code (Vehicle Identification Number, vehicle identification code), and user information such as a user ID. Further, the cloud server verifies that the vehicle information and user information in the download request are correct, and sends the updated voice style information set and corpus set to the vehicle voice system for storage. Among them, the preset time interval can be automatically set by the Lua engine, or it can be set by the user according to demand.

[0080] In another optional embodiment, the Lua engine can call a communication interface to subscribe to a communication data channel of the cloud server related to the updated voice style information set and corpus set. If an updated voice style information set and corpus set are uploaded to the cloud server, the cloud server can send a broadcast message to the vehicle voice system by broadcasting to notify the Lua engine that an updated voice style information set and corpus set are available for download. The broadcast message may include vehicle information and user information, and the Lua engine verifies that the vehicle information and user information in the broadcast message are correct, and can call a communication interface to send a download request to the cloud server for downloading.

[0081] Step S200 : determining candidate corpora corresponding to each piece of speech style information according to the speech style information set and the corpus set.

[0082] In this embodiment, the vehicle voice system parses the downloaded voice style information set and corpus set to obtain multiple voice style information and multiple candidate corpora, each voice style information includes a corresponding voice style identifier and a voice style configuration file, and each candidate corpus includes a corresponding candidate corpus identifier, candidate corpus text, candidate scene identifier, and candidate scene information. Then the vehicle voice system can determine multiple candidate corpora corresponding to each voice style information based on the preset correspondence between each voice style identifier and each candidate corpus identifier. For example, the correspondence between the voice style information and the candidate corpus in this embodiment can refer to Figure 4 .

[0083] Figure 4 Schematic diagram of the corresponding relationship between speech style information and candidate corpus in an embodiment of the present invention. Figure 4 As shown, the speech style information includes a speech style identifier and a speech style profile. The candidate corpus includes a candidate corpus identifier, a candidate corpus text, a candidate scene identifier, and candidate scene information. Among them, the speech style identifier "A1" corresponds to the "cute speech style profile" and the candidate corpus identifiers "A11", "A12", and "A13". The speech style identifier "B2" corresponds to the "humorous speech style profile" and the candidate corpus identifiers "B21", "B22", and "B23". The cute speech style profile and the humorous speech style profile can have different parameters in terms of speech speed, pitch, volume, tone, etc. to achieve different speech styles.

[0084] In this embodiment, the candidate corpus identifier "A11" corresponds to the candidate corpus text "Brother, the air conditioner is turned on for you", the candidate scene identifier "A111" and the candidate scene information "the air conditioner is turned on".

[0085] The candidate corpus identifier "A12" corresponds to the candidate corpus text "Brother, I've turned down the air conditioner to Z for you", the candidate scene identifier "A121" and the candidate scene information "turn down the air conditioner".

[0086] The candidate corpus identifier "A13" corresponds to the candidate corpus text "Brother, I've turned up the air conditioner to Z for you", the candidate scene identifier "A131" and the candidate scene information "turn up the air conditioner".

[0087] The candidate corpus identifier "B21" corresponds to the candidate corpus text "My friend, the air conditioner is turned on for you", the candidate scene identifier "B211" and the candidate scene information "the air conditioner is turned on".

[0088] The candidate corpus identifier "B22" corresponds to the candidate corpus text "My friend, I'll turn down the air conditioner to Z for you", the candidate scene identifier "B221" and the candidate scene information "turn down the air conditioner".

[0089] The candidate corpus identifier "B23" corresponds to the candidate corpus text "My friend, I'll turn up the air conditioner to Z for you", the candidate scene identifier "B231" and the candidate scene information "turn up the air conditioner".

[0090] Thus, the vehicle voice system can determine a plurality of candidate corpora corresponding to each voice style information according to the downloaded voice style information set and corpus set.

[0091] Step S300”: store each sound style information and each candidate corpus corresponding to the voice style information in the database respectively.

[0092] In this embodiment, each tone style information and the candidate corpus corresponding to each voice style information can be stored in the database of the vehicle voice system. For example, a corresponding folder can be created in the database for each tone style information, and the tone style information and the corresponding multiple candidate corpus are stored in the folder.

[0093] Step S400 : downloading an update file, and updating the correspondence between the candidate corpus and the candidate scene identifier of the Lua script according to the update file.

[0094] In an optional embodiment, similar to step S100", a Lua update script can be developed based on the correspondence between the candidate scene identifier and the candidate corpus identifier and the correspondence between the candidate corpus identifier and the voice style identifier, and then the Lua update script can be uploaded to the cloud server. Thus, after receiving the user input information, the target corpus can be directly determined based on the target scene identifier and the current voice style information. That is, the Lua script directly determines the target corpus based on the correspondence between the current voice style identifier and the candidate corpus identifier and the correspondence between the candidate corpus identifier and the candidate scene identifier, without obtaining multiple candidate corpora corresponding to the current voice style information from the database. For example, if the target scene identifier is 121 and the identifier of the current voice style information is A1, the Lua script can directly determine the target corpus as "Little brother, I've turned down the air conditioner to Z for you." Thus, the processing process can be simplified, the efficiency of voice reply can be improved, and the user experience can be improved.

[0095] In another optional embodiment, a Lua update script can be developed based on the correspondence between the candidate scene identifiers and the candidate corpus identifiers, and then the Lua update script can be uploaded to the cloud server. Thus, the development speed of the Lua update script can be increased, the development cycle can be reduced, and the user experience can be improved. Among them, the update file is also the Lua update script. In the following description, the development of the Lua update script based on the correspondence between the candidate scene identifiers and the candidate corpus identifiers is used as an example for explanation.

[0096] Optionally, the Lua update script can also be developed in an online editing manner on the cloud server system platform, thereby further improving the development speed of the Lua update script, reducing the development cycle, and improving the user experience.

[0097] In this embodiment, if the Lua engine detects that there is a Lua update script on the cloud server, the Lua engine can call the communication interface of the vehicle voice system to automatically download the Lua update script. The vehicle voice system then updates the correspondence between the candidate corpus identifier and the candidate scene identifier of the current Lua script according to the Lua update script. Among them, the Lua engine can automatically call the communication interface at a preset time interval to send a polling signal to the cloud server, or the Lua engine can also call the communication interface to subscribe to the communication data channel related to the cloud server and the Lua update script to determine whether there is a Lua update script on the cloud server. The specific implementation method is similar to step S100", and this embodiment will not be repeated here.

[0098] Step S200': determine whether a voice style switching instruction is received.

[0099] In this embodiment, the user can control the vehicle voice system through voice style switching instructions according to needs to achieve switching of voice style information.

[0100] In an optional implementation, the voice style switching instruction may be the user's audio information. The vehicle voice system may obtain the user's audio information through a microphone disposed in the vehicle, and then perform ASR processing, NLU processing, etc. on the audio information input by the user to determine the target voice style information and the intention information, and the target voice style information includes the target voice style identifier. For example, if the current voice style information is a cute voice style, and the user's audio information is "switch the voice style information to a humorous voice style", the target voice style identifier is "B2", and the intention information is "switch voice style information".

[0101] In another optional embodiment, the voice style switching instruction can be a user's text information. The vehicle voice system can obtain the user's text information through an input device, a touch screen, etc., and then import the user's text information into a Lua script through a Lua engine, and the Lua script calls a semantic recognition library to determine the target voice style information and intent information.

[0102] In this embodiment, if a voice style switching instruction is received, step S300' is executed. If no voice style switching instruction is received, step S500' is executed.

[0103] Step S300': determining target voice style information from a plurality of voice style information according to the voice style switching instruction.

[0104] In this embodiment, the vehicle voice system can determine the target voice style information from multiple voice style information in the database according to the target voice style identifier corresponding to the voice style switching instruction, and the target voice style information includes a target voice style profile. For example, if the target voice style identifier is "B2", the vehicle voice system can determine its corresponding humorous voice style profile.

[0105] Step S400': determining the target speech style information as the current speech style information.

[0106] In this embodiment, the vehicle voice system switches the current voice style information to the target voice style information. Specifically, the vehicle voice system obtains the target voice style configuration file from the database, and configures it according to the parameters of the speech speed, pitch, volume, tone, etc. in the target voice style configuration file. In this way, the switching of voice style information can be achieved.

[0107] Step S500': determine the information interaction interface.

[0108] In this embodiment, the Lua engine can call a pre-set information interaction interface according to the resource information (such as current voice style information, vehicle status information, etc.) that needs to be obtained. The Lua engine can also register an information interaction interface according to the information that needs to be obtained. Among them, the information interaction interface is used for the Lua engine to establish a communication connection with the vehicle voice system, so that the Lua engine can access and use the resource information of the vehicle voice system. In the following description, the Lua engine can also register an information interaction interface according to the information that needs to be obtained as an example for explanation.

[0109] In this embodiment, the Lua engine can register the information interaction interface based on a variety of communication protocols or customized non-standard communication protocols, such as Modbus communication protocol, RTP communication protocol (Realtime Transport Protocol), json communication protocol, Protobuf communication protocol library, etc.

[0110] In an optional embodiment, the Lua engine encapsulates the target scene identification, current voice style information, vehicle status information, etc. that need to be obtained into predetermined format data, and then the Lua engine registers the information interaction interface according to the predetermined format data. For example, the predetermined format data can be data in NLU json format, etc., and the data in NLU json format belongs to the Rasa training data format. Correspondingly, the information interaction interface can be a custom json interface, etc. Specifically, the data to be trained (such as target scene identification, current voice style information, vehicle status information, etc.) is split to obtain one or more YAML files, each YAML file can include any combination of NLU data, stories data and rules data, and then the type of data to be trained is determined by TLK (Top Level Keys), and then one or more YAML files are converted into data in NLU json format. Thus, the registration of the information interaction interface can be achieved.

[0111] Optionally, the information interaction interface can also be implemented through bus interfaces such as CAN (Controller Area Network), LIN (Local Interconnect Network), RS-485, UART (Universal Asynchronous Receiver / Transmitter). Among them, CAN is a serial communication protocol of the ISO International Organization for Standardization. LIN (bus is a low-cost serial communication protocol based on UART / SCI (Universal Asynchronous Receiver / Serial Interface), mainly used for serial communication between sensors and controllers. The RS-485 bus standard is a very widely used bidirectional, balanced transmission standard interface in industry (attendance, monitoring, data acquisition systems), supporting multi-point connection. UART is a universal serial data bus used for asynchronous communication. The bus has bidirectional communication and can achieve full-duplex transmission and reception.

[0112] In another optional implementation, the Lua engine may load a Lua script and determine a preset interface in the Lua script as the vehicle control interface, the preset interface being, for example, a Lua API interface.

[0113] In another optional embodiment, the Lua engine can register the information interaction interface through a preset Lua interface library. For example, a developer can upload a static file library to a cloud server, and the static file library can be written in C / C++ language. The Lua engine can download the static file library while downloading the Lua script, and then the Lua engine calls the preset interface function (such as dl.open function) in the static file library through the Lua script to synthesize the Lua interface library according to the static file library. The Lua engine can then register the information interaction interface according to the Lua interface library.

[0114] Step S600': import the target scene identification, current voice style information, vehicle status information and user input information into the preset script through the information interaction interface.

[0115] In this embodiment, the Lua engine imports the target scene identification, current voice style information, vehicle status information at the current moment corresponding to the user input information, and intention information determined by the user input information into the Lua script through the information interaction interface.

[0116] In an optional embodiment, there can be multiple target scene identifiers imported into the Lua script, that is, there can be multiple target scene identifiers determined by the user input information. Therefore, the Lua script can determine the corresponding target scene identifier based on the current voice style information. For example, the current voice style identifier is "A1", the user input information is "turn up the air conditioner by 2 degrees", the corresponding target scene information is "turn up the air conditioner", and the corresponding target scene identifiers are "A131" and "B231". Then the Lua script can determine that the corresponding target scene identifier is "A131" based on the current voice style identifier "A1".

[0117] In another optional implementation, the vehicle voice system may determine the target scene identifier corresponding to the current voice style information from a plurality of target scene identifiers determined by the user input information.

[0118] Step S700': execute a preset script to obtain multiple candidate corpora corresponding to the current voice style information from the database.

[0119] Step S800': based on the predetermined correspondence between the candidate corpora and the candidate scene identifiers, determine the target corpus from the plurality of candidate corpora according to the target scene identifier.

[0120] In this embodiment, the Lua engine executes a Lua script, and the Lua script obtains the corresponding multiple candidate corpora from the database according to the current voice style identifier of the current voice style information, and determines the candidate corpus text corresponding to the candidate corpus identifier that is the same as the target scene identifier as the target corpus based on the correspondence between the candidate corpus identifier and the candidate scene identifier. For example, the current voice style identifier is "A1" and the target scene identifier is "A121". The Lua script obtains multiple candidate corpora whose candidate corpora are "A11", "A12" and "A13" from the database, and then determines the candidate corpus text "Little brother, I've turned down the air conditioner to Z for you" corresponding to the candidate corpus identifier that is the same as the target scene identifier "A121" in the multiple candidate corpora as the target corpus.

[0121] Step S900': determining reply information according to vehicle status information, user input information and target corpus.

[0122] In this embodiment, the Lua script generates a reply message by combining the vehicle status information at the current moment corresponding to the user input information, the intention information determined by the user input information, and the target corpus. Specifically, the Lua script generates the adjusted vehicle status information based on the intention information determined by the user input information and the vehicle status information at the current moment. Then the Lua script generates a reply message by combining the adjusted vehicle status information and the target corpus. For example, the vehicle status information at the current moment is "air conditioning temperature 24 degrees", the intention information determined by the user input information is "increase by 2 degrees", and the target corpus is "little brother, the air conditioning is lowered to Z for you". Then the Lua script generates the adjusted vehicle status information as "air conditioning temperature 26 degrees", and then combines the adjusted vehicle status information and the target corpus to generate a reply message as "little brother, the air conditioning is lowered to 26 degrees for you".

[0123] Optionally, the Lua script can perform predetermined functional operations according to the target scene information and intention information corresponding to the user input information while generating the reply information. Specifically, the Lua script can determine the target interface information corresponding to the predetermined functional operation that needs to be performed according to the target scene information and the intention information. The target interface information includes the target interface name, the target interface number, etc. Then the Lua script calls the target interface through the vehicle system according to the target interface information to perform the predetermined functional operation. For example, the target scene information is "air conditioning up", the intention information is "up 2 degrees", and the target interface can be a slot interface. The Lua script determines that the target interface information may include the target interface name "slot name" and the target interface number "slot1". Then the Lua script sends a communication signal to the control system in the vehicle system (such as ECU, MCU (Microcontroller Unit, micro control unit), etc.), and the communication signal includes the target interface information, so that the control system in the vehicle system calls the slot interface to control the air conditioner in the vehicle to perform a heating operation.

[0124] In this embodiment, the Lua script can transmit the reply information to the vehicle voice system through the information interaction interface determined by the Lua engine, and the vehicle voice system displays the reply information. The vehicle voice system can display the reply information after the vehicle system calls the target interface to perform a predetermined function operation, or it can display the reply information during the process of the vehicle system calling the target interface to perform a predetermined function operation.

[0125] In this embodiment, the vehicle voice system can play the reply information through the speaker, or display the reply information on the touch screen or display through a pop-up window, status bar, etc. Figure 5 .

[0126] Figure 5 is a data flow diagram in an embodiment of the present invention. Figure 5 As shown, after the vehicle voice system in the information interaction device receives the user input information 3, it determines the target scene identifier 4 according to the user input information 3, and obtains the current voice style information 1 and the vehicle state information 6 corresponding to the user input information 3, and the current voice style information 1 includes the current voice style identifier. Then the vehicle voice system executes the Lua script through the Lua engine, and the Lua script obtains the corresponding multiple candidate corpora 2 from the database according to the current voice style identifier in the current voice style information 1. Then the Lua script determines the target corpus 5 from the multiple candidate corpora 2 according to the target scene identifier 4. Then the Lua script generates reply information according to the user input information 3, the target corpus 5 and the vehicle state information 6, and the information interaction is completed.

[0127] In this embodiment, the application scenario of the information interaction method is referenced Figure 6 .

[0128] Figure 6 Schematic diagram of an application scenario of the information interaction method of an embodiment of the present invention. Figure 5 As shown, the vehicle includes a vehicle body 110, a sensor 120 and an information interaction device 200. The information interaction device 200 includes multiple speakers, multiple microphones and a controller 231. The multiple speakers include 211 and 212, and the multiple microphones include 221 and 222. A vehicle voice system may be provided in the controller 231. The controller 231 is connected to the multiple speakers, multiple microphones and the sensor 120. The sensor 120 is used to obtain vehicle status information, and the controller 231 is used to generate reply information according to user input information, so as to play the reply information through the speaker.

[0129] In this embodiment, the controller 231 is an electronic device having functions such as data transmission, data processing, data storage and human-computer interaction. The controller 231 may include a general computer hardware structure, which includes at least a processor and a memory. The memory is suitable for storing instructions or programs executable by the processor. The processor may be an independent microprocessor or a collection of one or more microprocessors. Thus, the processor executes the instructions stored in the memory, thereby executing the method flow of the embodiment of the present invention as described above to realize the processing of data and the control of other devices.

[0130] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, devices (equipment) or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present application may adopt a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0131] The present application is described with reference to flowcharts of methods, apparatuses (devices) and computer program products according to embodiments of the present application. It should be understood that each process in the flowchart can be implemented by computer program instructions.

[0132] These computer program instructions may be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device that implements the process Figure 1 A function specified in a process or multiple processes.

[0133] These computer program instructions may also be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce the instructions for implementing the process Figure 1 A device that specifies functions in a process or multiple processes.

[0134] Another embodiment of the present invention relates to a non-volatile storage medium for storing a computer-readable program, wherein the computer-readable program is used for a computer to execute part or all of the above method embodiments.

[0135] That is, those skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by specifying relevant hardware through a program, and the program is stored in a storage medium, including several instructions for a device (which can be a single-chip microcomputer, chip, etc.) or a processor (processor) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0136] For example, see Figure 6 and Figure 4. The current voice style identifier corresponding to the current voice style information is A1. At this time, the vehicle voice system in the controller 231 obtains the voice style switching instruction "Switch the voice style information to the humorous voice style" of the user 10a through the microphone 221, and the vehicle voice system determines that the target voice style identifier is "B2", and the intention information is "switch voice style information". Then the vehicle voice system obtains the "humorous voice style configuration file" corresponding to the target voice style identifier "B2" for configuration. Further, the vehicle voice system obtains the input information of the user 10b "Turn up the air conditioner by 2 degrees" through the microphone 222, then the target scene information of turning up the air conditioner by 2 degrees is determined to be "turn up the air conditioner", the target scene identifiers are "A131" and "B231", and the intention information is "turn up by 2 degrees". At the same time, the vehicle voice system obtains the vehicle status information at the current moment as "air conditioning temperature 24 degrees" and the current voice style identifier is "B2". Then, the vehicle voice system imports the target scene identifiers "A131" and "B231", the vehicle status information "air conditioning temperature 24 degrees" and the current voice style identifier "B2" into the Lua script. Then the Lua engine executes the Lua script, determines that the target scene identifier corresponding to the current voice style identifier "B2" is "B231", and obtains multiple candidate corpus identifiers "B21", "B22" and "B23" and candidate corpus texts corresponding to the current voice style identifier "B2". Then, the Lua script determines that the target corpus is "My friend, the air conditioning is turned up to Z for you" from multiple candidate corpus texts based on the target scene identifier "B231". Finally, the Lua script generates a reply message "My friend, the air conditioning is turned up to 26 for you" based on the target corpus, user input information and vehicle status information, and plays it through speakers 211 and 212, and the information interaction is completed.

[0137] The embodiment of the present invention obtains the Lua script transmitted by the cloud server in real time in the vehicle voice system for updating. After receiving the user input information, the target scene identifier is determined according to the user input information, and the current voice style information and vehicle status information are obtained. The Lua script is executed by the Lua engine to obtain multiple candidate corpora corresponding to the current voice style information, and based on the correspondence between the predetermined candidate corpora and the candidate scene identifier, the target corpus is determined from the multiple candidate corpora according to the target scene identifier, and then the Lua script determines the reply information according to the vehicle status information, the user input information and the target corpus. Therefore, without relying on the overall upgrade of the vehicle system, the user can experience the updated reply information with different voice styles, and the operation is simple, which can improve the user experience.

[0138] Figure 7 Schematic diagram of an information interaction device according to an embodiment of the present invention. Figure 7As shown, the information interaction device includes an information acquisition unit 310, a candidate corpus determination unit 320, a target corpus determination unit 330 and a reply information determination unit 340. Among them, the information acquisition unit 310 is used to determine the target scene identification according to the user input information, and obtain the current voice style information and vehicle status information. The candidate corpus determination unit 320 is used to execute a preset script to obtain multiple candidate corpora corresponding to the current voice style information from the database. The target corpus determination unit 330 is used to determine the target corpus from the multiple candidate corpora according to the target scene identification based on the predetermined correspondence between the candidate corpora and the candidate scene identification. The reply information determination unit 340 is used to determine the reply information according to the vehicle status information, the user input information and the target corpus.

[0139] The embodiment of the present invention determines the target scene identifier according to the user input information, obtains the current voice style information and the vehicle status information, executes the preset script to obtain multiple candidate corpora corresponding to the current voice style information, and determines the target corpora from the multiple candidate corpora according to the target scene identifier based on the predetermined correspondence between the candidate corpora and the candidate scene identifier, and then the preset script determines the reply information according to the vehicle status information, the user input information and the target corpora. Therefore, without relying on the overall upgrade of the vehicle system, the user can experience the reply information of different voice styles, and the operation is simple, which can improve the user experience.

[0140] The solutions described in this specification and in the examples, if they involve the processing of personal information, will be processed on the premise of having a legal basis (such as obtaining the consent of the subject of personal information, or being necessary for the performance of a contract, etc.), and will only be processed within the scope of regulations or agreements. If a user refuses to process personal information other than the necessary information for basic functions, it will not affect the user's use of basic functions.

[0141] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An information interaction method, characterized in that: The method comprises: Determine the target scene identifier based on the user input information, and obtain the current voice style information and vehicle status information; Execute a preset script to obtain multiple candidate corpora corresponding to the current voice style information from a database; Based on the predetermined correspondence between the candidate corpora and the candidate scene identifiers, determining a target corpus from the plurality of candidate corpora according to the target scene identifier; Reply information is determined according to the vehicle state information, the user input information and the target corpus.

2. The method according to claim 1, characterized in that The method further comprises: Determine the information interaction interface; The target scene identifier, the current voice style information, the vehicle status information and the user input information are imported into the preset script through the information interaction interface.

3. The method according to claim 1, characterized in that The method further comprises: Downloading a speech style information set and a corpus set, wherein the speech style information set includes a plurality of speech style information, and the corpus set includes a plurality of candidate corpora; Determine candidate corpora corresponding to each piece of speech style information according to the speech style information set and the corpus set; The candidate corpora corresponding to each of the tone style information and each of the voice style information are stored in a database respectively.

4. The method according to claim 3, characterized in that The preset script is a Lua script, and the method further includes: An update file is downloaded, and the correspondence between the candidate corpus and the candidate scene identifier of the Lua script is updated according to the update file.

5. The method according to claim 3, characterized in that: The method further comprises: In response to receiving a voice style switching instruction, determining target voice style information from the plurality of voice style information according to the voice style switching instruction; The current voice style information is switched to the target voice style information.

6. The method according to claim 1, characterized in that Determining the target scene identifier according to the user input information includes: The user input information is semantically understood to determine the target scene identifier.

7. An information interaction device, characterized in that: The device comprises: An information acquisition unit, used to determine a target scene identifier according to user input information, and to acquire current voice style information and vehicle status information; A candidate corpus determining unit, configured to execute a preset script to obtain a plurality of candidate corpora corresponding to the current voice style information from a database; a target corpus determining unit, configured to determine a target corpus from a plurality of the candidate corpora according to the target scene identifier based on a predetermined correspondence between the candidate corpora and the candidate scene identifiers; A reply information determination unit is used to determine reply information according to the vehicle state information, the user input information and the target corpus.

8. A computer-readable storage medium storing computer program instructions, characterized in that: The computer program instructions implement the method according to any one of claims 1 to 6 when executed by a processor.

9. An information interaction device, characterized in that: The device comprises: a microphone configured to obtain user input information; A controller comprises a memory and a processor, wherein the memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method according to any one of claims 1 to 6.

10. A vehicle, characterized in that: The vehicle comprises: Vehicle body; A sensor configured to obtain vehicle status information; The information interaction device as claimed in claim 9.

Images