An interactive method based on smart earphones

Abstract

An interactive method based on smart headphones, the method comprising the following steps: Step 1), the user uses a wake-up phrase or a wireless device to activate the smart headphones, mainly through natural language to activate the smart headphones; the wake-up phrase includes a non-universal wake-up phrase and a universal wake-up phrase; Step 2), after the smart headphones are activated, the wake-up phrase is identified, the user's intention is inferred, and the relevant information is queried and returned through the phrases spoken by the user; Step 3), the sound control and playback module of the smart headphones determines the information playback mode and speed according to different user intentions. The present invention provides an interactive method based on smart headphones, using non-universal wake-up phrases or a wireless device to reduce the embarrassment of users in situations with listeners, form a natural language flow, and change the playback mode and speed of the smart headphones according to different user intentions.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent headset interaction, and in particular to an interaction method based on an intelligent headset. Background Art

[0002] Smart headphones mainly refer to 1) Bluetooth headphones including a small microphone and a small speaker and a terminal system (such as a smartphone) connected to the Bluetooth headphones, or 2) a miniature smartphone including a small microphone and a small speaker and all other components of the phone are made into the shape of headphones and placed in the user's external auditory canal to achieve sound interaction with the user.

[0003] The main limitations of current smart headset technology and major products on the market are:

[0004] 1) Method of activating smart headsets: Activation here means that after the smart headset is turned on, it waits for a certain sound or signal from the user to start receiving the user's voice input. The activation methods of current smart headset technology and major products on the market are: 1A) The user needs to activate the smart headset by operating the mobile phone or the button on the smart headset, or 1B) Use a common wake-up word (such as "Xiaodu Xiaodu") to activate the smartphone.

[0005] For 1A, the user needs to operate it by hand, which increases the inconvenience of the user.

[0006] For 1B, with the popularity of mobile phones and voice-interactive smart assistants from large companies (such as Siri on Apple phones, Xiao Ai on Xiaomi phones, Xiaodu at home, etc.), common wake-up words are often known to other listeners near the user (such as "Xiao Ai", "Xiaodu Xiaodu"). When users say the relevant common wake-up words with their mouths, they will experience a certain degree of embarrassment and cannot hide that they are using smart assistants to obtain relevant information in the presence of listeners. However, in many actual scenarios, users need to query relevant information without the listeners knowing that they are using smart assistants, including specific scenarios such as users telling stories, jokes, reciting poems, singing, etc. to the listeners.

[0007] 2) Playback method and speed: Current smart headsets only play the content that users want to hear (such as query content), and the playback speed is the same regardless of the user's current scenario. This design will cause problems in some important application scenarios: when users are searching for information only for themselves, the speed of playing information by smart headsets can be normal, but if users need to tell the information they hear (such as jokes and stories) to the audience in real time, then the user needs a certain pause between the sentences played by the smart headset so that the user can have time to repeat the information they heard to the audience. The existing interaction method of smart headsets does not take this into account.

[0008] 3) Playback content: The current playback content of smart headphones is generally based on existing information. Users cannot customize or pre-upload relevant information and listen to the content they uploaded. Summary of the invention

[0009] In order to overcome the defects of the existing smart headset interaction methods, the present invention provides an interaction method based on smart headsets, which uses non-universal wake-up words or some wireless devices to reduce the embarrassment of users in situations with listeners, form a natural language flow, and change the playback mode and speed of the smart headset according to different user intentions.

[0010] The technical solution adopted by the present invention is:

[0011] An interactive method based on a smart headset, the method comprising the following steps:

[0012] Step 1), the user uses a wake-up phrase or a wireless device to activate the smart headset, mainly through natural language to activate the smart headset; the wake-up phrase includes a non-universal wake-up phrase and a universal wake-up phrase; the natural language refers to the user activating the smart headset through a non-universal wake-up phrase in the presence of an audience, avoiding the embarrassment of the user using a universal wake-up phrase in the presence of an audience;

[0013] Step 2), after the smart headset is activated, it identifies the wake-up words, infers the user's intention, and queries and returns relevant information through the words spoken by the user;

[0014] Step 2.1), when the smart headset is activated, the voice activation and recognition module of the smart headset recognizes these wake-up words and records the wake-up method, that is, activation by voice or activation by some wireless device. At the same time, the sound control and playback module of the smart headset plays the feedback sound or feedback voice to the user, prompting the user that the smart headset has received the user's wake-up words and is waiting for the user to enter words; for example, "I'm here, please say what you need to play."

[0015] Step 2.2), the user says a phrase containing a keyword, such as "Li Bai's poems";

[0016] Step 2.3), the voice activation and recognition module recognizes the user's words and extracts keywords from the words spoken by the user;

[0017] Step 2.4), the user intention inference and information library service interaction module of the smart headset searches for relevant information library content according to the keywords obtained in step 2.3), including existing information libraries or user-built information libraries, and plays feedback sounds or feedback voices, such as "searching". If the query can be found, the queried content and the activated information library are returned; if the query is not found, the user is fed back that the information was not found; the user intention inference and information library service interaction module of the smart headset will activate and query relevant information libraries according to the user's keywords. For example, if the keyword is "joke", the joke library will be activated, and the query content and the name of the activated information library will be returned; if the keyword contains relevant information about the song, such as the name of the song, the name of a singer, or a line of lyrics of the song, the information library containing the song will be activated, and the query content and the name of the activated information library will be returned; the information library refers to various text or audio information libraries containing songs, which can be existing established A good information database can also be an information database created by the user himself; the information database is divided into two categories, the one containing song information such as lyrics or audio of songs is called a song database, and the one not containing song information is called a non-song database; the non-song database can specifically include but is not limited to: jokes, stories, poems, riddles, brain teasers, love words, food and cooking, encyclopedia knowledge, safety and emergency handling, problem solving, various professional knowledge and dictionaries, life experience, famous quotes and good sentences, conversation skills and suggestions, travel information, health and disease diagnosis and treatment, beauty and weight loss, film and television, cute pets, celebrities, entertainment, sports, technology, home furnishings, constellations, art, history, geography, military, news, etc. as well as information created by users themselves.

[0018] Step 2.5) This step is performed simultaneously with step 2.4. The user intention inference and information base service interaction module of the smart headset infers the user's intention based on the user's way of waking up the smart headset (via voice activation or via a wireless device) and the wake-up phrase (non-universal or universal wake-up phrase):

[0019] meaning Figure 1 : If the user says one of the five types of non-universal wake-up words or presses a certain wireless device, it is inferred that the user's intention is to extract the content of an existing or self-built information library to speak to the audience or sing to the audience; the user intention inference and information library service interaction module will send the intention to the sound control and playback module;

[0020] Intent 2: If the user says a common wake-up phrase, such as the name of the smart headset, it is inferred that the user intends to query information for himself. The user intention inference and information library service interaction module will send the intention to the sound control and playback module.

[0021] Step 3), the sound control and playback module of the smart headset determines the information playback mode and speed according to different user intentions and the activated information library; according to the result of step 2.5), if it is intentional Figure 1 : If the user wants to speak or sing to the audience, go to step 3.1); if it is intention 2, go to step 3.4);

[0022] Step 3.1), determine whether the activated information library is a song library or a non-song library through the user's keywords in step 2.4); if a non-song library is activated, proceed to step 3.2); if a song library is activated, proceed to step 3.3);

[0023] Step 3.2) If it is non-song library activation, then according to the query content returned in step 2.4), the sound control and playback module of the smart headset will process according to the following steps:

[0024] Step 1: Use punctuation marks in the text as markers to segment the paragraphs and sentences into multiple clauses;

[0025] Step 2: Check the length of each sentence. If the length of a sentence exceeds the maximum word limit parameter X, then analyze the grammatical structure of the sentence and automatically split the sentence into multiple sentences. Ensure that the split sentences meet the following requirements: 1) the subject, predicate, and object are as complete as possible; 2) pronouns, nouns, verbs, adjectives, adverbs, phrases (such as idioms), etc. are not split; 3) the total word count is less than the maximum word limit parameter X;

[0026] Step 3: pause intentionally for N seconds after playing each small sentence, where N = the number of words in the small sentence × C; wherein parameter C is the time parameter (in words / second) for the intentional pause after playing a single small sentence. When the user uses the smart headset for the first time, the parameter C has a default value, such as C = 0.4 words / second. The user can also modify the value of the parameter through natural language (see step 4) or the user definition and input module of the smart headset;

[0027] Step 3.3), if the song library is activated, the user phrases obtained in step 2.4) and the results of the search in the song library are divided into two query situations:

[0028] Query situation 1: If the audio of the song can be found, play the audio;

[0029] Query situation 2: If only the lyrics of the song can be found: Query situation 2A) If the voice activation and recognition module of the smart headset can accurately identify the user's current singing progress, the sound control and playback module will play the next line of the song to the user M seconds in advance. M is the advance time for playing lyrics, in seconds. The smart headset has a default value (such as 1 second) when it leaves the factory, and the user can adjust it through the user customization and input module of the smart headset; Query situation 2B) If the smart headset cannot identify the user's singing progress or the recognition error is high, the lyrics will be played according to the lyrics playback table of the song, including the playback time of each line of lyrics.

[0030] Step 3.4): If the user's wake-up word is the name of the smart headset, that is, a universal wake-up word, it is inferred that the user's intention is to query information for himself; the smart headset will play these sounds at normal speed: if a non-song library is activated, the time parameter C for the intentional pause after a single sentence is played is 0; if the lyrics in the song library are played, M=0.

[0031] Furthermore, the interaction method further includes the following steps: Step 4), the user controls the playback of the smart headset through keywords:

[0032] Users can control the playback of smart headphones through the following relatively natural voice interaction methods, especially when there are listeners, to avoid embarrassment.

[0033] Playback situation 1: If the user is not satisfied with the content played by the smart headset and wants to change the content or go to the next message or the next song:

[0034] Step 4.1.1): The user can reactivate the smart headset using any non-universal activation word from category 1 to category 5 or a wireless device.

[0035] Step 4.1.2): The user speaks the keywords that need the information database (such as stories, jokes), or speaks the synonyms of the meaning of change (such as again, next, change, etc.), such as "I will tell you another story".

[0036] Step 4.1.3): After the voice activation and recognition module obtains the keywords of these words and sentences, the smart headset will repeat steps 2) and 3) above.

[0037] Playback situation 2: If the user feels that the pause time between sentences played by the smart headset is too short, that is, the playback speed of the smart headset is too fast.

[0038] Step 4.2.1): The user can reactivate the smart headset using any non-universal activation word from category 1 to category 5 or a wireless device.

[0039] Step 4.2.2): The user can use his mouth to say words and phrases that he speaks too fast and synonymous words and phrases, such as "I may speak too fast."

[0040] Step 4.2.3): The voice activation and recognition module identifies the key words containing "too fast" or its synonyms in the sentence, identifies the user's current intention through the intention inference and information base service interaction module, and then extends the pause time between sentences through the sound control and playback module, that is, increases the value of the time parameter C of the intentional pause after playing a single sentence. At the same time, the smart headset will give the user relevant feedback, such as "Got it, I will slow down the playback speed."

[0041] Playback situation 3: If the user feels that the pauses between sentences played by the smart headset are too long, that is, the playback speed is too slow:

[0042] Step 4.3.1): The user can reactivate the smart headset using any non-universal activation word from Category 1 to Category 5 or a wireless device.

[0043] Step 4.3.2): The user can use his mouth to say words and phrases that he speaks too slowly and synonymous words and phrases, such as "I may speak too slowly."

[0044] Step 4.3.3): The voice activation and recognition module identifies the words and sentences containing the keyword "too slow" or its synonyms, identifies the user's current intention through the intention inference and information library service interaction module, and then reduces the pause time between sentences through the sound control and playback module, that is, reduces the value of the time parameter C of the intentional pause after a single sentence is played. At the same time, the smart headset will give the user relevant feedback, such as "Got it, I will speed up the playback."

[0045] Playback situation 4: If the user wants to pause the playback of the smart headset:

[0046] Step 4.4.1): Condition 1) The intention inference and information library service interaction module of the smart headset extracts that the current user's intention is to extract the content of an existing or self-built information library and tell it to the audience or sing to the audience; and condition 2) is met at the same time) The voice activation and recognition module does not obtain any user's voice within the time range of D seconds, including wake-up words, singing or other words. Where D is the time threshold parameter for pausing the smart headset when there is no signal. There is a default value (such as 10 seconds) when the smart headset leaves the factory. The user can also modify the value of this parameter through natural language or the user definition and input module of the smart headset. When conditions 1) and 2) are met at the same time, the intention inference and information library service interaction module determines that the user's intention is to pause the playback of the headset, and sends the intention to the sound control and playback module.

[0047] Step 4.4.2): The sound control and playback module receives the intention to pause the headset playback and pauses the playback of the sound; at the same time, the playback pause is fed back to the user and the user is prompted with the activation method of the smart headset, such as "I'm in sleep now, you can say, let me think about it, etc., to reactivate me."

[0048] The user can also use multiple buttons of a wireless device to achieve the above functions of changing content, fast forward, slow forward, and pause. Conventional keywords In addition, when the user's intention is to play the content for himself, that is, after step 3.4) is executed, the user can also use conventional keywords, such as pause, close, stop playing, etc. to control the playback of the smart headset.

[0049] Furthermore, the non-universal wake-up phrases include the following five categories:

[0050] Category 1: Words and phrases spoken by the user in the first person that require thinking or recall;

[0051] Category 2: words and sentences that users may know in the first person;

[0052] Category 3: Phrases and words that users ask others;

[0053] Category 4: The user uses the first person to say the words he is about to speak or wants to tell others;

[0054] Category 5: User-defined wake-up phrases;

[0055] The universal wake-up phrase is usually the name of the smart headset.

[0056] Furthermore, the first category of the non-universal wake-up phrases may adopt the following phrases: “Let me think about it”, “Let me think about it”, “I want to think about it”, “Let me think about it”, “Let me think about it”, “Let me think about it”, “Let me recall”, “Let me recall” or synonyms of the above phrases;

[0057] The second category of non-universal wake-up phrases may adopt the following phrases: "I may know", "I may know", "I may know this", "I may know that" or synonyms of the above phrases;

[0058] The third category of non-universal wake-up phrases can use the following phrases: "You know", "I want to ask you", "I want to know", "I want to ask", "Excuse me", "Can you please tell me"

[0059] / you", "Do you know" or synonyms of the above words and phrases;

[0060] The fourth category of non-universal wake-up words and phrases can use the following words and phrases: "Now let me speak", "Let me start the report", "Let me do it", "Let me say it", "Let me say it", "I start now", "I want to say it", "I want to speak", "I want to report", "I want to tell everyone", "I want to tell you", "I want to tell you", "I want to tell you", "I want to report to everyone", "I want to tell everyone" or synonyms of the above words and phrases.

[0061] The fifth category of non-universal wake-up phrases is user-defined non-universal wake-up phrases.

[0062] The user-defined and input module includes the following functions:

[0063] 1) Displaying a list of information libraries and methods for activating each information library, and allowing the user to modify or add or reduce the methods for activating information libraries, and also allowing the user to modify or add or reduce existing information libraries and their contents; the display includes presenting relevant information in a visual manner, and also includes informing the user through voice and other means.

[0064] 2) Display and allow the user to modify or add or delete all wake-up words of the smart headset, and also allow the user to modify the feedback sound of the smart headset;

[0065] 3) Display various parameters of the smart headset and their factory default values, and allow users to customize or adjust these parameters, including the maximum word limit parameter X for a single sentence, the time parameter C for intentional pause after playing a single sentence, the advance time parameter M for playing lyrics, and the time threshold parameter D for pausing when the headset has no signal;

[0066] 4) Allow users to input and modify the information to be played, that is, to build a self-built information library, and allow users to specify one or more activation keywords corresponding to each self-built information library;

[0067] The module may specifically be software or a web page of a terminal to which the smart headset is connected.

[0068] The beneficial effects of the present invention are mainly manifested in: the user can use non-universal wake-up words or other wireless devices to activate the smart headset, avoiding the embarrassment of the user using universal wake-up words to wake up the smart assistant in the presence of listeners; the present invention will form a relatively natural language flow between the user and the smart headset, and can adjust the playback mode and speed of the sound played by the smart headset according to the user's intention, effectively cooperating with the user to speak text content or sing to the audience in the presence of listeners. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a flow chart of the interaction method of the present invention. DETAILED DESCRIPTION

[0070] The present invention will be further described below in conjunction with the accompanying drawings.

[0071] Reference Figure 1 The main functional modules of a smart headset include voice activation and recognition module, user intention inference and information base service interaction module and sound control and playback module. Optional modules are user customization and input module and some wireless device.

[0072] The voice activation and recognition module is used to receive the user's wake-up words and recognize the user's voice, and convert the voice into text information. The module includes a microphone and related voice wake-up and recognition components and wired or wireless connections between them.

[0073] The user intention inference and information library service interaction module is used to infer the user's intention based on the user's wake-up words and subsequent words, and send the recognized information to the relevant information library for query, and send the returned query information to the sound control and playback module. These information libraries can be stored in the local storage of the smart headset or on the cloud server. Users can also build their own information library. If it is in text form, it will be classified into the non-song library; if it is in song form, it will be classified into the song library.

[0074] The sound control and playback module is used to play and control the sound for the user to hear, including a sound control component and a micro speaker, and a wired or wireless connection between them; the sound control component can be a part of the playback program or software.

[0075] User definition and input module, the functions of this module are as follows: 1) Display the list of information libraries and the method of activating each information library (for example, the method of activating the poetry library is the name of a certain author, a certain sentence of poetry, a certain type of poetry, etc.), and allow the user to modify or increase or decrease the method of activating the information library, and also allow the user to modify or increase or decrease the existing information library and its content. The display includes presenting relevant information in a visual way, and also includes informing the user through voice and other means. 2) Display and allow the user to modify or increase or decrease all wake-up words and phrases of the smart headset, and also allow the user to modify the feedback sound of the smart headset. 3) Display various parameters of the smart headset and their factory default values, and allow the user to customize or adjust these parameters, including the maximum word limit parameter X (unit: word) of a single sentence, the time parameter C (unit: word / second) of intentional pause after playing a single sentence, the advance time parameter M (unit: second) of playing lyrics, and the time threshold parameter D (unit: second) of pausing when the headset has no signal; 4) Allow the user to input and modify the information to be played, that is, self-built information library, and let the user specify one or more activation keywords and phrases corresponding to each self-built information library, such as "My report today". The module can specifically be software or a web page of a terminal to which the smart headset is connected, such as an APP or web page on a mobile phone or a WeChat applet, computer software or a web page, etc.

[0076] It also includes a wireless device that activates the smart headset: any device that can wirelessly communicate with the smart headset or smart headset terminal and activate the smart headset, which can be a wireless device independent of the Bluetooth headset, such as a remote control.

[0077] There are many hardware solutions to implement these functional modules, and specific hardware such as processors, memories, batteries, and buttons are not listed.

[0078] Just like existing smart headsets, the above modules can be implemented in a variety of ways: For example, Method 1: The microphone in the voice activation and recognition module and the micro-speaker in the sound control and playback module can be made into a Bluetooth headset and placed in the user's external auditory canal, and connected to a terminal (such as a smartphone) containing other modules and components through wireless communication methods such as Bluetooth. The modules in the terminal include: a voice wake-up and recognition component in the voice activation and recognition module, a user intention inference and information library service interaction module, a sound control component in the sound control and playback module, and an optional user customization and input module, which can be implemented through APP, web pages, mini-programs, or voice interaction without visual display. In addition, the voice wake-up component in the voice activation and recognition module can be implanted in a Bluetooth headset or in a terminal (such as a smartphone).

[0079] Another example is method 2: the voice activation and recognition module, the user intention inference and information library service interaction module, and the sound control and playback module are all integrated into a whole piece of hardware, made into the shape of earphones and placed in the user's external auditory canal. The optional user customization and input module can be implemented through APP, web pages, WeChat applets, or voice interaction without visual display, and connected to the whole hardware through wireless methods such as Bluetooth.

[0080] An interactive method based on a smart headset, the method comprising the following steps:

[0081] Step 1), the user activates the smart headset using a wake-up phrase or a wireless device, mainly through natural language to activate the smart headset; the natural language refers to the user activating the smart headset by a non-universal wake-up phrase in the presence of an audience, avoiding the embarrassment of the user using a universal wake-up phrase in the presence of an audience;

[0082] Step 2) After the smart headset is activated, it identifies the wake-up words, infers the user's intention, and queries and returns relevant information through the words spoken by the user.

[0083] Step 2.1), when the smart headset is activated, the voice activation and recognition module of the smart headset recognizes these wake-up words, and records the wake-up method, that is, whether it is activated by voice or by some wireless device. At the same time, the sound control and playback module of the smart headset plays the feedback sound or feedback voice to the user, prompting the user that the smart headset has received the user's wake-up words and is waiting for the user to enter words; for example, "I'm here, please say what you need to play."

[0084] Step 2.2), the user speaks a phrase containing a keyword, such as "Li Bai's poems".

[0085] Step 2.3), the voice activation and recognition module recognizes the user's words and extracts keywords from the words spoken by the user;

[0086] Step 2.4), the user intention inference and information library service interaction module of the smart headset searches for relevant information library content according to the keywords obtained in step 2.3), including existing information libraries or user-built information libraries, and plays feedback sounds or feedback voices, such as "searching". If the query can be found, the queried content and the activated information library are returned; if the query is not found, the user is fed back that the information has not been found. The user intention inference and information library service interaction module of the smart headset will determine the activated information library based on the user's keywords. For example, if the keyword is "joke", the joke library is activated, and the query content and the name of the activated information library are returned. If the keyword contains relevant information about the song, such as the name of the song, the name of a singer, or a line of lyrics of the song, the song library is activated, and the query content and the name of the activated information library are returned. The information database refers to an information database of various texts or audios containing songs, which can be an existing information database or an information database created by the user. The information database is divided into two categories: the audio containing song information such as lyrics or songs is called a song database, and the audio without song information is called a non-song database. The non-song database can specifically include but is not limited to: jokes, stories, poems, riddles, brain teasers, love words, food and cooking, encyclopedia knowledge, safety and emergency handling, problem solving, various professional knowledge and dictionaries, life experience, famous sayings and good sentences, conversation skills and suggestions, travel information, health and disease diagnosis and treatment, beauty and weight loss, movies, cute pets, celebrities, entertainment, sports, technology, home furnishings, constellations, art, history, geography, military, news, etc., as well as information created by users.

[0087] Step 2.5) This step is performed simultaneously with step 2.4. The user intention inference and information base service interaction module of the smart headset infers the user's intention based on the user's way of waking up the smart headset (via voice activation or via a wireless device) and the wake-up phrase (non-universal or universal wake-up phrase):

[0088] meaning Figure 1 : If the user says one of the five types of non-universal wake-up words or presses a wireless device, it is inferred that the user's intention is to extract the content of an existing or self-built information library to speak to the audience or sing to the audience. The user intention inference and information library service interaction module will send the intention to the sound control and playback module.

[0089] Intent 2: If the user says a common wake-up phrase, such as the name of the smart headset, it is inferred that the user's intention is to query information for himself. The user intention inference and information library service interaction module will send the intention to the sound control and playback module.

[0090] Step 3), the sound control and playback module of the smart headset determines the information playback mode and speed according to different user intentions and the activated information library. According to the result of step 2.5), if it is intentional Figure 1 : If the user wants to speak or sing to the audience, go to step 3.1); if it is intention 2, go to step 3.4).

[0091] Step 3.1), determine whether the activated information library is a song library or a non-song library through the user's keywords in step 2.4); if a non-song library is activated, go to step 3.2); if a song library is activated, go to step 3.3).

[0092] Step 3.2) If it is non-song library activation, then according to the query content returned in step 2.4), the sound control and playback module of the smart headset will process according to the following steps:

[0093] Step 1: Use punctuation marks in the text as markers to segment the paragraphs and sentences into multiple clauses;

[0094] Step 2: Check the length of each sentence. If the length of a sentence exceeds the maximum word limit parameter X, then analyze the grammatical structure of the sentence and automatically split the sentence into multiple sentences. Ensure that the split sentences meet the following requirements: 1) the subject, predicate, and object are as complete as possible; 2) pronouns, nouns, verbs, adjectives, adverbs, phrases (such as idioms), etc. are not split; 3) the total word count is less than the maximum word limit parameter X;

[0095] Step 3: pause intentionally for N seconds after playing each small sentence, where N = the number of words in the small sentence × C; wherein parameter C is the time parameter (in words / second) for the intentional pause after playing a single small sentence. When the user uses the smart headset for the first time, the parameter C has a default value, such as C = 0.4 words / second. The user can also modify the value of the parameter through natural language (see step 4) or the user definition and input module of the smart headset;

[0096] Step 3.3), if the song library is activated, the user phrases obtained in step 2.4) and the results of the search in the song library are divided into two query situations:

[0097] Case 1: If the audio of the song can be found, play the audio;

[0098] Case 2: If only the lyrics of the song can be found: Case 2A) If the voice activation and recognition module of the smart headset can accurately identify the user's current singing progress, the sound control and playback module will play the next line of the song to the user M seconds in advance. M is the advance time for playing lyrics, in seconds. The smart headset has a default value (such as 1 second) when it leaves the factory, and the user can adjust it through the user customization and input module of the smart headset; Case 2B) If the smart headset cannot identify the user's singing progress or the recognition error is high, the lyrics will be played according to the lyrics playback table of the song, including the playback time of each line of lyrics.

[0099] Step 3.4): If the user's wake-up word is the name of the smart headset, that is, a universal wake-up word, it is inferred that the user's intention is to query information for himself; the smart headset will play these sounds at normal speed: if a non-song library is activated, the time parameter C for the intentional pause after a single sentence is played is 0; if the lyrics in the song library are played, M=0.

[0100] Furthermore, the interaction method further comprises the following steps: Step 4), the user controls the playback of the smart headset through keywords:

[0101] Users can control the playback of smart headphones through the following relatively natural voice interaction methods, especially when there are listeners, to avoid embarrassment.

[0102] Case 1: If the user is not satisfied with the content played by the smart headset and wants to change the content or go to the next message or the next song:

[0103] Step 4.1.1): The user can reactivate the smart headset using any non-universal activation word from category 1 to category 5 or a wireless device.

[0104] Step 4.1.2): The user speaks a word containing the keywords in the required information library (such as a story or a joke). Or the user speaks a synonym for "change" (such as "again", "next", "change", etc.), such as "I'll tell you another story".

[0105] Step 4.1.3): After the voice activation and recognition module obtains the keywords of these words and sentences, the smart headset will repeat steps 2) and 3) above.

[0106] Case 2: If the user feels that the pause time between sentences played by the smart headset is too short, that is, the smart headset plays too fast:

[0107] Step 4.2.1): The user can reactivate the smart headset using any non-universal activation word from category 1 to category 5 or a wireless device.

[0108] Step 4.2.2): The user can use his mouth to say words and phrases that he speaks too fast and synonymous words and phrases, such as "I may speak too fast."

[0109] Step 4.2.3): The voice activation and recognition module identifies the key words containing "too fast" or its synonyms in the sentence, identifies the user's current intention through the intention inference and information base service interaction module, and then extends the pause time between sentences through the sound control and playback module, that is, increases the value of the time parameter C of the intentional pause after playing a single sentence. At the same time, the smart headset will give the user relevant feedback, such as "Got it, I will slow down the playback speed."

[0110] Case 3: If the user feels that the pauses between sentences played by the smart headset are too long, that is, the playback speed is too slow:

[0111] Step 4.3.1): The user can reactivate the smart headset using any non-universal activation word from Category 1 to Category 5 or a wireless device.

[0112] Step 4.3.2): The user can use his mouth to say words and phrases that he speaks too slowly and synonymous words and phrases, such as "I may speak too slowly."

[0113] Step 4.3.3): The voice activation and recognition module identifies the key words containing "too slow" or its synonyms in the sentence, identifies the user's current intention through the intention inference and information base service interaction module, and then reduces the pause time between sentences through the sound control and playback module, that is, reduces the value of the time parameter C of the intentional pause after a single sentence is played. At the same time, the smart headset will give the user relevant feedback, such as "Got it, I will speed up the playback."

[0114] Case 4: If the user wants to pause the playback of the smart headset:

[0115] Step 4.4.1): Condition 1) The intention inference and information library service interaction module of the smart headset extracts that the current user's intention is to extract the content of an existing or self-built information library and tell it to the audience or sing to the audience; and condition 2) is met at the same time) The voice activation and recognition module does not obtain any user's voice within the time range of D seconds, including wake-up words, singing or other words. Where D is the time threshold parameter for pausing the smart headset when there is no signal. There is a default value (such as 10 seconds) when the smart headset leaves the factory. The user can also modify the value of this parameter through natural language or the user definition and input module of the smart headset. When conditions 1) and 2) are met at the same time, the intention inference and information library service interaction module determines that the user's intention is to pause the playback of the headset, and sends the intention to the sound control and playback module.

[0116] Step 4.4.2): The sound control and playback module receives the intention to pause the headset playback and pauses the playback of the sound; at the same time, the playback pause is fed back to the user and the user is prompted with the activation method of the smart headset, such as "I'm in sleep now, you can say, let me think about it, etc., to reactivate me."

[0117] The user can also use multiple buttons of a wireless device to achieve the above functions of changing content, fast forward, slow forward, and pause. In addition, when the user intends to play the content for himself, that is, after step 3.4) is executed, the user can also use conventional key words, such as pause, close, stop playing, etc. to control the playback of the smart headset.

[0118] Furthermore, the wake-up words in step 1) include non-universal wake-up words and universal wake-up words, and the non-universal wake-up words include the following five categories:

[0119] Category 1: Words and phrases spoken by the user in the first person that require thinking or recall;

[0120] Category 2: words and sentences that users may know in the first person;

[0121] Category 3: Phrases and words that users ask others;

[0122] Category 4: The user uses the first person to say the words he is about to speak or wants to tell others;

[0123] Category 5: User-defined wake-up phrases;

[0124] The universal wake-up phrase is usually the name of the smart headset.

[0125] The first category of the non-universal wake-up words can adopt the following words: "Let me think about it", "Let me think about it", "I'll think about it", "Let me think about it", "Let me think about it", "Let me consider it", "Let me recall", "Let me recall" or synonyms of the above words.

[0126] The second category of non-universal wake-up phrases may adopt the following phrases: "I may know", "I may know", "I may know this", "I may know that" or synonyms of the above phrases;

[0127] The third category of non-universal wake-up phrases can use the following phrases: "You know", "I want to ask you", "I want to know", "I want to ask", "Excuse me", "Can you please tell me"

[0128] / you", "Do you know" or synonyms of the above words and phrases;

[0129] The fourth category of non-universal wake-up words and phrases can use the following words and phrases: "Now let me speak", "Let me start the report", "Let me do it", "Let me say it", "Let me say it", "I start now", "I want to say it", "I want to speak", "I want to report", "I want to tell everyone", "I want to tell you", "I want to tell you", "I want to tell you", "I want to report to everyone", "I want to tell everyone" or synonyms of the above words and phrases.

[0130] The fifth category of non-universal wake-up phrases is user-defined non-universal wake-up phrases.

[0131] The following is an example of a natural language flow in which a user uses a non-universal wake-up phrase to activate and interact with a smart headset:

[0132] Example 1) A user searches for existing information and tells the audience:

[0133] 1. The user says to the audience: "Let me think about it";

[0134] 2. The user hears the smart headset say: "I'm here, please tell me what to play";

[0135] 3. The user says to the audience: "Let me tell you a good story";

[0136] 4. The user hears the smart headset say: "Searching for..., the story takes place in the far Middle East", the headset pauses for N seconds, where N = the number of words in the sentence × C, "the protagonist of the story is Fetoga";

[0137] 5. The user told the audience: "The story takes place in the faraway Middle East"...

[0138] If the user feels that the playback speed is too slow, the user says to the audience: "I'm sorry, maybe I speak too slowly", then the smart headset reduces the pause time between sentences and plays, and the user hears the smart headset say: "I will speed up the playback speed."

[0139] If condition 1 is met, the current user's intention is to extract the content of an existing or self-built information database and tell it to the audience or sing to the audience; and condition 2 is met, the voice activation and recognition module does not obtain any user's voice within D seconds, including wake-up words, singing or other words, then the smart headset pauses playing the sound, and the user hears the smart headset say: "I'm in sleep now, you can say, let me think about it, etc., to reactivate me."

[0140] Example 2) A user creates and presents the contents of his own information database to an audience:

[0141] Step 1: Before speaking to the audience, the user first uses the user customization and input module of the smart headset to input the information to be played and its activation keywords, such as "today's report content", into the user intention inference and information library service interaction module of the smart headset.

[0142] Step 2: The user interacts with the smart headset and tells the audience the following steps:

[0143] 1. The user says to the audience: "I want to report something";

[0144] 2. The user hears a ding-dong sound from the smart headset;

[0145] 3. The user says to the audience: "Today's report content";

[0146] 4. The user hears a ding-dong sound from the smart headset, and starts playing "In this bright spring", the headset pauses for N seconds, where N = the number of words in the sentence × C, "We gather together to celebrate the successful sale of our company's one millionth product," the headset pauses for N seconds, where N = the number of words in the sentence × C, ...

[0147] 5. The user said to the audience: "In this bright spring"...

Claims

1. An interactive method based on a smart headset, characterized in that: The method comprises the following steps: Step 1), the user uses a wake-up phrase or a wireless device to activate the smart headset, and activates the smart headset through natural language; the wake-up phrase includes a non-universal wake-up phrase and a universal wake-up phrase; the natural language means that the user activates the smart headset through a non-universal wake-up phrase in the presence of an audience, avoiding the embarrassment of the user using a universal wake-up phrase in the presence of an audience; Step 2) After the smart headset is activated, it recognizes the wake-up words and infers the user's intention. And query and return relevant information through the words and sentences spoken by the user; Step 2.1), when the smart headset is activated, the voice activation and recognition module of the smart headset recognizes these wake-up words and records the wake-up method, that is, activation by voice or activation by a wireless device. At the same time, the sound control and playback module of the smart headset plays the feedback sound or feedback voice to the user, prompting the user that the smart headset has received the user's wake-up words and is waiting for the user to enter words; Step 2.2), the user speaks a sentence containing a keyword; Step 2.3), the voice activation and recognition module recognizes the user's words and extracts keywords from the words spoken by the user; Step 2.4), the user intention inference and information library service interaction module of the smart headset searches for relevant information library content according to the keywords obtained in step 2.3), including existing information libraries or user-built information libraries, and plays feedback sound or feedback voice at the same time; if the query can be found, the queried content and the activated information library are returned; if the query is not found, the user is fed back feedback that the information was not found; the user intention inference and information library service interaction module of the smart headset activates and queries the relevant information library according to the user's keywords, and returns the query content and the name of the activated information library; Step 2.5) This step is performed simultaneously with step 2.4), and the user intention inference and information base service interaction module of the smart headset infers the user's intention based on the user's way of waking up the smart headset and the wake-up phrase; the user's intention includes the following two types: Intent 1: If the user says a non-universal wake-up phrase or presses a wireless device, it is inferred that the user's intention is to extract the content of an existing or self-built information library to speak to the audience or sing to the audience; the user intention inference and information library service interaction module will send the intention to the sound control and playback module; Intent 2: If the user says a common wake-up phrase, it is inferred that the user intends to query information for himself. The user intention inference and information library service interaction module will send the intention to the sound control and playback module. Step 3), the sound control and playback module of the smart headset determines the information playback mode and speed according to different user intentions and the activated information library; according to the result of step 2.5), if it is intention 1: the user wants to speak or sing to the audience, then go to step 3.1); if it is intention 2, then go to step 3.4); Step 3.1), determine whether the activated information library is a song library or a non-song library through the user's keywords in step 2.4); if a non-song library is activated, proceed to step 3.2); if a song library is activated, proceed to step 3.3); Step 3.2) If it is non-song library activation, then according to the query content returned in step 2.4), the sound control and playback module of the smart headset will process according to the following steps: Step 1: Use punctuation marks in the text as markers to segment the paragraphs and sentences into multiple clauses; Step 2: Check the length of each sentence. If the length of a sentence exceeds the maximum word limit parameter X, then analyze the grammatical structure of the sentence and automatically split the sentence into multiple sentences. Ensure that the split sentences meet the following requirements: 1) the subject, predicate, and object are as complete as possible; 2) pronouns, nouns, verbs, adjectives, adverbs, and phrases are not split; 3) the total word count is less than the maximum word limit parameter X. Step 3: pause intentionally for N seconds after playing each small sentence, where N = the number of words in the small sentence × C; parameter C is the time parameter of the intentional pause after playing a single small sentence, and the unit is words / second. When the user uses the smart headset for the first time, the parameter C has a default value, and the user can modify the value of the parameter through natural language or the user definition and input module of the smart headset; Step 3.3), if the song library is activated, the user phrases and the results of the search in the song library obtained in step 2.4) are divided into two query situations: Query situation 1: If the audio of the song can be found, play the audio; Query situation 2: If only the lyrics of the song can be queried: Query situation 2A) If the voice activation and recognition module of the smart headset accurately recognizes the user's current singing progress, the sound control and playback module will play the next line of the song to the user M seconds in advance, where M is the advance time for playing the lyrics, in seconds. The smart headset has a default value when it leaves the factory, and the user can adjust it through the user customization and input module of the smart headset; Query situation 2B) If the smart headset cannot recognize the user's singing progress or the recognition error is high, the lyrics will be played according to the lyrics playback table of the song, including the playback time of each line of lyrics; Step 3.4): If the user's wake-up phrase is the name of the smart headset, that is, a universal wake-up phrase, it is inferred that the user's intention is to query information for himself; the smart headset will play these sounds at normal speed: if a non-song library is activated, the time parameter C for the intentional pause after a single sentence is played is 0; if the lyrics in the song library are played, M = 0; The non-universal wake-up phrases include the following five categories: Category 1: Words and phrases spoken by the user in the first person that require thinking or recall; Category 2: words and sentences that users may know in the first person; Category 3: Phrases and words that users ask others; Category 4: The user uses the first person to say the words he is about to speak or wants to tell others; Category 5: User-defined wake-up phrases.

2. The interactive method based on smart earphone according to claim 1, characterized in that: The interactive method further comprises the following steps: Step 4), the user controls the playback of the smart headset through keywords: Users can control the playback of smart headphones through the following relatively natural voice interaction methods; Playback situation 1: If the user is not satisfied with the content played by the smart headset and wants to change the content or go to the next message or the next song: Step 4.1.1): The user can reactivate the smart headset using any non-universal activation word from category 1 to category 5 or a wireless device; Step 4.1.2): The user speaks the keywords that need the information database, or speaks a synonym with a different meaning; Step 4.1.3): After the voice activation and recognition module obtains the keywords of these words and sentences, the smart headset will repeat the above steps 2) and 3); Playback situation 2: If the user feels that the pause time between sentences played by the smart headset is too short, that is, the playback speed of the smart headset is too fast: Step 4.2.1): The user can reactivate the smart headset using any non-universal activation word from category 1 to category 5 or a wireless device; Step 4.2.2): The user can use his / her mouth to say the words and phrases that he / she speaks too fast and the synonymous words and phrases; Step 4.2.3): The voice activation and recognition module identifies the keyword "too fast" or its synonyms in the sentence, identifies the user's current intention through the intention inference and information base service interaction module, and then extends the pause time between sentences through the sound control and playback module, that is, increases the value of the time parameter C of the intentional pause after a single sentence is played; at the same time, the smart headset will give the user relevant feedback; Playback situation 3: If the user feels that the pauses between sentences played by the smart headset are too long, that is, the playback speed is too slow: Step 4.3.1): The user can reactivate the smart headset using any non-universal activation word from category 1 to category 5 or a wireless device; Step 4.3.2): The user can use his mouth to say the words and sentences that he speaks too slowly and the synonymous words and sentences; Step 4.3.3): The voice activation and recognition module identifies the keyword "too slow" or its synonyms in the sentence, identifies the user's current intention through the intention inference and information library service interaction module, and then reduces the pause time between sentences through the sound control and playback module, that is, reduces the value of the time parameter C of the intentional pause after a single sentence is played; at the same time, the smart headset will give the user relevant feedback; Playback situation 4: If the user wants to pause the playback of the smart headset: Step 4.4.1): Condition 1) The intention inference and information library service interaction module of the smart headset extracts the intention of the current user to extract the content of an existing or self-built information library to tell the audience or sing to the audience; and condition 2) is met at the same time) The voice activation and recognition module does not obtain any sound of the user within the time range of D seconds, including wake-up words, singing or other words; wherein D is the time threshold parameter for pausing the smart headset when there is no signal, and there is a default value when the smart headset leaves the factory. The user can also modify the value of the parameter through natural language or the user definition and input module of the smart headset; when conditions 1) and 2) are met at the same time, the intention inference and information library service interaction module determines that the user's intention is to pause the playback of the headset, and sends the intention to the sound control and playback module; Step 4.4.2): The sound control and playback module receives the intention to pause the headset playback and pauses the playback of the sound; at the same time, the playback pause is fed back to the user and the user is prompted with the activation method of the smart headset.

3. An interactive method based on a smart headset as claimed in claim 1 or 2, characterized in that: The first category of the non-universal wake-up phrases may adopt the following phrases: "Let me think about it", "Let me think about it", "I want to think about it", "Let me think about it", "Let me think about it", "Let me think about it", "Let me recall", "Let me think about it" or synonyms of the above phrases; The second category of non-universal wake-up phrases may adopt the following phrases: "I may know", "I may know", "I may know this", "I may know that" or synonyms of the above phrases; The third category of non-universal wake-up phrases may use the following phrases: "you know", "I want to ask you", "I want to know", "I want to ask", "Excuse me", "Please teach you", "Do you know" or synonyms of the above phrases; The fourth category of non-universal wake-up phrases may use the following phrases: "Now I will speak", "I will start the report", "Let me do it", "Let me speak", "Let me speak", "I will start now", "I want to speak", "I want to speak", "I want to report", "I want to tell everyone", "I want to tell you", "I want to tell you", "I want to tell you", "I want to report to everyone", "I want to tell everyone" or synonyms of the above phrases; The fifth category of non-universal wake-up phrases is user-defined non-universal wake-up phrases.

4. The interactive method based on smart earphone according to claim 1 or 2, characterized in that: The user-defined and input module includes the following functions: 1) Display a list of information libraries and the method of activating each information library, and allow the user to modify or add or delete the method of activating information libraries, and also allow the user to modify or add or delete existing information libraries and their contents; 2) Display and allow the user to modify or add or delete all wake-up words of the smart headset, and also allow the user to modify the feedback sound of the smart headset; 3) Display various parameters of the smart headset and their factory default values, and allow users to customize or adjust these parameters, including the maximum word limit parameter X for a single sentence, the time parameter C for intentional pause after playing a single sentence, the advance time parameter M for playing lyrics, and the time threshold parameter D for pausing when the headset has no signal; 4) Allowing users to input and modify the information to be played, i.e., self-building information libraries, and allowing users to specify one or more activation key words corresponding to each self-building information library; The module can be implemented by software, web page or voice interaction of the terminal to which the smart headset is connected.

5. The interactive method based on smart earphone according to claim 1 or 2, characterized in that: The information database refers to an information database of various text or audio containing songs, which is an existing information database or an information database established by the user himself; the information database is divided into two categories, a song database containing song information and a non-song database without song information; the non-song database may specifically include but is not limited to: jokes, stories, poems, riddles, brain teasers, love words, food and cooking, encyclopedia knowledge, safety and emergency handling, problem solving, various professional knowledge and dictionaries, life experience, famous sayings and good sentences, conversation skills and suggestions, travel information, health and disease diagnosis and treatment, beauty and weight loss, film and television, cute pets, celebrities, entertainment, sports, technology, home furnishings, constellations, art, history, geography, military, news and information established by users themselves.

Images