Voice recognition device, voice recognition device cooperation system, and voice recognition device cooperation method

Abstract

In order to use a plurality of speech recognition devices in a cooperative manner, the speech recognition device includes: a speech input unit; a speech output unit; a communication control unit for transmitting and receiving data with other speech recognition devices; a dialogue mode execution unit for transmitting speech data input from the speech input unit to other speech recognition devices and outputting speech data received from other speech recognition devices from the speech output unit; a speech recognition unit for converting speech input from the speech input unit into text data; a hot word detection unit for detecting a dialogue start hot word from text data, wherein the dialogue start hot word is used to instruct to start the dialogue mode execution unit; and a command sending unit for sending a control command to other speech recognition devices. When the hot word detection unit detects the dialogue start hot word, the command sending unit sends a control command to other speech recognition devices to start the dialogue mode execution unit of the other speech recognition devices.

Description

Technical Field

[0001] The invention relates to a speech recognition device, a collaboration system of speech recognition devices and a collaboration method of speech recognition devices. Background Art

[0002] In recent years, there have been a number of voice recognition devices called so-called smart speakers or AI speakers that use voice recognition technology and artificial intelligence technology. They perform voice recognition on the voice content uttered by the speaker, and output a response (answer) corresponding to the voice content in the form of sound from an attached speaker by analyzing the speech content. For example, in Patent Document 1, as an example of voice recognition technology, the following content is recorded: "When a voice recognition algorithm is applied to a voice input to obtain an interpretation candidate, in order to reduce redundancy and to present the interpretation candidate in an integrated manner, the user is provided with an opportunity to select from the interpretation candidates, and the options are presented to the user in a non-repetitive manner (selected from the abstract)". In addition, Patent Document 2 records that "as a function for managing the use of voice recognition models and data in an automatic voice recognition system, tracking can be performed to predict when a user is likely to use the system (selected from the abstract)".

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: Japanese Patent Application Publication No. 2013-68952

[0006] Patent Document 2: Japanese Patent Application No. 2015-537258 Summary of the invention

[0007] Technical problem to be solved by the invention

[0008] The above-mentioned voice recognition devices such as smart speakers can obtain responses by processing the devices when people around the devices give instructions to the devices by voice, etc., but none of these documents disclose the content of using multiple devices in a collaborative way. That is, taking the use in a residence as an example, when family members share the device in the living room to request weather forecasts, news, music and other information on the Internet, the content corresponding to the request can be output in the form of voice, but it does not take into account certain communication and collaborative actions with other voice recognition devices outside the living room, such as in the children's room. Therefore, it is impossible to use multiple voice recognition devices in a collaborative way, and there is room for improvement in new ways of using voice recognition devices.

[0009] The present invention has been made in view of the above-mentioned problems, and an object of the present invention is to provide a speech recognition device capable of cooperatively using a plurality of speech recognition devices, a speech recognition device cooperation system, and a speech recognition device cooperation method.

[0010] Technical means of solving problems

[0011] To achieve the above-mentioned purpose, the following technical solution of the present invention can be adopted.

[0012] Effects of the Invention

[0013] According to the present invention, a speech recognition device capable of collaboratively using multiple speech recognition devices, a speech recognition device collaboration system, and a speech recognition device collaboration method can be provided. Other technical objectives, technical features, and technical effects than those described above will be described in the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a hardware structure diagram of the speech recognition device of this embodiment.

[0015] Figure 2 It is a functional block diagram of a speech recognition device.

[0016] Figure 3A It is a diagram showing an example of hot word data stored in the reference memory.

[0017] Figure 3B FIG. 1 is a diagram showing an example of voice authentication data stored in the reference memory.

[0018] Figure 4 This is a diagram showing a cooperative system constructed using a voice recognition device.

[0019] Figure 5 This is a flowchart showing the first call action of the collaboration system.

[0020] Figure 6 This is a diagram showing an example of position estimation data.

[0021] Figure 7 This is a flowchart showing the flow of the first switching process between the normal mode and the interactive mode.

[0022] Figure 8 1 is a flowchart showing the flow of the second switching process between the normal mode and the interactive mode.

[0023] Fig. 9 This is a conceptual diagram showing the first setting process when installing a new machine.

[0024] Fig.10 This is a conceptual diagram showing the second setting process when installing a new machine.

[0025] Fig.11 This is a flowchart showing the flow of the second setting process when a new machine is installed.

[0026] Fig.12This is a diagram showing a schematic configuration of a system for cooperating with voice recognition devices in a house according to a second embodiment.

[0027] Fig.13 This is a diagram showing an example of usage status data recording the usage status of a person for each time period.

[0028] Fig.14 This is a diagram showing an example of call priority data recording the call priority of a person in each time period.

[0029] Fig.15 This is a diagram showing an example in which a voice recognition device is connected to a portable communication terminal using a docking station.

[0030] Fig.16 : is a flowchart showing the switching process to the home-watching mode. DETAILED DESCRIPTION

[0031] Hereinafter, an example of the embodiment of the present invention will be described using the drawings. In each of the drawings, the same functions are denoted by the same reference numerals and repeated descriptions are omitted.

[0032] <First embodiment>

[0033] Figure 1 : is a hardware structure diagram of the speech recognition device 1 of this embodiment. The speech recognition device 1 can be a device dedicated to speech recognition, or it can be an existing electronic device with communication function such as a portable telephone terminal, a smart phone, a personal computer, a game console, etc. In addition, in the speech recognition device 1, as a communication function, a common communication function such as a wired LAN, a wireless LAN, wireless communication via a portable telephone line, Bluetooth (registered trademark), RFID, etc. can be used, and one or more communication interfaces corresponding to these communication functions can be included.

[0034] Specifically, in the speech recognition device 1, a CPU 101, a memory 103, a wired LAN I / F 104, a wireless LAN I / F 105 and a wireless communication I / F 106 as external interfaces, a speech input unit 107 (e.g., a microphone), a speech output unit 108 (e.g., a speaker), and a display output unit 109 (e.g., a liquid crystal screen) are connected to each other via a bus 102. In addition, a human detection sensor I / F 110, a timer 111, an RTC 112, and a camera 113 may also be connected to the bus 102.

[0035] The memory 103 includes an internal memory 1031 formed of a volatile memory and a reference memory 1032 formed of a nonvolatile memory.

[0036] The person detection sensor I / F 110 is an I / F for an external person detection sensor, which is not limited to the type of person detection sensor, such as a human body sensor, a sound collection sensor, etc.

[0037] Figure 2 is a functional block diagram of the speech recognition device 1.

[0038] The speech recognition device 1 includes a speech processing engine 120 . The speech processing engine 120 mainly includes a speech processing unit 1201 , a speech recognition unit 1202 , a hot word detection unit 1203 , a speech analysis unit 1204 and a feature extraction unit 1205 .

[0039] The speech processing engine 120 realizes its functions by the CPU 101 reading a speech processing program stored in the reference memory 1032 , loading the program into the internal memory 1031 , and executing processing according to the speech processing program.

[0040] Specifically, when a person speaks toward the voice recognition device 1 , the voice is collected by the voice input unit 107 , and the voice (analog data) is converted into voice data composed of digital data.

[0041] The speech processing unit 1201 performs some adjustments, such as removing environmental noise included in the speech data.

[0042] The speech recognition unit 1202 performs speech recognition processing to convert speech data into character string data.

[0043] The hotword detection unit 1203 determines whether the character string data is character string data containing a prescribed sentence (hereinafter referred to as a "hotword"), wherein the prescribed sentence is a sentence requiring the speech recognition device 1 to start working or start up, such as resuming from a standby state.

[0044] Figure 3A 1 is a diagram showing an example of hot word data 150 stored in the reference memory 1032. The hot word data 150 is data obtained by associating a registered hot word 1501, a category 1502, and a call target 1503, wherein the category 1502 specifies the action of the voice recognition device 1 for the registered hot word 1501, and the call target 1503 specifies the device-specific information for determining the voice recognition device 1 to be called using the registered hot word 1501. The hot word detection unit 1203 detects hot words based on whether the character string data is recorded in the hot word data 150.

[0045] The speech analysis unit 1204 determines how to interpret the content spoken by the user to the speech recognition device 1, that is, determines the response data and control command corresponding to the character string data, outputs the response data from the speech output unit 108, or outputs the control command to the CPU 101, so that the CPU 101 executes the processing represented by the speech data. For example, as a control command, it can be a command to play a specific music.

[0046] The speech recognition unit 1202 may be provided not on the speech recognition device 1 but on an external server 201 connected to the speech recognition device 1 (see Figure 4 ), the speech recognition process is performed by the external server 201. As a result, the load of the speech recognition device 1 can be reduced. When the speech recognition process is performed by the speech recognition device 1, the amount of data communication with the external server 201 can be reduced.

[0047] Furthermore, as a characteristic function of the voice recognition device 1 of the first embodiment, in addition to operating independently like the existing voice recognition device, it also has a function of being able to transmit the voice spoken by the user to a private communication network (hereinafter referred to as the residential LAN 210, see Figure 4 For example, the other speech recognition device 1 on the home) is used to conduct a conversation. Therefore, the hot word detection unit 1203 detects a hot word, which is a starting point (i.e., a trigger) from which the conversation mode for conducting a conversation is transferred. Furthermore, the speech recognition device 1 has a feature extraction unit 1205, which extracts features of sound and image from people who can participate in the conversation via the LAN 210 in the house, and determines whether it is consistent with the registered data.

[0048] Figure 3B 1032 is a diagram showing an example of voice authentication data 160 stored in the reference memory 1032 .

[0049] The voice authentication data 160 is data obtained by associating a speaker 1601, a speaker category 1602 and a speaker template 1603, wherein the speaker 1601 is information used to uniquely identify a person who can participate in a conversation via the LAN 210 within the house, the speaker category 1602 indicates the attributes of the speaker, such as whether the speaker is a "main user" who has setting authority for the voice recognition device 1 and the communication network, or a "normal user" who does not have setting authority and only has the right to participate in conversations on the LAN 210 within the house, and the speaker template 1603 indicates the characteristics of each person's voice.

[0050] Figure 41 is a diagram showing a cooperative system 100 constructed using a voice recognition device 1, and shows an example of a case where the system is used in a house. Although this example shows a house, the system is not limited to a house, and the present embodiment can also be applied to a specific space where a crowd of people gather, such as an office or a classroom.

[0051] (First call action)

[0052] Figure 4 In the room 1, room 2, room 3, and room 4 in the house, the first voice recognition device 1A, the second voice recognition device 1B, the third voice recognition device 1C, and the fourth voice recognition device 1D are respectively provided with Figure 1 Here, the first to fourth voice recognition devices 1A to 1D in each room are connected to the router 202 installed in room 1 via the second AP 2032 to the fourth AP 2034 as access points or wireless repeaters installed in each room 2 to 4. Moreover, the first to fourth voice recognition devices 1A to 1D are connected to the external Internet 200 via the router 202.

[0053] Next, press Figure 5 The first call action of the collaboration system 100 is described in the order of each step. It is assumed that at the beginning of the processing of this step, the first to fourth speech recognition devices 1A to 1D are all powered on, and the speech input unit 107 and the speech processing engine 120 are in the started state. This state is called standby mode.

[0054] When person A in room 1 calls the first voice recognition device 1A, the voice of person A is collected by the voice input unit 107 of the first voice recognition device 1A, and the hot word detection unit 1203 determines whether it is the first hot word indicating the meaning of requesting to start. When the hot word detection unit 1203 determines that the calling voice is the first hot word (S101 / Yes), the voice response data of a prescribed format sentence such as "What's the matter?" is played from the voice output unit 108. The hot word detection unit 1203 outputs the first hot word to the voice analysis unit 1204, and the voice analysis unit 1204 outputs an execution command to the normal mode execution unit 1406. As a result, the first voice recognition device 1A switches to the normal mode. The normal mode is an action mode that can execute all functions of the first voice recognition device 1A.

[0055] When the hot word detection unit 1203 determines that it is not a hot word (S101 / No), the standby mode is maintained.

[0056] Next, when character A calls out "Little B!" to character B in another room, his voice is collected as voice data via the voice input unit 107. After the voice processing unit 1201 of the voice processing engine 120 performs adjustments such as removing environmental noise, the hot word detection unit 1203 determines whether it is the second hot word (a hot word requesting an instruction to switch to conversation mode) (S102).

[0057] The calling voice data "Little B!" has been pre-registered as a hot word in the hot word data 150. When the hot word detection unit 1203 determines that the word "Little B!" is the second hot word (S102 / Yes), the second hot word is output to the voice analysis unit 1204. In this embodiment, the name of the person in the house is used as the hot word, but it is not limited to this. It can also be other format words, such as words such as "dialogue", "connection", etc. or other words. Here, the hot word is the starting point (trigger) for transferring to the dialogue mode.

[0058] The speech analysis unit 1204 analyzes that the second hot word is a request for switching to the dialogue mode, and selects the control command required for it (S103). In this example, it is equivalent to a speech transmission command and a speech playback command, which is used to send a mode switching command for switching the first to fourth speech recognition devices 1A to the dialogue mode and the speech data "Little B!" from the first speech recognition device 1A to each of the second to fourth speech recognition devices 1B to 1D, respectively, so that they are output from their respective speech output units 108.

[0059] In the first speech recognition device 1A, the conversation mode execution unit 1403 is also activated to switch to the conversation mode (S104).

[0060] Based on the voice transmission command, the command transmission unit 1402 of the first voice recognition device 1A transmits the call voice data "Little B!" to the second to fourth voice recognition devices 1B to 1D via the communication control unit 1410, and transmits a mode switching command for switching to the conversation mode and a voice playback command to each voice recognition device. In addition, the command transmission unit 1402 of the first voice recognition device 1A starts measuring the time elapsed from the transmission of the call voice data to the second to fourth voice recognition devices 1B to 1D (S105).

[0061] Each of the second to fourth voice recognition devices 1B to 1D plays the call voice data as the voice "Little B!" from the voice output unit 108, and activates the dialogue mode execution unit 1403 to switch to the dialogue mode. Regarding the playback order of the voice data, the second to fourth voice recognition devices 1B to 1D can play the voice at the same time, or the second to fourth voice recognition devices 1B to 1D can output the voice in a specified order. The specified order can be, for example, the order in which the voice recognition devices are installed, or the order in which the priorities are assigned to each room (see Fig.14 )wait.

[0062] When person B in room 2 responds, for example, by saying "Here!", the second voice recognition device 1B collects the voice data through the voice input unit 107 and sends the response voice data back to the first voice recognition device 1A as the caller. At this time, the second voice recognition device 1B is associated (bound) with person B. Furthermore, the associated information is commonly registered in the first, third, and fourth voice recognition devices 1A, 1C, and 1D, where the associated information indicates that the person in room 2 where the second voice recognition device 1B is located is person B.

[0063] Figure 6 An example of position estimation data is shown.

[0064] As an example of registration of the association between person B and the second voice recognition device 1B, when the first voice recognition device 1A receives the above-mentioned response voice data, "person B = second voice recognition device 1B" can be added to the position estimation data (reference memory 1032) registered in advance. Figure 6 ) The first to fourth voice recognition devices 1A to 1D in the house can be identified by using a specific identifier such as a Mac address of the device or an IP address assigned to the device in the LAN 210 in the house.

[0065] Furthermore, the first voice recognition device 1A requests the other voice recognition devices in the house, namely the second to fourth voice recognition devices 1B to 1D, to register the above-mentioned position estimation data and store it in the reference memory 1032 of each of the second to fourth voice recognition devices 1B to 1D. However, if the above-mentioned position estimation data has already been stored in the reference memory 1032 in each of the first to fourth voice recognition devices 1A to 1D, it is determined that it has been registered and no association operation is performed. Regarding the registration of the person, if the voice processing engine 120 or the external server 201 Figure 2 By extracting part of the name contained in the calling voice "Little B!" and using it as name data, an association such as "name of person B = second voice recognition device 1B" can be achieved.

[0066] In addition, regarding the method of registering a hot word for determining whether to switch to the conversation mode, the method described below is performed when the voice recognition device is set up, such as at the time of initial setting, or when a newly registered hot word is set.

[0067] The command sending unit 1402 of the first voice recognition device 1A sends a command to return from the dialogue mode to the standby mode to the device among the second to fourth voice recognition devices 1B to 1D that has not responded, when the elapsed time measured by the timer 111 is greater than or equal to the standby time threshold for determining whether there is a response (S106 / Yes) (S107).

[0068] When a response to voice data is received from at least one of the second to fourth voice recognition devices 1B to 1D and the elapsed time of the response is less than the standby time threshold (S106 / No), the dialogue mode between the first voice recognition device 1A and the device that issues the response, such as the second voice recognition device 1B, is maintained (S108).

[0069] After switching to the conversation mode in step S104, the conversation mode execution unit 1403 performs speech transmission and reception between the first speech recognition device 1A and the second speech recognition device 1B for the speech input to the speech input unit 107 of the first speech recognition device 1A and the speech input to the speech input unit 107 of the second speech recognition device 1B in step S105.

[0070] When any one of the first voice recognition device 1A and the second voice recognition device 1B detects the third hot word for ending the conversation mode (S109 / Yes), the device that detects the hot word, such as the second voice recognition device 1B, switches to the standby mode (S110), and sends a command to switch to the standby mode to the first voice recognition device 1A (S107). Then, the first voice recognition device 1A also switches to the standby mode, and the conversation mode ends.

[0071] Furthermore, if the second hot word is not detected in step S102 ( S102 / No), the first voice recognition device 1A does not shift to the conversation mode but maintains the normal mode ( S111 ) and ends the process.

[0072] In the above example, regarding the associated operation, the first voice recognition device 1A that first implemented the call becomes the main device, and it implements control instructions to the called second to fourth voice recognition devices 1B to 1D in the house 2 respectively, but this is not limited to this. The second voice recognition device 1B that is called and responds can also become the main device.

[0073] As another association method, for example, when the fourth voice recognition device 1D is first installed in a residence, it is also possible to register in advance that the main user or operator of the fourth voice recognition device 1D is person D, and immediately after the installation, the association data is sent to the first to third voice recognition devices 1A to 1C via the LAN 210 in the residence, and "person D = fourth voice recognition device 1D" is registered in the reference memory 1032 in the first to fourth voice recognition devices 1A to 1D.

[0074] The above example is an example of making a call from the first voice recognition device 1A to the second voice recognition device 1B. The steps can be applied between all voice recognition devices in a house, and can also be applied to calls made from voice recognition devices other than the above example to other voice recognition devices.

[0075] In addition, in the above example, the first to fourth voice recognition devices 1A to 1D communicate with each room using a wireless LAN, but a connection based on a wired LAN or a connection using a portable telephone line can also be used. Furthermore, the communication mode can also be changed to another interface only in the conversation mode. For example, a wireless LAN can be used in the normal mode, and another wireless system such as Bluetooth can be used in the conversation mode.

[0076] (Second call action)

[0077] After the association between the voice recognition device 1 and the person is completed as in the above-mentioned embodiment, when person A and person B have a conversation thereafter, only the communication between the first voice recognition device 1A used by person A and the second voice recognition device 1B used by person B is opened, and the communication with the voice recognition device 1 other than this is closed, thereby improving the confidentiality of the conversation content.

[0078] For example, when person A calls person B for the second time and thereafter, the command sending unit 1402 of the first voice recognition device 1A establishes and registers the position estimation data ( Figure 6 ) communicates with the second voice recognition device 1B associated with the character B in the communication (establishing communication by sending a mode switching command and receiving a response to the command), and sends the voice data to the second voice recognition device 1B.

[0079] The conversation mode execution unit 1403 then communicates voice data only between the first voice recognition device 1A and the second voice recognition device 1B, so that person A and person B can directly talk to each other. This eliminates the need to send voice data to all voice recognition devices in the house.

[0080] During the execution of the dialogue mode, for example, if no response voice data is received from person B within a specified time from the call, the dialogue mode execution unit 1403 of the first voice recognition device 1A determines that the other party is absent. Then, the dialogue mode execution unit 1403 instructs the command sending unit 1402 to open the communication channel with other voice recognition devices in the house that has been in a communication closed state, that is, to send voice data to other voice recognition devices (third and fourth voice recognition devices 1C, 1D) and wait for a response.

[0081] Here, for example, if a response is received from the fourth voice recognition device 1D located in room 4, the first voice recognition device 1A starts communicating with the fourth voice recognition device 1D and starts a conversation again. In this case, the first voice recognition device 1A may not store the association information between person B and the fourth voice recognition device 1D in the position estimation data 190 of the internal reference memory 1032, but may regard person B as having temporarily moved to another place, or may generate the association information between person B and the fourth voice recognition device 1D and assign a priority, set the priority order according to the association between person B and the second voice recognition device 1B as the superior and the association between person B and the fourth voice recognition device 1D as the subordinate, connect according to the priority order, and wait for the response.

[0082] Furthermore, if no response is received from the fourth voice recognition device 1D, voice data is sent to other voice recognition devices in the house (the third voice recognition device 1C in this example) in turn to wait for a response. Then, voice data is sent to the third voice recognition device 1C to wait for a response. When no voice recognition devices in all rooms respond within a specified time, the first voice recognition device 1A determines that there is no response and responds to the operator, i.e., person A, with a specified word such as "No one responded." Alternatively, instead of the first voice recognition device 1A determining that other voice recognition devices have no response as described above, other voice recognition devices may determine that there is no response voice from person B within a specified time, and send the no response information back to the first voice recognition device 1A, so that the first voice recognition device 1A recognizes that there is no response and outputs a specified response voice such as "No one responded."

[0083] The predetermined answer voice data may be stored in the memory 103 in advance, or data stored in an external server 201 on the Internet 200 may be used.

[0084] (Third call action)

[0085] In this example, Figure 1The person detection sensor I / F 110 of the voice recognition device 1 is connected to a person detection sensor such as a camera sensor that can confirm a person, or a human body sensor that determines whether a person exists, and a call action is performed based on its detection result. In addition, the built-in camera 113 of the voice recognition device 1 can also be used.

[0086] For example, when responding to a call from person A and sending the call voice data of person A from the first voice recognition device 1A to the second to fourth voice recognition devices 1B to 1D located in each room, the person detection sensors respectively provided in the second to fourth voice recognition devices 1B to 1D are used to determine the presence or absence of a person. In a room that can be determined as empty, the voice recognition device installed in the room returns an absence notification indicating that no one is present to the first voice recognition device 1A, and the command sending unit 1402 receives the absence notification.

[0087] Then, the command transmission unit 1402 of the first voice recognition device 1A does not output a calling voice to the voice recognition device that has transmitted the absence notification to the first voice recognition device 1A.

[0088] On the other hand, the command sending unit 1402 of the first voice recognition device 1A sends voice data to the voice recognition device that has not sent the absence notification to the first voice recognition device 1A, and the voice recognition device that has received the voice data plays the voice data to make a call. The subsequent actions may be the same as those in the above embodiments.

[0089] The above-mentioned identification and judgment of a person can be performed using a common method. The presence or absence of a person can be detected by detecting the movement of a person using a human body sensor composed of an infrared sensor or the like.

[0090] Furthermore, the camera 113 may also be used as a person detection sensor. Thus, the facial recognition unit 1404 may extract the features of a person (e.g., a facial image) from the image captured by the camera 113 to determine the presence or absence of a person. The facial authentication data obtained by associating a facial image with a person may be stored in the reference memory 1032 in advance, and compared with the correspondence information between the person and the facial image stored in advance to determine whether the called person is in the room. If the camera 113 captures the called person and can determine that the person is in the room, a conversation between person A and person B can be conducted through the communication connection between the first voice recognition device 1A and the second voice recognition device 1B.

[0091] (Fourth Call Action)

[0092] In another embodiment, Figure 4 When character A calls character B and a character other than character B responds, a communication connection between the first voice recognition device 1A and the second voice recognition device 1B is not established, but the call is implemented again.

[0093] First, each of the first to fourth voice recognition devices 1A to 1D installed in each room receives voice authentication data 160 ( Figure 3B ) is stored in the reference memory 1032 in advance. The voice authentication data 160 is generated in such a way that the feature extraction unit 1205 of the voice processing engine 120 generates a speaker template 1603 using the voice print of the person, the intonation of the voice, or the frequency characteristics of the voice, and stores it in advance as the voice authentication data 160 in the reference memory 1032 of the first to fourth voice recognition devices 1A to 1D.

[0094] The voice authentication data 160 can be registered in each device when the first to fourth voice recognition devices 1A to 1D are initially set. On this basis, the voice recognition unit 1405 of the first voice recognition device 1A compares the voice feature data of person B registered in the voice authentication data 160 with the features of the voice of the person who answered, and if they match, it is determined to be person B, and the judgment result is given to the command sending unit 1402. Then, the command sending unit 1402 sends a command for transferring to the dialogue mode to the second voice recognition device 1B.

[0095] If the voice recognition unit 1405 determines that the characteristics of the voice do not match, it will process it as if it is not person B.

[0096] As described above, the voice authentication data 160 can be stored in advance in each of the first to fourth voice recognition devices 1A to 1D located in each room to compare whether the voices of the characters in each room are consistent with the voices of character B called by character A. However, instead of this method, the voice authentication data 160 can be stored only in the reference memory 1032 of the first voice recognition device 1A serving as the main device to determine whether the characteristics of the voices are consistent on the first voice recognition device 1A.

[0097] Alternatively, the voice authentication data 160 may be stored in advance in a server or the like installed in a house, and the voice data transmitted from the first to fourth voice recognition devices 1A to 1D may be compared with the voice authentication data 160 to determine whether they match or not.

[0098] Furthermore, the voice authentication data 160 may be stored in an external server installed outside the house, and the characteristics of the voices may be compared.

[0099] By determining the characteristics of the voice of speaker 1601 as in the above example, it is possible to prevent persons other than family members living in the house and outsiders other than persons permitted by family members to participate in the conversation from participating in the conversation, thereby improving security.

[0100] In addition, when the characteristics of a voice other than a registered speaker are detected, for example, the display output unit 109 of the first voice recognition device 1A of the calling party can be made to display an alarm, or a voice such as "an outsider is responding" can be output from the voice output unit 108 to issue a warning.

[0101] (Switch between normal mode and dialogue mode 1)

[0102] Figure 7 is a flowchart showing the flow of the first switching process between the normal mode and the dialogue mode, and Figure 5 In contrast to the example above, the default setting is conversation mode.

[0103] In this example, after the first to fourth speech recognition devices 1A to 1D are installed, the dialogue mode execution unit 1403 is activated when the main power is turned on, and the dialogue mode is waiting for the operator to call other characters. At this time, when the operator says a hot word for mode switching, such as a prescribed word such as "mode switching" (S201 / Yes), the mode switching unit 1401 switches to the normal mode (S202), and the normal mode execution unit 1406 is activated (S203).

[0104] In step S201 , when no hot word for mode switching is detected ( S201 / No), the interactive mode execution unit 1403 maintains the interactive mode.

[0105] The normal mode execution unit 1406 maintains the normal mode while the return condition to the interactive mode is not satisfied ( S204 / No).

[0106] When the return condition from the normal mode to the interactive mode is met (S204 / Yes), the interactive mode execution unit 1403 is restarted to return to the interactive mode. As the return condition, a hot word for returning can be set, or if no response from the operator is received within a specified time, the interactive mode is returned.

[0107] (Switch between normal mode and dialogue mode 2)

[0108] Figure 8 1 is a flowchart showing the flow of the second switching process between the normal mode and the interactive mode.

[0109] In this example, the voice recognition device 1 first starts the voice input unit 107 and only continuously monitors the presence of voice (S301 / No). When the voice input unit 107 detects voice (S301 / Yes), the hot word detection unit 1203 determines whether the detected voice is the first hot word requesting the voice recognition device 1 to start (requesting to start the normal mode) or the second hot word requesting to start the conversation mode (S302). If it does not match any of the first hot word and the second hot word (S302 / No), it returns to the voice detection process.

[0110] When the first hot word is detected (S302 / first hot word), the normal mode execution unit 1406 is activated (S303). For example, when the user calls the nickname set to activate the voice recognition device 1, the subsequent processing is performed in the normal mode.

[0111] When the second hot word is detected (S302 / second hot word), the conversation mode execution unit 1403 is activated (S304). For example, when the name of a family member or a person in the house is called, it is determined to be a conversation mode and the subsequent processing is performed.

[0112] The first hot word and the second hot word can be set in advance, or they can be changed to hot words that match the operator's ease of use after setting. In order to change the setting of the hot word, a dedicated application software for setting can be installed on a smart phone or a personal computer to implement it. Alternatively, the main operator 501 can be determined in advance, and the voice of the main operator 501 can be used to instruct the first to fourth voice recognition devices 1A to 1D to change the hot word, or the main voice recognition device, such as the first voice recognition device 1A, can issue an instruction to change to other slave devices located in the residence, such as the second to fourth voice recognition devices 1B to 1D. At this time, the hot word can be changed only when the voice of the main operator 501 is recognized, thereby preventing the hot word from being easily changed. The method for setting the voice of the main operator 501 is to prepare a dedicated application for setting of a smart phone or a personal computer, thereby registering the voice. For the voice of the main operator 501, the feature extraction unit 1205 in the voice processing engine 120 of the first voice recognition device 1A can generate voice feature data and register the generated data in the reference memory 1032, or the voice authentication data 160 can be generated on a smart phone or a personal computer and registered in the reference memory 1032. Furthermore, in the case of a home server 601 connected to the network environment in the house, the voice feature data can be stored on the home server 601, and the server can cooperate with the voice recognition device to compare and confirm the features of the voice being called with the features of the voice in the stored data.

[0113] (Setting 1 for new machine installation)

[0114] Fig. 9 This is a conceptual diagram showing the first setting process when installing a new machine.

[0115] When a new voice recognition device 1 is installed in a house, a dedicated application software (initial setting application software) 410 is installed on an electronic device 401 such as a smart phone or a personal computer. Then, these settings are made, for example, a network connection is set with a router 202 in the house, a device is registered if an existing voice recognition device 1 exists, and associated data about the correspondence between family members in the house and the voice recognition device 1, voice authentication data 160 of family members, etc. are set. In this example, the method used for setting the connection between the newly installed voice recognition device and a communication device such as a wireless router in the house is to use the above-mentioned smart phone, personal computer, etc. to set the connection, or to use an automatic setting method such as WPS (WiFi Protected Setup) to set the connection.

[0116] In addition, as the correspondence between the above-mentioned characters and the voice recognition devices 1, the following association settings are made on the application software of the smartphone, personal computer, etc.: the device mainly used by character A is the first voice recognition device 1A, and the device mainly used by character B is the second voice recognition device 1B.

[0117] Furthermore, regarding the person's voice authentication data 160, the data stored in an existing device such as the third voice recognition device 1C can be read to share the data, or the data can be managed and set on the above-mentioned electronic device.

[0118] (Setting 2 for new machine installation)

[0119] Fig.10 This is a conceptual diagram showing the second setting process when installing a new machine. Fig.11 This is a flowchart showing the flow of the second setting process when a new machine is installed.

[0120] In this example, the main operator 501 who has the authority to set up the in-home communication device performs the connection setting of the new voice recognition device 1S by voice.

[0121] First, the main operator 501 starts speaking toward the new voice recognition device 1S, and the voice input unit 107 starts the connection setting process by receiving the voice input, and starts counting of the timer 111 (S401).

[0122] When the hot word detection unit 1203 of the new voice recognition device 1S detects the fourth hot word W501 for initial setting (S402 / OK), the initial setting unit 1408 of the new voice recognition device 1S starts the initial setting process. Specifically, the initial setting unit 1408 sends the voice data of the main operator 501 and the initial setting request data to the existing devices in the house, such as the first to fourth voice recognition devices 1A to 1D and the home server 601 (S403). The sending process so far must be performed within a specified time (S404).

[0123] The transmission process is limited to a predetermined time in order to reduce the possibility of the initial setting request data and the voice data of the main operator 501 being intercepted, which may spread outside the residence when the data is transmitted.

[0124] The transmitted request data and voice data are received by the first to fourth voice recognition devices 1A to 1D already in the house (S403).

[0125] The feature extraction unit 1205 of each of the first to fourth voice recognition devices 1A to 1D verifies whether the voice data sent is the voice data of the main operator 501 based on the broadcasted initial setting request data (S405). The feature extraction unit 1205 extracts voice feature data from the speaker template representing the voice feature of the main operator 501 stored in the reference memory 1032 of each of the first to fourth voice recognition devices 1A to 1D and the broadcasted voice data, respectively, and compares them. If they are consistent (S405 / OK), the initial setting is performed on the new voice recognition device 1S (S406), and the connection setting process ends.

[0126] When the determination result in steps S402 and S405 is NG (S402 / NG), (S405 / NG), this processing is also terminated.

[0127] The initial setting is performed by the main voice recognition device among the first to fourth voice recognition devices 1A to 1D, which can uniformly control all voice recognition devices in the house. The role of the main voice recognition device is performed by, for example, a device installed in the living room of the house (a voice recognition device that family members use more frequently) or the first device installed in the house. Alternatively, the main voice recognition device can be set by the main operator 501.

[0128] Alternatively, the home server 601 in the house can be used as the main device to perform the above-mentioned voice data verification and initial setting of the newly connected device. When the home server 601 implements the voice data verification, the voice data and initial setting request data of the main operator 501 are received by the existing first to fourth voice recognition devices 1A to 1D and forwarded to the home server 601, or received by the home server 601 itself. The home server 601 stores the voice feature data of the main operator 501, that is, the voice template, in advance, and verifies whether the voice features are consistent. If they are consistent, the home server 601 sends instructions on various communication settings to the new voice recognition device 1S so that it can connect to the LAN 210 in the house.

[0129] <Second embodiment>

[0130] In the present embodiment, the in-home interactive system using the voice recognition device 1 is constituted by a system further including a home server device. Fig.12 1 is a diagram showing a schematic configuration of a cooperative system 100 a of voice recognition devices 1 in a house according to the second embodiment.

[0131] and Figure 4 The difference is that a home server 601 is provided on the LAN 210 in the house. The home server 601 stores voice data of the person in the house and voice authentication data 160 including feature points of the voice. Furthermore, the home server 601 always monitors whether the person in the house is near any of the first to fourth voice recognition devices 1A to 1D based on the voice data and data notifying the presence of the person sent from each of the first to fourth voice recognition devices 1A to 1D.

[0132] Therefore, in Fig.12 In the case where character A calls character D, the first voice recognition device 1A that receives the call from character A obtains information of the voice recognition device (fourth voice recognition device 1D) that is judged to be closest to character D from the home server 601.

[0133] Then, by sending voice data only to the fourth voice recognition device 1D from the first voice recognition device 1A, it is possible to communicate only with the voice recognition devices near the person to be called without confirming the location of the other party each time.

[0134] Fig.12In the example, the people in the room are person B in room 2 and person D in room 4. The first to fourth voice recognition devices 1A to 1D each use their own person detection sensor I / F 110 to grasp the indoor conditions and send the results to the home server 601. Thus, the first voice recognition device 1A can preferentially establish communication connections with the second voice recognition device 1B and the fourth voice recognition device 1D in room 2 and room 4 by asking the home server 601 which voice recognition device detected the person.

[0135] Furthermore, by collecting information such as the presence or absence of a person and the presence or absence of a voice through the home server 601, it is possible to always know which person is in which room. In this way, the home server 601 can confirm the destination of the call data sent from the first voice recognition device 1A to the person D, and send the call data to the fourth voice recognition device 1D installed in the room 4 where the person D is.

[0136] Next, character D responds to the call played on the fourth voice recognition device 1D, and a communication connection is established between the first voice recognition device 1A and the fourth voice recognition device 1D, enabling a conversation between character A and character D.

[0137] In the above example, the home server 601 grasps the indoor conditions through the person detection sensors respectively provided on the first to fourth voice recognition devices 1A to 1D, but it is also possible to use data on the usage status of each voice recognition device instead of or in addition to this method.

[0138] Fig.13 An example of the usage status data 170 in which the usage status of the person A is recorded for each time period is shown. Fig.14 An example of the call priority data 180 for each time period of the character A determined based on the usage status data 170 is shown. Similar data is generated for other characters, but is not shown here.

[0139] The reference memory 1032 of each of the first to fourth voice recognition devices 1A to 1D stores the usage status data 170 and the call priority data 180. For example, when the person A receives a call and responds through the first voice recognition device 1A, the first to fourth voice recognition devices 1A to 1D write and update the response score (history indicating that a response has been made) and the call priority in the usage status data 170 and the call priority data 180 stored in the device.

[0140] Furthermore, the first voice recognition device 1A broadcasts the updated usage status data 170 and call priority data 180 to the in-home LAN 210. The second to fourth voice recognition devices 1B to 1D each use the received updated usage status data 170 and call priority data 180 to update the usage status data 170 and call priority data 180 stored in the reference memory 1032 of the device.

[0141] In this state, it is assumed that person B calls person A from the fourth voice recognition device 1D at 7:00 on Monday. The command sending unit 1402 of the fourth voice recognition device 1D refers to the call priority data 180 and reads the first to third voice recognition devices 1A to 1C other than the fourth voice recognition device 1D in order of priority from high to low, that is, in the order of the second voice recognition device 1B, the first voice recognition device 1A, and the third voice recognition device 1C.

[0142] The call priority data 180 not only depends on the usage data 170, but can also be changed according to the user's designation. For example, if it is known that the call target is located next to a specific device during a certain period of time, the call priority data 180 can be temporarily changed to set the priority of the device to level 1.

[0143] In addition, for example, if the first usage status data 170 shows that person A frequently uses the second voice recognition device 1B between 20:00 on Saturdays and Sundays and 8:00 the next morning, then if it is during this time, it can be determined that person A is likely to be indoors, and a conversation mode connection can be attempted. In time periods other than the above, the home server 601 can also process as if person A is not indoors.

[0144] (Dialogue with people outside the residence)

[0145] In the case of the above example, if it is determined that the called person D is not indoors, the voice data can also be transmitted to a communication device such as a smartphone held by person D.

[0146] In this case, by registering the Internet address, line information, device ID and other device information of person D and the communication device he possesses in advance in the reference memory 1032 of the home server 601, call data can be transmitted to the communication device according to the information.

[0147] When the call information reaches the communication device held by person D, person D is notified by screen display, sound output, vibration, etc. Here, person D may respond and a call may be started between the first voice recognition device 1A in the house and the communication device held by person B outside the house.

[0148] If person D does not answer the call at this stage, absence notification data is sent from the home server 601 to the first voice recognition device 1A. In response to the fact that person D does not respond (did not answer the call), a specified voice such as "No one is responding at the moment" is output from the first voice recognition device 1A.

[0149] In the above-described embodiment, an example of calling the second to fourth voice recognition devices 1B to 1D from the first voice recognition device 1A is described, but the present invention is not limited thereto, and calling can also be made from any one of the second to fourth voice recognition devices 1B to 1D. Therefore, it is possible to call other voice recognition devices from any voice recognition device in a residence. In addition, multiple voice recognition devices of this embodiment can be set up, and in the case of a new setting, the above-mentioned setting method can be used to add a new device.

[0150] (Notification from outside the residence)

[0151] Fig.15 An example is shown in which the voice recognition apparatus 1 and the portable communication terminal 71 are connected via the docking station 701 .

[0152] The speech recognition device 1 also includes a docking station 701. The docking station 701 includes a charging control interface 711 for charging the portable communication terminal 71, and a communication control interface 712 for communicating via a connection terminal. As specific functions, a function of wired connection through a USB (Universal Serial Bus) or a specific portable communication terminal interface, a wireless charging function, and a wireless communication function can be realized.

[0153] When the portable communication terminal 71 receives an incoming call, the portable communication terminal 71 outputs an incoming call notification voice such as “A call has come.” or “An email has come.” to the voice recognition device 1 via the communication control interface 712 . The voice output unit 108 outputs an incoming call notification voice.

[0154] When the holder of the portable communication terminal 71 responds with "Please play.", "Who is it from?", "What's the content?", the portable communication terminal 71 can be instructed to answer the call, to transmit the content of the email, to inform the name of the other party who sent the notification, or to use the voice recognition device as a speaker of the phone. If it is an email, the email content can be output in the form of voice.

[0155] Furthermore, when the holder of the portable communication terminal 71 is not indoors, it is determined that the holder is in another place in the house based on whether there is a response within a specified time. Fig.12The home server 601 shown estimates the voice recognition device 1 closest to the current location of the owner of the portable communication terminal 71, and transmits an incoming call notification to the voice recognition device 1. Based on the transmitted incoming call notification, the same operation as the above-mentioned calling operation is performed.

[0156] In addition, as a method of identifying the current location of the holder of the portable communication terminal 71, the usage status of each voice recognition device 1 in the house, feature extraction data of the sound spoken by the user to each voice recognition device 1, the sound collected by each voice recognition device 1, the connection status of the docking station 701 and the portable communication terminal 71, etc. can be used in the home server 601 to determine which voice recognition device 1 the person in the house is near.

[0157] Furthermore, it is also possible that a device such as a portable communication terminal 71, which is usually carried by its holder (it may also be a wearable device), can use short-range communication - which can be judged to be close enough, for example, when connected to a docking station 701 - to communicate with the voice recognition device 1, or when it is possible to use the camera 113 of the voice recognition device 1 to confirm that the terminal is in the same room, it is inferred that the terminal holder is in the room, and measures are taken to increase the priority of the voice recognition device 1 in the room in response to the call.

[0158] <Third embodiment>

[0159] (Monitoring when no one is home)

[0160] The third embodiment will Fig.12 The structure of the second embodiment shown is applied in other ways, involving a housekeeping mode. Fig.16 : is a flowchart showing the switching process to the home-watching mode.

[0161] Among the people in the house, the person who has permission to access each voice recognition device 1 utters the fifth hot word for switching to the housekeeping mode, such as "Please housekeeping!" When any one of the first to fourth voice recognition devices 1A to 1D detects the fifth hot word (S501 / Yes), the feature extraction unit 1205 of the voice recognition device that detects the fifth hot word compares and determines the features of the voice of the person registered as a person with permission to access the LAN in the house in the voice authentication data 160 with the voice feature data extracted by the voice recognition device 1.

[0162] The method described in each of the above-mentioned embodiments can be used as the method for this determination. In addition, the order of confirming the hot word and determining whether to allow access is not determined.

[0163] When the feature extraction unit 1205 determines that they are consistent (S502 / Yes), the command transmission unit 1402 issues an instruction to the home server 601 to shift to the home-watching mode (S503).

[0164] When the home server 601 receives a transfer instruction to the home-watching mode, it sends a home-watching mode transfer instruction (S504) to all the first to fourth voice recognition devices 1A to 1D in the house to transfer to the home-watching mode after a specified time has passed since the receipt of the instruction. In the home-watching mode, when a sound above a specified volume is detected, a notification indicating the presence of an abnormal sound is sent to the home server 601.

[0165] Each of the first to fourth voice recognition devices 1A to 1D is switched to the home-keeping mode by the mode switching unit 1401 provided therein, and the home-keeping mode execution unit 1407 executes the processing of the home-keeping mode ( S505 ).

[0166] When at least one of the first to fourth voice recognition devices 1A to 1D detects voice in the house-keeping mode (S506 / Yes), the feature extraction unit 1205 of the first to fourth voice recognition devices 1A to 1D that detects the voice compares the detected voice with the voice feature data of the person (family member) registered in the voice authentication data 160. When the detected voice is consistent with the voice feature data registered in the voice authentication data 160 (S507 / Yes), the mode switching unit 1401 returns from the house-keeping mode to the normal mode (S508).

[0167] In step S507, when it is determined that the voice data detected by the feature extraction unit 1205 is inconsistent with the voice feature data of the person (family member) registered in the voice authentication data 160 (S507 / No), the mode switching unit 1401 switches to the alert mode (S509) and the alert mode execution unit 1409 is started.

[0168] The alert mode execution unit 1409 activates the alert mode, for example, starts the camera 113 provided in each of the first to fourth voice recognition devices 1A to 1D to record indoor image data, or performs recording processing of the voice detected by the voice input unit 107. In addition, abnormal information may be sent to the home server 601. When receiving the abnormal information, the home server 601 sends an alert such as an email to the portable communication terminal 71 such as a portable phone or a smart phone held by the family members registered in advance.

[0169] Family members who receive the alarm can also communicate with the home server 601. The home server 601 receives the sound collected by the voice input unit 107 of the voice recognition device 1 that detects the abnormal sound as voice data, and sends the data directly to the family members so that they can confirm the situation in the house.

[0170] During the execution of the alert mode, when the release condition of the alert mode is met, for example, when the first hot word indicating returning to the normal mode from a family member who has returned home is detected (S510 / Yes), the mode switching unit 1401 releases the alert mode and returns to the normal mode (S508).

[0171] The above describes an example of using multiple voice recognition devices 1, but the voice recognition device 1 can also perform monitoring operations alone. In the case of independent operation, the voice recognition device 1 has the functions performed by the home server 601. Alternatively, a cloud server on the Internet can be used instead of the home server 601.

[0172] According to this embodiment, the voice recognition devices 1 in the house can be made to work together to start the alert mode based on whether there is voice input when no one is at home and whether the characteristics of the voice are registered voices. In this way, the voice recognition devices 1 installed in multiple rooms can be used to monitor illegal intrusions in the house. After detecting an abnormality, all the voice recognition devices 1 in the house can be used to execute the alert mode, and the intruder can be tracked, the action and facial image can be recorded, and the family members can be notified.

[0173] In addition, when the above-mentioned abnormal sound is detected, a warning sound and a warning voice can also be output from the voice recognition device 1. The warning sound can be pre-registered as a siren, and the words such as "who is it" can be pre-registered as the warning voice, so that it can be played after the abnormal sound is detected, which is expected to improve the anti-theft effect.

[0174] In the above embodiment, a one-to-one conversation between voice recognition devices is described, but it is not limited thereto, and a one-to-many or many-to-many conversation mode is also possible. In this case, the voice data of the characters in the conversation mode is sent to multiple voice recognition devices.

[0175] According to this embodiment, by using a plurality of voice recognition devices in cooperation, communication within a residence can be performed via a network such as a home network. That is, voice recognition devices installed in different rooms and different locations can communicate with each other as if they were in the same room and the same location, so that smooth communication with others can be performed via voice recognition devices installed in other locations.

[0176] Description of Reference Numerals

[0177] 1: Voice recognition equipment

[0178] 100: Collaboration System

[0179] 100a: Collaborative Systems

[0180] 101: CPU

[0181] 102: Bus

[0182] 103: Memory

[0183] 104: Wired LAN I / F

[0184] 105: Wireless LAN I / F

[0185] 106: Wireless communication I / F

[0186] 107: Voice input unit

[0187] 108: Voice output unit

[0188] 109: Display output unit

[0189] 111: Timer

[0190] 113: Camera

Claims

1. A speech recognition device connected to a plurality of other speech recognition devices via a communication network, characterized in that: include: Voice input unit; Voice output unit; A communication control unit for sending and receiving data with other speech recognition devices; a conversation mode execution unit that transmits the voice data input from the voice input unit to the other voice recognition device and outputs the voice data received from the other voice recognition device from the voice output unit; a speech recognition unit that converts the speech input from the speech input unit into text data; A hot word detection unit for detecting a conversation start hot word from the text data, wherein the conversation start hot word is used to instruct to start the conversation mode execution unit; a command sending unit for sending a control command to the other speech recognition devices; a storage unit for storing voice authentication data and first usage status data, wherein the voice authentication data is obtained by associating a person who has been allowed to use the voice recognition device for conversation with voice feature data of the person, and the first usage status data is obtained by associating a person who has been allowed to have the conversation with the number of responses made by the person from each voice recognition device; and a feature extraction unit that extracts voice feature data of the voice data input from the voice input unit and detects whether the voice feature data is consistent with the voice authentication data, The voice input unit receives a voice input for calling a specific person. The command transmission unit transmits the control command to each of the other voice recognition devices in descending order of the number of responses of the person detected by the feature extraction unit as being consistent with the voice authentication data in the first usage status data.

2. The speech recognition device according to claim 1, characterized in that: When the hot word detection unit detects the conversation start hot word, the command sending unit sends the voice data of the detected conversation start hot word and the play command of the voice data to the other voice recognition device.

3. The speech recognition device according to claim 1, characterized in that: The command transmission unit transmits the control command to the other voice recognition device when the feature extraction unit detects that the voice feature data is consistent with the voice authentication data.

4. The speech recognition device according to claim 1, characterized in that Also includes: Camera Department, The storage unit further stores facial authentication data, wherein the facial authentication data is obtained by associating a person who has been allowed to have a conversation using the voice recognition device with a photographed image of the person, The feature extraction unit further detects whether the image captured by the camera unit is consistent with the facial authentication data. When the feature extraction unit detects that the captured image is consistent with the facial authentication data, the command transmission unit transmits the control command to the other voice recognition device.

5. The speech recognition device according to claim 1, characterized in that: The command transmission unit transmits the control command to the other voice recognition device that has output the human detection information, based on the human detection information indicating that the other voice recognition device has detected the presence of a person.

6. The speech recognition device according to claim 1, characterized in that: Also includes a timer, The voice input unit receives a voice input for calling a specific person. The command sending unit sends the voice for calling the specific person and a play command for playing the voice on the other voice recognition device to the other voice recognition device, Obtaining from the timer the elapsed time since the voice for calling the specific person and the playback command were sent, When the elapsed time is equal to or longer than a predetermined waiting time, a response message is output from the voice output unit to notify that there is no response message from the specific person.

7. The speech recognition device according to claim 1, characterized in that: The voice input unit receives a voice input for calling a specific person. The command sending unit sends the voice for calling the specific person and the play command for playing the voice on the other voice recognition devices to all the other voice recognition devices. When receiving a response message from the specific person from one of the other voice recognition devices, the communication control unit maintains the communication connection with the other voice recognition device that sent the response message and disconnects the communication with the remaining other voice recognition devices.

8. The speech recognition device according to claim 1, characterized in that: The speech recognition device also includes Time measurement department, The storage unit further stores second usage status data, wherein the order of calling the plurality of other voice recognition devices from the voice recognition device is determined according to time periods in the second usage status data, The command sending unit obtains the time when the conversation start hot word is detected from the time measuring unit, and sends the control command to each of the plurality of other voice recognition devices according to the call priority determined for the time period to which the time belongs in the second usage status data.

9. The speech recognition device according to claim 1, characterized in that: The speech recognition device also includes: Human body sensors; and a housekeeping mode execution unit for monitoring whether the human body sensor detects a person, The hot word detection unit further detects a housekeeping hot word, and the housekeeping hot word is used to instruct to start the housekeeping mode execution unit. When the feature extraction unit detects that the voice feature data and the voice authentication data are consistent with each other during the execution of the house-keeping mode execution unit, the house-keeping mode execution unit is stopped.

10. A voice recognition device collaboration system, using a communication network to connect a first voice recognition device with a plurality of second voice recognition devices, characterized in that: The first speech recognition device and the second speech recognition device each include: Voice input unit; Voice output unit; A communication control unit for sending and receiving data with other speech recognition devices; a conversation mode execution unit that transmits the voice data input from the voice input unit to the other voice recognition device and outputs the voice data received from the other voice recognition device from the voice output unit; a speech recognition unit that converts the speech input from the speech input unit into text data; A hot word detection unit for detecting a conversation start hot word from the text data, wherein the conversation start hot word is used to instruct to start the conversation mode execution unit; a command sending unit for sending a control command to the other speech recognition devices; a storage unit for storing voice authentication data and first usage status data, wherein the voice authentication data is obtained by associating a person who has been allowed to have a conversation using a voice recognition device with voice feature data of the person, and the first usage status data is obtained by associating a person who has been allowed to have the conversation with the number of responses made by the person from each voice recognition device; and a feature extraction unit that extracts voice feature data of the voice data input from the voice input unit and detects whether the voice feature data is consistent with the voice authentication data, Among them, in the first voice recognition device, the voice input unit receives the input of the voice calling a specific person, and the command sending unit sends the control command to each of the multiple second voice recognition devices in the order from high to low according to the number of responses of the person consistent with the sound authentication data detected by the feature extraction unit in the first usage status data.

11. A method for cooperating with a speech recognition device, executed on a speech recognition device connected to a plurality of other speech recognition devices via a communication network, characterized in that: include: The step of receiving an input of spoken speech; The step of converting the speech into text data; The step of detecting a conversation start hot word from the text data, wherein the conversation start hot word is used to indicate starting a conversation mode; The step of sending a control command for causing the other speech recognition device to transfer to a conversation mode to the other speech recognition device; and a step of starting a dialogue mode, wherein in the dialogue mode, the voice data received from the other voice recognition device is output in the form of voice, and the received input voice is sent to the other voice recognition device, Also includes: When the voice recognition device receives a voice input for calling a specific person, the voice recognition device extracts the voice feature data of the voice for calling the specific person by referring to the voice authentication data obtained by associating the voice feature data of the person who has been allowed to use the voice recognition device for conversation with the person, and detects whether the voice feature data is consistent with the voice authentication data; and Referring to the first usage status data obtained by associating the character who has been allowed to conduct the conversation and the number of responses made by the character from each voice recognition device, the step of sending the control command to each of the multiple other voice recognition devices is carried out in descending order of the number of responses detected in the first usage status data by the character consistent with the voice authentication data.

Images