JP6391386B2 - Server, server control method, and server control program - Google Patents

Description

  The present invention relates to an interactive apparatus, an interactive system, an interactive program, a server, a server control method, and a server control program that provide pseudo communication.

As an interactive device, a speech interactive device capable of interacting with a user has been proposed (Patent Document 1).
The apparatus employs a method in which emotion parameters are provided so that a natural conversation like humans is possible, and the response content is changed according to the value of the emotion parameters.

  As another example, an interactive device that suppresses unnatural interaction with a person has been proposed (Patent Document 2). The apparatus employs a method of changing the attitude expressed in accordance with the level of understanding of the dialogue.

JP 2002-123289 A JP 2013-154458 A

  On the other hand, in the method according to Patent Document 1, the emotion parameter is adjusted due to the user's emotion, but it is not necessary for the user's emotion and the dialogue device to be related in a natural dialogue, and conversely, There is also the possibility of a natural dialogue.

  In the method according to Japanese Patent Laid-Open No. 2004-228688, the correspondence is made according to the degree of understanding, so the user side needs to match the degree of understanding of the device, and a natural conversation does not occur.

  The present disclosure has been made to solve the above-described problem, and provides an interactive apparatus, an interactive system, an interactive program, a server, a server control method, and a server control program capable of natural interaction with a user. It is intended.

  An interactive device according to an aspect of the present invention includes an input receiving unit that receives input from a user, a parameter management unit that manages a mood parameter indicating emotion of the device, and a value of a mood parameter as an index when each is selected And a plurality of responses based on the mood parameters managed by the parameter management unit with reference to the storage unit, and a storage unit that stores a plurality of response information related to responses from the user received by the input reception unit A selection unit that selects one of the information; and a response processing execution unit that executes a response process based on the response information selected by the selection unit.

  An interactive system according to an aspect of the present invention includes an input receiving unit that receives an input from a user, a parameter management unit that manages a mood parameter indicating an emotion of the device, and a value of the mood parameter as an index when each is selected And a plurality of responses based on the mood parameters managed by the parameter management unit with reference to the storage unit, and a storage unit that stores a plurality of response information related to responses from the user received by the input reception unit A selection unit that selects one of the information; and a response processing execution unit that executes a response process based on the response information selected by the selection unit.

  An interactive program executed in a computer according to an aspect of the present invention, the interactive program receiving a user's input to the computer, managing a mood parameter indicating an emotion of the device, , Referring to a storage unit that is associated with the value of the mood parameter as an index at the time of selection, and stores a plurality of response information related to the response to the input from the received user, a plurality of response information based on the mood parameter A process including a step of selecting one and a step of executing a response process based on the selected response information is executed.

  An interactive system including an interactive device according to an aspect of the present invention and a server provided so as to be communicable, the interactive device receiving an input from a user, and a response executing response processing in accordance with an instruction from the server A processing execution unit. The server includes a parameter management unit that manages mood parameters indicating emotions of the device, a plurality of parameters related to responses from the user received by the input receiving unit, each of which is associated with a value of the mood parameter as an index for selection. A storage unit that stores the response information, a selection unit that selects one of a plurality of response information based on the mood parameters managed by the parameter management unit with reference to the storage unit, and a response selected by the selection unit A response process execution instructing unit that instructs the dialogue apparatus to execute the response process based on the information.

  An interactive apparatus provided to be able to communicate with a server according to another aspect of the present invention, managed by a server according to an input receiving unit that receives an input from a user and an input from the user that is received by the input receiving unit A response processing execution unit that executes response processing based on response information selected from a plurality of response information related to a response to an input from a user stored in the server based on a mood parameter indicating an emotion of the device .

  An interactive program executed in a computer of an interactive device provided to be able to communicate with a server according to another aspect of the present invention, the interactive program receiving an input from a user to the computer, and an accepted user In response to an input from, a response is made based on response information selected from a plurality of response information regarding a response to the input from the user stored in the server based on the mood parameter indicating the emotion of the device managed by the server And executing a process including the step of executing the process.

  A server provided to be able to communicate with an interactive device according to an aspect of the present invention, a receiving unit that receives an input from a user received by the interactive device, a parameter management unit that manages a mood parameter indicating the emotion of the device, , Each of which is associated with the value of the mood parameter as an index at the time of selection, a storage unit that stores a plurality of response information related to responses from the user received by the reception unit, and a parameter with reference to the storage unit A selection unit that selects one of a plurality of response information based on the mood parameter managed by the management unit, and instructs the dialog device to execute response processing based on the response information selected by the selection unit A response processing execution instruction unit.

Preferably, the parameter management unit updates the mood parameter.
Preferably, the parameter management unit updates the mood parameters based on information acquired from the outside.

  Preferably, the parameter management unit updates the mood parameters based on environmental information acquired from the outside.

  Preferably, the parameter management unit updates the mood parameter when information acquired from the outside satisfies a predetermined condition.

  Preferably, a parameter management part updates a mood parameter based on the input mode from the user received by the receiving part.

  Preferably, the parameter management unit updates the mood parameter based on the content of the voice input from the user received by the receiving unit.

  Preferably, the parameter management unit updates the mood parameter based on the sensor input information detected by the user received by the receiving unit.

  Preferably, the server further includes a history storage unit that stores a history of input from the user, and the parameter management unit updates the mood parameters based on a plurality of input modes from the user with reference to the history storage unit. .

  Preferably, the response process execution instructing unit instructs the dialogue apparatus to execute the speech synthesis process, and the plurality of response information includes different parameter information when executing the speech synthesis process.

  Preferably, the response process execution instructing unit instructs the dialogue apparatus to execute the audio output process, and the plurality of response information includes different output contents when executing the audio output process.

  A method for controlling a server provided to be able to communicate with an interactive device according to an aspect of the present invention, the step of receiving input from a user accepted by the interactive device, and the step of managing mood parameters indicating emotions of the device; , Each is associated with the value of the mood parameter as an index at the time of selection, and a plurality of response information regarding the response to the input from the received user is stored, and a plurality of response information is stored based on the mood parameter. Selecting one of the response information, and instructing the dialog device to execute a response process based on the selected response information.

  A server control program executed in a computer of a server provided to be able to communicate with an interactive apparatus according to an aspect of the present invention, the server control program receiving an input from a user received by the interactive apparatus with respect to the computer And a step of managing mood parameters indicating emotions of the device, each of which is associated with a value of the mood parameter as an index when selected, and a plurality of response information relating to responses to received inputs from the user is stored Referring to the stored storage unit, selecting one of a plurality of response information based on the mood parameter, and instructing the dialogue apparatus to execute a response process based on the selected response information The process is executed.

  According to one aspect of the present disclosure, natural interaction with a user is possible.

It is a figure explaining dialogue system 1 based on an embodiment. It is a figure explaining the principal part structure of the dialogue system 1 which concerns on embodiment. It is a figure explaining the response content database 232 based on Embodiment 1. FIG. It is a figure explaining the log | history memory | storage part 233 based on Embodiment 1. FIG. It is a figure explaining the sensor information database 234 based on Embodiment 1. FIG. It is a figure explaining the change table database 231 based on Embodiment 1. FIG. It is a sequence diagram which shows the flow of the response process in the dialogue system 1 based on Embodiment 1. It is a flowchart which performs the dialog output process of the server 20 based on Embodiment 1. FIG. It is a flowchart which performs the mood parameter acquisition process of the server 20 based on Embodiment 1. FIG. It is a figure explaining the speech synthesis parameter table based on Embodiment 2. FIG. It is a figure explaining the response content database 232A based on Embodiment 3. FIG. It is a figure explaining the character database based on Embodiment 4. FIG. It is a figure explaining the structure of the server based on Embodiment 5. FIG.

  The present embodiment will be described below with reference to the drawings. In the description of the embodiments, when the number and amount are referred to, the scope of the present invention is not necessarily limited to the number and amount unless otherwise specified. In the description of the embodiments, the same parts and corresponding parts are denoted by the same reference numerals, and redundant description may not be repeated. Unless there is a restriction | limiting in particular, it is planned from the beginning to use suitably combining the structure shown to the structure shown to embodiment.

<Embodiment 1>
(Configuration of Dialog System 1)
FIG. 1 is a diagram illustrating a dialogue system 1 based on the embodiment.

  With reference to FIG. 1, a dialogue system 1 based on the embodiment is configured by a doll (dialogue device) 10, a network 5, and a server 20.

  The doll 10 is provided to be able to communicate with the server 20 via the network 5. In this example, the case of communicating with the server 20 via the network 5 will be described, but a method of directly communicating with the server 20 may be used.

  As an example, when a voice is input from a human (user) to the doll 10 (hereinafter also referred to as “user utterance”), the dialogue system 1 recognizes the voice in the server 20 and responds to the input voice. Is output from the doll 10 (hereinafter also referred to as “audio playback”). By repeating the process, the interactive system 1 according to the embodiment realizes a pseudo conversation between the user and the doll 10. In this example, a case where a voice is input from the user to the doll 10 will be described as an example, but a voice is output from the doll 10 to the user. The voice may be input, and the order is not particularly limited.

  In the embodiment, a doll 10 that recognizes a voice and outputs a voice response to the user will be described as an example of an interactive device. However, the present invention is not limited to this. For example, home appliances (for example, a television, a microwave oven, etc.) other than the doll 10 having an interactive function may be employed as the interactive device.

  In the embodiment, a configuration in which the server 20 is realized by one server will be described as an example. However, the present invention is not limited to this, and at least one of the units (functions) included in the server 20 is described. A configuration may be adopted in which the unit is realized by another server.

(Main components of the dialogue system 1)
FIG. 2 is a diagram illustrating a configuration of main parts of the interactive system 1 according to the embodiment.

With reference to FIG. 2, the structure of the doll 10 is demonstrated first.
The doll 10 based on the embodiment includes a communication unit 101, a control unit 102, a microphone 103, a speaker 104, a driving unit 106, a sensor 108, and a storage unit 109.

  The communication unit 101 is means for performing communication with the outside. Specifically, the communication unit 101 communicates with the server 20 via a network 5 such as the Internet. Note that either wireless or wired communication is possible.

  The microphone 103 receives sound input from the outside. In the embodiment, a case will be described in which input of sound data (also referred to as sound data) included in a frequency band of sound mainly produced by humans is received as sound data indicating sound that the microphone 103 receives input. Sound data including a frequency band other than the frequency band of the audio data may be included. The microphone 103 outputs audio data indicating the input sound to the control unit 102.

  The speaker 104 outputs a voice response representing the response content output from the control unit 102. Hereinafter, the output of the voice response performed by the doll 10 via the speaker 104 is also referred to as “reproduction”. Details of the response contents will be described later.

  The drive unit 106 operates the doll 10 based on an instruction from the control unit 102. For example, it is possible to drive parts such as movable hands and feet provided on the doll 10 by the driving unit 106.

The operation unit 107 receives instructions for performing various operations with respect to the doll 10.
The sensor 108 detects various information. In this example, as an example, the environmental information includes a temperature sensor and a humidity sensor capable of detecting temperature and humidity. Further, a sensor (vibration sensor or the like) that detects that the user touches the doll 10 is also included. Further, the sensor 108 includes a camera that acquires imaging information, and includes a sensor that can capture a user's face by the camera and detect a smile as a user's facial expression included in the imaging information. And

  The storage unit 109 is a storage device such as a RAM (Random Access Memory) and a flash memory, and stores programs for realizing various functions of the doll 10.

  The control unit 102 is mainly composed of a CPU (Central Processing Unit), and realizes the function of each unit that the CPU executes a program stored in the storage unit 109.

  The control unit 102 comprehensively controls each unit of the doll 10. Specifically, the control unit 102 controls the operation of the doll 10 by controlling the driving unit 106. Further, the control unit 102 transmits audio data indicating sound acquired from the outside by the microphone 103 to the server 20 via the communication unit 101.

  Further, the control unit 102 receives the reply phrase data recognized by the server 20 via the communication unit 101 regarding the voice data transmitted to the server 20. And the control part 102 can output the audio | voice response showing the response content from the speaker 104 according to the received reply phrase data.

The main functional configuration of the control unit 102 will be described.
The control unit 102 includes a response process execution unit 112, a voice input reception unit 114, and an information acquisition unit 115.

  The voice input reception unit 114 detects (extracts) voice data. In other words, the voice input receiving unit 114 detects the sound data (voice data) by extracting the frequency band of the voice uttered by a person from the sound data received from the outside.

  As a method of detecting the voice data from the sound data in the voice input reception unit 114, for example, the voice data is extracted by extracting the frequency band (for example, the frequency band of 100 Hz or more and 1 kHz or less) of the voice generated by humans from the sound data. The method of detecting can be mentioned. In this case, the voice input receiving unit 114 may be provided with, for example, a bandpass filter or a filter combining a high-pass filter and a low-pass filter in order to extract a frequency band of voice uttered by humans from sound data. That's fine.

  The voice input reception unit 114 transmits voice data detected from the sound data to the server 20 via the communication unit 101.

  The response processing execution unit 112 performs speech synthesis processing based on the answer phrase data from the server 20 and utters the user via the speaker 104 as an example.

  The information acquisition unit 115 acquires information such as temperature and humidity detected by the sensor 108 and detection information detected by the user's contact and smile, and transmits them to the server 20 via the communication unit 101.

Next, the configuration of the server 20 based on the embodiment will be described.
The server 20 based on the embodiment includes a communication unit 201, a control unit 202, and a storage unit 203.

  The communication unit 201 is a means for performing communication with the outside. Specifically, the communication unit 201 communicates with the doll 10 via a network 5 such as the Internet. Moreover, it is provided so that communication with external apparatuses other than the doll 10 is also possible. Note that wireless or wired communication is possible.

  The storage unit 203 is a storage device such as a RAM (Random Access Memory) and a flash memory, and stores programs for realizing various functions of the server 20. In addition, the storage unit 203 includes, as an example, a change table database 231 used when updating the mood parameters, a response content database 232 that is information (also referred to as response information) regarding a response to voice input, and a history that manages the history It has the memory | storage part 233 and the sensor information database 234 which memorize | stores sensor information.

  The control unit 202 is mainly configured by a CPU (Central Processing Unit), and is realized by the CPU executing a program stored in the storage unit 203.

  The control unit 202 comprehensively controls each unit of the server 20. Specifically, the control unit 202 outputs answer phrase data to the doll 10 via the communication unit 201 as a result of performing voice recognition on the voice data received from the doll 10 via the communication unit 201.

Next, the main functional configuration of the control unit 202 of the server 20 will be described.
The control unit 202 includes a voice input reception unit 221, a management unit 222, a voice recognition unit 223, a selection unit 224, and a response process execution instruction unit 225.

  The voice input reception unit 221 receives voice data transmitted from the doll 10 via the communication unit 201. The voice input receiving unit 221 outputs the received voice data to the voice recognition unit 223.

  The management unit 222 manages mood parameters related to the doll 10 (device). In this example, the management unit 222 calculates and updates the mood parameters of the doll 10 based on the change table database 231 and the like as an example.

  The voice recognition unit 223 recognizes the voice content (sound content) indicated by the voice data received by the voice input reception unit 221 as the recognition content. Specifically, a recognition phrase for voice data is acquired using a dictionary used for voice recognition provided in advance in the storage unit 203. In addition, when the recognition phrase with respect to audio | voice data cannot be acquired using the dictionary utilized for the said audio | voice recognition, it determines with audio | voice recognition having failed.

  The selection unit 224 determines the response content based on the speech content recognition result of the speech recognition unit 223. Specifically, the selection unit 224 refers to the response content database 232 stored in the storage unit 203 and based on the mood parameters managed by the management unit 222, the response corresponding to the audio content indicated by the audio data Select (determine) the content (response information).

  The response process execution instructing unit 225 transmits the response phrase data that is the response content (response information) selected by the selection unit 224 to the doll 10 via the communication unit 201.

(Response database)
FIG. 3 is a diagram illustrating the response content database 232 based on the first embodiment.

  With reference to FIG. 3, the response content database 232 is stored in the storage unit 203 of the server 20 based on the embodiment as an example.

  Specifically, a plurality of response information is registered in the response content database 232. Specifically, the recognition content (recognition phrase) and the response content (answer phrase) are registered in association with each other. In this example, an identification number (response ID) is assigned to each combination of recognition phrase and answer phrase. As an example, it is assumed that the recognition phrase registered in the response content database 232 in this example is also registered in the dictionary used for speech recognition.

  As an example, here, as the recognition phrases, answer phrases are stored in association with "Good morning", "Cute", "No", "Tell me about Tokyo" ... Yes.

  For example, a case is shown in which an answer phrase “Good morning! Let's do our best today also” is registered in association with the recognition phrase “Good morning” of the response ID “1”.

  In addition, a case is shown in which the answer phrase “good morning” is associated and registered in correspondence with the recognition phrase “good morning” of the response ID “2”.

  Further, a case is shown in which an answer phrase “Fuwaa. Still sleepy yo” is associated and registered in correspondence with the recognition phrase “Good morning” of the response ID “3”.

  In addition, a case is shown in which an answer phrase “Thank you! I am happy.” Is associated and registered in correspondence with the recognition phrase “cute” of the response ID “4”.

  In addition, a case is shown in which an answer phrase “Fun. Anyway flattering?” Is associated and registered in correspondence with the recognition phrase “cute” of the response ID “5”.

  For example, a case is shown in which an answer phrase “Yeah, the flatter is already good” is associated and registered in correspondence with the recognition phrase “cute” of the response ID “6”.

  In addition, a case is shown in which an answer phrase “I'm sorry. I'll study.” Is registered in association with the recognition phrase “no good” of the response ID “7”.

  In addition, a case is shown in which the answer phrase “I'm sorry” is registered in association with the recognition phrase “no good” of the response ID “8”.

  In addition, a case is shown in which the response phrase “ha. I'm going to sleep” is registered in association with the recognition phrase “no good” of the response ID “9”.

  In addition, there is a case in which the response phrase “TOKYO is famous for the fireworks display of Sumida River” is registered in association with the recognition phrase “Tell me about Tokyo” of the response ID “10”. It is shown.

  Also, in response to the recognition phrase “Tell me about Tokyo” of the response ID “11”, the response phrase “There is Meiji Jingu, the first place for first visitor in Tokyo” is registered in association with it. The case is shown.

  In addition, a case is shown in which an answer phrase “How much do you investigate yourself?” Is registered in association with the recognition phrase “Tell me about Tokyo” of the response ID “12”.

  And in this example, the mood range is matched with the answer phrase corresponding to each recognition phrase.

  The mood range is used as an index to be selected when there are a plurality of answer phrases for the same recognition phrase. In this example, an answer phrase is selected based on the mood parameter P managed by the management unit 222. An answer phrase in a mood range corresponding to the mood parameter P is selected from the plurality of answer phrases.

  It should be noted that an answer phrase included in the mood range is not necessarily selected, and the probability of being relatively selected may be increased. As a result, natural communication with the user can be achieved without becoming patterned answer contents.

  In addition, an answer phrase (re-answer answer phrase) for requesting a re-response in response to a case where there is no recognized phrase (null) is provided. Here, the case where there is no recognition phrase means a case where speech recognition fails. If the recognition phrase registered in the dictionary used for speech recognition is not registered in the response content database 232, that is, the speech recognition has succeeded but the corresponding recognition phrase is registered in the response content database 232. Even when there is no recognition phrase, it may be processed as a case where there is no recognition phrase.

  Specifically, regarding the response ID “100”, when there is no recognized phrase (null), the answer phrase “what is” is associated and registered.

  In addition, regarding the response ID “101”, a case where the answer phrase “say again” is associated and registered when there is no recognized phrase (null) is shown. By providing a plurality of answer phrases when there is no such recognition phrase (null), it is possible to avoid a patterned response.

  In this example, when a response process corresponding to the user such as an utterance to the user or an answer to the user is performed, a response such as an utterance or an answer to the user is determined based on the mood parameter P of the device. The method to do is demonstrated.

  And in this example, the case where conversation classification and a mood change degree are registered corresponding to each recognition phrase is shown. The conversation type indicates the type of recognition phrase in the conversation, and is used when the mood parameter P is updated. The mood change degree is also a parameter used when the mood parameter P is updated.

  As an example, the conversation type “greeting” and the mood change degree “0” are associated and registered in correspondence with the recognition phrase “good morning”.

  In addition, the conversation type “evaluation” is associated and registered corresponding to the recognition phrases “cute ne” and “no use”, and different mood change degrees are respectively associated and registered according to the contents of the recognition phrase. Yes.

  Corresponding to the recognition phrase “Tell me about Tokyo”, a conversation type “question” and a mood change degree “0” are associated and registered.

The other recognition phrases are basically registered in the same manner.
Regarding the mood change degree, if the content of the recognition phrase is positive, the mood change degree is given so that the value of the mood parameter P increases. On the other hand, if the content of the recognition phrase is passive content, the mood change degree is given so that the value of the mood parameter P becomes low.

  For example, when the value of the mood parameter P of the device managed by the management unit 222 is large, the selection unit 224 selects a positive or cheerful answer phrase corresponding to the recognition phrase.

  On the other hand, when the value of the mood parameter P of the device managed by the management unit 222 is low, the selection unit 224 selects a passive or gloomy answer phrase corresponding to the recognized phrase.

  A response process corresponding to the change in the mood parameter is executed by the mood parameter calculation process based on the mood change degree or the like. Thereby, natural communication with the user can be achieved.

(History storage)
FIG. 4 is a diagram illustrating the history storage unit 233 based on the first embodiment.

  With reference to FIG. 4, the history storage unit 233 is stored in the storage unit 203 of the server 20 based on the first embodiment as an example.

  Specifically, the history storage unit 233 stores information related to the history of the doll 10 or the like. In this example, when response phrase data, which is the response content (response information) selected by the response processing execution instruction unit 225 by the selection unit 224, is transmitted to the doll 10 via the communication unit 201, the history storage unit 233 stores the response phrase data. Information on history is stored. In addition, the history storage unit 233 stores information related to the history of sensor information transmitted from the information acquisition unit 115 of the doll 10. Furthermore, information related to the history of information acquired from the external device is stored.

  Referring to FIG. 4A, here, the history information of the response content stored in history storage unit 233 is shown.

  The server 20 can manage a plurality of dolls, and a unique identification number is assigned to each doll. In this example, as an example, a case where histories of dolls to which “10” is assigned as a doll ID (device ID) is registered.

  Here, the case where the information of the uttered "time", "response ID", and "degree of change in mood" is registered is shown.

  “Time” means a time when the server 20 speaks to the user via the doll 10. In this example, it means the time when the user speaks, but is not limited to the time, and any time may be used as long as the time related to the speech process (response process) can be specified. For example, it may be the time when the server 20 receives input of voice data from the user, or may be the time when the voice is recognized.

  The “response ID” is information for identifying the response information uttered by the server 20 to the user via the doll 10 as an example, and basically corresponds to the response ID in the response content database 232.

  The “moment change degree” is a parameter used when changing the mood parameter P, and indicates the mood change degree associated with the response ID.

In this example, as an example, an utterance based on the response information corresponding to the response ID “2” is executed at the time “2014-07-01 06:30:42” with respect to the device ID “10” of the doll. It is shown. The mood change degree at that time is “0”. The utterance based on the response information corresponding to the response ID “110” at the time “2014-07-01 06:31:05” with respect to the device ID “10” of the doll. Is shown to have been executed. In addition, the mood change degree at that time is “0”. Also, regarding the doll device ID “10”, the response information corresponding to the response ID “130” at the time “2014-07-01 06:31:20” It is shown that the utterance based on was executed. The mood change degree at that time is “0”.

  In addition, regarding the doll device ID “10”, it is shown that the utterance based on the response information corresponding to the response ID “4” was executed at the time “2014-07-01 06:30:44”. Here, it is shown that the utterance based on the response information corresponding to the response ID “4” has been executed. In addition, the mood change degree at that time is “+5”.

  Referring to FIG. 4B, here, history information of information (including sensor information) acquired from the outside stored in history storage unit 233 is shown.

  Sensor information acquired by the doll 10 via the communication unit 201 is stored in the sensor information database 234. The sensor information database 234 will be described later.

  Here, a case is shown in which “device ID”, “time”, and “content” information of the acquired device is registered.

  “Apparatus ID” is an assigned unique identification number. In this example, the history of dolls to which “10” is assigned is registered.

  “Time” means the time at which the server 20 receives input of information from the doll 10. In this example, it means the time when information is received, but is not limited to this time, and any time may be used as long as the time related to acquisition can be specified. For example, the time when the doll 10 is detected by the sensor 108 or the time when the information acquisition unit 115 transmits information may be used.

  The “content” is information for specifying sensor information received from the doll 10 by the server 20 as an example. In the case of sensor information, the “content” corresponds to the sensor ID in the sensor information database 232. In the case of information acquired from an external device, the acquired information may be included.

  In this example, as an example, it is shown that sensor information corresponding to the sensor ID “1” is acquired at the time “2014-07-01 09:54:10” regarding the device ID “10” of the doll. .

  In addition, regarding the doll device ID “10”, the sensor information corresponding to the sensor ID “2” is acquired at the time “2014-07-01 09:55:33”.

  In addition, regarding the doll device ID “10”, it is indicated that the sensor information corresponding to the sensor ID “3” is acquired at the time “2014-07-01 09:59:42”.

  In addition, regarding the doll device ID “10”, the sensor information corresponding to the sensor ID “4” is acquired at the time “2014-07-01 09:59:43”.

  In addition, regarding the doll device ID “10”, it is indicated that the sensor information corresponding to the sensor ID “5” is acquired at the time “2014-07-01 10:00:01”.

  In addition, regarding the doll device ID “10”, it is indicated that the sensor information corresponding to the sensor ID “6” is acquired at the time “2014-07-01 10:00:01”.

  Further, regarding the doll device ID “10”, it is shown that the information of “weather“ sunny ”” is acquired at the time “2014-07-01 12:00:00”.

The same applies to other histories.
(Sensor information database)
FIG. 5 is a diagram illustrating the sensor information database 234 based on the first embodiment.

  Referring to FIG. 5, the sensor information database 234 is stored in the storage unit 203 of the server 20 based on the embodiment as an example.

  Specifically, the sensor information database 234 stores sensor information detected by the sensor 108 such as the doll 10. In this example, when the control unit 202 receives the sensor information transmitted from the information acquisition unit 115 via the communication unit 201, the sensor information is stored in the sensor information database 234 of the storage unit 203.

  When storing in the sensor information database, a unique identification number is assigned to each sensor information. In this example, as an example, a case where sensor IDs are assigned in ascending order is shown. The assigned sensor ID is stored in the history of information acquired from the outside of the history storage unit 233 as described with reference to FIG.

  As an example, “time”, “type”, and “value” are associated and registered in correspondence with the sensor ID.

  “Time” means the time when the server 20 receives input of sensor information from the doll 10. In this example, it means the time when sensor information is input, but is not limited to this time, and any time may be used as long as the time related to sensor acquisition can be specified. For example, the time when the doll 10 is detected by the sensor 108 or the time when the information acquisition unit 115 transmits the sensor information may be used.

  “Type” means the type of sensor information, and as an example, the type “touch” means that the user's contact with the doll 10 is detected. The type “smile” means that a smile of the user has been detected. The type “humidity” means the humidity acquired by the sensor 108. The type “temperature” means a temperature acquired by the sensor 108.

“Value” means a numerical value detected in the case of the types “humidity” and “temperature”.
In this example, as an example, it is indicated that sensor information of type “touch” is acquired at time “2014-07-01 09:54:10” as sensor ID “1”.

  Further, it is indicated that the sensor information of the type “touch” is acquired at the time “2014-07-01 09:55:33” as the sensor ID “2”.

  Further, it is indicated that the sensor information of the type “smile” is acquired at the time “2014-07-01 09:59:42” as the sensor ID “3”.

  Further, it is indicated that the sensor information of the type “touch” is acquired at the time “2014-07-01 09:59:43” as the sensor ID “4”.

  Further, it is indicated that the sensor information of the type “temperature” is acquired at the time “2014-07-01 10:00:01” as the sensor ID “5”. Further, it is indicated that the value “28” has been acquired.

  Further, it is indicated that the sensor information of the type “humidity” is acquired at the time “2014-07-01 10:00:01” as the sensor ID “6”. Further, it is indicated that the value “52” is acquired.

The same applies to other sensor information.
(Change table database)
FIG. 6 is a diagram illustrating the change table database 231 based on the first embodiment.

FIG. 6A shows the correspondence between the number of consecutive conversation types and the degree of change in mood.
It is set to have a negative mood change degree when the conversation types are continuous.

  As an example, when the number of consecutive conversation types is 10 or more, the mood change degree “−5” is associated.

When the number of consecutive times is 5 or more, the mood change degree “−1” is associated.
Further, when the number of continuous times is 0, the mood change degree “0” is associated.

FIG. 6B shows the correspondence between the discomfort index and the mood change degree.
As an example, the higher the discomfort index Q, the lower the mood change level.

  When the discomfort index Q is greater than 75 and equal to or less than 80 (75 <Q ≦ 80), the mood change degree “−1” is associated.

  When the discomfort index Q is greater than 80 and less than or equal to 85 (80 <Q ≦ 85), the mood change degree “−5” is associated.

  When the discomfort index Q is greater than 85 (85 <Q), the mood change degree “−10” is associated.

FIG. 6C shows a correspondence relationship between the time zone and the mood change degree.
It is set to have a negative mood change in the early morning and late night hours.

  As an example, in the case of the time zone “0: 0 to 3:59”, the mood change degree “−10” is associated.

In the case of the time zone “4: 00 to 5: 59”, the mood change degree “−5” is associated.
In the case of the time zone “6: 0 to 6:59”, the mood change degree “−2” is associated.

  In the case of the time zone “23: 00 to 23: 59”, the mood change degree “−2” is associated.

FIG. 6D shows the correspondence between the day of the week and the mood change degree.
As an example, a negative mood change degree is set at the beginning of the week, and a positive mood change degree is set at the weekend.

In the case of Monday, the mood change degree “−10” is associated.
In the case of Friday, the mood change degree “+10” is associated.

In the case of Saturday, the mood change degree “+10” is associated.
FIG. 6E shows a correspondence relationship between the weather and the mood change degree.

  As an example, a positive mood change degree is set on a day when the weather is good, and a negative mood change degree is set on a bad day.

When the weather is fine, the mood change degree “+10” is associated.
In the case of cloudy, the mood change degree “0” is associated.

In the case of rain, the mood change degree “−10” is associated.
(Response processing)
FIG. 7 is a sequence diagram showing a response process flow in the interactive system 1 based on the first embodiment.

  As shown in FIG. 7, the user utters (also referred to as user utterance) to doll 10 (sequence sq0).

  The doll 10 receives an input of voice in response to the user utterance (sequence sq1). Specifically, the voice input receiving unit 114 receives a sound input from the outside via the microphone 103.

  Next, the doll 10 outputs audio data to the server 20 (sequence sq2). Specifically, the voice input reception unit 114 outputs to the server 20 via the communication unit 101.

  Next, server 20 receives the voice data transmitted from doll 10 and performs voice recognition (sequence sq3). Specifically, the voice input reception unit 221 receives voice data via the communication unit 201 and outputs the voice data to the voice recognition unit 223. Then, the voice recognition unit 223 recognizes the voice content. Then, the voice recognition unit 223 outputs the recognition result to the selection unit 224.

  Next, the server 20 executes a dialog output process for determining an answer phrase based on the recognition result (sequence sq4). Specifically, the selection unit 224 determines an answer phrase and outputs it to the response process execution instruction unit 225. The dialog output process will be described later.

  Then, server 20 transmits the determined answer phrase data to doll 10 (sequence sq5). Specifically, the response process execution instruction unit 225 transmits the reply phrase data determined by the selection unit 224 to the doll 10 via the communication unit 201. In this example, the answer phrase is an audio file as an example. It may be a text file. The same applies to other examples.

  Next, the doll 10 executes voice dialogue output (sequence sq6). Specifically, the response process execution unit 112 responds (voice conversation) to the user via the speaker 104 based on the answer phrase data received via the communication unit 201. That is, the response process execution unit 112 reproduces an audio file that is answer phrase data and responds (speaks) the audio to the user through the speaker 104 (sequence sq6A).

  Then, the user speaks (answers) the doll 10 as a response to the response process from the doll 10 (sequence sq6B).

  The doll 10 receives a voice input from the user for the response process (sequence sq7). Specifically, the voice input receiving unit 114 receives a sound input from the outside via the microphone 103.

  Next, the doll 10 outputs audio data to the server 20 (sequence sq8). Specifically, the voice input reception unit 114 outputs to the server 20 via the communication unit 101.

  Next, server 20 receives the voice data transmitted from doll 10 and performs voice recognition (sequence sq9). Specifically, the voice input reception unit 221 receives voice data via the communication unit 201 and outputs the voice data to the voice recognition unit 223. Then, the voice recognition unit 223 recognizes the voice content. Then, the voice recognition unit 223 outputs the recognition result to the selection unit 224.

  Next, the server 20 executes a dialog output process for determining an answer phrase based on the recognition result (sequence sq10). Specifically, the selection unit 224 determines an answer phrase and outputs it to the response process execution instruction unit 225. The dialog output process will be described later.

  Then, server 20 transmits the determined answer phrase data to doll 10 (sequence sq11). Specifically, the response process execution instruction unit 225 transmits the reply phrase data determined by the selection unit 224 to the doll 10 via the communication unit 201.

  Next, the doll 10 executes voice dialogue output (sequence sq12). Specifically, the response process execution unit 112 responds (voice conversation) to the user via the speaker 104 based on the answer phrase data received via the communication unit 201. That is, the response process execution unit 112 reproduces an audio file that is answer phrase data and responds (speaks) the audio to the user through the speaker 104 (sequence sq12A). Thereafter, the same processing is repeated.

(Interactive output processing)
FIG. 8 is a flowchart for executing the dialogue output process of the server 20 based on the first embodiment.

  Referring to FIG. 8, the flowchart is a process executed by executing a program stored in storage unit 109 and causing each unit of control unit 102 to function.

  First, it is determined whether or not the voice recognition is successful (step S1). Specifically, the selection unit 224 determines whether or not a recognition phrase is notified from the voice recognition unit 223 as a voice recognition result.

  If it is determined in step S1 that the speech recognition is successful (YES in step S1), it is next determined whether or not there are a plurality of answer phrases (step S2). Specifically, the selection unit 224 refers to the response content database 232 (FIG. 3) and determines whether or not a plurality of answer phrases corresponding to the recognized phrase are registered.

  If it is determined in step S2 that there are a plurality of answer phrases (YES in step S2), a mood parameter is acquired (step S3). Specifically, the selection unit 224 instructs the management unit 222 to request a mood parameter. The management unit 222 calculates the mood parameter and outputs the calculation result to the selection unit 222. The mood parameter calculation process will be described later.

  Next, an answer phrase is selected (step S4). Specifically, as an example, the selection unit 224 selects an answer phrase corresponding to the recognized phrase with reference to the response content database 232 (FIG. 3) based on the mood parameter output from the management unit 222.

  Next, an output process is executed (step S5). Specifically, the selection unit 224 outputs the selected answer phrase to the response process execution instruction unit 225. The response process execution instructing unit 225 outputs the reply phrase data selected (determined) by the selection unit 224 to the doll 10 via the communication unit 201.

Then, the process ends (return).
On the other hand, if it is determined in step S2 that there are not a plurality of answer phrases (NO in step S2), an answer phrase is determined (step S7). Specifically, the selection unit 224 selects (determines) an answer phrase corresponding to the recognized phrase with reference to the response content database 232 (FIG. 3).

  Next, an output process is executed (step S5). Specifically, the selection unit 224 outputs the selected answer phrase to the response process execution instruction unit 225. The response process execution instructing unit 225 outputs the reply phrase data selected (determined) by the selection unit 224 to the doll 10 via the communication unit 201.

Then, the process ends (return).
On the other hand, if it is determined in step S1 that the speech recognition has not been successful (NO in step S1), a reresponse answer phrase is determined (step S6).

  Specifically, when the selection unit 224 determines that the speech recognition has not been successful, the selection unit 224 refers to the response content database 232 (FIG. 3) and selects the re-response reply phrase corresponding to the case where there is no recognized phrase (null). (decide. For example, in order to obtain a response from the user again, for example, a re-response reply phrase such as “what” or “go again” is selected (determined). In particular, the phrase is not limited to the re-response reply phrase, and may be another phrase such as “It should have been good”.

  Next, an output process is executed (step S5). Specifically, the selection unit 224 outputs the selected answer phrase to the response process execution instruction unit 225. The response process execution instructing unit 225 outputs the reply phrase data selected (determined) by the selection unit 224 to the doll 10 via the communication unit 201.

Then, the process ends (return).
With this process, an answer phrase is selected from the response content database 232 based on the mood parameters, and the answer phrase data is output from the server 20 to the doll 10 and uttered.

  Thereby, it is possible to achieve natural communication with the user by executing the response process (speech) according to the change in the mood parameter.

  In the above, the selection unit 224 refers to the response content database 232 (FIG. 3) and outputs a re-response reply phrase corresponding to the case where there is no recognized phrase (null) to the doll 10 via the communication unit 201. Although the case has been described, the answer phrase data output last time with reference to the history storage unit 233 may be output again.

  Note that the speech recognition unit 223 can also calculate the reliability indicating the accuracy of recognition (the degree of probability) obtained as a speech content recognition result (speech recognition result). When the reliability is low, it may be determined that there is no recognition phrase. As a method for determining the reliability of the speech recognition result in the speech recognition unit 223, for example, a plurality of speech waveform models (acoustic models) indicating a predetermined word (phrase) and a waveform indicating speech data are prepared in advance. A determination method of determining the degree of coincidence and using the highest degree of coincidence as the reliability can be used. In addition, this invention is not limited to this, Other systems can also be used.

  Further, when the response process is executed from the doll 10 based on the answer phrase data from the server 20, it is assumed that the response process takes time. Therefore, it is possible to reduce the feeling of discomfort to the user and to give the doll 10 a sense of familiarity by uttering a voice such as “Utto” or performing other response processing. . That is, more natural communication can be achieved. Response processing such as uttering such connected voice may be executed at regular intervals. Such a response process may be a predetermined response, may be selected from several patterns, or may be selected at random. It is more preferable that the doll 10 performs the response process such as uttering the voice of the connection in terms of the response speed, but a method of executing it by the instruction of the server 20 can also be adopted. Specifically, in the sequence sq2 in the sequence diagram of FIG. 7, when the server 20 receives the voice data from the doll 10, the server 20 performs a response process such as uttering the voice of the connection. A configuration instructing the doll 10 may be adopted. Note that the following embodiments are also applicable.

(Moody parameter acquisition process)
FIG. 9 is a flowchart for executing the mood parameter acquisition process of the server 20 based on the first embodiment.

  With reference to FIG. 9, the flowchart is a process executed by the management unit 222 of the control unit 202 by executing a program stored in the storage unit 203.

First, an evaluation value for calculating the mood parameter P is calculated (step S10).
First, the mood parameter is expressed by one parameter. Here, as an example, it is expressed by a value between −10 and +10. That is, −10 is the state with the worst mood, and +10 is the state with the best mood. Since mood is considered to be influenced by many factors, as an example, the mood parameter is calculated as a linear combination of various factors.

  For example, as the factors, five of the latest conversation frequency, the latest conversation content, the latest sensor information, the current date and time, and external information are used. Assume that the evaluation values of the latest conversation frequency, the latest conversation content, the latest sensor information, the current date and time, and the external information are V1, V2, V3, V4, and V5.

Next, the mood parameter P is calculated based on the evaluation value (step S12).
In this example, the mood parameter P is calculated by the following equation.

Mood parameter P = K1 × V1 + K2 × V2 + K3 × V3 + K4 × V4 + K5 × V5
Here, K1 to K5 are weighting values of the respective factors, and the weighting values can be arbitrarily changed.

  Then, the mood parameter is updated based on the calculation result (step S14). The management unit 222 outputs the updated mood parameter P to the selection unit 224.

Then, the process ends (END).
A method for calculating the evaluation values V1 to V5 of the management unit 222 will be described.

  For the most recent conversation frequency evaluation value V1, first, the history storage unit 233 is referred to, and the most recent (for example, 72 hours) conversation count T1 recorded from the response ID is acquired.

Using the conversation count T1, an evaluation value V1 based on the conversation frequency is calculated by the following equation.
Evaluation value V1 = T1 / 10−10
The upper limit value of V1 is 10 and the lower limit value is −10. The same applies to the other evaluation values V2 to V5.

  That is, the evaluation value V1 = 10 when the number of conversations is 200 or more, and V1 = 0 when the number of conversations is zero.

  For the evaluation value V2 of the latest conversation content, the history storage unit 233 is referred to, and the recorded value (total value) of the latest mood change degree (for example, 24 hours) is acquired.

  It should be noted that the latest conversation type corresponding to the response ID is acquired by referring to the history storage unit 233, and minus if the same conversation type is continuous with reference to the change table database 231 of FIG. You may make it add with the previous total value as evaluation.

  For the evaluation value V3 of the latest sensor information, the latest sensor information recorded (for example, 24 hours) is acquired by referring to the history storage unit 233 and the sensor information database 234.

  A positive evaluation is performed for the number of strokes (type “touch”), and a positive evaluation is performed for the number of times that a smile is detected (type “smile”). Further, the discomfort index Q is calculated based on the detected temperature and humidity, the mood change degree is calculated with reference to the change table database 231 in FIG. 6B, and the total value of these values is obtained. To do.

  For the evaluation value V4 of the current date and time, the current date and time are evaluated based on a specific date / time zone.

  Referring to FIGS. 6C and 6D, the mood change degree is calculated based on the corresponding specific date / time zone.

  The external information evaluation value V5 is calculated based on the information acquired from the external device with reference to the history storage unit 233. In this example, referring to FIG. 6E, the mood change degree based on the weather information is calculated.

  The mood parameter P is calculated based on the calculation results of the evaluation values V1 to V5. As an example, if the weight values K1 to K5 are all 0.2, the mood parameter P can be calculated as an average.

  In addition, the calculation method of the mood parameter P is not limited to the above method, and various methods can be adopted. For example, in this example, the method of calculating the mood parameter P based on the evaluation values V1 to V5 has been described. However, the mood parameter P may be calculated based on any one of the evaluation values, and a combination thereof. Is also optional. In the above description, the method of linearly adding the evaluation values has been described as the method of calculating the mood parameter P. However, each evaluation value is multiplied, and the multiplied value is calculated as the value of the mood parameter. good.

  In this example, the configuration of the interactive system 1 in which the doll 10 and the server 20 operate in cooperation has been described. However, the interactive operation that includes the functions of the server 20 such as voice recognition in the doll 10 and operates in a stand-alone manner. You may make it implement | achieve an apparatus.

  Moreover, in this example, although the system which performs the response process based on the mood parameter P with respect to the response process with respect to the utterance from a user was demonstrated, it is not restricted to this in particular, it is not an utterance from a user but self-speaking Also for the automatic response process in a predetermined cycle, a response process based on the mood parameter P may be executed.

<Embodiment 2>
In said Embodiment 1, the system which changes an answer phrase based on a mood parameter was demonstrated.

  In the second embodiment, a method for changing a speech synthesis parameter will be described.

FIG. 10 is a diagram illustrating a speech synthesis parameter table based on the second embodiment.
Assume that the table is stored in the response content database 232 as an example.

  Referring to FIG. 10, a case where “moody range” is associated with “volume”, “pitch”, and “speech speed” is shown. In this example, “volume”, “pitch”, and “speech speed” are set higher as the mood parameter is higher, and “volume”, “pitch”, and “speech speed” are set lower as the mood parameter is lower. Yes.

  In this example, it is assumed that the reference value is “1.0”, the sound is smaller as the volume value is smaller, and the sound is larger as the value is larger. Also, the pitch is lower as the pitch value is smaller, and the pitch is higher as the pitch value is larger. Further, it is assumed that the smaller the speech speed value, the slower the reproduction speed, and the larger the value, the faster the reproduction speed.

  Specifically, when the mood range of the mood parameter P is less than −6, the case where the volume is set to “0.9”, the pitch “0.9”, and the speech speed “0.9” is shown. ing.

  Further, when the mood range of the mood parameter P is −6 or more and less than −2, the volume is set to “0.95”, the pitch “0.95”, and the speech speed “1.0”. It is shown.

  Further, when the mood range P of the mood parameter P is −2 or more and less than 2, the volume “1.0”, the pitch “1.0”, and the speech speed “1.0” are set. It is shown.

  Also, when the mood range of the mood parameter P is 2 or more and less than 6, the case where the volume is set to “1.05”, the pitch “1.1”, and the speech speed “1.0” is shown. Has been.

  Further, when the mood range of the mood parameter P is 6 or more, the case where the volume is set to “1.1”, the pitch “1.2”, and the speech speed “1.1” is shown.

  Accordingly, when the mood parameter P is low, a dark tone is reproduced, and when the mood parameter P is high, a bright tone is reproduced.

  In the second embodiment, the selection unit 224 of the server 20 outputs the selected answer phrase to the response process execution instructing unit 225 and refers to the speech synthesis parameter table of FIG. (Volume, pitch, speech speed data) is output.

  The response process execution instructing unit 225 outputs the speech synthesis parameters to the doll 10 via the communication unit 201 together with the answer phrase data selected (determined) by the selection unit 224.

  The response processing execution unit 112 of the doll 10 responds (voice dialogue) to the user via the speaker 104 based on the answer phrase data and the voice synthesis parameter received via the communication unit 201.

  With this processing, the answer phrase based on the mood parameter is reproduced and output from the doll 10 by the speech synthesis parameter based on the mood parameter and is uttered.

  Thereby, it is possible to achieve natural communication with the user by executing the response process (speech) according to the change in the mood parameter.

  In this example, the case where both the answer phrase and the speech synthesis parameter are selected based on the mood parameter has been described. However, the present invention is not limited to this, and only the speech synthesis parameter is selected and the same answer phrase is used. It is also possible to do.

<Embodiment 3>
In the above-described embodiment, the case where the voice is reproduced as the response content to the user's utterance has been described.

  FIG. 11 is a diagram illustrating the response content database 232A based on the third embodiment.

  Referring to FIG. 11, the response content database 232A is stored in the storage unit 203 included in the server 20 based on the third embodiment as an example.

  Specifically, a plurality of response information is registered in the response content database 232A. Specifically, the recognition content (recognition phrase) and the response content (answer phrase) are registered in association with each other. In this example, an identification number (response ID) is assigned to each combination of recognition phrase and answer phrase. As an example, it is assumed that the recognition phrase registered in the response content database 232 in this example is also registered in the dictionary used for speech recognition.

  As an example, here, as a recognition phrase, answer phrases corresponding to “good morning”... Are stored in association with each other.

  For example, the case where the response mode “operation pattern 1” is associated and registered corresponding to the recognition phrase “good morning” of the response ID “1” is shown.

  Further, the case where the response mode “operation pattern 2” is associated and registered in correspondence with the recognition phrase “good morning” of the response ID “2” is shown.

  Further, the case where the response mode “operation pattern 3” is associated and registered corresponding to the recognition phrase “good morning” of the response ID “3” is shown.

  Here, the response modes “motion pattern 1” to “motion pattern 3” mean that the server 20 instructs the doll 10 to execute a motion (movement process) of a predetermined motion pattern.

  For example, “motion pattern 1” means a pattern in which the doll 10 moves the hand for a predetermined period by the driving unit 106.

  “Operation pattern 2” means a pattern in which the drive unit 106 performs a movement of moving the doll's 10 foot for a predetermined period.

  “Operation pattern 3” means a pattern in which the driving unit 106 performs a movement of moving the head of the doll 10 for a predetermined period.

  And in this example, the mood change degree is matched regarding the response mode corresponding to each recognition phrase.

  Regarding the mood change degree, if the content of the recognition phrase is positive, the mood change degree is given so that the value of the mood parameter P increases. On the other hand, if the content of the recognition phrase is passive content, the mood change degree is given so that the value of the mood parameter P becomes low. As described above, the response process according to the change in the mood parameter is executed by the mood parameter calculation process based on the mood change degree or the like.

  And in this example, the mood range is matched with the answer phrase corresponding to each recognition phrase.

  The mood range is used as an index to be selected when there are a plurality of answer phrases for the same recognition phrase. In this example, an answer phrase is selected based on the mood parameter P managed by the management unit 222. It is determined which mood range the mood parameter P belongs to, and a corresponding answer phrase is selected.

  For example, when the mood parameter Q is larger than 5 as the mood range, the response ID “1” is selected. If the mood parameter Q is greater than −5 and less than or equal to 5 as the mood range, the response ID “2” is selected. When the mood parameter Q is −5 or less as the mood range, the response ID “3” is selected.

  It should be noted that an answer phrase included in the mood range is not necessarily selected, and the probability of being relatively selected may be increased. As a result, natural communication with the user can be achieved without becoming patterned answer contents.

<Embodiment 4>
In order to achieve natural communication with the user, it is possible to assign a character corresponding to each device instead of uniform audio reproduction.

FIG. 12 is a diagram illustrating a character database based on the fourth embodiment.
Referring to FIG. 12, it is assumed that the character database in this example is stored in storage unit 203.

  A plurality of pieces of character information are registered in the character database. Specifically, character information (name, speaker, volume, pitch, speaking speed, support team) associated with the character ID is associated and registered.

  Specifically, “name”, “speaker”, “volume”, “pitch”, “speaking speed”, and “support team” are associated and registered in correspondence with the character ID.

  As an example, when the character ID is “1”, the name “here”, the speaker “Woman1”, the volume “1.0”, the pitch “1.0”, the speech speed “1.0”, the support team “ The case where “Null” is registered is shown.

  When the character ID is “2”, the name “Hanako”, the speaker “Woman2”, the volume “1.0”, the pitch “1.05”, the speech speed “0.9”, and the support team “Null” The registered case is shown.

  When the character ID is “3”, the name “Genki”, the speaker “man1”, the volume “1.2”, the pitch “1.0”, the speech speed “1.1”, and the support team “Hanshin” The registered case is shown.

  When the character ID of the doll 10 whose device ID is “10” is assigned to “1”, the name “Coco”, speaker “Woman1”, volume “1.0”, pitch “1.0”, It is possible to interact with the user as the character characteristics of the speaking speed “1.0” and the support team “Null”.

  The name is used when introducing yourself. The speaker means voice pattern data which is a reference for voice synthesis processing. The speaker “Woman1” points to the female pattern data 1. The support team indicates a support team of a baseball team that has character personality.

In this example, since it is “Null”, there is nothing in particular.
On the other hand, when the support team is “Hanshin”, Hanshin is set as the support team.

  The server 20 acquires information on the result of winning or losing the baseball team at a predetermined timing via the communication unit 201. For example, the winning / losing result of the previous day is acquired, and the mood parameter is increased when the support team wins. Or, when the support team loses, the mood parameter is lowered.

  With this process, response processing (utterance) according to changes in the character's mood parameters is executed by giving each character a personality, rather than uniform audio reproduction. Natural communication is possible.

<Embodiment 5>
FIG. 13 is a diagram illustrating a configuration of a server based on the fifth embodiment.

  Referring to FIG. 13, in this example, a case where a plurality of servers are provided is shown.

  In this example, the case where the server 20A and the server 20B are provided is shown as an example.

  In the above configuration, the case where the voice recognition and the process for determining the answer phrase (response mode) for voice recognition are executed by the same server has been described. On the other hand, the process is executed by an independent server. Is also possible.

  Specifically, the server 20A may perform voice recognition on the voice data, and the server 20B may output answer phrase data to the doll 10.

  For example, voice data is transmitted from the doll 10 to the server 20A (1). The server 20A executes voice recognition of the voice data (2). Then, the server 20A transmits the recognition phrase to the doll 10 (3).

  The doll 10 receives the recognition phrase from the server 20A, and transmits the recognition phrase to another server 20B (4).

  The server 20B receives the recognition phrase from the doll 10 and determines an answer phrase corresponding to the recognition phrase (5). And server 20B transmits reply phrase data to a doll (6).

  In addition, in this example, although the server 20A demonstrated the case where the recognition phrase which performed the speech recognition of audio | voice data was transmitted with respect to the doll 10, if it is the information which shows not only the recognition phrase but the speech recognition result, which Something like that. For example, it may be access information (URL (Uniform Resource Locator), etc.) necessary for accessing an answer phrase stored in the server 20B. For example, the doll 10 may receive the answer information from the server 20B by receiving the access information (URL) from the server 20A and accessing the server 20B. Further, not limited to access information, when an answer phrase stored in the server 20B is saved in a file format, it is information for designating a file name as information indicating the result of speech recognition from the server 20A. May be. For example, the doll 10 receives the file name from the server 20A, requests the server 20B by specifying the file name, and requests information, thereby acquiring a file related to the answer phrase from the server 20B. It is.

  Similarly, text information obtained by converting a recognition phrase into text may be transmitted as information indicating the result of speech recognition from the server 20A. The doll 10 may extract the recognition phrase from the text information and access the server 20B to obtain the answer phrase, or transmit the text information to the server 20B, and the server 20B determines the recognition phrase. It is good also as a structure which analyzes the text information to include, determines an answer phrase based on an analysis result, and transmits to the doll 10.

  Moreover, the structure which transmits reply phrase data to the doll 10 from the server 20B was demonstrated. Specifically, the case where the voice file which is the answer phrase data is transmitted and the doll 10 speaks according to the voice file has been described. However, the doll 10 is not limited to the voice file, but the text information is transmitted as the answer phrase data, and the doll 10 Then, the text information may be analyzed (so-called reading function or the like) and uttered (response processing).

  Moreover, in this example, although the case where the speech recognition was performed by the server 20 was described, the speech recognition is performed by the doll 10, the answer phrase for the result is determined in the doll 10, and the answer phrase is acquired from the server 20B. May be. In this case, it can be realized by providing the storage unit 109 with a URL correspondence table in which access information (URL) for accessing the reply phrase of the server 20B for the recognized phrase is associated.

  It is also possible to acquire an answer phrase for the result of speech recognition using information stored in the doll 10.

  For example, when the cache memory capable of temporarily storing information includes information on the answer phrase for the previously used recognition phrase, the answer phrase information stored in the cache memory is By using it, for example, it is possible to acquire an answer phrase and access the doll 10 (response processing) without accessing the server 20B. As a result, it is possible to utter early using the information stored in the cache memory.

  Further, when an audio file that is an answer phrase is stored in the doll 10, the server 20A may specify the audio file stored in the doll 10 as information indicating the result of the voice recognition. . Through this process, it is possible to speak early using an audio file stored in the doll 10 without accessing the server 20B. If the audio file is not stored in the doll 10, the designated audio file may be requested to the server 20B, and the audio file may be acquired from the server 20B and spoken. .

  The server configuration of the fifth embodiment can be applied to any of the first to fourth embodiments.

<Embodiment 6>
Control blocks such as the doll 10 and the server 20 may be realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like, or may be realized by software using a CPU (Central Processing Unit). Also good.

  In the latter case, the doll 10 and the server 20 include a CPU that executes instructions of a program that is software that realizes each function, and a ROM (Read Only Memory) in which the program and various data are recorded so as to be readable by a computer (or CPU). ) Or a storage device (these are referred to as “recording media”), a RAM (Random Access Memory) for expanding the program, and the like. And the objective of this invention is achieved when a computer (or CPU) reads the said program from the said recording medium and runs it. As the recording medium, a “non-temporary tangible medium” such as a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used. The program may be supplied to the computer via an arbitrary transmission medium (such as a communication network or a broadcast wave) that can transmit the program. The present invention can also be realized in the form of a data signal embedded in a carrier wave in which the program is embodied by electronic transmission.

(Summary)
The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is shown not by the above description of the embodiments but by the scope of claims for patent, and is intended to include meanings equivalent to the scope of claims for patent and all modifications within the scope.

  1 Dialog system, 5 network, 10 doll, 20, 20A, 20B server, 101, 201 communication unit, 102, 202 control unit, 103 microphone, 104 speaker, 106 drive unit, 107 operation unit, 109, 203 storage unit, 112 Response processing execution unit, 114 voice input reception unit, 221 voice input reception unit, 222 management unit, 223 voice recognition unit, 224 selection unit, 225 response processing execution unit, 231 change table database, 232 response content database, 233 history storage unit 234 Sensor information database.

Claims (13)

A server provided to be able to communicate with a dialog device,
A receiving unit for receiving voice input from a user received by the interactive device;
A parameter management unit that manages mood parameters indicating emotions of the device;
A storage unit that stores a plurality of response information relating to responses to voice input from the user received by the reception unit, each of which is associated with the value of the mood parameter as an index when selected,
A selection unit that refers to the storage unit and selects one of the plurality of response information based on the mood parameter managed by the parameter management unit;
A response process execution instructing unit that instructs the dialogue apparatus to execute a response process based on the response information selected by the selection unit;
The storage unit stores change information of the mood parameter provided corresponding to each of the plurality of response information,
The server, wherein the parameter management unit updates the mood parameter based on change information of the mood parameter corresponding to one of the plurality of response information selected by the selection unit.   The server according to claim 1, wherein the parameter management unit updates the mood parameter based on information acquired from the outside.   The server according to claim 2, wherein the parameter management unit updates the mood parameter based on environmental information acquired from the outside. The server according to claim 2 , wherein the parameter management unit updates the mood parameter when information acquired from the outside satisfies a predetermined condition.   The server according to claim 1, wherein the parameter management unit updates the mood parameter based on an input mode received from the user by the receiving unit.   The server according to claim 5, wherein the parameter management unit updates the mood parameter based on voice input content from the user received by the receiving unit.   The server according to claim 5, wherein the parameter management unit updates the mood parameter based on sensor input information detected by the user received by the receiving unit. The server further includes a history storage unit that stores a history of input from the user,
The server according to claim 1, wherein the parameter management unit refers to the history storage unit and updates the mood parameter based on a plurality of input modes from the user. The response process execution instructing unit instructs the dialogue apparatus to execute a speech synthesis process,
The server according to claim 1, wherein each of the plurality of response information includes different parameter information when executing the speech synthesis process. The response process execution instructing unit instructs the dialogue apparatus to execute a voice output process,
The server according to claim 1, wherein each of the plurality of response information includes different output contents when executing the voice output process. The storage unit further stores other response information related to the voice input from the user received by the receiving unit, which is not associated with the value of the mood parameter,
The server according to claim 1, wherein the selection unit selects the other response information when a predetermined condition is satisfied. A method of controlling a server provided to be able to communicate with a dialog device,
Receiving input from a user accepted by the interactive device;
Managing mood parameters indicating the emotion of the device;
Based on the mood parameters, each of which is associated with the value of the mood parameter as an index at the time of selection, and refers to a storage unit that stores a plurality of response information regarding responses to the input from the received user. Selecting one of the plurality of response information;
Instructing the interactive device to perform response processing based on selected response information, and
The storage unit stores change information of the mood parameters provided corresponding to the plurality of response information, respectively.
The step of managing the mood parameter includes a step of updating the mood parameter based on change information of the mood parameter corresponding to one of the selected plurality of response information. A server control program executed in a computer of a server provided so as to be communicable with a dialog device,
The server control program is for the computer.
Receiving input from a user accepted by the interactive device;
Managing mood parameters indicating the emotion of the device;
Based on the mood parameters, each of which is associated with the value of the mood parameter as an index at the time of selection, and refers to a storage unit that stores a plurality of response information regarding responses to the input from the received user. Selecting one of the plurality of response information;
Instructing the interactive device to perform response processing based on selected response information, and
The storage unit stores change information of the mood parameters provided corresponding to the plurality of response information, respectively.
The step of managing the mood parameter includes a step of updating the mood parameter based on change information of the mood parameter corresponding to one of the selected plurality of response information.

Images