JP2005062240A - Voice response system - Google Patents

Abstract

An object of the present invention is to provide a response required by a user who is born from a situation where the user is placed.
A voice response system for providing a voice response service to a user, a voice input unit for generating voice information from the user voice, and a situation for specifying a current situation where the user is placed A situation in which a specific unit, an emotion analysis unit that estimates a user's current emotion, a voice recognition unit that performs voice recognition processing on voice information generated by the voice input unit, and a user are placed A storage unit storing user characteristic information including personal characteristic information in which a user's emotion and a response pattern are associated with each other, and a current situation in which the user specified by the situation specifying unit is placed A response pattern corresponding to the user's current emotion estimated by the emotion analysis unit, a response pattern determination unit acquiring the response pattern from the personal characteristic information, and a response pattern acquired by the response pattern determination unit Wherein comprises a response processing unit that acquires a spoken sentence that corresponds to the recognition result by the voice recognition unit, and a sound output section which reads out a spoken sentence acquired by the response processing unit.
[Selection] Figure 6

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to the field of voice response services using voice.
[0002]
[Prior art]
In recent years, voice response services have been used in various fields. In this service, the system provides various services to the user by the user interacting with the system having the voice recognition technology (see, for example, FIG. 26). Specifically, for the digital divide generation, a personal computer that can be easily operated by voice without using a keyboard or mouse, and a voice portal that allows users to access and receive services from a personal computer or telephone. Etc.
[0003]
In general, the voice response service is provided in accordance with a scenario of the system.
The system recognizes the voice uttered by the user, compares the recognition result with the scenario, generates or selects the next utterance sentence, and responds.
[0004]
In recent years, it is also possible to analyze emotions from the state of voice input by the user and change the response according to the analysis result (see Patent Document 1).
[0005]
In addition, a dialog control method is disclosed that can be changed according to the user's system usage frequency, and can be adapted to the user's preference regarding dialog progress by having personal information of the user (Patent Document) 2).
[0006]
[Patent Document 1]
Japanese Patent Laid-Open No. 10-55194
[Patent Document 2]
JP 2002-99404 A
[0007]
[Problems to be solved by the invention]
However, when changing the voice response using only the user's system usage frequency and personal information, the situation (usage location, usage means, usage time, etc.) of the changing user is always known. Therefore, there is a problem that it is not possible to flexibly provide a response that a user is born from the situation where the user is placed. That is, in order to perform an appropriate response, it is difficult only to grasp the user's emotions, and it is difficult to use only the user's system frequency and personal information.
[0008]
In the prior art, attention is paid to the voice input of the user, and the user's emotion is estimated only from the voice input state (the pitch and volume of the user's voice), and the response is determined. Therefore, for example, it is not possible to distinguish between a case where an originally fast-speaking person speaks slowly, a case where an originally slowly-speaking person speaks quickly and a case where a normal person speaks at a normal speed. In this case, it seems reasonable to assume that people who are quick-hearted are relaxed, but that people who are slowly speaking originally are impatient. If these differences are not taken into account, there is a problem that an incorrect state can be responded to the user because it leads to grasping the wrong state.
[0009]
A first object of the present invention is to provide a response required by a user born from a situation where the user is placed. The second problem of the present invention is to give an appropriate response to the user in consideration of the characteristics of the user.
[0010]
[Means for Solving the Problems]
In order to solve the above-described problems, the present invention is a system that provides a voice response service to a user, and a voice input unit that generates voice information from the user voice and a user are placed at present. A situation identifying unit that identifies the situation of the user, an emotion analyzing unit that estimates a user's current emotion, a speech recognition unit that performs speech recognition processing on the speech information generated by the speech input unit, and a user For each situation where user characteristics information including personal characteristic information in which a user's emotion is associated with a response pattern is stored, and a user specified by the situation specifying unit is stored. A response pattern determination unit that acquires from the personal characteristic information, a response pattern determination unit that acquires a response pattern that corresponds to the current situation of the user and the user's current emotion estimated by the emotion analysis unit. A response processing unit that acquires a spoken sentence that the answering pattern corresponding to a recognition result by the voice recognition unit has, and a configuration and an audio output unit which reads out the utterance acquired by the response processing unit.
[0011]
According to the present invention, the uttered sentence read out by the voice output unit is determined by the response pattern (determined by the current situation where the user is placed and the user's current emotion) and the recognition result by the voice recognition unit. . Therefore, according to the present invention, it is possible to provide a response required by a user born from a situation where the user is placed.
[0012]
For example, if the current situation where the user is located is “home”, the user wants a “slow and calm response service” and the current situation where the user is located is “company”. In this case, the user can respond to the response service requested by the user that he / she wants a “quickly and speedy response service”. Note that the current situation where the user is placed can be specified (understood) by using, for example, a transmission number or by hearing from the user at an early stage of the voice response service.
[0013]
The voice response system further includes, for example, an update unit that updates a correspondence relationship between the user's emotion and a response pattern.
[0014]
In this way, the next voice response service for the user can be further optimized. For example, it is conceivable to update the correspondence between the two based on the response history of the user from the start to the end of the voice response service.
[0015]
In the voice response system, for example, the emotion analysis unit estimates a user's current emotion by performing a comparative analysis using personal voice characteristic information stored in advance.
[0016]
In this way, it becomes possible to grasp the user's emotions in consideration of the user's personal voice characteristics (sound characteristics such as volume level and voice pitch), which is more appropriate for the user. It becomes possible to respond.
[0017]
In the voice response system, for example, the user characteristic information based on the user characteristic information, the voice information generated by the voice input unit, and the response history of the user from the start to the end of the voice response service. Is further provided with a user analysis unit for adding and updating.
[0018]
In this way, the next voice response service for the user can be further optimized.
[0019]
In the voice response system, for example, the response history includes timing information on the start and end of voice output by the voice response service and voice input of the user, and the user characteristic information is voice response in the past of the user. It includes the timing information of the voice output by voice response service when using the service and the start / end of the voice input by the user.
[0020]
In this way, the next voice response service for the user can be further optimized.
[0021]
Further, the present invention can be specified as follows.
[0022]
A voice response system that provides a voice response service to a user is a method of providing a voice response service that responds to a user's emotions inferred from the user's voice information. The user characteristics information estimated from the response history including changes in the user's emotions from the start to the end of the voice response service, and the user characteristics information A method for providing a voice response service in which the order of responses to users and the content of responses are changed based on emotion information that takes into account the user's situation.
[0023]
In this way, it is possible to respond flexibly to the occasion according to the situation where the user is placed and the emotion of the user at that time.
[0024]
Further, the present invention can be specified as follows.
[0025]
A voice response system that provides a voice response service to a user. The voice input unit generates voice information from the user's voice, and generates user characteristic information by estimating user characteristic information of the user. Identifies user's words from voice analysis, user analysis section, emotion analysis section that guesses user's emotion and generates emotion information, scenario storage section that stores sentences necessary to create spoken sentences A voice response system comprising: a response processing unit that creates an utterance sentence from the scenario storage unit based on the user characteristic information and the emotion information; and a voice output unit that reads the utterance sentence.
[0026]
In this way, the order of responses and the content of responses can be changed according to the situation where the user is placed and the emotion of the user at that time.
[0027]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, a voice response system according to an embodiment of the present invention will be described.
[0028]
(Principle of voice response system)
[0029]
1-4 is a figure for demonstrating the principle of the voice response system of this embodiment.
[0030]
(Outline system configuration of voice response system)
[0031]
As shown in FIG. 1, the voice response system 10 includes a voice input unit 110, a voice output unit 120, a response processing unit 130, a user authentication unit 140, an emotion analysis unit 150, a user analysis unit 160, a scenario storage unit 410, A response history storage unit 420, a user characteristic information storage unit 430, and the like are provided.
[0032]
(Outline operation of voice response system)
[0033]
The operation of the voice response system according to this embodiment will be described with reference to FIGS. 2, 3, and 4. FIG.
[0034]
FIG. 2 is a flowchart for explaining the operation of the voice response system of this embodiment. FIG. 3 is an operation principle diagram of the voice response system of the present embodiment. This also shows the data flow. FIG. 4 is a flowchart for explaining the operation of the user analysis unit.
[0035]
<User authentication (S100)>
[0036]
The user authentication unit 140 identifies the user of this system and the situation where the user is placed. For example, when the user accesses the system by telephone, this identification can be realized by using other methods that can identify the user, such as a calling number notification function and a voice authentication technique. The specified content is passed to the response processing unit 130 as user information.
<User characteristic information setting (S110)>
[0037]
The response processing unit 130 acquires the personal characteristic information of the user corresponding to the situation where the user is placed from the user characteristic information storage unit 430 based on the user information. In addition, the response processing unit 130 passes the user information to the emotion analysis unit 150.
[0038]
Based on the user information passed from the response processing unit 130, the emotion analysis unit 150 obtains the user's personal voice characteristic information corresponding to the situation where the user is placed from the user characteristic information storage unit 430. get.
<Voice input (S120)>
[0039]
The voice input unit 110 generates voice information from the user voice and passes it to the response processing unit 130 and the emotion analysis unit 150.
[0040]
<Emotion analysis (S130)>
[0041]
The emotion analysis unit 150 uses the personal voice characteristic information acquired from the user characteristic information storage unit 430 and performs comparison analysis with the voice characteristic information extracted from the voice information passed from the voice input unit 110. Thereby, the situation where the user is placed is taken into consideration, the emotion is estimated, and emotion information is created. The emotion information is passed to the response processing unit 130.
[0042]
<Voice recognition (S140)>
[0043]
The response processing unit 130 performs voice recognition processing on the voice information passed from the voice input unit 110. That is, the words uttered by the user are identified. Thereby, the result (identification information) of the voice recognition process is acquired.
[0044]
<Spoken sentence creation (S150)>
[0045]
The response processing unit 130 creates a scenario based on the emotion information passed from the emotion analysis unit 150, the personal characteristic information acquired from the user characteristic information storage unit 430, and the identification information acquired from the user's voice information. The utterance sentence with the optimal response content is selected from the storage unit 410 in the optimal response order. The selected utterance sentence is passed to the voice output unit 120.
[0046]
<Audio output (S160)>
[0047]
The voice output unit 120 notifies the user of the utterance sentence delivered from the response processing unit 130 by some method of transmitting a message to the user such as reading out by a voice synthesis technique.
[0048]
Note that the processing from voice input (S120) to voice output (S160) is repeated until the voice response service ends.
[0049]
<Response history registration (S180)>
[0050]
When the voice response service of the user and the voice response system 10 ends (S170: Yes), the response processing unit 130 determines the situation where the user is placed (use place, use time, use means, etc.), how the user speaks. Response history information including changes in the user's emotion between the sound response (volume information, pitch information, utterance timing information, etc.) is passed (stored) to the response history storage unit 420.
[0051]
<User analysis (S190)>
[0052]
As shown in FIG. 4, when the response history information is passed to the response history storage unit 420, the user analysis unit 160 acquires the response history information (S200). Further, the user analysis unit 160 requests user characteristic information (S210), and if there is user characteristic information of the user (S220: Yes), the user characteristic information storage unit 430 sends the user characteristic information. User characteristic information is acquired (S230). Based on the information, the user characteristic information of the user is analyzed (or estimated) again (S240) and created (S250).
[0053]
On the other hand, when there is no user characteristic information of the user (S260), the user analysis unit 160 acquires general user characteristic information (S260). And the user characteristic information of the user is newly created.
[0054]
<User characteristic information registration (S200)>
[0055]
The user analysis unit 160 passes the user characteristic information created in S250 to the user characteristic information storage unit 430. Thereby, user characteristic information is added or updated (S260). The user characteristic information created here is used when the user uses the voice response service next time.
[0056]
As described above, in the present embodiment, the order of responses and the content of responses can be changed according to the situation where the user is placed and the emotion of the user at that time.
Therefore, it is possible to provide a voice response service that can be flexibly changed.
[0057]
(Example)
[0058]
Hereinafter, an example in which the present invention is applied to a voice response system for hotel reservation will be described.
[0059]
(Outline of voice response system)
[0060]
FIG. 5 is a diagram for explaining a schematic system configuration of a voice response system for hotel reservation (hereinafter simply referred to as a voice response system).
[0061]
The voice response system 10 is connected to a public telephone network 40, and when a user receives an incoming call from the telephone 30, a voice response service for hotel reservation (reading of guidance voice) is given to the user. Etc.).
[0062]
This voice response service is performed in consideration of the current situation where the user is placed and the user's current emotion. In order to perform this voice response service, the voice response system 10 includes a user characteristic information storage unit 430 described later.
[0063]
Further, in order to further optimize the next voice response service for the same user, the stored contents of the user characteristic information storage unit 430 are updated.
[0064]
(Outline system configuration of voice response system)
[0065]
As shown in FIG. 6, the voice response system 10 includes a voice input unit 110, a voice output unit 120, a response processing unit 130, a user authentication unit 140, an emotion analysis unit 150, a user analysis unit 160, a scenario storage unit 410, A response history storage unit 420, a user characteristic information storage unit 430, and the like are provided.
[0066]
The voice input unit 110 is for generating voice information from user voice (voice uttered by the user) input via the telephone 30.
[0067]
The voice output unit 120 is for reading out an utterance sentence (response sentence) or the like acquired (selected or generated) by the response processing unit 130 using an existing voice synthesis technique. The utterance sentence read out by the voice output unit 120 is notified to the user via the telephone 30.
[0068]
The response processing unit 130 is for performing various processes shown in FIG.
[0069]
The user authentication unit 140 is for specifying (or authenticating) the user and the situation where the user is placed when a call is made from the telephone 30. These authentications can be specified by, for example, comparing the caller number notified by the caller number notification function with the user database.
[0070]
The emotion analysis unit 150 is for estimating the current emotion of the user.
[0071]
The user analysis unit 160 adds and updates the user characteristic information based on the user characteristic information, the voice information generated by the voice input unit 120, and the response history of the user from the start to the end of the voice response service. Is to do.
[0072]
In the scenario storage unit 410, the service flow and utterance (scenario) of the voice response system 10 are created and stored for each scenario flow and user emotion. The scenario may be accumulated as a table for each user's emotion.
[0073]
The response history storage unit 420 is for storing a user's response history from the start to the end of the voice response service. The response history includes the situation in which the user is placed (use location, use time, use method, etc.), how the user speaks (volume information, pitch information, utterance timing information, etc.), There are emotional changes, voice output of the voice response system 10 and timing information of start / end of voice input by the user.
[0074]
The user characteristic information storage unit 430 stores user characteristic information (FIG. 10) including personal characteristic information in which a user's emotion is associated with a response pattern for each situation where the user is placed. Is.
[0075]
(Operation of voice response system)
[0076]
The operation of the voice response system configured as described above will be described with reference to the drawings.
FIG. 7 is a flowchart for explaining the operation of the voice response system according to this embodiment.
[0077]
<Service start (S300)>
[0078]
The voice response system 10 starts the following voice response service for hotel reservation when the user A receives an incoming call from the telephone 30.
[0079]
<User authentication (S301)>
[0080]
When there is an incoming call from the telephone 30 to the user A, the voice response system 10 uses the user authentication unit 140 to identify (recognize) the user A and the situation where the user A is placed. The specific result is passed to the response processing unit 130 as user information ((1) in FIG. 6).
[0081]
Here, the voice response system 10 reads out “Thank you for using this. This is a hotel reservation system. Are you sure you want your name?” By the voice output unit 120 for user authentication. The uttered sentence read out is notified to the user A via the telephone 30. On the other hand, when the user A responds “user A”, the voice response system 10 identifies (or recognizes) the user name “user A” by executing voice recognition processing or the like. ) Further, the voice response system 10 specifies (or recognizes) the current situation in which the user A is placed. This specification is performed, for example, by collating a caller telephone number with a user database (a telephone number is associated with a usage means, a usage location, etc.).
Here, it is assumed that, as the current situation where the user A is placed, the usage means “landline phone” and the usage location “place” (location where the phone is called) are specified (or recognized). .
[0082]
The user authentication unit 140 passes these specific results (“user A”, “company”, “fixed phone”) to the response processing unit 130 as user information (FIG. 8) ((1) in FIG. 6). .
[0083]
<User characteristic information setting (S302)>
[0084]
Based on the user information, the response processing unit 130 changes from the user characteristic information storage unit 430 (FIG. 9) to the current situation (“company”, “fixed phone”) in which the user A is placed. Corresponding personal characteristic information (FIG. 10) is acquired ((2) in FIG. 6). This personal characteristic information is stored (or set) in a predetermined storage unit such as its own memory.
[0085]
The response processing unit 130 passes the user information to the emotion analysis unit 150 ((3) in FIG. 6). The emotion analysis unit 150 determines the current situation (“company”, “fixed”) from the user characteristic information storage unit 430 based on the user information passed from the response processing unit 130. Personal voice characteristic information (FIG. 11) corresponding to “telephone”) is acquired ((4) in FIG. 6). This personal voice characteristic information is stored (or set) in a predetermined storage unit such as its own memory.
[0086]
<Voice input (S303)>
[0087]
When the user A starts speaking, the voice response system 10 creates voice information (FIG. 12) from the voice of the user A input via the telephone 30 by the voice input unit 110. This voice information is passed to the response processing unit 130 and the emotion analysis unit 150 (FIG. 6 (5) (6)).
[0088]
<Emotion analysis (S304)>
[0089]
The emotion analysis unit 150 extracts “volume level: 6” and “sound pitch: 8” from the voice characteristic information (FIG. 13) extracted from the voice information passed from the voice input unit 110.
[0090]
Next, the emotion analysis unit 150 estimates the current emotion of the user A. For this estimation, the personal voice characteristic information of user A (FIG. 11) and the current voice characteristic information of user A (“volume level: 6”, “voice pitch: 8”) are compared and analyzed. Referring to the personal voice characteristic information of user A (FIG. 12), “normal” corresponds to the current voice characteristic information of user A (“volume level: 6”, “voice pitch: 8”). ing. Therefore, the emotion analysis unit 150 estimates (or analyzes) that the current emotion of the user A is “normal”.
[0091]
The emotion analysis unit 150 also estimates the current response characteristics of the user A. For this estimation, it is determined whether or not the speech time zone of the voice response system 10 and the user A's speech time zone overlap. Here, referring to the voice information of user A (FIG. 12), the speech time zone of voice response system 10 and the speech time zone of user A do not overlap. Therefore, the emotion analysis unit 150 estimates (or analyzes) that the response characteristic of the user A is “normal”.
[0092]
The results (“current emotion: normal”, “response characteristics: normal”) estimated (or analyzed) as described above are passed to the response processing unit 130 as emotion information (FIG. 14) (FIG. 6 (7)). ).
[0093]
<Voice recognition>
[0094]
The response processing unit 130 performs voice recognition processing on the voice information passed from the voice input unit 110. Here, as a result of the voice recognition, the response processing unit 130 identifies the words spoken by the user A, and acquires the identification information “I would like to reserve Hotel Onda on September 30” (FIG. 15). Shall.
[0095]
<Spoken sentence creation>
[0096]
Based on the emotion information “current emotion: normal” and the voice recognition result “identification information: I would like to reserve Hotel Onda on September 30”, the response processing unit 130 makes a scenario. From the accumulation unit 410, a scenario (FIG. 16) in which the emotion is normal (FIG. 16) matching the current scenario flow (in this case, step 3-1) is selected.
[0097]
Next, the response processing unit 130 obtains the current emotion (“normal”) of the user A from the personal characteristic information corresponding to the current situation (“company”, “fixed phone”) in which the user A is placed. Get the response pattern corresponding to.
[0098]
Here, personal characteristic information corresponding to the current situation (“company”, “landline telephone”) in which user A is placed (this is obtained in FIG. 6 (2) and stored in its own memory or the like. Referring to FIG. 10, the response pattern corresponding to the current emotion (“normal”) of user A is “normal response”.
[0099]
Therefore, the response processing unit 130 acquires “normal response” as a response pattern corresponding to the current emotion (“normal”) of the user A.
[0100]
For this reason, the response processing unit 130 finally “sorry. Full of rooms corresponding to the response pattern“ normal response ”from the selected scenario (FIG. 16). "Is acquired (selected). This spoken sentence is passed to the voice output unit 120.
[0101]
On the other hand, in the emotion analysis (S304), it is assumed that the utterance time zone of the voice response system 10 and the utterance time zone of the user A overlap as shown in FIG. In this case, the emotion analysis unit 150 estimates (or analyzes) that the response characteristic of the user A is “impatient”. When the response characteristic of the user A is estimated (or analyzed) as “impatient”, the response processing unit 130 finally determines the selected scenario (FIG. 1) regardless of the response pattern “ordinary correspondence”. 16) Select “I'm sorry. No, you are in a hurry?” Corresponding to the response pattern “Speedy response” (FIG. 18).
[0102]
<Audio output (S307)>
[0103]
The voice output unit 120 reads out the utterance sentence delivered from the response processing unit 130 using a voice synthesis technique. The uttered sentence read out is notified (notified) to the user A via the telephone 30.
[0104]
The processing from the voice input (S303) to the voice output (S307) is repeated until the voice response service is completed.
[0105]
<Response history registration (S309)>
[0106]
When the voice response service of the voice response system 10 for the user A is finished, the response processing unit 130 displays the situation where the user A is placed (use location, use time, use means, etc.), how the user speaks (volume information , Pitch information, utterance timing information, etc.), response history information including timing information of the user's emotion change during the voice response, voice output of the voice response system 10 and the start / end timing of voice input of the user A 19) to the response history storage unit 420.
[0107]
<User analysis (S310)>
[0108]
When the response history information is passed to the response history storage unit 420, the user analysis unit 160 acquires the response history information. Moreover, if there exists user characteristic information of the user A from the user characteristic information storage part 430, the user analysis part 160 will acquire this (FIG. 9).
[0109]
Next, the user analysis unit 160 updates the user characteristic information of the user A. For this update, the user analysis unit 160 compares and analyzes the response history information and the user characteristic information. Referring to the response history information of the user A (FIG. 19), the emotion of the user A is normal at first, but is disgusting at the end of the voice response service. It can be inferred that this was due to some reason for making the user A unhappy during the voice response service. Further, comparing the timing of voice output and voice input, it can be seen that voice input is performed during voice output from the start to the end of the voice response service.
It is considered that this is because user A felt disgust because of the long voice output message. That is, when the emotion is “normal”, it can be understood that the response pattern desired by the user A is “speedy response” (rule 2 described later).
[0110]
Therefore, the user analysis unit 160 corrects (recreates) the response pattern “normal response” in the user characteristic information (FIG. 10) of the user A to “speedy response” (FIG. 20).
[0111]
On the other hand, when there is no user characteristic information of the user A in the user characteristic information storage unit 430, the user analysis unit 160 acquires general user characteristic information (FIG. 21), and the same as above. Correction is performed and user characteristic information of the user A is newly created.
[0112]
FIG. 22 shows an example of a table used by the user analysis unit 160 for analysis. This is a knowledge base in which rules of the “IF to THEN” format are collected. The figure illustrates three rules. A line starting with # is a comment.
[0113]
Rule 1 is a rule applied to a new user. That is, when the number of uses is 0 (in the above description, when there is no user characteristic information of the user A in the user characteristic information storage unit 430), various parameters such as a volume level are set.
[0114]
Rule 2 is that when the user's emotion changes from “normal” to “disgust”, the user responds during voice output, and the response pattern is “ordinary reception”, the response pattern “normal” Correct "Response" to "Speedy response". This has already been explained.
[0115]
Rule 3 is a rule for performing the reverse correction to rule 2. That is, if there is no change in emotion, the user does not respond during voice output, and the response pattern is “ordinary reception”, the response pattern “ordinary reception” is corrected to “polite reception”. .
[0116]
<User characteristic information registration (S311)>
[0117]
The user analysis unit 160 adds or updates the user characteristic information created in S310 to the user characteristic information storage unit 430. This added or updated user characteristic information is used when the user uses the voice response service next time.
[0118]
In other words, if the situation where the user A is placed is “company” or “landline phone” next time and the emotion of the user A is “ordinary”, the modified response pattern “speedy reception” ”Will be handled.
[0119]
As a result, the next voice response service for the same user A can be further optimized.
[0120]
Further, the voice response service by the voice response system 10 can be performed in consideration of the current situation where the user A is placed and the current emotion of the user A.
Note that the user characteristic information (FIG. 9) described in the present embodiment is merely an example, and instead of this, for example, user characteristic information defined as shown in FIG. 23 may be used. In the user characteristic information, the sound level and the sound pitch are defined by the width (for example, 0 to 4 or 0 to 7). For this reason, it becomes possible to guess a user's emotion more appropriately.
When the user characteristic information of FIG. 22 is used, this user characteristic information is corrected (recreated) as shown in FIG. 24, for example. In this case, for example, the general user characteristic information shown in FIG. 25 is used.
[0121]
The present invention can be implemented in various other forms without departing from the spirit or main features thereof. For this reason, said embodiment is only a mere illustration in all points, and is not interpreted limitedly.
[0122]
The present invention can also be specified as follows.
(Supplementary note 1) A voice response system that provides voice response services to users, a voice input unit that generates voice information from user voices, and a situation that identifies the current situation in which the user is placed A situation where a specific unit, an emotion analysis unit that estimates a user's current emotion, a voice recognition unit that performs voice recognition processing on voice information generated by the voice input unit, and a user are placed A storage unit storing user characteristic information including personal characteristic information in which a user's emotion and a response pattern are associated with each other, and a current situation in which the user specified by the situation specifying unit is placed A response pattern corresponding to the user's current emotion estimated by the emotion analysis unit, a response pattern determination unit that acquires from the personal characteristic information, and a response pattern acquired by the response pattern determination unit A voice response system comprising: a response processing unit that acquires an utterance sentence corresponding to a recognition result by the voice recognition unit; and a voice output unit that reads out the utterance sentence acquired by the response processing unit.
(Additional remark 2) The voice response system of Additional remark 1 further provided with the update part which updates the correspondence of the said user's emotion and a response pattern.
(Supplementary note 3) The voice response system according to supplementary note 1, wherein the emotion analysis unit estimates a user's current emotion by performing comparative analysis using personal voice characteristic information stored in advance.
(Additional remark 4) Based on the said user characteristic information, the audio | voice information produced | generated by the said audio | voice input part, and the user's response history from the start to the end of the said voice response service, user characteristic information is added and updated The voice response system according to attachment 3, further comprising a user analysis unit.
(Additional remark 5) The said response history contains the timing information of the audio | voice output by the said voice response service, and the start / end of a user's voice input, and user characteristic information is when the voice response service was used in the past of the user. The voice response system according to appendix 4, including timing information of voice output by the voice response service and start / end timing of voice input by the user.
(Supplementary note 6) A voice response method using a voice response system that provides voice response services to users, generating voice information from user voices, and specifying the current situation in which the user is placed Performing a step of estimating a user's current emotion, performing a speech recognition process on the generated speech information, a current situation in which the identified user is placed, and the step Obtaining a response pattern corresponding to the inferred user's current emotion from personal characteristic information in which the user's emotion and response pattern are associated for each situation in which the user is placed; and A voice response method comprising: obtaining an utterance sentence corresponding to the received response pattern and the voice recognition result; and reading the acquired utterance sentence.
(Supplementary note 7) A voice response system that provides a voice response service to a user is a method of providing a voice response service that performs a response according to a user's emotion estimated from the voice information of the user, User characteristics information inferred from response history including changes in user's emotions from the start to the end of the voice response service, the user's situation, the user's situation, and the user characteristics A voice response service providing method for changing a response order and response contents to a user based on emotion information in consideration of a situation of the user obtained based on information.
(Supplementary note 8) A voice response system that provides a voice response service to a user, including a voice input unit that generates voice information from the user's voice, and a user characteristic by estimating user characteristic information of the user A user analysis unit that generates information, an emotion analysis unit that estimates emotions of users and generates emotion information, a scenario storage unit that stores sentences required to create spoken sentences, and a user based on voice information A voice response comprising: a response processing unit that creates a spoken sentence from the scenario storage unit based on the user characteristic information and the emotion information; and a voice output unit that reads the spoken sentence system.
[0123]
【The invention's effect】
As described above, according to the present invention, since it is possible to respond according to emotion information obtained by taking into account user characteristic information and characteristic information, the following effects are obtained.
[0124]
First, it is possible to grasp the situation where the user is placed and respond flexibly according to the situation where the user is placed. Second, it is possible to grasp the user's emotions without being influenced by the user's way of speaking and to respond flexibly according to the emotions.
[0125]
It is also possible to respond flexibly according to the situation where the user is placed and the emotion of the user at that time. In addition, the user's emotion can be estimated more accurately by taking into account changes in user characteristics depending on the situation where the user is placed. Furthermore, it is possible to provide a voice response service that is nifty for the user.
These lead to improvement in customer satisfaction of the service, acquisition of repeaters, and increase in service users (customers).
[Brief description of the drawings]
FIG. 1 is a diagram for explaining the principle of a voice response system according to an embodiment;
FIG. 2 is a flowchart for explaining the operation of the voice response system according to the embodiment.
FIG. 3 is an operation principle diagram of the voice response system according to the present embodiment.
FIG. 4 is a flowchart for explaining an operation of a user analysis unit.
FIG. 5 is a diagram for explaining a schematic system configuration of a voice response system for hotel reservation.
FIG. 6 is a diagram for explaining a schematic system configuration of a voice response system for hotel reservation.
FIG. 7 is a flowchart for explaining the operation of the hotel reservation voice response system.
FIG. 8 is an example of user information.
FIG. 9 is an example of user characteristic information.
FIG. 10 is an example of personal characteristic information.
FIG. 11 is an example of personal voice characteristic information.
FIG. 12 is an example of audio information.
FIG. 13 is an example of audio characteristic information.
FIG. 14 is an example of emotion information.
FIG. 15 is an example of identification information.
FIG. 16 is an example of a scenario.
FIG. 17 is an example of audio information.
FIG. 18 is an example of an utterance sentence.
FIG. 19 is an example of response history information.
FIG. 20 is an example of re-created user characteristic information.
FIG. 21 is an example of general user characteristic information.
FIG. 22 is an example of a table used by the user analysis unit for analysis;
FIG. 23 is another example of user characteristic information.
FIG. 24 is another example of re-created user characteristic information.
FIG. 25 is another example of general user characteristic information.
FIG. 26 is a diagram for explaining the prior art.
[Explanation of symbols]
10 Voice response system
30 telephone
40 Public network
110 Voice input unit
120 Audio output unit
130 Response processing unit
140 User Authentication Department
150 Emotion Analysis Department
160 User Analysis Department
410 Scenario storage unit
420 Response history storage
430 User characteristic information storage

Claims (5)

A voice response system that provides voice response services to users,
A voice input unit for generating voice information from user voice;
A situation identification unit that identifies the current situation where the user is located,
An emotion analysis unit that guesses the user's current emotions,
A voice recognition unit that performs voice recognition processing on the voice information generated by the voice input unit;
A storage unit for storing user characteristic information including personal characteristic information in which a user's emotion and a response pattern are associated with each other for each situation in which the user is placed;
A response pattern for acquiring a response pattern corresponding to the current status of the user specified by the status specification unit and the current emotion of the user estimated by the emotion analysis unit from the personal characteristic information A decision unit;
A response processing unit that acquires an utterance corresponding to the response pattern acquired by the response pattern determination unit and the recognition result by the voice recognition unit;
A voice response system comprising: a voice output unit that reads out an utterance sentence acquired by the response processing unit. The voice response system according to claim 1, further comprising an update unit that updates a correspondence relationship between the user's emotion and a response pattern. The voice response system according to claim 1, wherein the emotion analysis unit estimates a user's current emotion by performing comparative analysis using personal voice characteristic information stored in advance. A user analysis unit that adds and updates user characteristic information based on the user characteristic information, the voice information generated by the voice input unit, and the response history of the user from the start to the end of the voice response service The voice response system according to claim 3, further comprising: The response history includes timing information of start and end of voice output and voice input of the user by the voice response service,
5. The voice response system according to claim 4, wherein the user characteristic information includes timing information of voice output by the voice response service when the voice response service is used by the user in the past and start / end timing of voice input by the user. .

Images