JP2018091911A - Voice interactive system and voice interactive method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice interactive system and a voice interactive method for performing smooth voice interaction with a user.SOLUTION: A voice interactive robot 10 switches a speech mode for outputting voice and a voice recognition mode for recognizing the voice of a user so as to perform operation. At the speech mode, the voice interactive robot 10 outputs the voice from a speaker, terminates the speech mode when the output of the voice is completed, and shifts to the voice recognition mode. The voice interactive robot 10 does not recognize the voice of the user at the speech mode, but collects sound by a microphone and calculates similarity with feature data of the voice of a self-device, which is previously registered. When the similarity becomes a threshold or less before the output of the voice is completed, the output of the voice is stopped halfway and the speech mode is terminated by considering that the speech of the user is detected.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a voice dialogue system and a voice dialogue method for receiving voice input and outputting voice according to the received voice.

  2. Description of the Related Art Conventionally, there has been known a voice dialogue system that recognizes a user's voice and outputs a voice having contents corresponding to the recognition result to perform a voice dialogue with the user. Such a voice interaction system can be used for automatic response via a communication line, user support on a portable terminal, and the like, and can also be mounted on a robot. A robot equipped with a voice interaction system can use conversation as an interface with a user, and can be used in various situations such as security, assistance of store staff, personal life support and entertainment.

  Here, when performing a voice dialogue with the user, it is important to separate the output voice and the input voice. This is because if the output sound from the system side is included in the input sound, there arises a problem that the output sound of the own system is erroneously recognized as the user's sound. Thus, erroneous recognition due to the output sound of the own system is performed by switching between the speech mode in which the system outputs sound and the speech recognition mode in which user speech is recognized.

  In the configuration in which the speech mode and the speech recognition mode are switched, even if the user speaks during the system-side speech mode, the speech is not recognized. Therefore, the user speaks after waiting for the completion of the voice output from the system side. However, when the user is unfamiliar with the voice interaction system, the user may speak during the output of the voice from the system side.

  Therefore, Patent Document 1 discloses a robot provided with a voice recognition device that recognizes a user's voice even while speaking. The speech recognition device disclosed in Patent Document 1 starts user speech recognition after a predetermined time from the start of speech output, and also outputs output speech corresponding to the output speech equivalent of the device itself from speech collected by a microphone. By providing the removing unit, voice output is performed in parallel with voice recognition.

JP 2007-155986 A

  However, even if the conventional technique represented by the above-mentioned Patent Document 1 is used, it is difficult to smoothly perform dialogue with the user. Even if it is configured to remove the equivalent portion of the output sound by correlation calculation as in Patent Document 1, the output sound cannot be completely removed due to factors such as sound reflection environment, noise situation, distortion, It is not possible to prevent misrecognition sufficiently.

  In addition, when the voice output and the voice recognition are performed in parallel, the user cannot grasp whether his / her speech is recognized on the system side, and whether the speech should be continued or the voice output from the system side. Cannot determine whether to wait for completion. In particular, when the dialogue becomes more sophisticated and the voice output from the system side becomes longer, allowing the user to wait until the output of the voice from the system side is completed greatly impairs the smooth dialogue.

  For these reasons, how to realize a smooth voice conversation with the user has become an important issue. Such a problem becomes conspicuous in the case where a voice dialogue system is installed in a robot that is difficult to install apart from a microphone and a speaker, but voice dialogue is used for automatic response via a communication line or user support on a portable terminal. The same occurs when using the system.

  The present invention has been made in order to solve the above-described problems of the prior art, and an object thereof is to provide a voice dialogue system and a voice dialogue method for performing a smooth voice dialogue with a user.

  In order to solve the above-described problems and achieve the object, the invention according to claim 1 is an input receiving unit that receives voice input, and an output that outputs voice according to the input voice received by the input receiving unit. A speech dialogue system comprising a processing unit, a registration unit for registering output speech output by the output processing unit as self-speech, and during output of speech by the output processing unit, the input speech and the self A similarity calculation unit that calculates the similarity to the sound, and an operation control unit that controls whether or not to stop the output of the sound by the output processing unit based on the similarity calculated by the similarity calculation unit; It is provided with.

  The invention according to claim 2 further includes a speech recognition unit that performs speech recognition on the input speech received by the input reception unit according to the invention according to claim 1, wherein the output processing unit includes: The content of the voice to be output is determined according to the result of the voice recognition by the voice recognition unit, and the operation control unit performs a voice recognition mode for performing voice recognition by the voice recognition unit and outputs a voice by the output processing unit. Control is performed to switch between speech modes.

  According to a third aspect of the present invention, in the second aspect of the present invention, the operation control unit outputs the voice based on the similarity or whether the output of the voice is completed by the output processing unit. When the operation stops, the speech mode is switched to the speech recognition mode.

  According to a fourth aspect of the present invention, in the invention according to any one of the first to third aspects, when the output processing unit stops outputting the voice based on the similarity, After outputting a specific sound corresponding to the stop of the output, the output of the sound is stopped.

  The invention according to claim 5 is the invention according to any one of claims 1 to 4, wherein the registration unit analyzes the feature data related to the frequency of the output speech and generates the feature data. It is registered as self-speech, and the similarity calculation unit calculates the similarity between feature data generated by analyzing features related to the frequency of the input sound and feature data registered as the self-speech .

  The invention according to claim 6 is the invention according to any one of claims 1 to 5, wherein the registration unit further registers a predetermined sound other than the self-speech as the other person's sound, The operation control unit stops outputting the sound by the output processing unit according to the degree of similarity between the input sound and the other person's sound when the degree of similarity between the input sound and the self sound is equal to or less than a threshold value. It is characterized by determining whether to do.

  The invention according to claim 7 is the invention according to claim 6, further comprising an actuator that is provided in the same housing as the input receiving unit and performs a physical operation, and the registration unit includes Sound generated by the operation of the actuator is registered as the other person's voice.

  The invention according to claim 8 is the invention according to any one of claims 1 to 6, further comprising an actuator that is provided in the same casing as the input receiving unit and performs a physical operation. The registration unit registers as a self-sound a synthesized voice of a sound generated by the operation of the actuator and an output voice output from the output processing unit.

  According to a ninth aspect of the present invention, there is provided a voice dialogue system comprising: an input receiving unit that receives voice input; and an output processing unit that outputs voice according to the input voice received by the input receiving unit. A registration method for registering an output sound output by the output processing unit as a self-sound, and calculating a similarity between the input sound and the self-sound during the sound output by the output processing unit A similarity calculation step, and an operation control step for controlling whether or not to stop the output of sound by the output processing unit based on the similarity calculated by the similarity calculation step. .

  According to the present invention, the output sound output by the output processing unit is registered as self-speech, the similarity between the input sound and the self-sound is calculated during the output of the sound by the output processing unit, and the similarity calculation unit Since it is configured to control whether or not to stop outputting the voice by the output processing unit based on the calculated similarity, a smooth voice conversation with the user can be performed.

FIG. 1 is an explanatory diagram of the concept of the voice interaction system according to the first embodiment. FIG. 2 is a block diagram showing the configuration of the voice interactive robot shown in FIG. FIG. 3 is an explanatory diagram of a decrease in similarity due to the user's utterance. FIG. 4 is a flowchart showing a processing procedure of self-voice feature data registration processing. FIG. 5 is a flowchart showing a processing procedure in the voice recognition mode. FIG. 6 is a flowchart showing the processing procedure of the speech mode. FIG. 7 is an explanatory diagram of the operation of the voice interactive robot according to the second embodiment. FIG. 8 is a block diagram showing the configuration of the voice interactive robot shown in FIG. FIG. 9 is a flowchart illustrating the processing procedure of the speech mode in the second embodiment.

  Exemplary embodiments of a voice interaction system and a voice interaction method according to the present invention will be described below in detail with reference to the accompanying drawings.

  First, the concept of the voice interaction system according to the first embodiment will be described. FIG. 1 is an explanatory diagram of the concept of the voice interaction system according to the first embodiment. In the first embodiment, a voice dialogue robot 10 which is a robot equipped with a voice dialogue system recognizes a user's voice and outputs a voice having contents corresponding to the recognition result, thereby performing a voice dialogue with the user.

  The voice interactive robot 10 includes a speaker 11 and a microphone 12 as will be described later, and an utterance mode in which voice is output from the speaker 11 and a voice recognition mode in which user's voice is collected by the microphone 12 and recognized. Switch to operate.

  In the utterance mode, the voice interaction robot 10 outputs a voice from the speaker 11, and ends the utterance mode and shifts to the voice recognition mode when the output of the voice is completed. The voice interactive robot 10 does not recognize the user's voice in the utterance mode, but collects the sound with the microphone 12 and calculates the similarity with the feature data of the voice of the own device registered in advance.

  When the voice dialogue robot 10 outputs a voice and the user is not speaking, the microphone 12 collects the voice of the voice dialogue robot 10 and is similar to the voice feature data registered in advance. The degree is high.

  On the other hand, if the user speaks during the output of the voice by the voice dialogue robot 10, the voice collected by the microphone 12 becomes a synthesized voice in which the voice of the voice dialogue robot 10 and the user's voice are mixed. The similarity with the registered voice feature data of the own device is lowered.

  If the similarity is equal to or less than the threshold before the completion of the voice output, the voice interactive robot 10 detects the user's utterance, stops the voice output halfway, and ends the utterance mode. That is, in this case, the utterance mode is terminated due to the interruption and shifts to the voice recognition mode.

  As described above, the voice interactive robot 10 registers in advance the feature data of the sound of the own device output from the speaker 11, and the sound collected by the microphone 12 during the output of the sound in the speech mode and the sound of the own device. The similarity with the feature data is calculated, and when the similarity is less than or equal to the threshold, the speech mode is interrupted and the speech recognition mode is entered. For this reason, when a user speaks, it can change to voice recognition mode quickly and can recognize a user's voice, and can carry out smooth voice conversation.

  Further, since the voice of the own device is not collected in the voice recognition mode, erroneous recognition due to the voice of the own device can be prevented. Furthermore, since the user can grasp that his / her speech is recognized by the voice interaction robot 10, he / she can speak without stress. When the voice output from the voice interaction robot 10 and the user's utterance are performed simultaneously, it is unclear to the user whether or not his / her speech is recognized by the voice dialogue robot 10. This is because if the output of the voice is interrupted, the user recognizes that the state has shifted to a state in which the user's speech is recognized.

  Next, the configuration of the voice interactive robot 10 shown in FIG. 1 will be described. FIG. 2 is a block diagram showing the configuration of the voice interactive robot 10 shown in FIG. As shown in FIG. 2, the voice interaction robot 10 includes a speaker 11, a microphone 12, an operation unit 13, an actuator 14, a storage unit 15, and a control unit 16.

  The speaker 11 is used for outputting voice by the voice interactive robot 10. The microphone 12 functions as an input reception unit that receives an input of a user's voice by collecting ambient sounds. The operation unit 13 receives an operation input using a button or the like. The operation input is not limited to button operation input, and remote operation from a tablet or the like, or operation reception by gesture recognition may be possible.

  The actuator 14 is used for causing the voice interactive robot 10 to perform a physical operation. Specifically, the operation of the member corresponding to the arm and neck of the voice interactive robot 10 and the operation of the member showing a facial expression are controlled by driving the actuator 14. Here, a human-type or animal-type robot is assumed, but the shape of the voice dialogue robot 10 can be arbitrarily designed, and the actuator 14 can be widely used for physical operations of the voice dialogue robot 10. .

  The storage unit 15 is a storage device composed of a hard disk device, a nonvolatile memory, or the like. The storage unit 15 stores the voice feature data of the own device output from the speaker 11 as the own voice feature data 15a.

  The control unit 16 is a control unit that controls the entire voice dialogue robot 10, and includes a voice recognition unit 16a, an utterance processing unit 16b, a voice registration unit 16c, a similarity calculation unit 16d, a similarity determination unit 16e, and a state transition unit 16f. And an actuator drive processing unit 16g.

  The voice recognition unit 16a is a processing unit that performs processing for recognizing the user's voice in the voice recognition mode. Specifically, the user's voice is extracted from the input voice collected by the microphone 12 and analyzed, and the content of the utterance by the user is specified.

  The utterance processing unit 16b is an output processing unit that outputs voice in the utterance mode. Specifically, when the content of the user's utterance is specified by the voice recognition unit 16a, the content of an appropriate response is determined for the specified content of the utterance, and the output sound of the determined content is sent from the speaker 11 Output. Moreover, it is also possible to output an output sound having specific contents from the speaker 11 in a state where the user does not speak.

  The voice registration unit 16c performs a process of storing the voice of the own device output from the speaker 11, that is, the feature data of the output voice in the storage unit 15 as the own voice feature data 15a. The feature data can be obtained, for example, by generating data indicating the feature by frequency analysis of the output speech. Specifically, an arbitrary method such as an LPC (Linear Predictive Coding) cepstrum coefficient or an MFCC (Mel-Frequency Cepstrum Coefficient) can be used.

  The similarity calculation unit 16d is a processing unit that calculates the similarity between the input sound collected by the microphone 12 in the speech mode and the self-speech feature data 15a. Specifically, the voice registration unit 16c generates the input voice feature data by performing the same process on the input voice as when the self-voice feature data 15a is generated from the output voice, and the feature data of the input voice The degree of similarity with the self-speech feature data 15a is calculated.

  Here, the similarity calculation unit 16d performs a frequency filter on the input sound collected by the microphone 12, thereby reducing the influence of sound other than the sound, and extracting the sound part, and then features of the input sound. Generate data. In addition, when generating feature data of the input voice, a plurality of partial voices for a predetermined time are cut out as voice frames from the input voice, and feature data is generated for each voice frame. Accordingly, the degree of similarity with the self-speech feature data 15a is also calculated for each of a plurality of speech frames.

  The similarity determination unit 16e performs a process of determining whether or not the similarity calculated by the similarity calculation unit 16d is equal to or less than a threshold value. It is assumed that the similarity determination unit 16e detects the user's utterance when a predetermined number of audio frames whose similarity is equal to or less than the threshold value continue. By adjusting the length of one voice frame and the number of voice frames for detecting the user's speech, it is possible to remove sudden noise and appropriately detect the user's speech.

  The state transition unit 16f is an operation control unit that controls switching between the speech mode and the voice recognition mode. Specifically, in the utterance mode, the state transition unit 16f ends the utterance mode when the output of the output speech determined by the utterance processing unit 16b is completed or when the user's utterance is detected, and the state transition unit 16f Switch to recognition mode.

  When the utterance mode is terminated by the user's utterance, the output of the output voice having the content determined by the utterance processing unit 16b is stopped halfway, and the utterance mode is terminated. It should be noted that the speech mode may be terminated after outputting the specific speech after stopping the output of the output speech of the content determined by the speech processing unit 16b. For this specific voice, for example, “What did you do?” Or the like, the voice dialogue robot 10 is informed of the transition to a state of recognizing the user's voice, and the voice of the content that prompts the user to speak is given. Use.

  Further, when the state transition unit 16f detects the end of the user's utterance in the voice recognition mode, the state transition unit 16f ends the voice recognition mode and shifts to the utterance mode. The end of the user's utterance may be detected based on a condition such as “silent state continues for a predetermined time”.

  The actuator drive processing unit 16g is a processing unit that controls driving of the actuator 14. The actuator 14 is driven in accordance with, for example, the content of the speech of the voice interactive robot 10. By controlling the actuator 14, it is possible to perform an operation that takes into account changes in gestures and facial expressions during speech. In addition, the drive control of the actuator 14 can be used for consideration of the user's utterance and the movement of the voice interactive robot 10.

  FIG. 3 is an explanatory diagram of a decrease in similarity due to the user's utterance. As shown in FIG. 3, in the section where the voice dialogue robot 10 is speaking, the output voice of the voice dialogue robot 10 is included in the input voice, and in the section where the user is speaking, the user's voice is included in the input voice. It is. For this reason, in the section where the speech section of the voice interactive robot 10 and the user's speech section overlap, both the output voice and the user's voice are included in the input voice.

  Therefore, when the similarity between the feature data of the input speech and the self-speech feature data 15a is calculated, the similarity exceeds a threshold value in the section where only the voice interactive robot 10 is speaking, but the user speaks In a certain section, the degree of similarity decreases and becomes below the threshold.

  Next, a processing procedure of the voice interactive robot 10 will be described. FIG. 4 is a flowchart showing a processing procedure of registration processing of the self-speech feature data 15a. First, the voice registration unit 16c starts a registration mode by an operation input to the operation unit 13 or the like (step S101).

  After the registration mode is started, the voice registration unit 16c acquires a registration target voice (step S102). For example, the registration target sound is acquired by outputting sound from the speaker 11 and collecting sound by the microphone 12. Moreover, you may receive the audio | voice data previously acquired with the other apparatus.

  When sound is output from the speaker 11 and collected by the speaker 11, it is desirable to perform in an environment with little noise. Alternatively, registration target speech may be acquired in an actual environment in which the voice interactive robot 10 is operated. Furthermore, by acquiring the registration target sound while operating the actuator 14, it is possible to register a sound in which the drive sound of the actuator 14 and the output sound are synthesized.

  The voice registration unit 16c calculates the feature data of the acquired voice (step S103), registers it as the self voice feature data 15a in the storage unit 15 (step S104), and ends the registration mode (step S105).

  FIG. 5 is a flowchart showing a processing procedure in the voice recognition mode. First, when the voice recognition mode is started by the state transition unit 16f (step S201), the voice recognition unit 16a acquires a sound collected by the microphone 12 as an input voice (step S202). Thereafter, the state transition unit 16f determines whether or not the user's utterance has ended (step S203). The end of the user's utterance may be detected based on a condition such as “silent state continues for a predetermined time”.

  If the user's utterance has not ended (step S203; No), the speech recognition unit 16a proceeds to step S202 and continues to acquire the input speech. On the other hand, if the user's utterance is completed (step S203; Yes), the speech recognition unit 16a performs speech recognition processing on the acquired input speech (step S204). The content of the utterance by the user is specified by this voice recognition processing. The utterance processing unit 16b determines an appropriate response content with respect to the content of the specified user's utterance (step S205).

  Thereafter, the voice recognition unit 16a ends the voice recognition mode (step S206), and the state transition unit 16f shifts from the voice recognition mode to the speech mode (step S207).

  FIG. 6 is a flowchart showing the processing procedure of the speech mode. First, when the utterance mode is started by the state transition unit 16f (step S301), the utterance processing unit 16b outputs sound from the speaker 11 (step S302). The content of the sound output from the speaker 11 is determined according to the content of the user's utterance. Alternatively, it is possible to use specific contents prepared in advance for output in a state where the user does not speak.

  Further, the similarity calculation unit 16d acquires the sound collected by the microphone 12 as the input sound (step S303), and calculates the similarity between the feature data of the input sound and the self-speech feature data 15a (step S304).

  The similarity determination unit 16e determines whether or not the similarity calculated by the similarity calculation unit 16d is equal to or less than a threshold value (step S305). As a result, when the similarity is equal to or lower than the threshold (step S306; Yes), more specifically, when a certain number of voice frames whose similarity is equal to or lower than the threshold continue, the state transition unit 16f has the utterance processing unit 16b. The output of the determined output voice is stopped halfway (step S310). You may output the specific audio | voice corresponding to having stopped utterance on the way after a stop.

  If the similarity is not less than or equal to the threshold (step S306; No), more specifically, if there is no constant continuation of the audio frames having the similarity less than or equal to the threshold, the utterance processing unit 16b completes the output of the audio It is determined whether or not (step S307). As a result, if the audio output is not completed (step S307; No), the process proceeds to step S302, and the audio output is continued.

  When the voice output is completed (step S307; Yes) or when the voice output is stopped halfway in step S310, the speech processing unit 16b ends the speech mode (step S308), and the state transition unit 16f is the speech mode. To the voice recognition mode (step S309), and the process ends.

  As described above, the voice interaction robot 10 according to the first embodiment registers the self-speech feature data 15a indicating the feature of the sound output from the speaker 11 from the own device in the storage unit 15, and outputs the sound in the utterance mode. The similarity between the input voice feature data collected by the microphone 12 and the self-speech feature data 15a is calculated, and when the similarity falls below a threshold, the speech mode is interrupted and the speech recognition mode is entered. . For this reason, when a user speaks, it can change to voice recognition mode quickly and can recognize a user's voice, and can carry out smooth voice conversation.

  Further, since the voice of the own device is not collected in the voice recognition mode, erroneous recognition due to the voice of the own device can be prevented. Furthermore, since the user can grasp that his / her speech is recognized by the voice interaction robot 10, he / she can speak without stress.

  In the first embodiment, a description will be given of a configuration in which self-speech feature data 15a indicating a feature of speech output from the speaker 11 is registered in the storage unit 15 and control related to utterance interruption is performed using the self-speech feature data 15a. However, it is also possible to perform control related to utterance interruption by further registering voice other than the voice output from the speaker 11 by the own apparatus.

  For example, when the voice conversation robot 10 is interacting with the user, if the in-house broadcast or background music (BGM: background music) is collected by the microphone 12, the similarity is lowered due to the in-house broadcast or background music. This may cause the user to misrecognize that he / she spoke and interrupt the output of the voice.

  Therefore, in the second embodiment, a description will be given of a configuration in which a sound that is predicted to be generated is registered in advance as an object to be excluded, and the output of sound is continued when a decrease in similarity occurs due to the object to be excluded.

  FIG. 7 is an explanatory diagram of the operation of the voice interactive robot 110 according to the second embodiment. In addition to the self-speech feature data 15a, the spoken dialogue robot 110 shown in FIG. 7 registers speech features to be excluded as speech feature data to be excluded.

  While outputting voice in the utterance mode, the voice interactive robot 110 calculates the similarity between the feature data of the input voice and the self-speech feature data 15a, and utters another person (user or exclusion target) by comparing the similarity. Detect.

  If the speech of another person is detected, the voice interactive robot 110 calculates the similarity between the feature data of the input voice and the exclusion target speech feature data, and determines whether or not it corresponds to the exclusion target. As a result, if it falls under the exclusion target, the speech mode is continued without stopping the voice output. On the other hand, when it corresponds to an out-of-field object, the output of the voice is stopped, the speech mode is interrupted, and the voice recognition mode is entered.

  Next, the configuration of the voice interactive robot 110 shown in FIG. 7 will be described with reference to FIG. FIG. 8 is a configuration diagram showing the configuration of the voice interactive robot 110 shown in FIG. As shown in FIG. 8, the voice interaction robot 110 further stores exclusion target voice feature data 15 b in the storage unit 15. Further, the operations of the voice registration unit 116c, the similarity calculation unit 116d, the similarity determination unit 116e, and the state transition unit 116f in the control unit 16 are different from those of the voice interactive robot 10 illustrated in FIG. Since other configurations and operations are the same as those of the voice interactive robot 10 shown in FIG. 2, the same components are denoted by the same reference numerals and description thereof is omitted.

  The exclusion target speech feature data 15b is data indicating features of speech to be excluded. For example, in-house broadcasting and background music can be registered as the exclusion target audio feature data 15b. It is also possible to register the voice of a specific person.

  The voice registration unit 116c performs a registration process of the exclusion target voice feature data 15b in addition to the registration process of the own voice feature data 15a. Specifically, at the start of the registration mode, registration may be performed by accepting an operation for selecting whether to register the own voice feature data 15a or the exclusion target voice feature data 15b.

  The similarity calculation unit 116d calculates the similarity between the input voice and the exclusion target voice feature data 15b in addition to calculating the similarity between the input voice and the own voice feature data 15a. The processing relating to the calculation of the similarity is the same as in the first embodiment, but when a plurality of exclusion target speech feature data 15b are registered, the similarity is calculated for each exclusion target speech feature data 15b.

  The similarity determination unit 116e compares the similarity between the input speech feature data and the self-speech feature data 15a with a threshold value, and compares the similarity between the input speech feature data and the exclusion target speech feature data 15b with a threshold value. Make a comparison. The comparison of the similarity between the feature data of the input speech and the self-speech feature data 15a and the threshold value is used to detect the speech of others. The comparison between the similarity between the input voice feature data and the exclusion target voice feature data 15b and the threshold value is used to identify whether or not the detected voice of the other person is the exclusion target. These threshold values are not the same value, but are set appropriately.

  The state transition unit 116f detects the voice of the other person in the utterance mode, and interrupts the utterance mode when the detected voice of the other person does not correspond to the exclusion target, but the detected voice of the other person is the exclusion target. In the case, the speech mode is continued. Note that the end of the speech mode and the end of the speech recognition mode when the output of the voice is completed are the same as in the first embodiment.

  FIG. 9 is a flowchart illustrating the processing procedure of the speech mode in the second embodiment. First, when the utterance mode is started by the state transition unit 116f (step S401), the utterance processing unit 16b outputs sound from the speaker 11 (step S402). The content of the sound output from the speaker 11 is determined according to the content of the user's utterance. Alternatively, it is possible to use specific contents prepared in advance for output in a state where the user does not speak.

  In addition, the similarity calculation unit 116d acquires the sound collected by the microphone 12 as the input sound (step S403), and calculates the similarity between the feature data of the input sound and the self-speech feature data 15a (step S404).

  The similarity determination unit 116e determines whether or not the similarity calculated by the similarity calculation unit 116d is equal to or less than a threshold value (step S405). As a result, when the similarity is equal to or less than the threshold (step S406; Yes), more specifically, when a certain number of audio frames having the similarity equal to or less than the threshold continue, the similarity calculation unit 116d determines the feature of the input sound. The similarity between the data and the exclusion target speech feature data 15b is calculated (step S410).

  If the similarity between the input speech feature data and the exclusion target speech feature data 15b is less than the threshold value, the state transition unit 116f determines that the speech processing unit 16b has not determined that it is not an exclusion target (step S411; No). The output of the output sound is stopped halfway (step S412). You may output the specific audio | voice corresponding to having stopped utterance on the way after a stop.

  When the similarity with the self-speech feature data is not less than the threshold (step S406; No), or the similarity with the self-speech feature data is less than the threshold and the similarity with the exclusion target speech feature data 15b is more than the threshold In the case (step S411; Yes), the speech processing unit 16b determines whether or not the output of the voice has been completed (step S407). As a result, if the audio output is not completed (step S407; No), the process proceeds to step S402, and the audio output is continued.

  When the voice output is completed (step S407; Yes) or when the voice output is stopped halfway in step S412, the speech processing unit 16b ends the speech mode (step S408), and the state transition unit 116f is the speech mode. To the voice recognition mode (step S409), and the process ends.

  As described above, the voice interaction robot 110 according to the second embodiment registers the self-speech feature data 15a and further registers the exclusion target speech feature data 15b indicating the feature of the speech to be excluded. If the drop in the similarity between the voice feature data and the self-voice feature data 15a occurs due to the exclusion target, the voice output is continued. For this reason, it is possible to prevent a situation where voice output is interrupted due to erroneous recognition of in-house broadcasting or background music as a user's utterance.

  In the second embodiment, an example has been described in which a feature of a voice to be excluded is registered, but a voice to be recognized with priority such as a security guard or a doctor is registered as a priority target. It is good also as a structure. If the voices of other voice interactive robots are registered as priority targets, a plurality of voice interactive robots can be linked using voice recognition. In addition, the user's voice may be registered as a priority object at the start of the dialogue with the user. Furthermore, the drive sound of the actuator of the own device may be registered as the exclusion target sound feature data 15b.

  In the first and second embodiments, the case where the voice interactive system is mounted on the robot has been described. However, the present invention is not limited to this, and an automatic response via a communication line or a mobile terminal can be used. It can be used for any spoken dialogue system such as user support.

  As described above, the voice dialogue system and the voice dialogue method according to the present invention are suitable for realizing a smooth voice dialogue with the user.

DESCRIPTION OF SYMBOLS 10,110 Voice dialogue robot 11 Speaker 12 Microphone 13 Operation part 14 Actuator 15 Storage part 15a Self-speech feature data 15b Exclusion target voice feature data 16 Control part 16a Speech recognition part 16b Speech processing part 16c, 116c Voice registration part 16d, 116d Similar Degree calculation unit 16e, 116e Similarity determination unit 16f, 116f State transition unit 16g Actuator drive processing unit

Claims (9)

A voice dialogue system comprising an input receiving unit that receives voice input and an output processing unit that outputs voice according to the input voice received by the input receiving unit,
A registration unit for registering the output sound output by the output processing unit as a self-sound;
A similarity calculation unit that calculates the similarity between the input voice and the self-sound during the output of the voice by the output processing unit;
A voice dialogue system comprising: an operation control unit that controls whether or not to stop outputting voice by the output processing unit based on the similarity calculated by the similarity calculation unit. A voice recognition unit that performs voice recognition on the input voice received by the input reception unit;
The output processing unit determines the content of the voice to be output according to the result of the voice recognition by the voice recognition unit,
The voice according to claim 1, wherein the operation control unit performs control to switch between a voice recognition mode in which voice recognition is performed by the voice recognition unit and an utterance mode in which voice is output by the output processing unit. Dialog system.   The operation control unit switches from the speech mode to the speech recognition mode when the output of the speech by the output processing unit is completed or when the output of the speech is stopped based on the similarity. Item 3. The voice interaction system according to item 2.   The output processing unit, when stopping the output of the sound based on the similarity, stops outputting the sound after outputting a specific sound corresponding to the stop of the output of the sound. Item 4. The spoken dialogue system according to any one of Items 1 to 3. The registration unit registers feature data generated by analyzing features related to the frequency of the output speech as the self-speech,
The said similarity calculation part calculates the similarity of the feature data registered by analyzing the characteristic concerning the frequency of the said input audio | voice, and the feature data registered as the said own audio | voice. The spoken dialogue system according to any one of the above. The registration unit further registers a predetermined voice other than the self-voice as the other person's voice,
When the similarity between the input voice and the self voice is equal to or less than a threshold value, the operation control unit outputs the voice by the output processing unit according to the similarity between the input voice and the other person's voice. It is determined whether to stop. The spoken dialogue system according to any one of claims 1 to 5, wherein the voice dialogue system is determined. Provided in the same housing as the input receiving unit, further comprising an actuator for performing a physical operation;
The voice registration system according to claim 6, wherein the registration unit registers a sound generated by an operation of the actuator as the other person's voice. Provided in the same housing as the input receiving unit, further comprising an actuator for performing a physical operation;
The said registration part registers the audio | voice with which the sound produced by the operation | movement of the said actuator and the output audio | voice output by the said output process part were synthesize | combined as a self audio | voice. The spoken dialogue system described in 1.
A voice dialogue method of a voice dialogue system comprising: an input reception unit that receives voice input; and an output processing unit that outputs voice according to the input voice received by the input reception unit,
A registration step of registering the output sound output by the output processing unit as a self-sound;
A similarity calculation step of calculating a similarity between the input sound and the self-sound during output of the sound by the output processing unit;
A voice interaction method comprising: an operation control step for controlling whether or not to stop outputting voice by the output processing unit based on the similarity calculated in the similarity calculation step.

Images