JP4624825B2 - Voice dialogue apparatus and voice dialogue method - Google Patents

Description

  The present invention relates to a voice dialogue apparatus and a voice dialogue method, and in particular, a voice dialogue apparatus which is mounted on an information device and reduces anxiety of a user when various conditions are interactively set in stages. The present invention relates to a voice interaction method.

As a conventional voice interactive device, for example, there is a “vehicle navigation device” disclosed in Japanese Patent Laid-Open No. 9-292896. In this conventional example, the vehicle driving guidance is performed using the voice input from the occupant, but in order to increase the voice recognition rate even in a noisy environment such as in the passenger compartment, the past voice input is performed. The word input by voice is specified in consideration of the frequency of use.
JP-A-9-292896

  However, in the technique disclosed in Patent Document 1 described above, since the topic that can be accepted at each time point of the dialogue is presented on the screen to prompt the user to speak, the topic of the conversation that is currently being performed There is a problem that it is difficult to understand where is seen from the whole dialogue. In addition, because it is difficult to understand, users may become anxious about how far they should speak, and they may give up before using the dialogue. There was also a situation where the voice dialogue function itself was not used.

  The present invention has been made in view of the above-described conventional circumstances, and has reduced the user's anxiety by presenting in an easy-to-understand manner the position of the topic of the current conversation as viewed from the entire conversation. An object of the present invention is to provide a voice interaction apparatus and a voice interaction method.

In order to solve the above-described object, the present invention includes a presentation unit that presents topic information and prompts the user to speak, and a voice recognition unit that recognizes the input user's voice. met complete voice dialogue system to interact with the user through an audio recognition by presentation and speech recognition means by presenting means, the Hisage 示手 stage, the current topic information, information topics be around a topic of current as a topic transition process has Hisage Shimesuru presentation control means together with the connection relationship between the presentation control unit, when the transition of the topic to reach the present topic there are a plurality, in fact it led to the current topic Other transition processes are also presented in addition to the transition process .
In order to solve the above-described object, the present invention includes a presentation unit that presents topic information and prompts the user to speak, and a voice recognition unit that recognizes the input user's voice. A speech dialogue method for completing dialogue with a user through presentation by a presentation means and voice recognition by a voice recognition means, wherein the presentation means includes current topic information, topic information before and after the current topic, and The presentation control step presents as a topic transition process together with the connection relationship of the current, and the presentation control step is an actual transition process that has led to the current topic when there are multiple topic transitions to reach the current topic. In addition, other transition processes are also presented.

  In the voice dialogue apparatus and the voice dialogue method according to the present invention, the presentation control means (presentation control step) presents the current topic information as a topic transition process together with connection relations with topic information before and after the current topic. Therefore, the user can know where the topic of the conversation that is currently being performed is seen from the whole conversation, and can complete the conversation without feeling uneasy.

  Hereinafter, with respect to the embodiments of the voice dialogue apparatus and the voice dialogue method of the present invention, the drawings are in order of [Embodiment 1], [Embodiment 2], [Embodiment 3], [Embodiment 4], and [Embodiment 5]. Details will be described with reference to FIG. In the description of the embodiment, a configuration applied to a navigation device mounted on a vehicle, a portable terminal, or the like is shown as an example. However, the present invention is mounted on an information device and interactively sets various conditions and the like. It can be applied to any application as long as it does.

  FIG. 1 is a configuration diagram of a voice interactive apparatus according to Embodiment 1 of the present invention. In FIG. 1, the voice interactive apparatus of the present embodiment includes a signal processing unit 1, a microphone 2, a speaker 3, an input device 4 and a display 5. Although not shown in the figure because it is not a main component related to the present invention, it is provided with a configuration of other transmission / reception means if it is a mobile terminal, and with a configuration provided with transmission / reception means or communication connection means if it is for in-vehicle use. There may be.

  The microphone 2 converts the user's speech into an audio signal and supplies it to the signal processing unit 1. Moreover, the speaker 3 outputs the audio | voice signal output from the signal processing unit 1 as an audio | voice as a system utterance. The input device 4 includes a speech switch 4a and a correction switch 4b. Here, the utterance switch 4a is a switch for instructing the start of voice recognition, and the correction switch 4b is a switch for performing correction when erroneously recognized. By pressing and holding the correction switch 4b for a certain period of time. Voice operation can be terminated halfway. Further, the display 5 is realized by, for example, an LCD (Liquid Crystal Display) or the like, and displays a standby screen when setting a navigation destination or a search condition, or displays a screen such as route guidance.

  The signal processing unit 1 includes an A / D converter 11, a D / A converter 12, an output amplifier 13, a signal processing device 14 and an external storage device 15, and the signal processing device 14 includes a CPU 21 and a memory 22. The signal processing unit 1 performs signal processing in response to pressing of the utterance switch 4a or the correction switch 4b of the input device 4 or in response to a user's voice input from the microphone 2, and the system response from the speaker 3 as a system response. An utterance is output, or a response image is displayed on the display 5.

  The external storage device 15 stores data used for navigation, recognition target data used for voice recognition processing, and the like, and the memory 22 stores processing programs and various frequently used data. In the speech recognition processing, the signal processing device 14 performs recognition processing with reference to the external storage device 15 for speech input from the microphone 2 converted into a digital signal via the A / D converter 11, and the speech input The signal processing device 14 converts the system utterance as a system response to the above to an analog signal via the D / A converter 12 and amplifies the signal by the output amplifier 13, and then outputs the sound from the speaker 3.

  The voice interaction apparatus according to the present embodiment includes a presentation unit that presents topic information and prompts the user to speak, and a voice recognition unit that recognizes the input user's voice, and includes a plurality of topic presentation units. The dialogue with the user is completed through the presentation and voice recognition by the voice recognition means. The presentation means is realized by the display 5, and the voice recognition means is the A / D converter 11, the signal processing device 14, and the external storage device. 15 is realized. Further, the present invention is characterized by the presenting control means for presenting the current topic information as a topic transition process together with connection information with topics information before and after the current topic on the presenting means (display 5). This is realized by the signal processing device 14.

  Next, with reference to FIG. 2, FIG. 3, and FIG. 4, a voice dialogue method in the voice dialogue apparatus of the present embodiment will be described. FIG. 2 is a flowchart for explaining the voice dialogue method, FIG. 3 is a diagram for explaining the data structure of the recognition target word, and FIG. 4 is a diagram for explaining the basic configuration of the standby screen displayed on the display 5.

  There are a plurality of methods for speech recognition. A method of recognizing speech uttered by separating words into words, such as command input to a computer, is called `` word speech recognition '', and on the other hand, it is uttered by connecting multiple words without dividing each word. The method of recognizing the voice is called “continuous word recognition”. Depending on the purpose of use or interface design, each of them may be used separately, and both may be implemented at the same time, and there are various forms. I can not say. In the present embodiment, the following description will be given on the assumption that implementation using “word speech recognition” is performed.

  In the flowchart of FIG. 2, first, when the user operates the speech switch 4a to instruct the start of speech (step S101), the signal processing device 14 detects the speech start instruction and starts processing.

  Next, the signal processing device 14 performs standby setting for voice recognition processing (step S102). The signal processing device 14 reads the recognition target word from the external storage device 15 into the memory 22. Here, the recognition target word relates to a navigation device for vehicles and an electronic device for vehicles, and is as shown in FIG. As shown in the figure, the data structure of the recognition target word has a hierarchical structure of three or more levels. The topic of the first hierarchy is function selection, and five functions of “destination”, “search condition”, “facility display”, “telephone”, and “help” can be selected. Each recognition target word in the first hierarchy has a connection relationship between a plurality of recognition target words in the second hierarchy and a topic, and the recognition target word in the jth hierarchy is a plurality of recognition target words in the j + 1 hierarchy lower than the recognition target word. And have a connection relationship. The number of hierarchies as a whole is determined according to the function to be selected and the content to be uttered after the selection.

  Next, the signal processing device 14 switches and changes the display content of the display 5 to the standby screen so that the user can easily input voice (step S103).

  FIG. 4 exemplifies a standby screen in the middle (prefecture name selection) when interactively setting the navigation destination. The basic configuration of the standby screen is a configuration including an area A, an area B, and an area C.

  Area A is an area for displaying what standby settings are currently set, and what kind of recognition target words should the user speak by looking at the contents displayed in area A? Can know. In the example of FIG. 4, it can be easily understood that the name of the prefecture should be uttered next. Area B is an area for displaying what is recognized by the user's utterance. In the example of FIG. 4, when the user selects a genre for destination setting one level before (topic 2), the user utters “station” and the speech recognition result is continuously displayed.

  Furthermore, the area C is an area that is presented together with a connection relationship with topics that precede and follow the current topic, and is a characteristic display configuration of the present invention. In the example of FIG. 4, since it is currently in the hierarchy for inputting the prefecture name, the prefecture name selection C03 of that hierarchy (topic 3) is identified (display color is different from other topics C01, C02, C04 and C05). (Displayed in color). Thus, after presenting together with the connection relationship between the current topic and the topic before and after, function selection is performed on topic 1, genre selection is performed on topic 2, and the current (topic 3) prefecture name is selected. Can easily understand a series of flow of ending after selecting topic name of topic 4 and selecting search condition of topic 5, and the user should perform the work of selecting and setting this flow. Can do.

  In addition to the method of changing the display color, there are various methods for the identification display such as changing the gradation (gray scale) or putting a pattern as the background of the character. Further, when the level from the function selection of topic 1 to the end exceeds the number of levels that can be displayed in area C, for example, the level from the current topic to the end may be preferentially displayed. . Moreover, it is good also as a structure which can scroll the display of the area | region C right and left by the specific switch of the input device 4, or a user's fingertip press.

  Next, in order to notify the user that the processing has started, the signal processing device 14 outputs a notification sound signal stored in the external storage device 15 to the D / A converter 12, and the amplifier 13 and Announcement voice is output through the speaker 3. When the user hears the notification voice, the user utters the speech included in the recognition target, so the signal processing device 14 captures the voice signal from the microphone 2 converted into a digital signal by the A / D converter 11 (step S104). ).

  The signal processing device 14 calculates the power of the output (digital audio signal) of the D / A converter 12 until the utterance switch 4a is operated. After the utterance switch 4a is operated, the previous average is calculated. The section in which the instantaneous power of the digital voice signal is larger than a predetermined threshold value as compared with the power is regarded as the voice section, and the speech section is detected. Detection of such speech utterance sections is a known technique, and is disclosed in, for example, Japanese Patent Application Laid-Open No. 63-226692 “Pattern Comparison Method”.

  Next, the signal processing device 14 starts the degree of coincidence calculation with the recognition target stored in the memory 22 (step S105). The degree of coincidence indicates how similar each recognition target is to the digital speech signal in the detected utterance section, and is calculated as a score. Here, it is evaluated that the greater the score value, the higher the matching degree. It should be noted that while the coincidence calculation is performed, the voice capturing is continued by the parallel processing.

  Next, when the instantaneous power of the digital audio signal becomes equal to or less than a predetermined threshold, the signal processing device 14 determines that the user's utterance has ended, and ends the audio capturing (step S107). This termination determination is also the same as the “pattern comparison method” disclosed in Japanese Patent Laid-Open No. 63-226692, for example.

  Next, the signal processing device 14 waits for the completion of the coincidence calculation (step S105), and outputs the recognition target having the highest coincidence as a recognition result as a result of the coincidence calculation. In the example of FIG. 4, the recognition target word having the highest degree of coincidence is “station”, and “station” is displayed in the region B as the recognition result. The recognition result is output not only on the display 5 but also on the sound output from the speaker 3. The recognition result “station” is converted into a speech signal using the speech synthesis processing function, and the D / A converter. After being converted into an analog audio signal by the signal 12, the signal is amplified by the amplifier 13 and then output from the speaker 3.

  For example, when the user utters “station” but erroneous recognition occurs and the recognition result is “address”, the user operates the correction switch 4b to perform recognition processing ( Step S102) It is possible to return to the previous state. That is, after outputting the recognition result in step S107, the user waits for a predetermined time whether or not the correction switch 4b has been pressed. If the correction switch 4b has been pressed, the process returns to step S102 and again from step S102 to step S102. The processing up to S107 is repeated, and if the correction switch 4b has not been pressed within a predetermined time, the user accepts the recognition result and proceeds to step S109 (step S108).

  Next, the recognition target output as the recognition result in step S107 is received, and it is determined whether there is a lower hierarchy (step S109). If there is a lower hierarchy, the process returns to step S102, and standby processing according to the lower hierarchy, that is, a representative vocabulary instruction and a recognition target are set. In the example of FIG. 4, there is a lower hierarchy with respect to the recognition target word “station”, and “prefecture name” is set as a standby word of the lower hierarchy and displayed in the area A.

  In step S109, if the recognition target word output as the recognition result does not have a lower hierarchy, the process proceeds to step S110, and the function associated with the recognition target word is executed. In the example of FIG. 4, a function for setting a destination to a place finally determined by speech and a function for searching a route are associated with each other, and these functions are executed.

  As described above, the processing procedure has been described with reference to the flowchart of FIG. 2, but in the voice dialogue method in the voice dialogue apparatus of the present embodiment, the process is recursively executed every time the hierarchy is advanced by the user's utterance. Next, more specifically, referring to the explanatory diagrams from FIG. 5 to FIG. 10, the user utters “Destination” → “Station” → “Kanagawa” → “Yokohama Station” → “Toll road priority”. The transition of the display screen of the display 5 when proceeding through the hierarchy will be described.

  FIG. 5 is a standby screen that is first displayed on the display 5 after the user presses the speech switch 4a. In FIG. 5, it can be easily understood from the identification display of the function selection C01 in the area C that the current hierarchy is a screen for performing function selection (topic 1). It is also easy to understand that there are multiple topics that must be passed in the future.

  Next, FIG. 6 is a standby screen when the user selects from the representative vocabulary (of Topic 1) displayed in the area A of the screen of FIG. 5, and the user speaks “Destination” in response to this. is doing. Further, through the processing from step S104 to step S107 in FIG. 2 based on this utterance, the recognition result “destination” is displayed in the recognition result display area (area B).

  Next, FIG. 7 is a standby screen for the next layer (topic 3) of the recognition target word “destination”. From the identification display of the area C, it can be easily understood that the current hierarchy is a hierarchy for inputting the topic 2 of “genre selection”. Further, the recognition result “destination” of the user's utterance is displayed in the area B. A lower hierarchy exists for the recognition target word “destination”, and a plurality of representative vocabularies of the hierarchy are displayed in the area A. That is, it can be easily understood that the user can then select a genre from the representative vocabulary displayed in the area A on the standby screen of FIG.

  Next, FIG. 8 is a standby screen after the user selects “station” from the representative vocabulary displayed in area A of the screen of FIG. 7 and speaks “station”. In response to the utterance “station” by the user, the recognition result is displayed in area B, and there is a lower hierarchy with respect to the recognition target word “station”, and the representative vocabulary “prefecture name” is the waiting word in area A. Is displayed. Further, from the display of the area C, it can be easily understood that the current hierarchy is a hierarchy for inputting the topic 3 of “prefecture name selection”.

  Next, FIG. 9 is a standby screen when the user speaks “Kanagawa Prefecture” with respect to the representative vocabulary “prefecture name” displayed in the area A of the screen of FIG. In response to the user's utterance “Kanagawa”, the recognition result is displayed in the region B, and there is a lower hierarchy with respect to the recognition target word “Kanagawa”, and the representative vocabulary “station name” is the standby word in the region A. Is displayed. Further, from the display of the area C, it can be easily understood that the current hierarchy is a hierarchy for inputting the topic 4 of “station name selection”.

  Further, FIG. 10 is a standby screen when the user speaks “Yokohama Station” with respect to the representative vocabulary “station name” displayed in the area A of the screen of FIG. 9. In response to the user's utterance “Yokohama Station”, the recognition result is additionally displayed in the area B, and there is a lower hierarchy for the recognition target word “Yokohama Station”. It is displayed. Further, from the display of the area C, it can be easily understood that the current hierarchy is a hierarchy for inputting the topic 5 of “search condition selection”.

  In the specific example described above, we spoke in the order of “Destination” → “Station” → “Kanagawa” → “Yokohama Station” → “Toll Road Priority” and proceeded through the hierarchy in order. It is also possible to accept utterances of different levels. For example, in the standby screen of FIG. 5, when the user utters “station” two levels ahead, the screen transitions to the standby screen of FIG.

  As described above, in the voice interaction apparatus and the voice interaction method of the present embodiment, the presentation means (display 5) that presents topic information and prompts the user to speak, and the voice that recognizes the input user's voice. And a recognition means (11, 14), and a presentation control means (14) (14), which completes a dialogue with a user through presentation by a presentation means and voice recognition by a voice recognition means for a plurality of topics. In the presentation control step, the current topic information is presented as a topic transition process (in the area C on the screen of the display 5) together with the connection relationship with the topic information before and after the current topic. Thereby, the user can know where the topic of the conversation that is currently being performed is located from the whole conversation, and can complete the conversation without feeling uneasy.

  In addition, the current topic information in the topic transition process (for example, in the area C on the screen of the display 5 shown in FIG. 4, the display color of the prefecture name selection C03 of topic 3 is changed from the other display color). Therefore, it is possible to easily understand what (where) the current topic (hierarchy) is.

  Furthermore, since the topic transition process includes the connection relationship of topics until the completion of the conversation, the user can recognize the transition of the topics from the current topic to the completion, and imagine the burden of the conversation until the completion. This makes it easier to conduct conversations while feeling anxiety as in the past.

  Next, a voice dialogue apparatus and a voice dialogue method according to Embodiment 2 of the present invention will be described.

  The configuration of the voice interactive apparatus according to the present embodiment is the same as that illustrated in FIG. Also in the present embodiment, as in the first embodiment, the presentation control means realized by the signal processing device 14 sends the current topic information to the presentation means (display 5), and information on topics around the current topic. In this embodiment, the presentation control means (14) calculates the difficulty of recognition by the speech recognition means (11, 14) for each topic and combines it with the topic transition process. The point is to present.

  Here, the recognition difficulty level represents the difficulty of voice recognition in the voice recognition processing, and when the recognition difficulty level is low, in order to increase the voice recognition rate, utter clearly or speak with a large volume of voice, etc. A large burden is demanded on the user. The low speech recognition rate is generally proportional to the number of words included in the hierarchy of recognition target words, and therefore it can be determined that the speech recognition rate is low when the number of words included in the layer is a predetermined value or more. . The low speech recognition rate can also be judged by whether or not low recognition words are included, but this is based on the technical literature “Terahiro;“ Prediction technology of low recognition words in speech recognition ”, Technical Report (Toyota Central Research). R & D Review), Vol 13, No. 2, June 2001) ”. According to this technical document, it can be determined that the speech recognition rate is low when a predetermined number or more of low recognition words are included.

  Next, the voice dialogue method in the voice dialogue apparatus according to the present embodiment also follows the procedure shown in the flowchart of FIG. However, the processing contents in the standby screen display in step S103 are different and will be described in detail below.

  FIG. 11 is an explanatory diagram illustrating the display contents of the area C on the standby screen according to the present embodiment. The display of the areas A and B on the standby screen is the same as in the first embodiment. That is, the display content of the area C in FIG. 4 or FIG. 8 of the first embodiment is replaced with the content shown in FIG.

  In FIG. 11, the prefecture name selection C03 of topic 3 is identified and displayed (display color is displayed in a color different from other topics C01, C02, C04, and C05), and is currently in a hierarchy for inputting the prefecture name. It is easy to see. In addition, from the display of the connection relation, after selecting the current prefecture name, it is possible to easily understand a series of flows that end after selecting the prefecture name of topic 4 and the search condition selection of topic 5.

  Further, in FIG. 11, the difficulty of recognition when selecting the station name of the next layer is difficult (the voice recognition rate is low), which is indicated by the identification display by the frame C04b larger than other topics. It can be recognized that the burden of the user's utterance is heavy, such as sometimes speaking clearly or speaking with a loud voice volume. That is, since the recognition difficulty level is presented together with the topic transition process, the user can recognize in advance how much the burden will be.

  FIG. 12 is an explanatory diagram showing another example of the display contents of the area C on the standby screen of the present embodiment. As an identification method different from FIG. 11, it is possible to identify that the difficulty level of recognition at the time of selecting a station name in the next hierarchy is low (the voice recognition rate is low) by a color or pattern different from other topics. FIG. 12 shows an identification display based on a background C04c having a pattern different from that of other topics. This allows the user to recognize in advance how much the burden will be in the future. When the recognition difficulty level is identified by color, for example, red is used when the recognition difficulty level is difficult and the user burden is large, and blue is used when the recognition difficulty level is easy and the user burden is small. A method is conceivable. Furthermore, as a method of presenting the recognition difficulty level together with the topic transition process, various methods such as a method of displaying the fact in words and a method of displaying the value by quantification can be considered.

  As described above, in the voice dialogue apparatus and the voice dialogue method of the present embodiment, the presentation control means (14) calculates the recognition difficulty level by the voice recognition means (11, 14) for each topic, and enters the topic transition process. Since they are also presented, it is easy to imagine the burden derived from the difficulty of voice recognition, and the user's anxiety in voice conversation can be further reduced.

  Next, a voice dialogue apparatus and a voice dialogue method according to Embodiment 3 of the present invention will be described.

  The configuration of the voice interactive apparatus according to the present embodiment is the same as that illustrated in FIG. Also in the present embodiment, as in the first embodiment, the presentation control means realized by the signal processing device 14 sends the current topic information to the presentation means (display 5), and information on topics around the current topic. In this embodiment, the presentation control means (14) uses a continuous word voice for the current topic and other topics that have a connection relationship with the current topic. In the topic transition process, the current topic, other topics, and their connection relations are identified and presented. For this reason, in this embodiment, unlike the first embodiment, both “word speech recognition” and “continuous word recognition” are implemented at the same time. Here, the topic 3 “prefecture name selection” and the topic 4 “station name selection” will be described on the assumption that continuous word speech can be recognized.

  Next, the voice dialogue method in the voice dialogue apparatus according to the present embodiment also follows the procedure shown in the flowchart of FIG. However, the processing content in the standby screen display in step S103 is different, and will be described below with a specific example.

  FIG. 13 is an explanatory diagram illustrating the display contents of the area C on the standby screen according to the present embodiment. The display of the areas A and B on the standby screen is the same as in the first embodiment. That is, the display content of the region C in FIG. 7 of the first embodiment is replaced with the content shown in FIG. From the identification display of topic 2 in area C, it can be easily understood that the current hierarchy is a hierarchy for inputting “genre selection”. In addition, from the connection relation display, after selecting a genre (by selecting “station”), a series of flows is easily completed after selecting the prefecture name and station name of Topic 3 and Topic 4 and selecting search conditions for Topic 5. Can understand.

  In particular, in FIG. 13, the prefecture name and station name selection in the next hierarchy are highlighted with a thick frame C34. As a result, the user can recognize that utterances that connect both prefecture name selection and station name selection are possible in the next hierarchy, and easily understand how much the future utterance burden will be reduced. can do.

  As explained above, in the voice interaction apparatus and the voice interaction method of the present embodiment, the presentation control means (14) uses the continuous word of the current topic and other topics having a connection relationship with the current topic. When speech recognition is possible, the current topic, other topics, and their connection relations are identified and presented in the topic transition process. The user's anxiety in voice conversation can be further reduced.

  Next, a voice dialogue apparatus and a voice dialogue method according to Embodiment 4 of the present invention will be described.

  The configuration of the voice interactive apparatus according to the present embodiment is the same as that illustrated in FIG. Also in the present embodiment, as in the first embodiment, the presentation control means realized by the signal processing device 14 sends the current topic information to the presentation means (display 5), and information on topics around the current topic. This example presents the current topic when there are a plurality of topic transitions to reach the current topic in the presentation control means (14). Another feature is that other transition processes are presented together with the actual transition process.

  Next, the voice dialogue method in the voice dialogue apparatus according to the present embodiment also follows the procedure shown in the flowchart of FIG. However, the processing contents in the standby screen display in step S103 are different and will be described in detail below.

  FIG. 14 is an explanatory diagram illustrating the display contents of the area C on the standby screen according to the present embodiment. The display of the areas A and B on the standby screen is the same as in the first embodiment. That is, the display content of the area C in FIG. 9 of the first embodiment is replaced with the content shown in FIG.

  In FIG. 14, the station name selection C04 of Topic 4 is identified and displayed, and it can be easily understood that the station name is currently in a hierarchy for inputting the station name. In addition, from the display of the connection relationship, after selecting the current station name, it is possible to easily understand a series of flows that end after selecting the search condition of topic 5. However, in FIG. 14, unlike the first embodiment, two topic transition processes of prefecture name selection C031 and route selection C032 are displayed as topic connection relations between topic 2 genre selection and topic 4 station name selection. In Topic 3, it indicates that either prefecture name selection or route selection was possible.

  In other words, in the topic level 3 of the previous level, it is possible to select “route selection” in addition to “prefecture name selection”. For example, in addition to the utterance “Kanagawa prefecture”, the user utters “Keihin Express”. Also means that it was possible. In the standby screen of the topic 3 hierarchy, “prefecture name” and “route name” are displayed as representative vocabulary displayed in the area A.

  As described above, in the voice interaction apparatus and the voice interaction method of the present embodiment, in the presentation control means (14), when there are a plurality of topic transitions for reaching the current topic, It also shows other transition processes in addition to the actual transition process that arrived, so it can be compared with other topic transition processes that are different from the topic transition process that has actually transitioned, It is easier to understand where it is from the whole dialogue.

  Next, a voice interaction apparatus and a voice interaction method according to Embodiment 5 of the present invention will be described.

  The configuration of the voice interactive apparatus according to the present embodiment is the same as that illustrated in FIG. Also in the present embodiment, as in the first embodiment, the presentation control means realized by the signal processing device 14 sends the current topic information to the presentation means (display 5), and information on topics around the current topic. In this embodiment, the topic transition process of conversations completed in the past is stored in the storage means (memory 22 or external storage device 15), and the presentation control means (14 ), The topic transition process of the conversation completed in the past is presented together with the topic transition process of the current topic.

  Next, the voice dialogue method in the voice dialogue apparatus according to the present embodiment also follows the procedure shown in the flowchart of FIG. However, the processing contents in the standby screen display in step S103 are different and will be described in detail below.

  FIG. 15 is an explanatory diagram illustrating the display contents of the area C on the standby screen according to the present embodiment. The display of the areas A and B on the standby screen is the same as in the first embodiment. That is, the display content of the area C in FIG. 8 of the first embodiment is replaced with the content shown in FIG.

  In FIG. 15, the prefecture name selection C03 of Topic 3 is identified and displayed, and it can be easily understood that it is currently in the hierarchy for inputting the prefecture name. In addition, from the display of the connection relation, after selecting the current prefecture name, it is possible to easily understand a series of flows that end after selecting the prefecture name of topic 4 and the search condition selection of topic 5. However, in FIG. 15, unlike the first embodiment, the topic connection relationship between the genre selection of topic 2 and the search condition selection of topic 5 is the prefecture name selection C06, city name selection C07, capital letter selection C08, and address selection C09. Other topic transition processes for transitioning the topic are displayed. This other topic transition process (C02 → C06 to C09 → C05) is a transition process of other input topics performed by the user in the past, and by displaying this together, the user is responsible for the current input burden. And the input burden experienced in the past can be compared.

  As described above, in the voice dialogue apparatus and voice dialogue method of the present embodiment, the presentation control means (14) presents the topic transition process of the conversation completed in the past in addition to the topic transition process related to the current topic. By comparing with the burden of successful dialogue, it is possible to imagine the burden of the current dialogue and to further reduce the user's anxiety in the voice dialogue.

It is a block diagram of the voice interactive apparatus which concerns on the Example of this invention. It is a flowchart explaining a voice dialogue method. It is explanatory drawing of the data structure of a recognition target word. FIG. 3 is an explanatory diagram illustrating a basic configuration of a standby screen in the first embodiment. FIG. 6 is an explanatory diagram illustrating a standby screen when an utterance switch is pressed in the first embodiment. 6 is an explanatory diagram illustrating a standby screen when speaking “destination” in Embodiment 1. FIG. 6 is an explanatory diagram illustrating a standby screen after speaking “destination” in Embodiment 1. FIG. FIG. 6 is an explanatory diagram illustrating a standby screen after uttering “station” in the first embodiment; 6 is an explanatory diagram illustrating a standby screen after speaking “Kanagawa Prefecture” in Embodiment 1. FIG. FIG. 4 is an explanatory diagram illustrating a standby screen after speaking “Yokohama Station” in the first embodiment. FIG. 10 is an explanatory diagram illustrating display contents of a region C on a standby screen according to a second embodiment. FIG. 10 is an explanatory diagram illustrating display contents of a region C on a standby screen according to a second embodiment. FIG. 10 is an explanatory diagram illustrating display contents of a region C on a standby screen according to a third embodiment. FIG. 10 is an explanatory diagram illustrating display contents of a region C on a standby screen according to a fourth embodiment. FIG. 10 is an explanatory diagram illustrating display contents of a region C on a standby screen according to a fifth embodiment.

Explanation of symbols

1 signal processing unit 2 microphone 3 speaker 4 input device 5 display (presentation means)
11 A / D converter (voice recognition means)
12 D / A converter 13 Output amplifier 14 Signal processing device (voice recognition means or presentation control means)
15 External storage device 21 CPU
22 memory

Claims (11)

Presenting means for presenting topic information to prompt the user to speak, and speech recognition means for recognizing the input user's voice, and presenting a plurality of topics by the presenting means and by the speech recognition means A voice dialogue device that completes a dialogue with a user through voice recognition,
It said presentation means, have a current topic information, the presentation control means for presenting a topic transition processes with the connection relationship between the topic of information back and forth to the present topic,
The presenting control means, when there are a plurality of topic transitions to reach the current topic, presents other transition processes together with the actual transition process that has reached the current topic, A voice interaction device.   The spoken dialogue apparatus according to claim 1, wherein the presentation control unit identifies and presents the current topic information in the topic transition process.   3. The voice dialogue according to claim 1, wherein the presentation control unit calculates a difficulty level of recognition by the voice recognition unit for each topic and presents it together with the topic transition process. 4. apparatus.   The presentation control means can recognize the current topic in the topic transition process when the current topic and other topics having a connection relationship with the current topic are capable of recognizing continuous words. The spoken dialogue apparatus according to any one of claims 1 to 3, wherein the other topics and their connection relations are identified and presented. Having storage means for storing topic transition processes of dialogues completed in the past,
The presentation control means according to any one of claims 1 to 4, characterized in that presenting together topic transition process of dialogue completed in the past a topic transition process related to said current topic Spoken dialogue device. The voice conversation apparatus according to any one of claims 1 to 5, wherein the topic transition process includes a connection relationship of topics until the completion of the conversation. Presenting means for presenting topic information to prompt the user to speak, and speech recognition means for recognizing the input user's voice, and presenting a plurality of topics by the presenting means and by the speech recognition means A voice dialogue method for completing dialogue with a user through voice recognition,
The presenting means has a presentation control step for presenting current topic information as a topic transition process together with connection information with topic information before and after the current topic,
The presenting control step is characterized in that when there are a plurality of topic transitions for reaching the current topic, the actual transition process leading to the current topic is also presented together with other transition processes. It is Ruoto voice interactive method. 8. The spoken dialogue method according to claim 7, wherein the presentation control step identifies and presents the current topic information in the topic transition process . The presentation control step, the topic recognition difficulty by the speech recognition means calculates for each voice interaction according to claim 7 or claim 8, characterized in that presented in conjunction with the topic transition phase Method. In the presentation control step, when the current topic and other topics having a connection relationship with the current topic are capable of recognizing continuous words, the current topic in the topic transition process, other topics and voice dialogue method according to any one of claims 7 to claim 9, characterized by presenting to identify their connectivity. Having a memory step for memorizing topic transition processes of dialogues completed in the past,
The voice according to any one of claims 7 to 10, wherein the presentation control step presents a topic transition process of a dialog completed in the past in addition to a topic transition process related to the current topic. How to interact.

Images