JP2006133296A - Voice interactive device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an easy-to-use voice interactive device with which the user would not feel troublesome. <P>SOLUTION: A interactive control part 12 makes a dictionary control part 14 create a voice recognition dictionary, including possible phrases, such as "I do not understand" which a user may utter, when the user cannot find a reply to a question, and a voice recognition part 11 performs voice recognition, according to this voice recognition dictionary and outputs a result to the interactive control part 12. The interactive control part 12 decides it by an unknown expression word deciding part 18 whether the result of the voice recognition is, in agreement with an unknown expression word registered in an unknown expression word dictionary 17; and if the result is in agreement, a counter for counting the frequency in which the unknown expression word corresponding to the question was replied, is made to count up another count, and when a counter value exceeds pre-set value, the question is deleted from a scenario so as not to be made from the next time thereafter. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a voice interaction apparatus that recognizes a user's voice, synthesizes and outputs a voice corresponding to the recognized voice, and processes a user's request by interacting with the user using the voice.

  2. Description of the Related Art Conventionally, there is a voice dialogue apparatus that achieves a user's purpose by asking a user a question and repeating a question corresponding to the answer.

  In such a speech dialogue apparatus, when the user cannot answer the question or makes an ambiguous answer, the dialogue cannot be continued and the user's purpose cannot be achieved.

In order to solve such problems, if the user makes an answer that means unknown, such as “I don't know.”, It is assumed that there is no answer to that question, and the dialogue continues as if the next question was asked. Has been proposed (see, for example, Patent Document 1).
Japanese Patent No. 3576511

  However, in such a voice interaction device, the user may not be able to answer the same question frequently due to the user's habits or the situation in which it is used, and in such a case, the user cannot answer. I was bothered to ask questions and there was room for improvement in this regard.

  The present invention has been made to solve the conventional problems, and it is an object of the present invention to provide an easy-to-use spoken dialogue apparatus that does not cause trouble for the user.

  The speech dialogue apparatus according to the present invention includes speech recognition means for recognizing input speech, and an unknown expression for determining whether or not the speech recognition result output by the speech recognition means is an unknown expression word expressing that an answer to a question is unknown When an unknown expression word is input more than the number of times set for the same question, a dialogue control means for preventing the question from being performed next time is provided.

  With this configuration, when an unknown expression word is input many times for the same question, the question is not asked. Therefore, questions that the user does not know the answer will not be asked.

  Further, the speech dialogue apparatus of the present invention determines speech recognition means for recognizing input speech and whether or not the speech recognition result output by the speech recognition means is an unknown expression word expressing that the answer to the question is unknown. A configuration including an unknown expression word determination unit and a dialogue control unit that, when an unknown expression word is input more than the set number of times for the same question, performs another question instead of the question from the next time have.

  With this configuration, when an unknown word is input many times for the same question, the question is not asked and another question is asked. Therefore, questions that the user does not know the answer will not be asked.

  According to the present invention, the unknown expression word determination means determines whether the user has input an unknown expression word that expresses that the answer to the question is unknown, and the unknown expression word exceeds the number of times set for the same question. By preventing the question from being asked the next time when it is entered, it is possible to prevent the user from asking the question that the user does not know the answer. Can be improved.

  Embodiments of the present invention will be described below with reference to the drawings.

(First embodiment)
FIG. 1 is a diagram showing a voice interactive apparatus according to a first embodiment of the present invention.

  As shown in FIG. 1, the voice interaction apparatus according to the present embodiment includes a voice recognition unit 11 that recognizes a voice input by a user, a dialog control unit 12 that controls a voice conversation with the user, A dictionary storage unit 13 that stores necessary speech recognition dictionaries for each dialogue hierarchy (type, progress, etc.) for all dialogue layers, and a voice that is stored in the dictionary storage unit 13 in response to a command from the dialogue control unit 12 A dictionary control unit 14 that creates a speech recognition dictionary used by the speech recognition unit 11 by selecting and combining one or more recognition dictionaries, and a question voice or response that prompts the user to speak according to a command from the dialogue control unit 12 The response voice output unit 15 that outputs voice, the response voice storage unit 16 that stores a plurality of voices used in the response voice output unit 15, and words (unknown expression words) that are unknown are registered as items. Been An unknown expression word dictionary 17 and an unknown expression word determination unit 18 that refers to the unknown expression word dictionary 17 in response to an inquiry from the dialogue control unit 12 and determines whether the speech recognition result is unknown. For example, it is mounted on a navigation device and assists operations such as various searches and destination setting by voice input.

  In such a speech dialogue apparatus, the dialogue control unit 12 controls dialogue based on a scenario stored in advance in a built-in storage unit (not shown). This scenario is described by, for example, VoiceXML (the Voice eXtensible Markup Language, see http://radiofly.to/nishi/voicexml-sdoc/voicexml.html).

  FIG. 2 shows a facility search scenario according to the present embodiment described in VoiceXML. The operation according to this scenario will be described with reference to the interactive flowchart shown in FIG.

  First, when a voice dialogue is started by a user instruction, the dialogue control unit 12 instructs the dictionary control unit 14 to create a dictionary including words representing the type of operation. In response to this command, the dictionary control unit 14 creates a speech recognition dictionary including words representing the types of operations as shown in FIG. 4 from the dictionary storage unit 13.

  Then, the dialog control unit 12 instructs the response voice output unit 15 to output a message “What is your business?” To the user. In response to this command, the response voice output unit 15 selects and outputs a message “What is your use?” From the response voice storage unit 16.

  Further, the dialogue control unit 12 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When the user who has heard the message “What is your business?” Utters “Facilities search” and inputs it to the voice interaction device in order to search for facilities, the voice recognition unit 11 Recognize and output “facility search” as a recognition result to the dialogue control unit 12.

  Upon receiving “facility search.” As the recognition result, the dialog control unit 12 selects a scenario as shown in FIG. 2 as the scenario for facility search, and the dictionary control unit 14 “not sure” along with a word indicating the type of facility. If the user does not know the type of facility, the creation of a dictionary that includes words that may be spoken is ordered. In response to this command, the dictionary control unit 14 creates a speech recognition dictionary including words indicating the type of facility as shown in FIG. 5 and words such as “I don't know” from the dictionary storage unit 13 and the unknown expression word dictionary 17. create.

  Then, in accordance with the scenario, the dialogue control unit 12 instructs the response voice output unit 15 to output a message “Please tell us the type of facility.” To the user. In response to this command, the response voice output unit 15 selects and outputs a message “Please tell us the type of facility” from the response voice storage unit 16 (S11).

  Further, the dialogue control unit 12 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When a user who has heard the message “Please tell us the type of facility.” Speaks “Golf course.” As the facility to search for and inputs it to the voice interaction device, the voice recognition unit 11 recognizes this. Then, “golf course” is output to the dialogue control unit 12 as a recognition result.

  Upon receiving “golf course.” As the recognition result, the dialogue control unit 12 instructs the unknown expression word determination unit 18 to determine whether “golf course.” Is an unknown expression word. In response to this command, the unknown expression word determination unit 18 refers to the unknown expression word dictionary 17 whose item is an unknown word (unknown expression word) as shown in FIG. Judge whether or not and return the result. Since “Golf course.” Is not registered as an unknown expression word, it is returned that it is not an unknown expression word.

  When the dialogue control unit 12 returns that the unknown expression word is not an unknown expression word, a user such as “I don't know” is sent to the dictionary control unit 14 along with the prefecture name to narrow down the golf course by location. Instructs the creation of a dictionary that includes words that may be spoken when the name of the prefecture where the golf course is located is unknown. In response to this command, the dictionary control unit 14 creates a speech recognition dictionary including the prefecture name as shown in FIG. 7 and words such as “I don't know” from the dictionary storage unit 13 and the unknown expression word dictionary 17.

  Then, the dialogue control unit 12 gives a message to the user according to the scenario, which is a combination of “golf course.” Recognized by the voice recognition unit 11 and “Please tell me the name of a certain prefecture.” Command to output. In response to this command, the response voice output unit 15 selects and outputs a message “Please tell me the prefecture name where the golf course is located” from the response voice storage unit 16 (S12).

  Further, the dialogue control unit 12 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When a user who has heard the message “Please tell us the name of the prefecture where the golf course is located” utters “Tokyo” as the name of the prefecture where the golf course is located and inputs it to the voice interactive device, the voice recognition unit 11 However, this is recognized, and “Tokyo” is output to the dialogue control unit 12 as a recognition result.

  Upon receiving “Tokyo” as the recognition result, the dialogue control unit 12 instructs the unknown expression word determination unit 18 to determine whether “Tokyo.” Is an unknown expression word. In response to this command, the unknown expression word determination unit 18 determines whether it is an unknown expression word with reference to the unknown expression word dictionary 17 and returns the result in the same manner as described above. Since "Tokyo." Is not registered as an unknown expression word, it is returned that it is not an unknown expression word.

  When the dialogue control unit 12 returns that the unknown expression word is not an unknown expression word, the dialogue control unit 12 instructs the dictionary control unit 14 to create a dictionary including all names of golf courses in Tokyo. In response to this instruction, the dictionary control unit 14 creates a speech recognition dictionary including all the names of golf courses in Tokyo from the dictionary storage unit 13.

  Then, in accordance with the scenario, the dialogue control unit 12 sends a message combining “golf course.” And “Tell me your name” input to the response voice output unit 15 as the type of facility to the user. Command to output. In response to this command, the response voice output unit 15 selects and outputs the message “Please tell the name of the golf course” from the response voice storage unit 16 (S13).

  Further, the dialogue control unit 12 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When the user who heard the previous message “Please tell us the name of the golf course.” Utters “XX golf course.” As the name of the golf course and inputs it to the voice interaction device, the voice recognition unit 11 This is recognized, and “XX golf course” is output to the dialogue control unit 12 as a recognition result.

  Upon receiving “XX golf course.” As a recognition result, the dialogue control unit 12 determines the search target “XX golf course.” And “ Instructs the user to output a message combining "Display map." In response to this command, the response voice output unit 15 selects and outputs a message “display a golf course map” from the response voice storage unit 16 (S14).

  In addition, the dialogue control unit 12 instructs the navigation device to display a map of “XX golf course.”

  In this way, a map around the facility to be searched can be displayed on the display screen of the navigation device.

  In such a voice interaction device, as shown in FIG. 8, when a user utters “I don't know” in response to a message such as “Please tell me the type of facility” and inputs it to the voice interaction device, for example, The recognition unit 11 recognizes this, and outputs “I do not know” as the recognition result to the dialogue control unit 12.

  When the dialogue control unit 12 receives “I don't know” as the recognition result, it instructs the unknown expression word determination unit 18 to determine whether “I don't know” is an unknown expression word. In response to this command, the unknown expression word determination unit 18 determines whether it is an unknown expression word with reference to the unknown expression word dictionary 17 and returns the result in the same manner as described above. Since "I don't know" is registered as an unknown expression word, it is returned as an unknown expression word.

  When it is returned that the unknown expression word is an unknown expression word as the determination result of the unknown expression word, the dialog control unit 12 increments a counter that counts the number of times the unknown expression word provided for each question is returned, and Proceed to the next step, in order to narrow down the facilities by location, there is a possibility that the dictionary control unit 14 will speak if the user does not know the name of the prefecture where the golf course is located, along with the name of the prefecture. Commands the creation of a dictionary that also includes words. In response to this command, the dictionary control unit 14 creates a speech recognition dictionary including the prefecture name as shown in FIG. 7 and words such as “I don't know” from the dictionary storage unit 13 and the unknown expression word dictionary 17.

  Then, in accordance with the scenario, the dialogue control unit 12 instructs the response voice output unit 15 to output to the user a message that combines “facility.” And “Please tell me the name of the prefecture where there is.” . In response to this command, the response voice output unit 15 selects and outputs a message “Please tell me the name of the prefecture where the facility is located” from the response voice storage unit 16 (S12).

  Further, the dialogue control unit 12 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When the user who heard the previous message “Please tell me the name of the prefecture where the facility is located” utters “Tokyo” as the name of the prefecture where the facility is located and inputs it to the voice interactive device, the speech recognition unit 11 This is recognized, and “Tokyo” is output to the dialogue control unit 12 as a recognition result.

  Upon receiving “Tokyo” as the recognition result, the dialogue control unit 12 instructs the unknown expression word determination unit 18 to determine whether “Tokyo.” Is an unknown expression word. In response to this command, the unknown expression word determination unit 18 refers to the unknown expression word dictionary 17 to determine whether the word represents unknown and returns the result. Since “Tokyo” is not registered as an unknown word, it is returned that it is not an unknown word.

  When the dialogue control unit 12 returns that the unknown expression word is not an unknown expression word, the dialogue control unit 12 instructs the dictionary control unit 14 to create a dictionary including the names of all facilities in Tokyo. In response to this instruction, the dictionary control unit 14 creates a speech recognition dictionary including the names of all facilities in Tokyo from the dictionary storage unit 13.

  Then, in accordance with the scenario, the dialogue control unit 12 instructs the response voice output unit 15 to output to the user a message that combines “facility” and “Tell me your name.” In response to this command, the response voice output unit 15 selects and outputs a message “Please tell me the name of the facility” from the response voice storage unit 16 (S13).

  Further, the dialogue control unit 12 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When the user who heard the previous message “Please tell us the name of the facility” utters “Golf course” as the name of the golf course and inputs it to the voice interaction device, the voice recognition unit 11 Is output to the dialogue control unit 12 as a recognition result.

  Upon receiving “XX golf course.” As a recognition result, the dialogue control unit 12 determines the search target “XX golf course.” And “ Instructs the user to output a message combining "Display map." In response to this command, the response voice output unit 15 selects and outputs a message “display a golf course map” from the response voice storage unit 16 (S14).

  In addition, the dialogue control unit 12 instructs the navigation device to display a map of “XX golf course.”

  In this way, even when an unknown word such as “I don't know” is entered during the dialogue, the dialogue can be continued and a map around the facility to be searched can be displayed on the display screen of the navigation device. it can.

  In addition, the dialogue control unit 12 is provided for each question of the questions used in the current dialogue after the completion of the dialogue (for example, after the facility search dialogue is finished and the search target is determined). Compare the value of the counter that counts the number of times the unknown word is returned with a preset value, and if there is a question whose counter value is larger than the preset value, fix that to delete that question from the scenario Do.

  For example, in the facility search scenario of FIG. 2, when the number of times an unknown expression word is returned for a question inquiring about the type of facility becomes larger than a preset value as shown in FIG. 8, the dialogue control unit 12. As shown in FIG. 9, the question for inquiring about the type of facility is deleted, and the part that uses the response for inquiring about the type of facility is changed to the word “facility” that represents it. Process according to the scenario corrected from the facility search.

  Specifically, as shown in the flow chart of the dialogue shown in FIG. 10, first, when a voice dialogue is started according to a user instruction, the dialogue control unit 12 sends a word indicating the type of operation to the dictionary control unit 14. Directs creation of a dictionary containing it. In response to this command, the dictionary control unit 14 creates a speech recognition dictionary including words representing the types of operations as shown in FIG. 4 from the dictionary storage unit 13.

  Then, the dialog control unit 12 instructs the response voice output unit 15 to output a message “What is your business?” To the user. In response to this command, the response voice output unit 15 selects and outputs a message “What is your use?” From the response voice storage unit 16.

  Further, the dialogue control unit 12 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When the user who has heard the message “What is your business?” Utters “Facilities search” and inputs it to the voice interaction device in order to search for facilities, the voice recognition unit 11 Recognize and output “facility search” as a recognition result to the dialogue control unit 12.

  Upon receiving “facility search” as the recognition result, the dialogue control unit 12 selects a scenario as shown in FIG. 9 as a facility search scenario, and in order to narrow down the facility by location, the dictionary control unit 14 together with the prefecture name “ If you don't know the name of the prefecture where the facility is located, you can create a dictionary that includes words that may be spoken. In response to this command, the dictionary control unit 14 creates a speech recognition dictionary including the prefecture name as shown in FIG. 7 and words such as “I don't know” from the dictionary storage unit 13 and the unknown expression word dictionary 17.

  Then, the dialogue control unit 12 instructs the response voice output unit 15 to output a message “Please tell the name of the prefecture where the facility is located” to the user according to the scenario. In response to this command, the response voice output unit 15 selects and outputs a message “Please tell me the name of the prefecture where the facility is located” from the response voice storage unit 16 (S21).

  Further, the dialogue control unit 12 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When the user who heard the previous message “Please tell me the name of the prefecture where the facility is located” utters “Tokyo” as the name of the prefecture where the facility is located and inputs it to the voice interactive device, the speech recognition unit 11 This is recognized, and “Tokyo” is output to the dialogue control unit 12 as a recognition result.

  Upon receiving “Tokyo” as the recognition result, the dialogue control unit 12 instructs the unknown expression word determination unit 18 to determine whether “Tokyo.” Is an unknown expression word. In response to this command, the unknown expression word determination unit 18 determines whether it is an unknown expression word with reference to the unknown expression word dictionary 17 and returns the result in the same manner as described above. Since "Tokyo." Is not registered as an unknown expression word, it is returned that it is not an unknown expression word.

  When the dialogue control unit 12 returns that the unknown expression word is not an unknown expression word, the dialogue control unit 12 instructs the dictionary control unit 14 to create a dictionary including the names of all facilities in Tokyo. In response to this instruction, the dictionary control unit 14 creates a speech recognition dictionary including the names of all facilities in Tokyo from the dictionary storage unit 13.

  Then, the dialogue control unit 12 instructs the response voice output unit 15 to output a message “Please tell me the name of the facility” to the user according to the scenario. In response to this command, the response voice output unit 15 selects and outputs a message “Please tell me the name of the facility” from the response voice storage unit 16 (S22).

  Further, the dialogue control unit 12 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When the user who heard the previous message “Please tell us the name of the facility” utters “XX golf course” as the name of the facility and inputs it to the voice interaction device, the voice recognition unit 11 It recognizes and outputs “XX golf course” to the dialogue control unit 12 as a recognition result.

  Upon receiving “XX golf course.” As a recognition result, the dialogue control unit 12 determines the search target “XX golf course.” And “ Instructs the user to output a message combining "Display map." In response to this command, the response voice output unit 15 selects and outputs a message “display a golf course map” from the response voice storage unit 16 (S23).

  In addition, the dialogue control unit 12 instructs the navigation device to display a map of “XX golf course.”

  As described above, in the present embodiment, a word expressing unknown is registered in the unknown expression word dictionary 17, and the user is registered in the unknown expression word dictionary 17 with respect to the question issued by the apparatus. When answering in, count up the counter provided for each question, and if the counter value becomes larger than the preset value, the corresponding question is deleted from the scenario, so do not ask that question from the next time Therefore, it is possible to eliminate the user's annoyance and improve usability.

  In the present embodiment, the case where the system is installed in a navigation device with limited users has been described. Therefore, the user is not identified, but is used in an apparatus used by a plurality of users. A user may be identified, a counter for each question may be provided for each user, and the question may be deleted for each user. Specifically, authentication by voice, fingerprint, face image, IC card or the like is used.

  Further, in this embodiment, the counter is counted up when a word registered in the unknown expression word dictionary 17 is answered, but when a question is issued, a timer is set, and when the timer times out The counter for each question may be counted up (when the user cannot answer and is silent).

  Also, the counter for each question may be counted up even when the user's answer is not in the speech recognition dictionary and speech recognition is not possible.

  Further, in this embodiment, when a word expressing unknown to the same question is answered more than the set number of times, the corresponding question is deleted from the scenario, but the user is troublesome. The question may be revived at such a frequency that it does not feel, and when the answer is correctly answered, the question may be left as it is, and a word expressing an unknown thing may be deleted if the answer is answered.

  Moreover, in this embodiment, when a word expressing unknown to the same question is answered more than the set number of times, the corresponding question is deleted from the scenario. A deleted scenario may be prepared and replaced with the scenario.

(Second Embodiment)
Next, FIG. 11 is a diagram showing a voice interactive apparatus according to the second embodiment of the present invention. Since the present embodiment is configured in the same manner as the above-described first embodiment, the same reference numerals are given to the same configurations, and only characteristic portions will be described.

  The dialogue control unit 21 of the voice dialogue apparatus according to the present embodiment, as shown in FIG. 12, creates a scenario in which another preset question is asked when a user expresses an unknown word. If the number of times you have answered this question with a word expressing unknown exceeds the number of times set in advance, this scenario will be deleted, and another question written in the scenario will be asked instead. It is characterized by rewriting.

  Specifically, as shown in the flowchart of the dialogue shown in FIG. 13, first, when a voice dialogue is started according to a user instruction, the dialogue control unit 21 sends a word indicating the type of operation to the dictionary control unit 14. Directs creation of a dictionary containing it. In response to this command, the dictionary control unit 14 creates a speech recognition dictionary including words representing the types of operations as shown in FIG. 4 from the dictionary storage unit 13.

  Then, the dialogue control unit 21 instructs the response voice output unit 15 to output a message “What is your use?” To the user. In response to this command, the response voice output unit 15 selects and outputs a message “What is your use?” From the response voice storage unit 16.

  Further, the dialogue control unit 21 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When the user who has heard the message “What is your business?” Utters “Facilities search” and inputs it to the voice interaction device in order to search for facilities, the voice recognition unit 11 Recognize and output “facility search” as a recognition result to the dialogue control unit 21.

  Upon receiving “facility search” as the recognition result, the dialogue control unit 21 selects a scenario as shown in FIG. 12 as the facility search scenario, and in order to narrow down the facility by location, the dictionary control unit 14 together with the prefecture name “ If you don't know the name of the prefecture where the facility is located, you can create a dictionary that includes words that may be spoken. In response to this command, the dictionary control unit 14 creates a speech recognition dictionary including the prefecture name as shown in FIG. 7 and words such as “I don't know” from the dictionary storage unit 13 and the unknown expression word dictionary 17.

  Then, the dialogue control unit 21 instructs the response voice output unit 15 to output a message “Please tell the name of the prefecture where the facility is located” to the user according to the scenario. In response to this command, the response voice output unit 15 selects and outputs a message “Please tell me the name of the prefecture where the facility is located” from the response voice storage unit 16 (S31).

  Further, the dialogue control unit 21 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When the user who heard the previous message “Please tell me the name of the prefecture where the facility is located” utters “Tokyo” as the name of the prefecture where the facility is located and inputs it to the voice interactive device, the speech recognition unit 11 This is recognized, and “Tokyo” is output to the dialogue control unit 21 as a recognition result.

  When the dialogue control unit 21 receives “Tokyo” as the recognition result, it instructs the unknown expression word determination unit 18 to determine whether “Tokyo.” Is an unknown expression word. In response to this command, the unknown expression word determination unit 18 determines whether the expression is an unknown expression word with reference to the unknown expression word dictionary 17 and returns the result, as in the above-described embodiment. Since "Tokyo." Is not registered as an unknown expression word, it is returned that it is not an unknown expression word.

  The dialogue control unit 21 determines the determination result of the unknown expression word according to the scenario (S32), and if it is determined that it is not the unknown expression word, instructs the dictionary control unit 14 to create a dictionary including the names of all facilities in Tokyo. To do. In response to this instruction, the dictionary control unit 14 creates a speech recognition dictionary including the names of all facilities in Tokyo from the dictionary storage unit 13.

  Then, the dialogue control unit 21 instructs the response voice output unit 15 to output a message “Please tell me the name of the facility” to the user according to the scenario. In response to this command, the response voice output unit 15 selects and outputs a message “Please tell me the name of the facility” from the response voice storage unit 16 (S35).

  Further, the dialogue control unit 21 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When the user who heard the previous message “Please tell us the name of the facility” utters “XX golf course” as the name of the facility and inputs it to the voice interaction device, the voice recognition unit 11 It recognizes and outputs “XX golf course” to the dialogue control unit 21 as a recognition result.

  Upon receiving “XX golf course.” As the recognition result, the dialogue control unit 21 determines the search target “XX golf course.” And “ Instructs the user to output a message combining "Display map." In response to this command, the response voice output unit 15 selects and outputs a message “display a golf course map” from the response voice storage unit 16 (S36).

  In addition, the dialogue control unit 21 instructs the navigation device to display a map of “XX golf course.”

  Thus, in the facility search, the location of the facility can be narrowed down first, and then the map around the facility to be searched can be displayed on the display screen of the navigation device by the name of the facility.

  In such a voice interaction device, as shown in FIG. 14, when the user utters “I don't know” in response to the message “Please tell me the name of the prefecture where the facility is located.” The voice recognition unit 11 recognizes this and outputs “I don't know” as the recognition result to the dialogue control unit 21.

  Upon receiving “I don't know” as the recognition result, the dialogue control unit 21 instructs the unknown expression word determination unit 18 to determine whether “I don't know” is an unknown expression word. In response to this command, the unknown expression word determination unit 18 determines whether it is an unknown expression word with reference to the unknown expression word dictionary 17 and returns the result in the same manner as described above. Since “I don't know” is registered as an unknown expression word, it is returned as an unknown expression word.

  The dialogue control unit 21 determines the determination result of the unknown expression word according to the scenario (S32), and determines that the unknown expression word is an unknown expression word, the counter that counts the number of times the unknown expression word provided for each question is returned. When the user does not know the type of the facility, such as “I don't know” along with the word indicating the type of the facility in the dictionary control unit 14 in order to count up by one, advance to the next step of the scenario, and narrow down the facility by type Directs the creation of a dictionary that also includes words that may be spoken. In response to this command, the dictionary control unit 14 creates a speech recognition dictionary including words indicating the type of facility as shown in FIG. 5 and words such as “I don't know” from the dictionary storage unit 13 and the unknown expression word dictionary 17. create.

  Then, in accordance with the scenario, the dialogue control unit 21 instructs the response voice output unit 15 to output a message “Please tell us the type of facility.” To the user. In response to this command, the response voice output unit 15 selects and outputs a message “Please tell us the type of facility” from the response voice storage unit 16 (S33).

  Further, the dialogue control unit 21 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When a user who has heard the message “Please tell us the type of facility.” Speaks “Golf course.” As the facility to search for and inputs it to the voice interaction device, the voice recognition unit 11 recognizes this. Then, “golf course.” Is output to the dialogue control unit 21 as a recognition result.

  Upon receiving “golf course.” As the recognition result, the dialogue control unit 21 instructs the unknown expression word determination unit 18 to determine whether “golf course.” Is an unknown expression word. In response to this command, the unknown expression word determination unit 18 determines whether it is an unknown expression word with reference to the unknown expression word dictionary 17 and returns the result in the same manner as described above. Since “Golf course.” Is not registered as an unknown expression word, it is returned that it is not an unknown expression word.

  When the dialogue control unit 21 returns that the unknown expression word is not an unknown expression word, the dialogue control unit 21 instructs the dictionary control unit 14 to create a dictionary including all golf course names nationwide. In response to this command, the dictionary control unit 14 creates a speech recognition dictionary including all golf course names nationwide from the dictionary storage unit 13.

  Then, in accordance with the scenario, the dialogue control unit 21 sends a message combining “golf course.” And “Tell me your name” input to the response voice output unit 15 as the type of facility. Command to output. In response to this command, the response voice output unit 15 selects and outputs a message “Please tell the name of the golf course” from the response voice storage unit 16 (S34).

  Further, the dialogue control unit 21 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When the user who heard the previous message “Please tell us the name of the golf course.” Utters “XX golf course.” As the name of the golf course and inputs it to the voice interaction device, the voice recognition unit 11 This is recognized, and “XX golf course” is output to the dialogue control unit 21 as a recognition result.

  Upon receiving “XX golf course.” As the recognition result, the dialogue control unit 21 determines the search target “XX golf course.” And “ Instructs the user to output a message combining "Display map." In response to this command, the response voice output unit 15 selects and outputs a message “display a golf course map” from the response voice storage unit 16 (S36).

  In addition, the dialogue control unit 21 instructs the navigation device to display a map of “XX golf course.”

  In this way, if an unknown expression such as “I don't know” is entered for the question of the prefecture name where the facility is located, the facility type will be asked instead and the dialogue will continue. A map around the facility can be displayed on the display screen of the navigation device.

  In addition, the dialogue control unit 21 is provided for each question of the questions used in the current dialogue after the completion of the dialogue (for example, after the facility search dialogue is finished and the search target is determined). Compare the value of the counter that counts the number of times the unknown word is returned with a preset value, and if there is a question whose counter value is greater than the preset value, delete the question from the scenario and Modify to replace with the question.

  For example, in the facility search scenario of FIG. 12, when the number of times an unknown word is returned in response to a question asking for the name of a prefecture with a facility is greater than a preset value as shown in FIG. As shown in FIG. 15, the unit 21 deletes the question for inquiring about the name of the prefecture where the facility is located, and performs a modification to change the question to inquire about the type of facility written in the scenario instead. Process according to the revised scenario from the facility search.

  Specifically, as shown in the flow chart of the dialogue shown in FIG. 16, first, when a voice dialogue is started by a user's instruction, the dialogue control unit 21 sends a word indicating the type of operation to the dictionary control unit 14. Directs creation of a dictionary containing it. In response to this command, the dictionary control unit 14 creates a speech recognition dictionary including words representing the types of operations as shown in FIG. 4 from the dictionary storage unit 13.

  Then, the dialogue control unit 21 instructs the response voice output unit 15 to output a message “What is your use?” To the user. In response to this command, the response voice output unit 15 selects and outputs a message “What is your use?” From the response voice storage unit 16.

  Further, the dialogue control unit 21 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When the user who has heard the message “What is your business?” Utters “Facilities search” and inputs it to the voice interaction device in order to search for facilities, the voice recognition unit 11 Recognize and output “facility search” as a recognition result to the dialogue control unit 21.

  Upon receiving “facility search” as the recognition result, the dialogue control unit 21 selects a scenario as shown in FIG. 15 as a scenario for facility search, and the dictionary control unit 14 “is not understood” together with a word indicating the type of facility. If the user does not know the type of facility, the creation of a dictionary that includes words that may be spoken is ordered. In response to this command, the dictionary control unit 14 creates a speech recognition dictionary including words indicating the type of facility as shown in FIG. 5 and words such as “I don't know” from the dictionary storage unit 13 and the unknown expression word dictionary 17. create.

  Then, in accordance with the scenario, the dialogue control unit 21 instructs the response voice output unit 15 to output a message “Please tell us the type of facility.” To the user. In response to this command, the response voice output unit 15 selects and outputs a message “Please tell us the type of facility” from the response voice storage unit 16 (41).

  Further, the dialogue control unit 21 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When a user who has heard the message “Please tell us the type of facility.” Speaks “Golf course.” As the facility to search for and inputs it to the voice interaction device, the voice recognition unit 11 recognizes this. Then, “golf course.” Is output to the dialogue control unit 21 as a recognition result.

  Upon receiving “golf course.” As the recognition result, the dialogue control unit 21 instructs the unknown expression word determination unit 18 to determine whether “golf course.” Is an unknown expression word. In response to this command, the unknown expression word determination unit 18 determines whether the expression is an unknown expression word with reference to the unknown expression word dictionary 17 and returns the result as described above. Since “golf course.” Is not registered as an unknown expression word, it is returned that it is not an unknown expression word.

  When the conversation control unit 21 returns that the unknown expression word is not an unknown expression word, the dialogue control unit 21 instructs the dictionary control unit 14 to create a dictionary including all golf course names nationwide. In response to this command, the dictionary control unit 14 creates a speech recognition dictionary including all golf course names nationwide from the dictionary storage unit 13.

  Then, in accordance with the scenario, the dialogue control unit 21 sends a message combining “golf course.” And “Tell me your name” input to the response voice output unit 15 as the type of facility. Command to output. In response to this command, the response voice output unit 15 selects and outputs a message “Please tell the name of the golf course” from the response voice storage unit 16 (S42).

  Further, the dialogue control unit 21 instructs the voice recognition unit 11 to execute voice recognition using the dictionary created by the dictionary control unit 14.

  When the user who heard the previous message “Please tell us the name of the golf course.” Utters “XX golf course.” As the name of the golf course and inputs it to the voice interaction device, the voice recognition unit 11 This is recognized, and “XX golf course” is output to the dialogue control unit 21 as a recognition result.

  Upon receiving “XX golf course.” As the recognition result, the dialogue control unit 21 determines the search target “XX golf course.” And “ Instructs the user to output a message combining "Display map." In response to this command, the response voice output unit 15 selects and outputs a message “display a golf course map” from the response voice storage unit 16 (S43).

  In addition, the dialogue control unit 21 instructs the navigation device to display a map of “XX golf course.”

  As described above, in the present embodiment, a word expressing unknown is registered in the unknown expression word dictionary 17, and the user is registered in the unknown expression word dictionary 17 with respect to the question issued by the apparatus. When answering in, count up the counter provided for each question, and when the counter value becomes larger than the preset value, the corresponding question is deleted from the scenario and replaced with another question, so next time Therefore, the question can be avoided, the user can be prevented from feeling bothersome, and the usability can be improved.

  Also, in the present embodiment, when a word expressing unknown to the same question is answered more than the set number of times, the scenario is such that the corresponding question is deleted from the scenario and replaced with another question. However, it is also possible to prepare a scenario in which the corresponding question is deleted and replaced with another question, and replace that scenario.

  As described above, the voice interaction apparatus according to the present invention has an effect of improving usability without making the user feel bothersome, recognizes the voice of the user, and supports the recognized voice. It is useful as a voice dialogue device or the like that synthesizes and outputs a voice to be processed and interacts with the user by voice to process a user request.

The block diagram of the voice interactive apparatus in the 1st Embodiment of this invention The figure which shows the example of the scenario of the voice interactive apparatus in the 1st Embodiment of this invention Dialog flow diagram for explaining the operation of the voice interaction apparatus according to the first embodiment of the present invention. The figure which shows the speech recognition dictionary for recognizing the kind of operation of the voice interactive apparatus in the 1st Embodiment of this invention. The figure which shows the speech recognition dictionary for recognizing the classification | type of the facility of the voice interactive apparatus in the 1st Embodiment of this invention. The figure which shows the content of the unknown expression word dictionary of the voice interactive apparatus in the 1st Embodiment of this invention The figure which shows the speech recognition dictionary for recognizing the prefecture name of the speech dialogue apparatus in the 1st Embodiment of this invention. Flow chart of dialogue when answering unknown expression word of voice dialogue apparatus in the first exemplary embodiment of the present invention The figure which shows the example of the changed scenario of the voice interactive apparatus in the 1st Embodiment of this invention Dialog flow diagram for explaining the operation of the voice interaction device according to the changed scenario in the first exemplary embodiment of the present invention The block diagram of the voice interactive apparatus in the 2nd Embodiment of this invention The figure which shows the example of the scenario of the voice interactive apparatus in the 2nd Embodiment of this invention Dialogue flow diagram for explaining the operation of the speech dialogue apparatus according to the second embodiment of the present invention Flow diagram of dialogue when unknown expression word of voice dialogue apparatus in second embodiment of present invention is answered The figure which shows the example of the changed scenario of the voice interactive apparatus in the 2nd Embodiment of this invention Dialog flow diagram for explaining the operation according to the changed scenario of the voice interaction apparatus according to the second embodiment of the present invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 Speech recognition part 12 Dialog control part 13 Dictionary storage part 14 Dictionary control part 15 Response voice output part 16 Response voice storage part 17 Unknown expression word dictionary 18 Unknown expression word determination part 21 Dialog control part

Claims (2)

The same question as the voice recognition means for recognizing the input speech, and the unknown expression word determination means for determining whether the voice recognition result output by the voice recognition means is an unknown expression word expressing that the answer to the question is unknown. In contrast, when the unknown expression word is inputted more than the set number of times, the voice dialogue device includes dialogue control means for preventing the question from being performed next time. The same question as the voice recognition means for recognizing the input speech, and the unknown expression word determination means for determining whether the voice recognition result output by the voice recognition means is an unknown expression word expressing that the answer to the question is unknown. A spoken dialogue apparatus comprising dialogue control means for performing another question instead of the question when the unknown expression word is input more than the set number of times.

Images