JP2016118592A - Communication Robot - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a robot which re-adjusts its spoken voice while confirming whether a companion to talk with can catch the spoken voice, and can smoothly communicate without giving any stress to the companion to talk with.SOLUTION: A communication robot comprises: an input part which inputs a companion's voice; a database in which a plurality of conversation patterns that the companion and robot speak are registered; a voice recognition part which recognizes a conversation pattern from a voice of the companion by reference to the database; a conversation pattern determination part which determines a conversation pattern of the robot answering to a conversation pattern of the companion by reference to the database; an output part which outputs the conversation pattern of the robot as a spoken voice; and a companion response detection part which determines whether the companion can catch the spoken voice of the robot, When the companion response detection part determines that the companion can catch the spoken voice, the output part adjusts and outputs the spoken voice again.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a communication robot that interacts with a person.

  In recent years, the number of elderly people living alone has increased, and there has been an increasing demand for communication robots that can talk with each other and add to their loneliness. However, in elderly people, the auditory function has deteriorated, and there are cases in which it is impossible to hear the speech of the robot.

  Conventionally, as a conventional technique, a speech synthesis application device (Patent Document 1) that changes the pitch and tone of speech sound according to the situation, or a hearing aid that converts a speech signal into an appropriate speech speed (Patent Document) 2), and a speech speed conversion device (Patent Document 3) that slows the speech speed if the hurry of the conversation person is estimated and not rushed.

  In addition, as a technique for reducing the burden on the interlocutor due to misrecognition of speech processing, an information processing apparatus that performs restoration processing by determining that there is misrecognition when detecting a specific utterance (pitch frequency, power) of the interlocutor (Patent Document 4) is also available.

JP-A-6-318100 Japanese Patent Laid-Open No. 10-70790 JP 2010-212783 A JP-A-8-30290

  However, the techniques of Patent Documents 1 to 3 change the pitch, tone, and speech speed of the uttered voice to make it easier for the talker to hear, but have not confirmed whether the talker has actually heard it. The speech was not adjusted according to the auditory function of the person.

  Japanese Patent Laid-Open No. 2004-228867 determines whether the conversation person feels uncomfortable due to misrecognition on the robot side, and does not determine whether the conversation person can hear the speech voice of the robot.

  Therefore, in the conventional technology, there is a problem that the robot may unilaterally advance the conversation even though the conversation person cannot hear the voice of the robot, and the conversation person feels stress in the conversation with the robot. .

  The present invention has been made in view of the above-mentioned problems, and re-adjusted while confirming whether the utterance voice of the robot could be heard by the interlocutor, so that smooth communication is possible without causing the interlocutor to feel stress. It is to provide a robot that can be planned.

  The communication robot according to the present invention recognizes a conversation pattern from a conversation person's voice by referring to the input section for inputting the conversation person's voice, a database in which a plurality of conversation patterns spoken by the conversation person and the robot are registered, and the database. A speech recognition unit, a conversation pattern determination unit that determines a conversation pattern of a robot that responds to a conversation pattern of a talker by referring to a database, an output unit that outputs the conversation pattern of the robot as an utterance speech, and an utterance speech of the robot A dialogue response detection unit for determining whether or not a dialogue person can hear is provided, and when it is judged by the dialogue response detection unit that a dialogue person cannot be heard, the output voice is adjusted and re-outputted.

  In the communication robot, the conversation pattern of the conversation person includes an NG word that is uttered when the conversation person cannot hear, and the conversation person reaction detection unit cannot hear the conversation person when the speech recognition unit detects the NG word. It is characterized by judging.

  Further, in the communication robot, the conversational person reaction detection unit determines that the conversational person has not been heard when the speech recognition unit detects a silent time during which the voice of the conversational person is not recognized.

  Also, in the communication robot, when the conversation person reaction detection unit determines that the conversation person has not heard the utterance voice, the output section changes any of the volume of the utterance voice, the speech speed, the speech frequency, and the speech inflection. And

  Further, in the communication robot, when the conversation person reaction detection unit determines that the conversation person cannot hear the uttered voice, the conversation pattern of the robot is changed to a content for confirming whether or not the conversation person has heard.

  Further, in the communication robot, when the conversation person reaction detection unit determines that the conversation person cannot hear the uttered voice, the conversation pattern of the robot is changed to a content for proposing another conversation.

  ADVANTAGE OF THE INVENTION According to this invention, it can readjust the robot's speech sound while confirming whether it was able to hear, and it can provide the robot which can aim at smooth communication, without making a conversation person feel stress.

It is a block diagram which shows the basic composition of the communication robot 100 of this invention. 3 is a flowchart of audio processing in the first embodiment. 10 is a flowchart of audio processing in the second embodiment. 10 is a flowchart of audio processing in the third embodiment. 10 is a flowchart of audio processing in the fourth embodiment.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(Embodiment 1)
FIG. 1 is a block diagram showing a basic configuration of a communication robot 100 of the present invention. With reference to FIG. 1, the structure and function of the communication robot 100 of the present invention will be described.

  The communication robot 100 according to the present invention includes an input unit 10, a database 20, a voice recognition unit 21, a conversation pattern determination unit 22, a dialog person reaction detection unit 23, and an output unit 30.

  The input unit 10 includes an utterance speed detection unit 11, an audio frequency detection unit 12, an audio inflection detection unit 13, and a volume detection unit 14. With the above configuration, the input unit 10 inputs the voice uttered by the conversation person 1, and detects the speech speed, the voice frequency, the voice inflection, and the volume from the voice of the conversation person 1.

  The database 20 registers a plurality of conversation patterns that may be spoken by a dialog person and a robot. The voice recognition unit 21 refers to the database 20 and recognizes the voice uttered by the conversation person 1 as a conversation pattern. The conversation pattern of the conversation person 1 recognized by the voice recognition section 21 is sent to the conversation pattern determination section 22 or the conversation person reaction detection section 23.

  The conversation pattern determination unit 22 refers to the database 20 to determine a robot conversation pattern that responds to the conversation pattern of the conversation person 1 and outputs it to the output unit 30.

  The output unit 30 includes an utterance speed control unit 31, an audio frequency control unit 32, an audio inflection control unit 33, and a volume control unit 34. The output unit 30 adjusts the speech speed, speech frequency, speech inflection, and volume of the conversation pattern of the robot, and outputs it to the conversation person 1 as the speech speech of the robot.

  The conversation person reaction detection unit 23 detects the reaction of the conversation person 1 with respect to the speech voice of the robot. Details of the conversation person reaction detection unit 23 will be described later.

  FIG. 2 is a flowchart of voice processing in the communication robot according to the first embodiment. The voice processing in the communication robot will be described with reference to FIG.

  First, the voice (S11) uttered by the talker 1 is input to the input unit 10 (S12), and the speech speed, voice frequency, voice inflection, and volume are detected. Speech recognition (S13). The conversation pattern determination unit 22 refers to the database 20 to determine a conversation pattern for the robot to respond to the conversation pattern of the conversation person 1 (S14). The conversation pattern of the robot is adjusted in speech rate, speech frequency, speech inflection, and volume by the output unit 30, and is output to the conversation person 1 as speech speech from a speaker or the like (S15).

  Next, a description will be given of the case where the utterance voice of the robot cannot be heard by the conversation person 1.

  The conversation person reaction detection unit 23 detects the reaction of the conversation person 1 whether or not the conversation person 1 has heard the voice uttered by the robot (S16). Specifically, an NG word indicating that the interlocutor 1 could not be heard, for example, words such as “I could not hear”, “What did you say?”, “Tell me again”, etc. Is registered in the database 20, and when the speech recognition unit 21 recognizes the NG word by referring to the database 20, the conversation person reaction detection unit 23 determines that the conversation person 1 has not heard the speech of the robot. Yes.

  When the conversation person 1 cannot hear the voice of the robot (S17), the conversation person reaction detection unit 23 counts up (+1) the NG counter (S19), and transmits the NG count number to the output unit 30. The output unit 30 re-adjusts the speech voice of the robot (S25 to S28) according to the count number of the NG counter (S22 to S24) and re-outputs it from the speaker or the like (S15).

  In the voice processing flow shown in FIG. 2, until the number of NG counts reaches a predetermined number N1 (S22), the volume of the spoken voice is increased (S25), and re-output (S15) from the speaker, etc. The reaction is confirmed (S16). When the situation where it cannot be heard continues, until the NG count reaches the predetermined number N2 (S23), the process proceeds to the step (S26) of changing the speech speed of the speech.

  When the situation that cannot be heard continues further, until the NG count reaches the predetermined number N3 (S24), the step of changing the speech frequency of the speech (S27), and when the NG count exceeds the predetermined number N3, The adjustment is continued so that the conversation person 1 can hear the speech voice of the robot, such as shifting to the step (S28) of changing the speech inflection.

  When the conversation person 1 can hear the speech voice of the robot (S17), the NG counter is reset (S18), and the process proceeds to the next conversation with the conversation person 1 (S19).

  Note that when adjusting the utterance voice uttered by the robot, the order of changing the utterance volume, the utterance speed, the utterance frequency, and the utterance inflection is not limited to the above embodiment.

  As described above, the communication robot according to the present invention performs readjustment while confirming whether or not the conversation voice can be heard by the conversation person, so that the conversation from the robot becomes unilateral and stresses the conversation person. Smooth communication with the interlocutor can be achieved without feeling.

(Embodiment 2)
FIG. 3 is a flowchart of voice processing in the communication robot according to the second embodiment. With reference to FIG. 3, the voice processing in the communication robot of the second embodiment will be described.

  In the communication robot according to the second embodiment, when the conversation situation of the conversation person is not improved even if the utterance voice of the robot is readjusted, the conversation pattern of the robot is changed to a content for confirming whether or not the conversation person 1 has heard, and is adjusted. It is characterized by adding (S32, S33). Since the basic configuration of the communication robot and the voice book processing flow are the same as those in the first embodiment, detailed description thereof is omitted.

  In the communication robot of the second embodiment, as shown in FIG. 3, even if the utterance volume, the utterance speed, the utterance frequency, the utterance inflection, etc. are changed with respect to the utterance voice of the robot, the listening situation is not improved (S31). The conversation pattern of the robot is changed (S32) so as to confirm whether or not the conversation person 1 can hear it.

  The confirmation of whether or not it is heard is to cause the robot to utter a conversation pattern that asks the talker how to hear, such as “Did you understand?” Or “Have you heard?”, For example.

  According to the communication robot of the second embodiment, the dialogue pattern that confirms whether or not the robot can hear it makes it possible for the interlocutor to feel the care of the robot, and the stress caused by repeating the same conversation pattern indefinitely. Can be reduced.

(Embodiment 3)
FIG. 4 is a flowchart of voice processing in the communication robot of the third embodiment. The voice processing in the communication robot of the third embodiment will be described with reference to FIG.

  The communication robot according to the third embodiment is characterized in that a step (S34) for changing the conversation pattern of the robot to another conversation is added when the listening state of the conversation person is not improved even by adjusting the utterance voice. Since the basic configuration of the communication robot and the voice book processing flow are the same as those in the first embodiment, detailed description thereof is omitted.

  In the communication robot according to the third embodiment, as shown in FIG. 4, even if the speech volume, speech speed, speech frequency, and speech inflection are changed with respect to the speech voice uttered by the robot, the listening situation is not improved (S31). The conversation pattern of the robot is changed to a content that suggests another conversation to the interlocutor.

  The suggestion of another conversation is to cause the robot to utter a conversation pattern such as “Let's talk another story” or “Is there another story?”.

  According to the communication robot of the third embodiment, when a conversation person cannot be heard by any means, the stress caused by repeating the conversation pattern of the same content from the robot forever can be reduced by developing the conversation in another conversation.

(Embodiment 4)
FIG. 5 is a flowchart of voice processing in the communication robot of the fourth embodiment. The voice processing in the communication robot of the fourth embodiment will be described with reference to FIG.

  The communication robot according to the fourth embodiment is characterized in that whether or not a conversation person can hear is determined based on the silence time detected by the speech recognition unit 21. Since the basic configuration of the communication robot and the voice book processing flow are the same as those in the first embodiment, detailed description thereof is omitted.

  In the communication robot of the fourth embodiment, as shown in FIG. 5, the speech recognition unit 21 measures the silent time when the talker does not speak with respect to the spoken voice of the robot, and the silent time is a predetermined time T (for example, 5 (S36), the NG counter is counted up (S21), and the utterance volume, utterance speed, utterance frequency, and utterance inflection of the utterance voice of the robot are changed according to the NG count number. is there.

  According to the communication robot of the fourth embodiment, by determining whether or not the conversation person can hear it due to the silent time of the conversation person, even if the conversation person cannot hear and react at all, the conversation voice of the robot can be easily heard. Can be adjusted.

  The present invention is not limited to the above-described embodiments, and various modifications can be made within the scope shown in the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. Is also included in the technical scope of the present invention.

DESCRIPTION OF SYMBOLS 1 Dialoguer 10 Input part 11 Speech rate detection part 12 Voice frequency detection part 13 Voice inflection detection part 14 Volume detection part 20 Database 21 Speech recognition part 22 Conversation pattern determination part 23 Dialogue person reaction detection part 30 Output part 31 Speech speed control part 32 Voice Frequency Control Unit 33 Voice Inflection Control Unit 34 Volume Control Unit 100 Communication Robot

Claims (6)

An input unit for inputting the voice of the conversation person;
A database in which a plurality of conversation patterns uttered by the interlocutor and the robot are registered;
A voice recognition unit that recognizes a conversation pattern from the voice of the conversation person with reference to the database;
A conversation pattern determination unit that determines the conversation pattern of the robot that responds to the conversation pattern of the conversation person with reference to the database;
An output unit that outputs the conversation pattern of the robot as an utterance voice;
A conversational reaction detection unit for determining whether the conversational person can hear the utterance voice of the robot,
A communication robot characterized in that, when it is determined by the interlocutor reaction detection unit that the interlocutor has not been heard, the output unit adjusts and re-outputs the uttered voice.   The conversation pattern of the conversation person includes an NG word that is spoken when the conversation person cannot hear, and the conversation person reaction detection unit cannot hear the conversation person when the voice recognition unit detects the NG word. The communication robot according to claim 1, wherein the communication robot is determined.   2. The communication robot according to claim 1, wherein the interlocutor reaction detection unit determines that the interlocutor has not been heard when a silent time in which the speech of the interlocutor is not recognized by the speech recognition unit is detected. .   When the conversation person reaction detection unit determines that the conversation person has not been able to hear the uttered speech, the output unit changes any one of the volume of the utterance speech, the speech speed, the speech frequency, and the speech inflection. The communication robot according to any one of claims 1 to 3.   2. When the conversation person reaction detection unit determines that the conversation person cannot hear the uttered voice, the conversation pattern of the robot is changed to a content for confirming whether or not the conversation person has heard. 4. The communication robot according to any one of 3.   4. The robot according to claim 1, wherein when the interlocutor reaction detection unit determines that the interlocutor cannot hear the uttered voice, the conversation pattern of the robot is changed to a content for proposing another conversation. Crab communication robot.

Images