JP2006243555A - Response determination system, robot, event output server, and response determining method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To determine an appropriate response while taking the position of a speaker into consideration according to the voice of the speaker. <P>SOLUTION: A robot 100 inputs events indicating that data regarding voices input from a 1st microphone 306, a 2nd microphone 316, and a 3rd microphone 326 prepared for a 1st participant 300, a 2nd participant 310, and a 3rd participant 320 meet designated conditions together with voice identification information made to correspond to the speakers. The robot 100 acquires position information of a speaker specified with voice identification information made to correspond to each of the input events, and determines its speech and behavior based upon the event and the position information of the speaker. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a response determination system, a robot, an event output server, and a response determination method.

  Patent Document 1 discloses a communication robot that executes a communication action with a communication target having a unique wireless tag. This communication robot individually recognizes at least a tag information database that records tag information, an acquisition unit that acquires tag information from a communication target, and one or a plurality of communication targets that exist in the vicinity or surroundings based on the acquisition result of the acquisition unit Recognizing means, specifying means for specifying one communication target for executing the communication action based on the recognition result of the recognition means, and executing means for executing the communication action for the communication target specified by the specifying means.

The conventional communication robot having such a configuration operates as follows. When performing a communication action, the communication robot acquires tag information from a communication target by an acquisition unit. The recognizing means individually recognizes a person as a communication target. The specifying unit specifies one participant among the participants existing near or around the communication robot based on the recognition result of the recognition unit. The execution means executes communication behavior for the participant. Thus, it is described that participants existing in the vicinity or in the vicinity can be individually recognized and communication behavior suitable for the specified participant can be taken.
JP 2004-216513 A

  However, in the conventional robot (system), it is only possible to recognize a communication target existing in the vicinity of or around the robot, and the robot can only speak spontaneously based on history information and the like.

  Even if the partner of the dialogue can be grasped, the robot cannot communicate smoothly with the partner unless the utterance content and state of the partner can be recognized. In particular, when communicating with multiple partners, it was difficult to take appropriate actions by recognizing the utterances and circumstances of each partner.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a technique for taking an appropriate response in consideration of the position of the speaker according to the voice of the speaker.

According to the present invention,
An event indicating that data related to speech input from the speech input device assigned to the speaker matches a predetermined condition is input together with speech identification information associated with the speaker,
A correspondence determining unit for determining a correspondence to the speaker;
A location information acquisition unit that acquires location information of the speaker specified by the voice identification information associated with the event input by the correspondence determination unit;
Including
The correspondence determining unit determines the correspondence based on the event and the position information of the speaker, and provides a correspondence determining system.

  Here, the voice-related data includes voice data and a voice recognition result based on the voice data. The voice input device can be, for example, a microphone. The microphone may be a close-talking microphone.

  Here, for example, the response determination system is capable of responding to a speaker's voice such as a robot control system that controls an autonomously moving or interactive robot, a voice dialog system, or an information search system using voice recognition. It can be a system to do.

  According to the correspondence determination system of the present invention, it is possible to take a correspondence adapted to both the event based on the voice input from the voice input device assigned to the speaker and the speaker position. For example, when the correspondence determination system of the present invention is a robot control system, the correspondence determination unit can determine the behavior of the robot based on the event and the position information of the speaker. Thereby, according to the event, the speech adapted to the speaker position can be executed, such as the robot facing the speaker, approaching the speaker, or speaking based on the speaker position.

  According to the present invention, processing is performed on the basis of voice input from a voice input device such as a close-talking microphone, so that a robust correspondence determination system with high resistance to noise or the like is provided. Furthermore, by making the voice input device wirelessly connected to the correspondence determination system, the physical freedom of the speaker can be kept high. Further, according to the present invention, processing is performed according to the voice and position information of the speaker, so that a correspondence determination system with high interface transparency is provided.

  The correspondence determination system of the present invention may be provided in one system, or may be provided in a distributed manner in a plurality of systems connected to each other via a network. For example, when the response determination system of the present invention is a robot control system, the system may be provided in the robot or in a server that can communicate with the robot through a network such as wireless communication. In addition, some functions of the response determination system may be provided in the robot, and other functions may be provided in the server.

  In a robot that moves autonomously or interacts with a speaker, control is desired so that the speaker and the robot can communicate smoothly. Conventionally, when a robot interacts with a user or performs some operation on the user, attempts have been made to identify the user by image recognition or voice characteristics. However, with this method, it is difficult to recognize the speaker unless the speaker is in the immediate vicinity of the robot or the conditions such as no noise are satisfied. Also, as described in Patent Document 1, an attempt has been made to recognize a communication target by holding a wireless tag unique to the communication target. However, as described above, this method is similar to the other conventional techniques in that it can only recognize communication objects in the vicinity and it is difficult to grasp the content of the speaker's utterance.

  According to the correspondence determination system of the present invention, since the voice of the speaker is input from the voice input device and the voice identification information is given, it is easy to know which speaker has made what kind of utterance. can do. Moreover, even when communicating with a plurality of speakers, it is possible to grasp the utterance contents of each speaker.

  The correspondence determination system of the present invention inputs the data related to the voice together with the voice identification information, detects whether or not the data related to the voice meets a predetermined condition, and if the condition is met, And an event output unit that outputs an event indicating that it matches to the correspondence determination unit together with the voice identification information.

  When the correspondence determination system of the present invention is a robot control system, the event output unit may be provided in the robot, or may be provided in a server that can communicate with the robot wirelessly or the like. For example, in a form in which a robot interacts simultaneously with a large number of speakers, a plurality of servers are each provided with an event output unit function, and a plurality of servers are transferred to a robot or a single server having a behavior determining unit function. It can also be configured such that data such as events is input.

  In the response determination system of the present invention, the event output unit monitors the power of the voice input from the voice input device, and the voice input is performed when a state where the voice power is equal to or lower than a predetermined value continues for a predetermined time. An event indicating a failure can be output.

  With such a configuration, when voice input from the voice input device is interrupted due to some trouble, an appropriate response can be taken in consideration of the position of the speaker.

  In the response determination system of the present invention, the event output unit monitors the power of the voice input from the voice input device, and when the voice power becomes a predetermined value or more, an event indicating utterance detection is detected. Can be output.

  With such a configuration, when a speaker's utterance is input to the voice input device, an appropriate response can be taken in consideration of the speaker's position. For example, if the response determination system is a robot control system and an event indicating utterance detection is output, and the robot is controlled so that it faces the direction of the speaker, the speaker You can recognize that the robot is facing you. By such an adaptive operation, the interface transparency of the robot can be increased and the operation of the robot can be easily understood by the speaker. The utterance can also be detected based on the harmonic structure of the voice, the duration of the voice power being a predetermined value or more, and the event output unit detects such a state to indicate the utterance detection. Events can also be output.

  The correspondence determination system according to the present invention inputs a voice input from a voice input device assigned to a speaker together with voice identification information associated with the speaker, recognizes the voice, and obtains a voice recognition result. The speech recognition unit may further include a speech recognition unit that outputs the speech recognition information, and the event output unit may display an event indicating a speech recognition result when the speech recognition result is output from the speech recognition unit. Can be output together.

  With such a configuration, when a speaker's utterance is input to the voice input device and voice recognition is performed, an appropriate response can be taken in consideration of the position of the speaker.

  The correspondence determination system of the present invention may further include a tag reader that reads tag identification information from an identification tag given to the speaker, and the position information acquisition unit is based on the tag identification information read by the tag reader, The position information of the speaker can be acquired.

  The identification tag is, for example, an active / passive RFID (Radio Frequency Identification) tag, an ultrasonic tag, an infrared tag, or the like that can be read by radio waves, electromagnetic waves, ultrasonic waves, infrared rays, or the like.

  The correspondence determination system of the present invention may further include an identification information storage unit that associates the voice identification information with the tag identification information of the same speaker, and the position information acquisition unit includes the event output The position information of the identification tag having the corresponding tag identification information can be obtained by referring to the identification information storage unit based on the voice identification information associated with the event output by the unit.

  In the correspondence determination system according to the present invention, the behavior determination unit generates an event indicating that data related to speech input from a plurality of speech input devices respectively assigned to a plurality of speakers matches a predetermined condition. It can be input together with the voice identification information associated with the person.

  According to the present invention, even when communicating with a plurality of opponents, data related to the sound from each speaker and position information of each speaker are acquired in association with each speaker. Therefore, it is possible to take appropriate measures for each of a plurality of opponents.

  In the correspondence determination system of the present invention, the position information acquisition unit can acquire position information of the speaker with respect to the robot, and the behavior determination unit performs a behavior in which the robot is aware of the position of the speaker. The behavior of the robot can be determined to do.

  This facilitates acquisition of the speaker position information and facilitates control when the robot faces the speaker or approaches the speaker.

According to the present invention,
An autonomous mobile or interactive robot,
A response determination system according to any of the above,
A behavior execution unit that executes the correspondence determined by the correspondence determination unit as behavior;
Is provided.

  According to the robot of the present invention, it is possible to execute speech adapted to both the event based on the voice input from the voice input device given to the speaker and the speaker position. Further, the robot of the present invention can be configured to execute speech while facing a speaker. In such a robot, control is desired so that the speaker and the robot can communicate smoothly. According to the robot of the present invention, since the voice of the speaker is inputted from the voice input device and the voice identification information is given, it is easy to grasp which speaker has made what kind of utterance. Can do. Moreover, even when communicating with a plurality of speakers, it is possible to grasp the utterance contents of each speaker. As a result, the speaker and the robot can communicate smoothly.

According to the present invention,
An event output server connected via a network to a communication terminal device including an audio output unit that outputs audio input from an audio input device and an autonomous mobile or interactive robot, and relays these,
The robot is
An event indicating that the voice-related data input from the voice input device assigned to the speaker matches a predetermined condition is input together with the voice identification information associated with the speaker, and is associated with the event. Obtaining the position information of the speaker specified by the voice identification information, and executing the behavior determined based on the event and the position information of the speaker,
The event output server is
A voice input unit that receives the voice output by the voice output unit from the communication terminal device together with the voice identification information;
An event output unit that detects whether or not the voice input by the voice input unit matches a predetermined condition, and outputs an event corresponding to the condition together with the voice identification information when the condition is met;
A data output unit for transmitting the event output by the event output unit to the robot together with the voice identification information;
An event output server is provided.

  According to the event output server of the present invention, since the event is transmitted to the robot together with the voice identification information, in the robot, the event based on the voice input from the voice input device given to the speaker, the speaker position, It is possible to execute behavior adapted to both.

  In the event output server of the present invention, the event output unit monitors the power of the voice input from the voice input device, and the voice input is performed when the voice power is below a predetermined value for a predetermined time. An event indicating a failure can be output.

  In the event output server of the present invention, the event output unit monitors the power of the voice input from the voice input device, and when the voice power becomes equal to or higher than a predetermined value, an event indicating speech detection is detected. Can be output.

  In the event output server of the present invention, the voice input from the voice input device assigned to the speaker is input together with the voice identification information associated with the speaker, the voice is voice-recognized, and the voice recognition result is obtained. The speech recognition unit may further include a speech recognition unit that outputs the speech recognition information, and the event output unit may display an event indicating a speech recognition result when the speech recognition result is output from the speech recognition unit. Can be output together.

According to the present invention,
Inputting an event indicating that data related to voice input from a voice input device assigned to the speaker matches a predetermined condition, together with voice identification information associated with the speaker;
Obtaining positional information of the speaker specified by the voice identification information associated with the event input in the step of inputting the event together with the voice identification information;
Determining a response to the speaker based on the event and the location information of the speaker;
A correspondence determination method characterized by including:

  According to the correspondence determination method of the present invention, an autonomous mobile or interactive robot can be controlled, and in the step of determining the correspondence to the speaker, the behavior of the robot is determined based on the event and the position information of the speaker. be able to.

  According to the correspondence determination method of the present invention, it is possible to determine the correspondence adapted to both the event based on the voice input from the voice input device assigned to the speaker and the speaker position. For example, when the robot is controlled by the correspondence determination method of the present invention, the robot faces the speaker, approaches the speaker, or speaks based on the speaker position, depending on the event. Can be adapted to the person's position.

  In the correspondence determining method of the present invention, before the step of inputting the event together with the voice identification information, the voice-related data is input together with the voice identification information, and whether the voice-related data meets a predetermined condition or not. And detecting an event indicating that the condition is met together with the voice identification information when the condition is met.

  In the correspondence determining method of the present invention, the step of outputting the event together with the voice identification information includes the step of monitoring the power of the voice input from the voice input device, and the state where the voice power is equal to or lower than a predetermined value. And outputting an event indicating a voice input failure when continuing for a predetermined time.

  In the correspondence determination method of the present invention, the step of outputting the event together with the voice identification information includes the step of monitoring the power of the voice input from the voice input device, and the voice power is equal to or higher than a predetermined value. A step of outputting an event indicating utterance detection.

  In the correspondence determination method of the present invention, a voice input from a voice input device assigned to a speaker is input together with voice identification information associated with the speaker, the voice is voice-recognized, and a voice recognition result is obtained. The step of outputting together with the voice identification information may further include the step of outputting the event together with the voice identification information, when the voice recognition result is output, an event indicating a voice recognition result is output from the voice recognition result. Can be output together.

  In the correspondence determining method of the present invention, the step of inputting the event includes an event indicating that data relating to speech input from a plurality of speech input devices respectively assigned to a plurality of speakers matches a predetermined condition. It can be input together with the voice identification information associated with each speaker.

  It should be noted that any combination of the above-described constituent elements and a conversion of the expression of the present invention between a method, an apparatus, a system, a recording medium, a computer program, etc. are also effective as an aspect of the present invention.

  According to the present invention, an appropriate response can be taken in consideration of the position of the speaker according to the voice of the speaker.

  Next, the best mode for carrying out the invention will be described in detail with reference to the drawings. In the following drawings, configurations of parts not related to the essence of the present invention are omitted.

  In the following embodiments, the case where the correspondence determination system and the correspondence determination method are a robot control system and a robot control method for controlling a robot that moves autonomously and interacts with a speaker will be described as an example.

  In the following embodiments, for a person who wants to communicate with a robot, a microphone, a communication terminal device including a voice output unit that transmits voice input to the microphone to the robot, and identification for the robot to acquire position information A tag is given.

(First embodiment)
In the present embodiment, the correspondence determination system is incorporated in the robot.

FIG. 1 is a schematic diagram showing the relationship between a robot and a participant who is a speaker in the present embodiment.
Here, communication between the robot 100 and the first participant 300, the second participant 310, and the third participant 320 will be described as an example.

  The first participant 300, the second participant 310, and the third participant 320 are a first identification tag 302, a first audio output unit 304, a first microphone 306, a second microphone 306, respectively. The identification tag 312, the second audio output unit 314, and the second microphone 316, and the third identification tag 322, the third audio output unit 324, and the third microphone 326 are held.

Hereinafter, the first participant 300 will be described as an example.
The first microphone 306 inputs the voice of the first participant 300. The first microphone 306 can be a headset microphone so that participants can move easily. The first microphone 306 is connected to the first audio output unit 304 assigned to the first participant 300.

  The first audio output unit 304 is, for example, a portable wireless communication device. The first audio output unit 304 can be a PDA (Personal Digital Assistance), for example. The first sound output unit 304 transmits the sound input from the first microphone 306 to the robot 100. At this time, the first voice output unit 304 transmits voice identification information for identifying itself together with the voice to the robot 100. The first participant 300 can carry the first audio output unit 304 by putting the first audio output unit 304 in a bag having a mesh-like pocket or a backpack pocket, for example.

  The robot 100 and the first audio output unit 304 can communicate via a wireless LAN, for example. When the robot 100 and the first audio output unit 304 communicate via a wireless LAN, for example, communication can be performed using TCP (Transmission Control Protocol) / IP (Internet Protocol). In this case, for example, a port number or an IP address assigned to each device can be used as the voice identification information of the first voice output unit 304.

  The first identification tag 302 is, for example, an active / passive RFID (Radio Frequency Identification) tag, an ultrasonic tag, an infrared tag, or the like that can be read by radio waves, electromagnetic waves, ultrasonic waves, infrared rays, or the like. The first identification tag 302 stores tag identification information unique to the tag. The first participant 300 attaches the first identification tag 302 to a part of his / her body such as the chest. The robot 100 can identify the first participant 300 by reading the tag identification information from the first identification tag 302. Further, the robot 100 can acquire the position information of the first identification tag 302 based on the reading strength of the first identification tag 302 and the like. Here, the position information may be the position information of the first participant 300 with respect to the robot 100, such as the distance between the robot 100 and the first participant 300 and the direction of the first participant 300 with respect to the robot 100. it can.

  For example, when the first identification tag 302 is an ultrasonic tag, ultrasonic waves are periodically transmitted from the first identification tag 302 and received by the reader of the robot 100. The robot 100 can acquire the position information of the first participant 300 based on the arrival time and reception angle of the ultrasonic wave received by the reader.

  The second identification tag 312 of the second participant 310 and the third identification tag 322 of the third participant 320 have the same configuration as the first identification tag 302 of the first participant 300. The second audio output unit 314 of the second participant 310 and the third audio output unit 324 of the third participant 320 also have the same configuration as the first audio output unit 304 of the first participant 300. Have

  The robot 100 is configured to be able to simultaneously receive the voices of a plurality of participants, and is configured to be able to simultaneously acquire position information of the plurality of participants. In the present embodiment, the robot 100 is an autonomous mobile type.

FIG. 2 is an external configuration diagram showing an example of the robot in the present embodiment.
The robot 100 is configured, for example, by connecting the body 1 and the head 2. A wheel 3A and a wheel 3B are attached to the lower part of the body part 1 on the left and right, respectively, and these wheels can rotate back and forth independently.

  The head 2 can rotate within a predetermined range with respect to a vertical axis that is vertically attached to the body 1 and a horizontal axis that is installed at an angle of 90 degrees with respect to the vertical axis. The vertical axis is installed so as to pass through the center of the head 2, and the horizontal axis is installed horizontally through the center of the head 2 with the body 1 and the head 2 facing the front and horizontally in the left-right direction. Yes. That is, the head 2 can rotate within a predetermined range with two degrees of freedom, left and right and up and down.

  A speaker 12 and an internal microphone 13 are provided on the surface of the body portion 1. A CCD camera 21A and a CCD camera 21B are provided on the surface of the head 2.

FIG. 3 is a block diagram illustrating an example of the electrical configuration of the robot 100.
The body 1 includes a controller 10 for controlling the entire robot, a battery 11 serving as a power source for the robot, a speaker 12, an internal microphone 13, an actuator 14A and an actuator 14B for moving two wheels, and a communication interface (I / F). 24), the tag reader 26, and the like are accommodated.

  The internal microphone 13 collects ambient sounds including utterances from a specific speaker, and sends the obtained audio data to the controller 10.

  The controller 10 includes a CPU 10A and a memory 10B. The CPU 10A executes various processes by executing a control program stored in the memory 10B.

  The communication interface 24 is a communication terminal device (first audio output unit in FIG. 1) assigned to the participants (first participant 300, second participant 310, and third participant 320 in FIG. 1). 304, the second audio output unit 314, and the third audio output unit 324).

  The tag reader 26 is configured to identify identification tags (first identification tag 302, FIG. 1, first identification tag 302, second participant 310, and third participant 320 in FIG. 1). The tag identification information is read from the second identification tag 312 and the third identification tag 322).

  The head 2 houses a CCD camera 21A and a CCD camera 21B, an actuator 22A and an actuator 22B for rotating the head 2, and the like.

  The CCD camera 21 </ b> A and the CCD camera 21 </ b> B capture the surrounding situation and send the obtained image data to the controller 10.

  The actuator 22A and the actuator 22B rotate the head 2 of the robot 100 up and down and left and right.

  The controller 10 appropriately reads information from the memory 10B based on audio data obtained via the internal microphone 13 and the communication interface 24 and image data obtained from the CCD camera 21A and the CCD camera 21B, and the situation of the participants and the participants And the behavior of the corresponding robot 100 is determined.

  The controller 10 controls the actuator 14A, the actuator 14B, the actuator 22A, the actuator 22B, the tag reader 26, and the like to cause the robot 100 to execute the determined operation. In addition, the controller 10 generates a synthesized sound, supplies the synthesized sound to the speaker 12, and causes the robot 100 to output the determined utterance.

FIG. 4 is a block diagram showing in detail the configuration of the controller 10 of the robot 100 according to the present embodiment. FIG. 4 shows functional unit blocks, not hardware configuration of the controller 10.
The controller 10 (correspondence determination system) of the robot 100 receives predetermined data related to voice input from the first microphone 306, the second microphone 316, the third microphone 326, etc. (voice input device) assigned to the speaker. The event indicating that the above condition is met is input together with the voice identification information associated with the speaker, and the behavior determining unit 110 (correspondence determining unit) that determines the response to the speaker and the correspondence determining unit input A position information acquisition unit 108 that acquires the position information of the speaker specified by the voice identification information associated with the event, and the behavior determination unit 110 is based on the event and the position information of the speaker. 100 behaviors (correspondence) are determined.

  The controller 10 of the robot 100 includes a voice input unit 102, a voice recognition unit 104, an event output unit 106, a position information acquisition unit 108, a behavior determination unit 110, a part of the behavior execution unit 112, a voice recognition dictionary. 114, a condition storage unit 116, a corresponding behavior storage unit 118, an identification information storage unit 120, a robot behavior storage unit 130, and a scenario storage unit 132. The speech execution unit 112 includes a mechanical control unit 134, a voice synthesis unit 136, and an output unit 138 realized by the controller 10, and an actuator 14A, an actuator 14B, an actuator 22A, an actuator 22B, and the speaker 12.

  The communication interface 24 receives various data from a plurality of audio output units (the first audio output unit 304, the second audio output unit 314, and the third audio output unit 324) via the network 400.

  The voice input unit 102 inputs the voice data received by the communication interface 24 from the plurality of voice output units in association with each voice identification information. The voice input unit 102 outputs the input voice data to the voice recognition unit 104 and the event output unit 106 together with the voice identification information. The voice input unit 102 also receives voice data collected by the internal microphone 13 and outputs the voice data to the voice recognition unit 104.

  The voice input unit 102 can simultaneously input both the voice data input from the communication interface 24 or the voice data input from the internal microphone 13, but the input from either one is turned off and the other is started. It is also possible to turn on only the input.

  Processing of each component when the voice input unit 102 inputs voice data from the internal microphone 13 will be described later, and processing functions when the voice input unit 102 inputs voice data from the communication interface 24 will be described below.

  The voice recognition unit 104 recognizes the voice data input by the voice input unit 102. The speech recognition dictionary 114 includes a speech recognition word storage unit that stores speech recognition vocabulary that is a set of speech recognition words. The voice recognition unit 104 performs matching between the voice data input by the voice input unit 102 and the voice recognition vocabulary stored in the voice recognition dictionary 114. When the voice data is voice-recognized, the voice recognition unit 104 outputs a voice recognition result to the event output unit 106 in association with the voice identification information.

  The event output unit 106 detects whether or not the voice data meets a predetermined condition based on the voice data output from the voice input unit 102 and the voice recognition result output from the voice recognition unit 104. If they match, an event indicating that the condition is matched is output in association with the voice identification information. When the event output unit 106 acquires the speech recognition result from the speech recognition unit 104, the event output unit 106 also outputs the speech recognition result together with the event.

  The condition storage unit 116 stores a predetermined condition and an event indicating that the condition is met in association with each other. The event output unit 106 refers to the condition storage unit 116 to detect whether or not the audio data meets a predetermined condition.

  When the event output unit 106 outputs an event and voice identification information, the behavior determination unit 110 acquires the position information of the participant specified by the voice identification information from the position information acquisition unit 108. The identification information storage unit 120 stores voice identification information of each participant and tag identification information in association with each other. When the behavior determination unit 110 acquires the event and the voice identification information from the event output unit 106, the behavior determination unit 110 refers to the identification information storage unit 120, reads the tag identification information associated with the voice identification information, and determines the position of the tag identification information. The position information acquisition unit 108 is requested to acquire information. The position information acquisition unit 108 performs processing for reading the identification tag of the participant by the tag reader 26 and acquires the position information of the identification tag having the target tag identification information. Next, the position information acquisition unit 108 notifies the behavior determination unit 110 of this position information. The behavior determination unit 110 acquires the position information notified from the position information acquisition unit 108 as the position information of the intended participant.

  The behavior determination unit 110 determines the behavior to be performed by the behavior execution unit 112 based on the event acquired from the event output unit 106 and the location information of the participant. The corresponding behavior storage unit 118 stores the event and the behavior of the robot 100 corresponding to the event in association with each other. The behavior determining unit 110 refers to the corresponding behavior storage unit 118 and reads the behavior of the robot 100 corresponding to the event acquired from the event output unit 106.

  The robot behavior storage unit 130 stores utterance data and motion data of the robot in a specific situation. The scenario storage unit 132 stores scenario information.

  The behavior determining unit 110 refers to the robot behavior storage unit 130 and the scenario storage unit 132 as necessary based on the behavior of the robot 100 read from the corresponding behavior storage unit 118 and the position information of the speaker. To decide.

  The behavior determination unit 110 sends the determined behavior as a command to the mechanical control unit 134 and the speech synthesis unit 136. The mechanical control unit 134 generates a control signal for driving the actuator 14A, the actuator 14B, the actuator 22A, and the actuator 22B based on the command sent from the behavior determining unit 110, and outputs the control signal to the actuators 14A, 14B, 22A, And 22B. Thus, the actuators 14A, 14B, 22A, and 22B are driven according to the control signal.

  The speech synthesizer 136 generates a synthesized sound based on the command sent from the behavior determining unit 110. The output unit 138 is supplied with digital data of the synthesized sound from the voice synthesizer 136. The output unit 138 D / A converts the digital data into analog voice data, and the speaker. 12 for output.

Next, the processing function of each component when the voice input unit 102 inputs voice data from the internal microphone 13 will be described.
In this case, voice identification information is not associated with voice data input by the voice input unit 102. Although not shown here, the controller 10 can recognize the speaker using image data input from the CCD camera 21A or the CCD camera 21B, for example. When the speaker can be recognized, the voice recognition unit 104, the event output unit 106, and the behavior determination unit 110 can perform the same processing as when the voice input unit 102 inputs voice data from the communication interface 24 described above. .

  If the speaker cannot be recognized, the voice data input by the voice input unit 102 and the voice recognition result recognized by the voice recognition unit 104 are directly input from the voice input unit 102 and the voice recognition unit 104 to the behavior determining unit 110. The behavior determination unit 110 can determine the behavior of the robot 100 with reference to the robot behavior storage unit 130 and the scenario storage unit 132. Even when the speaker cannot be recognized, the voice data input by the voice input unit 102 and the voice recognition result recognized by the voice recognition unit 104 are input to the event output unit 106, and the event output unit 106 meets a predetermined condition. It is also possible to perform processing for detecting whether or not.

FIG. 5 is a diagram illustrating an example of an internal configuration of the condition storage unit 116. Hereinafter, a description will be given with reference to FIG.
The condition storage unit 116 includes an event column and a condition column. The event column includes a number column and a content column.

  For example, the content of the event with the number “1” is “voice input failure” and the condition is “no voice input for a predetermined time”. That is, this condition is met when the voice input unit 102 cannot input voice from the first participant 300, the second participant 310, and the third participant 320 due to some trouble.

  Possible causes of the voice input failure include, for example, a failure such as a power-off of the communication terminal device including each voice output unit, a disconnection between each microphone and the corresponding voice output unit, and the like. In the present embodiment, each microphone is configured such that the voice power does not continue to be zero due to surrounding voice or noise peculiar to the voice output unit even when there is no speech voice of each participant. Therefore, the event output unit 106 can detect that the condition of “no audio input for a predetermined time” is met when the state where the power of the audio output from the audio input unit 102 is zero continues for a predetermined time or longer. . Also, for example, whether test data is periodically transmitted from the robot 100 to each communication terminal device, and whether or not the condition “no voice input for a predetermined time” is met, depending on whether or not there is a response within a predetermined time. Can also be detected. The event output unit 106 can detect that the condition “no voice input for a predetermined time” is met when there is no response within a predetermined time after transmitting test data to each communication terminal device.

  For example, the content of the event with the number “2” is “speech detection”, and the condition is “sound level is equal to or higher than a predetermined threshold”. That is, this condition is met when there is an utterance from any of the first participant 300, the second participant 310, and the third participant 320.

  The event output unit 106 can detect that the condition that “the audio level is equal to or higher than the predetermined threshold” is met when the power of the audio output from the audio input unit 102 is equal to or higher than the predetermined threshold. The event output unit 106 collects and analyzes the voice of the identification information for a predetermined time when the power of the voice output from the voice input unit 102 exceeds a predetermined threshold value, It is also possible to determine whether or not it is included and detect whether or not the condition is met according to the determination result.

  Further, for example, the content of the event with the number “3” is “voice recognition result”, and the condition is “voice recognition result acquisition”. That is, this condition is met when the voice recognition unit 104 outputs a voice recognition result.

FIG. 6 is a diagram illustrating an example of the internal configuration of the corresponding behavior storage unit 118. Hereinafter, a description will be given with reference to FIG.
The corresponding behavior storage unit 118 has an event number column and a behavior column. In the event number column, a number corresponding to the number in the number column of the event column of the condition storage unit 116 is stored.

  For example, the behavior when the event with the number “1” is output is “(1) Approaching the corresponding speaker. (2) Switching to voice input by the internal microphone”. The behavior determination unit 110 causes the behavior execution unit 112 to perform an action in which the robot 100 approaches the corresponding speaker based on the position information of the participant associated with the event. At this time, the behavior determining unit 110 sequentially acquires the location information of the corresponding participant from the location information acquiring unit 108, and executes the behavior (2) when the distance from the participant is within a predetermined distance. . As the behavior of (2), the voice input unit 102 selectively inputs voice data from the internal microphone 13 and directly inputs the voice of the corresponding participant from the internal microphone 13 of the robot 100.

  Further, for example, the behavior when the event of the number “2” is output is “Look toward the speaker”. The behavior determination unit 110 causes the behavior execution unit 112 to perform an action in which the robot 100 faces the corresponding speaker based on the participant position information associated with the event.

  For example, when the event of the number “3” is output, the behavior is “(1) facing the speaker. (2) corresponding voice output”. The behavior determination unit 110 causes the behavior execution unit 112 to perform an action in which the robot 100 faces the corresponding speaker based on the participant position information associated with the event. Subsequently, the behavior determining unit 110 refers to the robot behavior storage unit 130 and the scenario storage unit 132 to determine the output of the response voice corresponding to the speech recognition result, and causes the behavior execution unit 112 to execute it.

FIG. 7 is a diagram illustrating an example of an internal configuration of the identification information storage unit 120.
The identification information storage unit 120 includes a name field, a voice identification information field, and a tag identification information field.
The name field stores the names of the participants. In the voice identification information column, identification information of each voice output unit is stored. Tag identification information for each identification tag is stored in the tag identification information column.

  Here, for example, voice identification information “1111” and tag identification information “0001” are associated with the participant named “Sakura”. Also, voice identification information “1112” and tag identification information “0002” are associated with the participant named “Momo”. Also, voice identification information “1113” and tag identification information “0003” are associated with the participant named “Taro”.

  Here, the identification information storage unit 120 does not necessarily have a name field, but by storing the name of the participant in the name field, the robot 100 can call the name to the participant. And the robot 100 can be more smoothly communicated. In addition, the identification information storage unit 120 may further include a column for storing information related to the participants such as the gender and age of each participant. Thereby, the robot 100 can respond according to a participant.

FIG. 8 is a flowchart showing a processing procedure of the controller 10 of the robot 100 in the present embodiment.
The processing procedure (correspondence determination method) of the controller 10 includes data relating to voice input from the first microphone 306, the second microphone 316, the third microphone 326, etc. (voice input device) assigned to the speaker. It is input together with the voice identification information, and it is detected whether or not the data related to the voice meets a predetermined condition (S100). If the condition is met (YES in S100), an event indicating that the condition is met is set to voice. A step of outputting together with the identification information (S102), and an event indicating that the predetermined condition is met together with the voice identification information associated with the speaker, and the voice identification information associated with the input event The robot for the speaker is obtained based on the step (S104) of acquiring the speaker position information specified by the step S104, the event, and the speaker position information. 00 behavior includes the steps of determining (corresponding) (S106), the.

This will be specifically described below.
The event output unit 106 constantly monitors the voice data input from the voice input unit 102 and the voice recognition result input from the voice recognition unit 104. When it is detected that the data relating to the voice matches any of the conditions stored in the condition storage unit 116 (YES in S100), the event output unit 106 associates the event corresponding to the condition with the voice identification information. It outputs to the behavior determination part 110 (S102).

  When the event output unit 106 outputs an event and voice identification information, the behavior determination unit 110 reads the tag identification information associated with the voice identification information with reference to the identification information storage unit 120. Next, the behavior determination unit 110 notifies the position information acquisition unit 108 of the tag identification information. The position information acquisition unit 108 detects an identification tag having the tag identification information read by the behavior determining unit 110, calculates its position, and notifies the behavior determining unit 110. Thereby, the behavior determining unit 110 acquires the location information of the participant having the tag identification information corresponding to the event (S104).

  The behavior determination unit 110 reads the behavior corresponding to the event with reference to the corresponding behavior storage unit 118. The behavior determining unit 110 also refers to the robot behavior storage unit 130 and the scenario storage unit 132 as necessary. The behavior determination unit 110 determines the behavior of the robot 100 based on the information read from the corresponding behavior storage unit 118, the robot behavior storage unit 130, the scenario storage unit 132, and the position information of the participants (S106). Subsequently, the behavior determining unit 110 notifies the mechanical control unit 134 and the speech synthesis unit 136 of the determined behavior.

  The behavior executing unit 112 such as the mechanical control unit 134 and the voice synthesizing unit 136 executes the behavior determined by the behavior determining unit 110 (S108).

  When the series of behaviors determined by the behavior determination unit 110 is finished, it is determined whether or not to end the processing of the controller 10 (S110). If not finished (NO in S110), the process returns to step S100. In step S110, when the process of the controller 10 is terminated (YES in S110), the process is terminated.

Next, a specific example will be described. Hereinafter, description will be made with reference to FIGS.
For example, an example in which the robot 100 has a conversation with the first participant 300 (Sakura), the second participant 310 (Momo), and the third participant 320 (Taro) will be described.

  The first participant 300 “Sakura” is assigned the first identification tag 302 having the tag identification information “0001” and the first voice output unit 304 having the voice identification information “1111”. ing. The voice uttered by “Sakura” is collected by the first microphone 306 and transmitted from the first voice output unit 304 to the robot 100 in association with the voice identification information “1111”.

  The second participant 310 “Momo” is provided with a second identification tag 312 having tag identification information “0002” and a second voice output unit 314 having voice identification information “1112”. Has been. The voice uttered by “Momo” is collected by the second microphone 316 and transmitted from the second voice output unit 314 to the robot 100 in association with the voice identification information “1112”.

  The third participant 320 “Taro” is provided with a third identification tag 322 having tag identification information “0003” and a third voice output unit 324 having voice identification information “1113”. ing. The voice uttered by “Taro” is collected by the third microphone 326 and transmitted from the third voice output unit 324 to the robot 100 in association with the voice identification information “1113”.

  For example, when there is no input of voice data associated with the voice identification information “1111” for a predetermined time, the event output unit 106 of the robot 100 determines that the voice data associated with the voice identification information “1111” It is detected that the condition associated with “input failure” is met. The event output unit 106 outputs the number “1” indicating the event together with the voice identification information “1111” to the behavior determining unit 110.

  The behavior determination unit 110 reads the tag identification information “0001” corresponding to the voice identification information “1111” with reference to the identification information storage unit 120 based on the voice identification information “1111”. Subsequently, the behavior determining unit 110 notifies the position information acquiring unit 108 of the tag identification information “0001”. The position information acquisition unit 108 acquires the position information of the first identification tag 302 that is an identification tag having the tag identification information “0001”, and notifies the behavior determination unit 110 of the position information.

  The behavior determining unit 110 determines to cause the robot 100 to execute the behavior associated with the number “1” indicating the event “voice input failure”. Specifically, based on the position information of the first identification tag 302 acquired from the position information acquisition unit 108, an operation for approaching “Sakura” is determined. The behavior determining unit 110 causes the behavior executing unit 112 to execute the determined operation. When the robot 100 approaches “Sakura”, the behavior determining unit 110 switches the voice input unit 102 to input voice from the internal microphone 13. As described above, when an event indicating “voice input failure” is output, when the robot 100 approaches the corresponding speaker, when the robot 100 inputs the speaker's voice from the internal microphone 13, ambient noise is generated. And long-distance speech can be avoided, and voice recognition and the like can be performed more accurately. However, this behavior does not necessarily have to be performed, it may be a behavior that the robot 100 just faces the speaker, or the speaker 100 faces the speaker and says “Sakura-chan, come over here” It is also possible to perform a behavior such that the speaker comes near the robot 100.

  It should be noted that when the robot 100 performs the above actions, it can be seen that the name of the participant who has the voice input problem is “Sakura”, so the robot 100 moves to “Sakura” before the robot 100 moves to “Sakura”. You can also say “Sakura-chan, wait a moment”. In addition, when the robot 100 approaches “Sakura”, the robot 100 can also make an utterance such as “Sakura-chan, say again”. Thereafter, voice recognition or the like can be performed based on the voice input from the internal microphone 13.

  As another example, you can turn to “Sakura” and say “You can't hear Sakura-chan. Listen to your brother nearby”.

  For example, when the voice power of the voice data associated with the voice identification information “1111” is equal to or greater than a predetermined threshold, the event output unit 106 of the robot 100 is associated with the voice identification information “1111”. It is detected that the voice data matches the condition associated with the event “speech detection”. The event output unit 106 outputs the number “2” indicating the event to the behavior determining unit 110 together with the voice identification information “1111”.

  The behavior determination unit 110 reads the tag identification information “0001” corresponding to the voice identification information “1111” with reference to the identification information storage unit 120 based on the voice identification information “1111”. Subsequently, the behavior determining unit 110 notifies the position information acquiring unit 108 of the tag identification information “0001”. The position information acquisition unit 108 acquires the position information of the first identification tag 302 that is an identification tag having the tag identification information “0001”, and notifies the behavior determination unit 110 of the position information.

  The behavior determining unit 110 determines to cause the robot 100 to execute the behavior associated with the number “2” indicating the event “speech detection”. Specifically, based on the position information of the first identification tag 302 acquired from the position information acquisition unit 108, the behavior that faces “Sakura” is determined. The behavior determination unit 110 causes the behavior execution unit 112 to execute the determined behavior.

  For example, when the voice recognition result of the voice data associated with the voice identification information “1111” is output, the event output unit 106 converts the voice data associated with the voice identification information “1111” to the event “ It is detected that the condition associated with “speech recognition result” is met. The event output unit 106 outputs the number “3” indicating the event together with the voice identification information “1111” to the behavior determining unit 110.

  The behavior determination unit 110 reads the tag identification information “0001” corresponding to the voice identification information “1111” with reference to the identification information storage unit 120 based on the voice identification information “1111”. Subsequently, the behavior determining unit 110 notifies the position information acquiring unit 108 of the tag identification information “0001”. The position information acquisition unit 108 acquires the position information of the first identification tag 302 that is an identification tag having the tag identification information “0001”, and notifies the behavior determination unit 110 of the position information.

  The behavior determining unit 110 determines to cause the robot 100 to execute the behavior associated with the number “3” indicating the event “speech recognition result”. Specifically, based on the position information of the first identification tag 302 acquired from the position information acquisition unit 108, the behavior that faces “Sakura” is determined. Subsequently, the behavior determination unit 110 refers to the robot behavior storage unit 130 and the scenario storage unit 132 to determine the behavior corresponding to the voice recognition result and causes the robot 100 to execute the behavior. For example, if the speech recognition result was a "Hello", the robot 100, "Sakura-chan, Hello" and the like and to speech.

  Further, for example, when the voice power of the voice data associated with the voice identification information “1111” and the voice data associated with the voice identification information “1112” become substantially equal to or greater than a predetermined threshold at the same time, the robot 100 The same processing can be performed in order for each corresponding participant. At this time, the robot 100 performs an operation facing the speaker. For example, the robot 100 can perform an operation alternately facing “Sakura” and “Momo”.

  Next, for example, let us explain an example in which the robot 100 is asked to give a quiz, and “Please answer the question to be answered with“ Seno ”” and get answers from three children at the same time. . Here, for example, “utterance detection” of event 2 is set not to be executed.

  At this time, when the voices of the three people are input via the communication interface 24, the voice recognition unit 104 performs voice recognition on the respective voice data. The robot 100 faces the child whose voice has been recognized. At this time, if there is a time difference in the output of the voice recognition results of the three persons, processing is performed on the child whose voice recognition result is first detected. Subsequently, processing for the next child is performed.

  For example, the robot 100 speaks, “A fish is a creature or a plant, answer either, Seno”, and in the order of Sakura, Momo, Taro, “Plant”, “Creature”, “Creature” If the voice is recognized, the robot 100 first faces “Sakura”, then “Momo”, and finally “Taro”. Subsequently, the robot 100 performs an operation corresponding to the voice recognition result. For example, say, “The correct answer is a creature. Momo-chan and Taro-chan answered correctly. Sakura-chan was disappointed. Then do your best.”

  Next, for example, let's ask the robot 100 to give a quiz, and say “Please answer the question you are about to say. You win fast” and get answers from the three children in order of speed. Will be explained. Here, for example, both event 2 “speech detection” and event 3 “speech recognition result” are set to be executed.

  In this case, for example, “utterance start time recording” can be set as the behavior corresponding to “utterance detection” of event 2. As a result, the robot 100 can recognize the participant who started speaking the earliest.

  For example, if the robot 100 utters “What is a black bird flying in the sky?” And says “Yes, it is a crow” or “Crow” in the order of Sakura and Momo, both are correct. become. Here is Sakura who started speaking the answer first, but the crow that uttered Momo has fewer syllables and the time required for speaking is shorter, so the speech recognition result is output first There is. In this case, if only before and after the output of the speech recognition result is considered, the correct answerer becomes “Momo”. However, since it is “Sakura” that has come up with the correct answer the earliest and started uttering, the robot 100 can determine that “Sakura”, which was the earliest utterance detection and was the correct answer, is the correct person.

FIG. 9 is a flowchart showing the procedure of the behavior determining unit 110 for performing this process.
Here, although not shown in the figure, the controller 10 includes an utterance start queue that stores the speech identification information associated with the utterance detection event in the order of event output, and the speech identification information associated with the speech recognition result event. Including a storage area for storing a speech recognition result cue for storing events in the order of event output.

  When there is an utterance detection event output from the event output unit 106 (YES in S200), the behavior determination unit 110 adds the voice identification information associated with the event to the utterance start queue (S202).

  In addition, when the event output unit 106 outputs an event of a speech recognition result (YES in S204), the behavior determination unit 110 and the speech identification information associated with the event and the speech identification of the head of the utterance start queue It is determined whether the information is the same (S206). When the voice identification information is the same as the head of the utterance start queue (YES in S206), the behavior determining unit 110 includes the corresponding behavior storage unit 118, the identification information storage unit 120, the robot behavior storage unit 130, the scenario storage unit 132, and the like. With reference to the speech identification information, the behavior of the robot 100 to the speaker associated with the voice identification information is determined (S208).

  Subsequently, the voice identification information is deleted from the utterance start queue (S210). Next, it is determined whether the voice identification information at the head of the utterance start queue is in the voice recognition result queue (S212). If there is (YES in S212), the voice identification information is deleted from the response recognition result queue (S214). ). Thereafter, the process returns to step S208, and the behavior of the robot 100 to the speaker associated with the voice identification information is determined.

  On the other hand, if the voice recognition result is not the same as the head of the utterance start queue in step S206 (NO in S206), the voice identification information is added to the voice recognition result queue (S216).

  In step S212, when the voice identification information at the head of the utterance start queue is not in the response recognition result queue (NO in S212), and after step S216, it is determined whether or not to end the process (S218). (NO in S218), the process returns to step S200 and waits.

  In addition, when the process ends in step S218 (YES in S218), the process ends.

  As described above, the robot 100 can determine the response to other participants in response to the utterances from a plurality of participants.

  The processing described above as a specific example can be realized by appropriately setting the corresponding behavior storage unit 118, the robot behavior storage unit 130, and the scenario storage unit 132.

  As described above, according to the robot 100 in the present embodiment, the robot 100 executes speech and behavior according to the event and the position of the speaker based on the output of the event triggered by the speaker's voice. Communication can be facilitated.

(Second embodiment)
This embodiment is different from the first embodiment in that the event output unit is provided not in the robot 100 but in an event output server provided outside the robot 100. In the present embodiment, a part of the correspondence determination system is incorporated in the robot 100, and the other part is incorporated in the event output server.

  FIG. 10 is a block diagram showing the configuration of the event output server in the present embodiment.

  In the present embodiment, the audio data and the audio identification information output from the first audio output unit 304, the second audio output unit 314, and the third audio output unit 324 are stored in the first event output server 200 or It is transmitted to the second event output server 220. The first event output server 200 or the second event output server 220 detects whether or not the received audio data meets a predetermined condition, and if it meets the predetermined condition, the first event output server 200 or the second event output server 220 satisfies the condition. Is transmitted to the robot 100 together with the voice identification information. The robot 100 determines and executes speech based on the event and voice identification information output from the first event output server 200 or the second event output server 220.

  Although only two event output servers are shown here, one event output server may be provided for each participant, or only one event output server may be provided for a plurality of participants. . Hereinafter, the audio data from the first audio output unit 304 is transmitted to the first event output server 200, and the audio data from the second audio output unit 314 and the third audio output unit 324 is output to the second event output. A case where data is transmitted to the server 220 will be described as an example.

  The first event output server 200 includes a voice input unit 202, a voice recognition unit 204, an event output unit 206, a data output unit 208, a voice recognition dictionary 210, a condition storage unit 212, and a communication interface (I / F) 214. . The second event output server 220 has the same configuration as the first event output server 200.

  The communication interface 214 receives data from the first audio output unit 304 via the network 400.

  The voice input unit 202 inputs the voice data received by the communication interface 214 from the first voice output unit 304 in association with the voice identification information. The voice input unit 202 outputs the input voice data to the voice recognition unit 204 and the event output unit 206 together with the voice identification information.

  The voice recognition unit 204 recognizes the voice data input by the voice input unit 202. The speech recognition dictionary 210 includes a speech recognition word storage unit that stores a speech recognition vocabulary that is a set of speech recognition words. The voice recognition unit 204 performs matching between the voice data input by the voice input unit 202 and the voice recognition vocabulary stored in the voice recognition dictionary 210. When voice data is voice-recognized, the voice recognition unit 204 outputs a voice recognition result to the event output unit 206 in association with the voice identification information.

  The event output unit 206 detects whether or not the voice data meets a predetermined condition based on the voice data output from the voice input unit 202 and the voice recognition result output from the voice recognition unit 204. If they match, an event indicating that the condition is met is output in association with the voice identification information. When the event output unit 206 acquires the speech recognition result from the speech recognition unit 204, the event output unit 206 also outputs the speech recognition result together with the event.

  The condition storage unit 212 has the same configuration as the condition storage unit 116 described with reference to FIG. 4 in the first embodiment. The event output unit 206 refers to the condition storage unit 212 and detects whether or not the audio data meets a predetermined condition.

  The data output unit 208 performs processing for transmitting data such as events output by the event output unit 206 to the robot 100. The communication interface 214 transmits data to the robot 100 based on an instruction from the data output unit 208.

FIG. 11 is a block diagram illustrating a configuration of the robot 100 according to the present embodiment.
The robot 100 is different in that it does not have the event output unit 106 and the condition storage unit 116 of the robot 100 shown in FIG.

  The communication interface 24 receives an event associated with the voice identification information from the first event output server 200 and the second event output server 220 via the network 400.

  The data input unit 122 inputs data from the plurality of first event output servers 200 and the second event output server 220 received by the communication interface 24. The data input unit 122 outputs the input data to the behavior determining unit 110.

  Based on the event output from the data input unit 122, the behavior determining unit 110 reads the behavior of the robot 100 corresponding to the event from the corresponding behavior storage unit 118. Subsequently, the behavior determining unit 110 refers to the identification information storage unit 120 and reads tag identification information corresponding to the target voice identification information. The subsequent processing is the same as that of the robot 100 in the first embodiment.

  In addition, the voice input unit 102 inputs voice data collected by the internal microphone 13 and outputs the voice data to the voice recognition unit 104. The voice recognition unit 104 recognizes the voice data.

  FIG. 12 is a flowchart showing a processing procedure of the first audio output unit 304, the first event output server 200, and the robot 100 in the present embodiment.

  The first audio output unit 304 always transmits the audio data of the first participant 300 input from the first microphone 306 to the first event output server 200 (S300).

  In the first event output server 200, the voice input unit 202 outputs the voice data transmitted from the first voice output unit 304 to the voice recognition unit 204 and the event output unit 206. When the speech recognition unit 204 recognizes speech data, the speech recognition unit 204 outputs the speech recognition result to the event output unit 206. In the event output unit 206, when the data related to the voice output from the voice input unit 202 or the voice recognition unit 204 meets a predetermined condition (S302), the event output unit 206 displays an event indicating that the condition is met. Output in correspondence with the voice identification information. The event, the voice identification information, and the voice recognition result are transmitted to the robot 100 (S304).

  In the robot 100, when the behavior determining unit 110 acquires an event from the first event output server 200, the behavior determining unit 110 obtains the corresponding participant location information based on the tag identification information corresponding to the voice identification information associated with the event. Obtain (S306). Subsequently, the behavior determination unit 110 refers to the corresponding behavior storage unit 118, the identification information storage unit 120, the robot behavior storage unit 130, and the scenario storage unit 132, and determines the behavior of the robot 100 based on the location information of the participants. Determine (S308). Next, the mechanical control unit 134, the voice synthesis unit 136, and the like are caused to execute speech (S310).

  Also in this embodiment, the same effect as that of the first embodiment can be obtained. Further, in the present embodiment, the event output unit is provided in the event output server outside the robot 100, and processing such as voice recognition is also performed in the event output server, so that the processing of the robot 100 itself can be simplified. . Further, by providing a plurality of event output servers and processing voices from a plurality of speakers in a distributed manner, processing such as voice recognition can be performed efficiently.

  Each component of the robot and the event output server described in the above embodiments includes a CPU, a memory, a program that realizes the components shown in the figure loaded in the memory, and a hard disk that stores the program. It is realized by any combination of hardware and software, centering on the storage unit and network connection interface. It will be understood by those skilled in the art that there are various modifications to the implementation method and apparatus.

  The embodiments of the present invention have been described above with reference to the drawings, but these are exemplifications of the present invention, and various configurations other than those described above can be adopted.

  In the above embodiment, for example, as shown in FIGS. 5 and 6, a plurality of conditions and events corresponding thereto are shown. In the robot or the event output server, whether or not these conditions are met is determined in parallel. It is also possible to adopt a configuration for determining whether or not only one of the conditions is met. Which condition is used to output an event can be set as appropriate according to the form in which the robot is used.

  In the above embodiment, the configuration in which the robot 100 includes the tag reader 26 has been described. However, the tag reader may be provided outside the robot 100. In this case, the robot 100 can acquire the position information of the speaker based on the information of the speaker identification tag read by the tag reader.

  For example, when the identification tag is an RFID tag, the robot 100 can acquire the position information of the speaker as follows. First, a plurality of tag readers are provided at predetermined positions in a room where the robot 100 and the speaker are present. The robot 100 stores the positions of these tag readers in advance. For example, three or more tag readers are provided, and the radio wave intensity of the speaker's identification tag read by each tag reader is transmitted to the robot 100. Based on these pieces of information, the robot 100 acquires the position information of the identification tag held by the speaker.

  Moreover, in the above embodiment, although the communication terminal device including an audio | voice output part and the form which provides an identification tag were shown to the speaker (participant), the position of the communication terminal device was not provided without providing an identification tag. Based on this, the position information of the speaker can be acquired. For example, the position information of the communication terminal device can be acquired by using a PDA terminal with a GPS function as the communication terminal device or using a radio wave transmitted from the communication terminal device.

  As described above, the method for acquiring the position information of the speaker can take various forms, and is not limited to the one described in the above embodiment.

  In the above embodiment, an example in which the robot 100 operates such as facing the speaker or approaching the speaker according to the position information of the speaker has been shown. It is also possible to perform only the utterance according to the position information of the speaker. For example, if a speaker utters "Let's go to the lake" and detects from the speaker's location information that the speaker is approaching a dangerous area, XX-chan, It ’s also possible to just say, “It ’s dangerous if you go.”

  In addition, for example, when the robot 100 communicates with a plurality of speakers, the robot 100 can be controlled to efficiently execute speech according to an event output for the plurality of speakers. For example, cherry, peach, children that are arranged in the order of Taro, cherry, Taro, if it is speech recognition and speech as "Hello" in the order of the thigh, the robot 100 once, turned to Sakura, also You must turn to the other side again after you turn to the side of the deaf. When there is an event output for one of the speakers, the robot 100 can wait for a predetermined time, confirm the presence / absence of an event output for another speaker, and then make a speech. When there is an event output for a plurality of speakers, the robot 100 can efficiently speak according to the position information of the speakers. For example, in the example above, Sakura within a predetermined time, peach, there is speech of "Hello" from all of Taro, if it is speech recognition, the robot 100 on the basis cherry, peach, the position information of Taro, cherry, peach, orientation towards the each in the order of Taro, it is possible to perform an utterance such as "Hello".

  In the above embodiment, the configuration in which the robot 100 has the internal microphone 13 has been described. However, the robot 100 may have a configuration in which the internal microphone 13 is not provided. In this case, for example, the robot 100 faces the speaker to which the voice input device in which the voice input failure has occurred is given, and makes some utterance, or brings a new voice input device to the speaker. It is possible to appropriately perform speech and actions without using the internal microphone 13 such as going.

  Although not specifically described in the above embodiment, the controller 10 of the robot 100 analyzes the surrounding situation based on the image data sent from the CCD camera 21A and the CCD camera 21B, and accordingly, the robot One hundred actions can be determined.

  In the above embodiment, a robot that moves autonomously and interacts with a speaker has been described as an example. However, the robot may have only one of the functions.

  In the above embodiment, the case where the response determination system and the response determination method are the robot control system and the robot control method has been described as an example. However, the response determination system and the response determination method include, for example, an anthropomorphic agent. The present invention can also be applied to the used voice dialogue system and voice dialogue method, or the information retrieval system and information retrieval method using voice recognition. Even in such a case, when the user inputs a sound to a sound input device such as a microphone, the system can take an appropriate action according to the data regarding the sound and the position of the user.

  In the above embodiment, the case where the network 400 is a wireless LAN has been described as an example. However, the network 400 may be other various wireless networks such as Bluetooth or a wired network.

  INDUSTRIAL APPLICABILITY The present invention can be applied to a system and a method for performing some correspondence with a speaker's voice, such as an autonomous mobile or interactive robot control system, a voice dialog system, and an information search system using voice recognition. .

It is a schematic diagram which shows the relationship between the robot in embodiment, and the participant who is a speaker. It is an external appearance block diagram which shows an example of the robot in embodiment. It is a block diagram which shows an example of the electrical structure of a robot. It is a block diagram which shows in detail the structure of the controller of the robot in embodiment. It is a figure which shows an example of an internal structure of a condition memory | storage part. It is a figure which shows an example of an internal structure of a corresponding speech memory | storage part. It is a figure which shows an example of the internal structure of an identification information storage part. It is a flowchart which shows the process sequence of the controller of the robot in embodiment. It is a flowchart which shows the process sequence of the behavior determination part in embodiment. It is a block diagram which shows the structure of the event output server in embodiment. It is a block diagram which shows the structure of the robot in embodiment. It is a flowchart which shows the process procedure of the 1st audio | voice output part in embodiment, a 1st event output server, and a robot.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Body part 2 Head 3A, 3B Wheel 10 Controller 10A CPU
10B Memory 11 Battery 12 Speaker 13 Internal microphone 14A, 14B Actuator 21A, 21B CCD camera 22A, 22B Actuator 24 Communication I / F
26 tag reader 100 robot 102 voice input unit 104 voice recognition unit 106 event output unit 108 position information acquisition unit 110 behavior determination unit 112 behavior execution unit 114 speech recognition dictionary 116 condition storage unit 118 corresponding behavior storage unit 120 identification information storage unit 122 data input Unit 130 robot behavior storage unit 132 scenario storage unit 134 mechanical control unit 136 speech synthesis unit 138 output unit 200 first event output server 202 speech input unit 204 speech recognition unit 206 event output unit 208 data output unit 210 speech recognition dictionary 212 condition Storage unit 214 Communication I / F
220 second event output server 222 voice input unit 224 voice recognition unit 226 data output unit 228 voice recognition dictionary 300 first participant 302 first identification tag 304 first voice output unit 306 first microphone 310 second Participant 312 Second identification tag 314 Second audio output unit 316 Second microphone 320 Third participant 322 Third identification tag 324 Third audio output unit 326 Third microphone

Claims (26)

An event indicating that the voice-related data input from the voice input device assigned to the speaker matches a predetermined condition is input together with the voice identification information associated with the speaker, and the response to the speaker is performed. A response determining unit to determine;
A location information acquisition unit that acquires location information of the speaker specified by the voice identification information associated with the event input by the correspondence determination unit;
Including
The correspondence determination unit determines the correspondence based on the event and the position information of the speaker. In the correspondence determination system according to claim 1,
Data related to the voice is input together with the voice identification information, whether or not the data related to the voice meets a predetermined condition, and an event indicating that the condition is met when the condition is met. A correspondence determination system, further comprising an event output unit that outputs the voice identification information to the correspondence determination unit. In the correspondence determination system according to claim 2,
The event output unit monitors the power of the voice input from the voice input device, and outputs an event indicating a voice input failure when the voice power is below a predetermined value for a predetermined time. Correspondence determination system characterized by In the correspondence determination system according to claim 2 or 3,
The event output unit monitors the power of the voice input from the voice input device, and outputs an event indicating utterance detection when the power of the voice exceeds a predetermined value. Decision system. In the correspondence determination system according to any one of claims 2 to 4,
The voice input from the voice input device assigned to the speaker is input together with the voice identification information associated with the speaker, the voice is voice recognized, and the voice recognition result is output together with the voice identification information. A speech recognition unit;
The event output unit outputs an event indicating a speech recognition result together with the speech recognition result when the speech recognition result is output from the speech recognition unit. In the correspondence determination system according to any one of claims 1 to 5,
A tag reader for reading tag identification information from an identification tag attached to the speaker;
The position information acquisition unit acquires position information of the speaker based on the tag identification information read by a tag reader. The correspondence determination system according to claim 6,
An identification information storage unit that associates the voice identification information with the tag identification information of the same speaker;
The position information acquisition unit refers to the identification information storage unit based on the voice identification information associated with the event output by the event output unit, and obtains the position information of the identification tag having the corresponding tag identification information. A correspondence determination system characterized by acquiring. In the correspondence determination system according to any one of claims 1 to 7,
The correspondence determination unit is configured to identify an event indicating that data related to speech input from a plurality of speech input devices assigned to a plurality of speakers matches a predetermined condition, and to identify the speech associated with each speaker. A correspondence determination system characterized by inputting together with information. In the correspondence determination system according to any one of claims 1 to 8,
The correspondence determining unit determines the behavior of an autonomously moving or interactive robot based on the event and the position information of the speaker. In the correspondence determination system according to claim 9,
The position information acquisition unit acquires position information of the speaker with respect to the robot;
The correspondence determining unit determines the behavior of the robot so that the robot performs a behavior that is conscious of the position of the speaker. An autonomous mobile or interactive robot,
A response determination system according to any one of claims 1 to 9,
A behavior execution unit that executes the correspondence determined by the correspondence determination unit as behavior;
A robot characterized by including: An event output server connected via a network to a communication terminal device including an audio output unit that outputs audio input from an audio input device and an autonomous mobile or interactive robot, and relays these,
The robot is
An event indicating that the voice-related data input from the voice input device assigned to the speaker matches a predetermined condition is input together with the voice identification information associated with the speaker, and is associated with the event. Obtaining the position information of the speaker specified by the voice identification information, and executing the behavior determined based on the event and the position information of the speaker,
The event output server is
A voice input unit that receives the voice output by the voice output unit from the communication terminal device together with the voice identification information;
An event output unit that detects whether or not the voice input by the voice input unit matches a predetermined condition, and outputs an event corresponding to the condition together with the voice identification information when the condition is met;
A data output unit for transmitting the event output by the event output unit to the robot together with the voice identification information;
An event output server comprising: In the event output server according to claim 12,
The event output unit monitors the power of the voice input from the voice input device, and outputs an event indicating a voice input failure when the voice power is below a predetermined value for a predetermined time. An event output server characterized by In the event output server according to claim 12 or 13,
The event output unit monitors the power of voice input from the voice input device, and outputs an event indicating utterance detection when the power of the voice exceeds a predetermined value. Output server. The event output server according to any one of claims 12 to 14,
A voice input from a voice input device assigned to the speaker is input together with the voice identification information associated with the speaker, the voice is recognized as a voice, and a voice recognition result is output together with the voice identification information. A speech recognition unit that
The event output server outputs an event indicating a speech recognition result together with the speech recognition result when the speech recognition result is output from the speech recognition unit. Inputting an event indicating that data related to voice input from a voice input device assigned to the speaker matches a predetermined condition, together with voice identification information associated with the speaker;
Obtaining positional information of the speaker specified by the voice identification information associated with the event input in the step of inputting the event together with the voice identification information;
Determining a response to the speaker based on the event and the location information of the speaker;
A correspondence determining method characterized by including: In the correspondence determination method according to claim 16,
Prior to the step of inputting the event together with the voice identification information, the voice-related data is input together with the voice identification information, and it is detected whether the voice-related data meets a predetermined condition. A method for determining correspondence, further comprising the step of outputting an event indicating that the condition is met together with the voice identification information when the condition is met. The correspondence determination method according to claim 17,
The step of outputting the event together with the voice identification information includes:
Monitoring the power of voice input from the voice input device;
Outputting an event indicating an audio input failure when a state where the power of the audio is a predetermined value or less continues for a predetermined time; and
A correspondence determining method characterized by including: The correspondence determination method according to claim 17 or 18,
The step of outputting the event together with the voice identification information includes:
Monitoring the power of audio input from the audio input device;
Outputting an event indicating utterance detection when the power of the voice is equal to or greater than a predetermined value;
A correspondence determining method characterized by including: The correspondence determining method according to any one of claims 17 to 19,
Inputting voice input from a voice input device assigned to a speaker together with voice identification information associated with the speaker, voice recognition of the voice, and outputting a voice recognition result together with the voice identification information Further including
The step of outputting the event together with the voice identification information includes outputting an event indicating the voice recognition result together with the voice recognition result when the voice recognition result is outputted. The correspondence determination method according to any one of claims 16 to 20,
In the step of inputting the event, an event indicating that data related to speech input from a plurality of speech input devices respectively assigned to a plurality of speakers matches a predetermined condition is associated with each speaker. A correspondence determination method characterized by inputting together with voice identification information. The correspondence determination method according to any one of claims 16 to 21,
In the step of determining the correspondence, the behavior determination method is characterized in that the behavior of the autonomous mobile or interactive robot is determined based on the event and the position information of the speaker. Computer
Means for inputting an event indicating that data relating to voice input from a voice input device assigned to a speaker matches a predetermined condition together with voice identification information associated with the speaker;
Position information acquisition means for acquiring position information of the speaker specified by the voice identification information associated with the event;
Correspondence determination means for determining a correspondence to the speaker based on the event and the position information of the speaker;
A program characterized by functioning as The program according to claim 23,
The correspondence determining means determines the behavior of an autonomously moving or interactive robot based on the event and the position information of the speaker. Computer
Means for inputting data relating to voice input from a voice input device assigned to a speaker together with voice identification information associated with the speaker;
An event output means for detecting whether or not the data relating to the voice meets a predetermined condition, and outputting an event indicating that the condition is met together with the voice identification information when the condition is met;
Position information acquisition means for acquiring position information of the speaker specified by the voice identification information associated with the event;
Correspondence determination means for determining a correspondence to the speaker based on the event and the position information of the speaker;
A program characterized by functioning as The program according to claim 25,
The correspondence determining means determines the behavior of an autonomously moving or interactive robot based on the event and the position information of the speaker.

Images