JP2023001299A - Interaction system and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an interaction system which can avoid breakdown of interaction as much as possible during interaction with an agent.

SOLUTION: Two agents (R1 and R2) are arranged in interaction places, and the robots interact with a person according to script (dialogue data). One agent (R1) utters a question sentence "Where do you want to go on a day off?" to the person, for example. When a response is uttered from the person, the response sentence is determined to hit a keyword or not. When the keyword hits, the other agent utters a previous phrase ("I guess Umeda.", for example) which develops interaction associated with the keyword, and one agent further utters a recognition response. When there is no response utterance, the other agent utters a proxy response.

SELECTED DRAWING: Figure 1

COPYRIGHT: (C)2023,JPO&INPIT

Description

The present invention relates to dialogue systems and programs, and more particularly to dialogue systems and programs in which, for example, at least one agent interacts with at least one person at a dialogue location.

With the expansion of the robot field, the focus of robot research is shifting to research on robots that work in everyday situations, and the development of robots that can interact with people in the environment where humans live is attracting attention.

With the recent development of speech recognition technology, robots with the ability to communicate with humans in verbal language have been developed. It is not easy to continue to give the person who interacts with the robot the same feeling of "dialogue feeling (the feeling of participating in the dialogue)" that a person has when participating in a dialogue with a human being. I didn't. In other words, people sometimes clearly lost the sense of dialogue.

Patent Literature 1, which is an example of the background art, discloses that in a dialogue between a robot and a person, the robot guesses the emotion of the person and determines the utterance of a response sentence and the action associated with the utterance.

Japanese Patent Laid-Open No. 2002-200000 describes a dialog system between a human and a robot that attempts to realize a continuous and natural interaction by synchronizing the two.

Japanese Unexamined Patent Publication No. 2004-90109 [B25J 13/00...] JP 2012-181697 A [G06F 3/16...]

In both the technique of Patent Document 1 and the technique of Patent Document 2, there is a limit to the processing based on speech recognition, and it is not easy for a person to maintain the above-described "sense of dialogue". In other words, it is easy to invite the failure of dialogue.

SUMMARY OF THE INVENTION Therefore, a primary object of the present invention is to provide a novel interactive system and program.

Another object of the present invention is to provide a dialogue system and program that can avoid breakdown of dialogue with people as much as possible.

In order to solve the above problems, the present invention employs the following configuration. It should be noted that reference numerals in parentheses, supplementary descriptions, etc. indicate the correspondence with the embodiments described to aid understanding of the invention, and do not limit the invention in any way.

A first invention is a dialogue system comprising at least one agent at a dialogue place, wherein the agent at the dialogue place interacts with a person according to a dialogue, comprising a question sentence utterance part for causing the agent to utter a question sentence, a question; A first judging part for judging whether or not a response sentence is uttered by a person to the sentence, when judging that the response sentence is uttered, judges whether or not the response sentence hits a predetermined keyword. A second determination unit, when the second determination unit determines that the response sentence hits a specific keyword, causes the agent to utter a pre-phrase that invites the agent's next utterance after the response from the person. The dialog system includes a phrase utterance unit and a recognition response sentence utterance unit that, following the utterance of the preliminary phrase by the preliminary phrase utterance unit, utters a recognition response sentence that creates a dialogue context for the preliminary phrase.

In the first invention, the dialogue system (10: a non-limiting reference sign exemplifying the corresponding part in the embodiment; hereinafter the same) includes at least one agent (R1, R2) in a dialogue place (12). ), and the agent interacts with the person (H) according to the dialog at the interaction place. The question utterance unit (20a, S7) makes the agent utter a question sentence. The person (H) utters a response sentence to the question sentence, and the first determination unit (20a, S9) determines whether or not the person has uttered a response sentence to the question sentence. When the first determination unit (20a, S9) determines that the response sentence is uttered, the second determination unit (20a, S11, S13) determines whether the response sentence hits a predetermined keyword. When the second determination unit (20a, S11, S13) determines that the response sentence hits a specific keyword, the pre-phrase utterance unit (20a, S17) instructs the agent to Invite the next utterance of the utterance of a pre-phrase. A recognition response sentence utterance unit (20a, S19) utters a recognition response sentence that creates a dialogue context for the preliminary phrase.

According to the first invention, by causing the recognized response sentence to be spoken, it is possible to avoid breakdown of the dialogue as much as possible.

A second invention is a dialogue system comprising at least one agent at a dialogue place, wherein the agent at the dialogue place interacts with a person according to a dialogue, comprising a question sentence utterance unit for causing the agent to utter a question sentence, a question; When the first judging unit for judging whether or not the person has uttered a response sentence to the sentence does not judge that the response sentence has been uttered, the agent is instructed to replace the person in response to the question sentence. a proxy response sentence utterance unit, and following the utterance of the proxy response sentence by the proxy response sentence utterance unit, recognition of uttering a response sentence that creates a dialogue context for the proxy response sentence A dialogue system comprising a response sentence utterance unit.

In the second invention, when the first determination unit (20a, S9) does not determine that the response sentence is uttered, the proxy response sentence utterance unit (20a, S27) instructs the agent to respond to the question sentence. A recognized response sentence utterance unit (20a, S29) utters a proxy response sentence for responding on behalf of a person, and furthermore, a recognition response sentence utterance unit (20a, S29) utters a dialogue regarding the proxy response sentence following the utterance of the proxy response sentence by the proxy response sentence utterance unit. Speak a recognition response sentence that creates a context.

According to the second invention, even if no response sentence is uttered, the breakdown of the dialogue can be avoided as much as possible by making the proxy response utterance or the recognition response utterance.

A third aspect of the invention comprises at least one agent at a dialogue location, and the agent at the dialogue location interacts with a person according to the dialogue. A first judging part for judging whether or not a response sentence is uttered by a person to the sentence, when judging that the response sentence is uttered, judges whether or not the response sentence hits a predetermined keyword. A second determination unit, when the second determination unit determines that the response sentence hits a specific keyword, causes the agent to utter a pre-phrase that invites the agent's next utterance after the response from the person. This is a dialogue system program that functions as a phrase utterance unit and a recognition response sentence utterance unit that, following the utterance of a preliminary phrase by the preliminary phrase utterance unit, utters a recognition response sentence that creates a dialogue context for the preliminary phrase.

According to the third invention, effects similar to those of the first invention can be expected.

A fourth invention is provided with at least one agent at a dialogue location, and the agent at the dialogue location interacts with a person according to the dialogue. When the first judging unit for judging whether or not the person has uttered a response sentence to the sentence does not judge that the response sentence has been uttered, the agent is instructed to replace the person in response to the question sentence. the proxy response sentence utterance unit, and following the utterance of the proxy response sentence by the proxy response sentence utterance unit, the recognition response sentence that creates the dialogue context for the proxy response sentence is uttered. This is a dialog system program that functions as a recognition response sentence utterance unit.

According to the fourth invention, effects similar to those of the second invention can be expected.

According to the present invention, breakdown of dialogue between an agent and a person can be avoided as much as possible.

The above object, other objects, features and advantages of the present invention will become more apparent from the following detailed description of the embodiments with reference to the drawings.

FIG. 1 is a schematic diagram showing an overview of a dialogue system according to one embodiment of the present invention. FIG. 2 is a block diagram showing an example of the configuration of the sensor manager in the FIG. 1 embodiment. FIG. 3 is a block diagram showing an example of the configuration of the group manager in the embodiment of FIG. 1; FIG. 4 is a schematic diagram showing an example of the robot in FIG. 1 embodiment. 5 is a block diagram showing an example of the configuration of a robot controller that controls the robot of FIG. 4. FIG. FIG. 6 is a flowchart showing an example of the operation of the group manager shown in FIG. 1;

Referring to FIG. 1, a first robot R1, a second robot R2 and one person H are present at a dialogue place 12 of the dialogue system 10 of this embodiment. However, the number of robots may be one, or three or more. Also, the number of people may be two or more. In addition, hereinafter, the first robot R1 and the second robot R2 may simply be referred to as robots R when there is no particular need to distinguish between them.

The dialogue system 10 of this embodiment, as indicated by arrow A in FIG. When the response utterance from H is appropriate, or when there is no response utterance from the person H to the question, as indicated by the arrow B, the robot R1 or R2 responds to the robot R1 or R2 that uttered the question. Then, according to the dialog, pre-phrase utterances, proxy response utterances, etc. are performed.

Even if the robot R1 or R2 utters a question sentence to the person H, there is no response utterance from the person H, or even if there is a response utterance, for example, "I don't know", "I don't know", "I forgot". , "I don't remember" or "I don't want to answer", it is likely to be difficult for the person H and the robot R1 or R2 to continue the dialogue. In other words, unless the human H's utterance is generated in response to the robot R1 or R2, the dialogue is likely to break down. Therefore, in this embodiment, when there is no appropriate response utterance from the person H, the robot R1 or R2 (same as the robot that uttered the question) is given to the robot R1 or R2 that has uttered the question sentence to the person H. or another robot) to make it possible to continue the dialogue as much as possible.

On the other hand, when there is an appropriate response utterance from the person H, the robot R1 or R2 (even if it is the same robot that uttered the question) is sent to the robot R1 or R2 that uttered the question sentence to the person H. or another robot) to utter a preliminary phrase to encourage continuation of the dialogue on that main topic. In other words, by continuing to ask questions to the person H, the robot R expresses the desire of the robot R to share the experience with the person H and provides the person H with a sense of dialogue.

A dialogue system like this embodiment can be used, for example, as a tool for extracting utterances from the elderly.

The dialogue location 12 of the dialogue system 10 is provided with a microphone 14 as an auditory sensor and a camera 16 as a visual sensor in this embodiment. The microphone 14 is for listening to voices uttered by the robot R or the person H, or acquiring environmental sounds, and may be a microphone array if necessary. The camera 16 is also a camera for photographing the situation of the dialogue place 12, especially the facial expressions and actions of the person H, and photographs moving images or still images. Two or more cameras 16 may be installed if necessary.

Furthermore, in addition to the microphone 14 and the camera 16 described above, sensors (not shown) include a wearable posture sensor, an acceleration sensor, a heartbeat state, a breathing state, and a body motion (body movement) state. Biometric sensors, motion capture systems, etc. may be provided to detect signals.

Sensor signals such as audio signals acquired by the microphone 14 and image signals captured by the camera 16 are input to the sensor manager 18 . The sensor manager 18 acquires these sensor signals, determines the situation of the dialogue place 12, and outputs the determination result to the group manager 20 as sensing data.

Referring to FIG. 2, the sensor manager 18 includes a CPU (Central Processing Unit) 18a, and a communication device 18c is connected to the CPU 18a via an internal bus 18b. The communication device 18c includes, for example, a network interface controller (NIC) and the like, and the CPU 18a can communicate with the group manager 20 and the like via this communication device 18c, and can exchange data therebetween.

A memory 18d is further connected to the CPU 18a via an internal bus 18b. The memory 18d includes ROM and RAM. For example, a sensor signal including an audio signal from the microphone 14 (FIG. 1) and an image signal from the camera 16 (FIG. 1) is input through a sensor I/F (interface) 18e configured by a DSP (Digital Signal Processor). . The memory 18d temporarily stores the sensor signal.

The sensor manager 18 is a kind of determiner, and the CPU 18a determines the state of the dialogue place 12 based on the sensor data stored in the memory 18d. The sensor manager 18 then sends data indicating the determined state to the group manager 20 .

Programs (OS, sensor signal acquisition program, etc.) required for the sensor manager 18 are stored in the memory 18d. The sensor manager 18 operates according to programs stored in the memory 18d.

Note that the CPU 18a may be further provided with a keyboard and a display (not shown).

The group manager 20 controls the speech operation (verbal operation) and behavior (nonverbal operation) of each of the two robots R1 and R2, for example, according to the flow diagram of FIG. 6, which will be described later.

The group manager 20 includes a CPU 20a, and a communication device 20c is connected to the CPU 20a via an internal bus 20b. The communication device 20c includes, for example, a network interface controller (NIC), etc., and the CPU 20a can communicate with the sensor manager 18, the robot R, and the like via this communication device 20c, and can exchange data therebetween.

A memory 20d is further connected to the CPU 20a via an internal bus 20b. The memory 20d includes ROM and RAM. Script data is read from a dialog database 22 through a memory I/F 20e and temporarily stored in a memory 20d.

Programs (OS, sensor signal acquisition program, etc.) required for the group manager 20 are stored in the memory 20d. Group manager 20 operates according to a program stored in memory 20d.

The CPU 20a of the group manager 20 controls the actions or behaviors of each robot as described above. do.

A memory 20d and an input device 20e are further connected to the CPU 20a via an internal bus 20b. The memory 20d includes ROM and RAM. A script (dialog) is read from the dialog database 22 through the memory I/F 20f and temporarily stored in the memory 20d.

However, "dialogue" means a sequence of commands for utterances and nonverbal actions to be performed during a dialogue, and the dialogue database 22 stores a collection of dialogues (for example, stories about childhood, stories about travel, stories about health, etc.). , which contains a series of directives for each main topic dialogue). "Script" is a character string representing the series of commands, and script data is a character string representing the command. Therefore, the series of script data becomes a script. Such scripts are sent by the group manager 20 from the dialogue queue 23c to the robots R1 and R2.

However, script data includes not only linguistic data but also non-linguistic data. The linguistic data is script data that instructs the utterances of the robots R1 and R2, and the non-linguistic data is, for example, actions of the robots R1 and R2, looking at the person H, nodding, shaking his head sideways, tilting his head, etc. It is script data that instructs non-verbal actions.

Programs (OS, sensor signal acquisition program, etc.) required for the group manager 20 are stored in the memory 20d. Group manager 20 operates according to a program stored in memory 20d.

The CPU 20a of the group manager 20 controls the actions or behaviors of each robot as described above. do.

Further, the dialog system 10 shown in FIG. 1 comprises a next dialog candidate pool 23a. The next dialog candidate pool 23a stores a group of scripts to be spoken next, which are candidates to be dynamically selected according to the response of the person H to the currently ongoing script stored in the dialog queue 23c. In particular, depending on whether or not there is an utterance of a response sentence from person H, whether a keyword included in the response sentence hits a specific keyword set in advance, etc. be prepared for

The non-response dialog pool 23b, for example, when there is no response sentence from the person H to the question sentence uttered by the robot R1 (not only when there is no voice input of the response sentence to the microphone 14, but also when the voice input is 14, the robot R1 and/or R2 should speak to the robot R1 and/or R2. Dialog pooling. That is, the unresponsive dialog pool 23b is loaded into the dialog queue 23c and a group of future dialogs loaded into the next dialog candidate pool 23a, and is loaded into the dialog queue 23c to handle exceptional cases that were not planned for the ongoing dialog. A storage area for storing a group of scripts that should perform speech or non-verbal actions.

The dialog queue 23c is also an area in the memory 20d, for example. Script data loaded in the next dialog candidate pool 23a and script data loaded in the non-response dialog pool 23b are stored in this dialog queue 23c. , can then be loaded in a queue for immediate execution by robots R1 and/or R2.

In the script, for example, the execution time (tnext) of the script data contained therein is written as a header, and the dialog queue 23c always sorts the script data according to the execution time (tnext). The script data is sent to each robot controller 24 so that the script data with the same time are executed at the same time. Therefore, for example, the robots R1 and R2 can perform the same action at the same time, such as looking at the person H, and the same robot R1 or R2 can look at the other robot or the person H at the same time as speaking, for example.

Here, the dialog in this embodiment consists of a main topic or main category and several subtopics or subcategories within each main topic or main category. In this embodiment, in order to deepen the depth of dialogue with person H, dialogues on relatively few main topics (or categories) such as childhood stories, travel stories, and health stories are prepared. do.

In the main topic (major topic) of the story of childhood, subtopics (middle topics) such as play, meals, life and housing are set. For the subtopic “play”, subtopics (small topics) such as places, parks, toys, hide-and-seek, playing house, tag, running, and Daruma-san fell are set. For the subtopic "meal", subtopics such as school lunch, snacks, side dishes, brown rice, favorite things, disliked things, milk, cake, fish and meat, rice and bread, curry and sushi, etc. are set. In the sub-topic "life/housing", for example, where you lived, house, well, father and mother, brothers and sisters, work, chickens, cows, horses, dogs and cats, fun times, hard times, etc. set.

Within the main topic of travel, set subtopics such as hot springs, Mt.

Within the main topic of health, subtopics such as exercise and golf are set, and further detailed subtopics are prepared for each subtopic.

Referring to FIG. 4, this FIG. 4 shows the appearance of the robot R of the embodiment. In other words, the body 32 has two degrees of freedom.

A right arm 34R and a left arm 34L are rotatably provided forward, backward, leftward and rightward by shoulder joints (not shown), respectively, from left and right positions of the body 32 corresponding to human shoulders. That is, two degrees of freedom are set for each of the right arm 34R and the left arm 34L.

A neck 36 is provided at the center of the upper end of the body 32, and a head 38 is provided thereon. A neck 36 or head 38 is attached to the body 32 so as to be rotatable forward, backward, left and right. That is, the neck 36, that is, the head 38, is set with three degrees of freedom of roll angle (tilt to the left and right), pitch angle (tilt to the front and back), and yaw (rotation to the left and right).

A right eye 40R and a left eye 40L are provided on the front surface of the head 38, that is, a surface corresponding to a human face, and eyeballs 42R and 42L are provided on the right eye 40R and the left eye 40L. The eyelids of the right eye 40R and the left eye 40L can be closed and opened, and the eyeballs 42R and 42L can be rotated up, down, left and right, respectively. That is, the right eye 40R and the left eye 40L, that is, the eyelids, have one degree of freedom, and the eyeballs 42R and 42L have two degrees of freedom.

The face is further provided with a mouth 44, which can be closed and opened. That is, the mouth 44 is set with one degree of freedom.

A speaker 46 for uttering speech to be heard by the person H in the dialogue system 10 and a microphone 48 for listening to the environment, especially the speech voice of the person H, are provided on the torso 32 at the chest position of the person.

A camera 50 capable of capturing moving images or still images is incorporated in a portion of the head 38 corresponding to the forehead of the face. This camera 50 is capable of photographing a person H facing him, and a camera signal (video signal) from this camera 50 is sent to the CPU 22a via the sensor I/F of the sensor manager 18, as with the environmental camera 16 (FIG. 1). , may be entered.

FIG. 5 is a block diagram showing a robot controller 24 that is built in the robot R and that controls the actions of the robot R (utterance, gestures, etc.). Referring to FIG. 5, the robot controller 24 includes a CPU 20a, and a communication device 24c is connected to the CPU 20a via an internal bus 24b. The communication device 24c includes, for example, a network interface controller (NIC), etc., and the CPU 20a communicates with the sensor manager 18, the group manager 20, an external computer, other robots (both not shown), etc. via this communication device 24c. They can communicate and send and receive data between them.

A memory 24d is further connected to the CPU 20a via an internal bus 24b. The memory 24d includes ROM and RAM. Control data and script data sent from the group manager 20 are temporarily stored in the memory 24d.

Programs (OS, sensor signal acquisition program, etc.) necessary for robot control are stored in the memory 24d. The robot controller 24 controls the motion of the robot R according to a program stored in the memory 24d.

That is, the CPU 20a of the robot controller 24 is further connected to an actuator control board 24e, which is composed of, for example, a DSP. controls the behavior of

The two-degree-of-freedom movement of the torso 32, namely forward, backward, left and right rotation, is controlled by the torso actuator 52 controlled by the CPU 20a through the actuator control board 24e.

The two-degree-of-freedom movement of the right arm 34R and the left arm 34L, that is, the forward/backward/leftward/rightward rotation is controlled by the CPU 20a controlling the arm actuator 54 through the actuator control board 24e.

Movement of the neck 36 or head 38 in three degrees of freedom, that is, rotation in the forward/backward/leftward/rightward direction is controlled by the head actuator 56 by the CPU 20a through the actuator control board 24e.

The right eye 40R and left eye 40L, ie, the opening and closing operations of the eyelids, are controlled by the CPU 20a controlling the eyelid actuator 58 through the actuator control board 24e. The two-degree-of-freedom movement of the eyeballs 42R and 42L, that is, the forward/backward/leftward/rightward rotation is controlled by the CPU 20a controlling the eyeball actuator 60 through the actuator control board 24e. The opening and closing operation of the mouth 44 is controlled by the CPU 20a controlling the mouth actuator 62 through the actuator control board 24e.

A speaker 46 of the robot R shown in FIG. 4 is connected to the CPU 24 a of the robot controller 24 . The CPU 24a causes the speaker 46 to speak (speak) according to the script data given from the group manager 20 and stored in the memory 24d as required.

With such a robot controller 24, the head and arms of the robot R move as required in the dialogue system 10, for example, when a non-verbal action is requested by a script. Specific control of actuators and the like can be easily guessed from the above description, and therefore will not necessarily be described.

As shown in FIG. 1, each robot R1 and R2 is provided with a built-in robot sensor 26 as well as a robot controller 24 . Robot sensors 26 include attitude sensors, acceleration sensors, and the like for detecting the state of each movable component of robots R1 and R2, and sensor signals from these sensors are input to sensor manager 18 . Therefore, sensor manager 18 can sense the states of robots R1 and R2 based on sensor signals from robot sensors 26 .

Note that the microphone 48 and camera 50 of the robot R shown in FIG. 4 are input to the sensor manager 18 via the robot sensor 26 . The sensor manager 18 stores voice data captured from the microphone 48 in the memory 18d (FIG. 2), and executes voice recognition processing as necessary. Sensor manager 18 also processes camera signals from camera 50 to sense the context of interaction location 12 .

Although only one sensor manager 18 is shown in the embodiment of FIG. 1, any number of two or more sensor managers may be provided. can be shared.

Similarly, if desired, two or more group managers 20 may be used, or conversely, the sensor manager 18 and group manager 20 may be implemented by one computer.

Also, the robot R used in the dialog system 10 of the embodiment of FIG. 1 is not limited to the robot described above with reference to FIG. 4, and may at least have the function of speaking according to a script.

Referring to FIG. 6, CPU 20a of group manager 20 of dialog system 10 of FIG. 1 performs initialization such as reading dialog data (script data) as described above from dialog database 26 (FIG. 1). do. The operation shown in FIG. 6 is repeatedly executed at a speed of about the frame rate, for example.

In the next step S3, the CPU 20a determines whether or not to change the main topic (major topic) of the dialog. Whether or not to change the main topic is determined in this step S3 by judging whether or not a predetermined period of time has elapsed in the case of changing according to the time schedule. In the specific description of the embodiment below, the case of following the dialog of the main topic "travel story" will be described as an example.

In addition, as a condition for changing the main topic determined in step S3, a predetermined number of times N times since the last change of the main topic (this is because the same main topic continues too long, making the dialogue boring). This is the maximum number of repetitions of the same main topic, which is set in order to avoid this.) When the above period has elapsed, the response utterance from person H hits a keyword of a main topic different from the main topic currently being spoken. and when the response utterance from the person H cannot be recognized a predetermined number of times.

If "NO" is determined in step S3, the main topic is changed in step S5. In response, a question such as "Where do you want to go on your day off?" is uttered. Here, "where to go" can be considered a subtopic of the main topic "travel story". The CPU 20a that executes this step S7 functions as a question sentence utterance unit.

In step S9, the CPU 20a determines, based on the voice data from the microphone 14 detected by the sensor manager 16, whether or not the person H has responded to the question uttered by the robot R1 in step S7. The CPU 20a that executes this step S9 functions as a first determination unit that determines whether or not there is a response utterance from a person.

When it is determined that there is a response utterance, in the next step S11, the CPU 20a determines whether or not the response utterance from the person H detected in step S9 has been recognized. Determines whether or not there is a hit for a keyword that is In other words, it is determined whether or not the preset keyword is included in the response sentence of the person H. This can be easily done by processing the voice data from the microphone 14 using any voice recognition technique.

However, even if a keyword is hit, if the response sentence of the person H hits a plurality of keywords at the same time, it is determined "NO" in step S11. This is because it is difficult to determine which keyword should be used to proceed with the dialogue. The CPU 20a that executes this step S11 (which may include the next step S13) functions as a second judgment unit that judges whether or not the response sentence from the person hits the keyword.

In step S13, the CPU 20a determines whether the keyword detected in step S11 is a negative remark. A negative remark is, as described above, a remark as if refusing to answer a question.

If "NO" is determined in step S13, in the next step S15, it is determined whether the keyword is defined in the main topic currently being spoken or set in another main topic. to decide.

If "YES" is determined in step S15, as described in relation to step S3, when the response utterance from person H reaches a keyword of a main topic different from the main topic currently being spoken. , the process returns to step S5 to execute the main topic change process, and then the process proceeds to step S7 again.

When it is determined in step S11 that the keyword is hit and in step S15 the keyword is of the main topic currently in progress, in subsequent step S17, the CPU 20a determines whether the robot that uttered the question sentence, in this example Then, the robot R1 and another robot, the robot 2, are made to utter a preliminary phrase, for example, "As expected, it's Umeda." Here, the pre-phrase is an utterance sentence with the meaning of inducing the next robot R1 (or robot R2) to utter. However, if the person H actually intends to utter the detected keyword, then the preliminary phrase uttered by the robot R2 is simply a listening feeling (a feeling that the robot R is listening to the utterance of the person H). Although it only contributes, if the pre-phrase is not intended by the person H, it has the effect of creating a context for the subsequent reply from the robot R1.

In other words, the pre-phrase serves as a so-called "talking point" (a word that means providing a topic so that dialogue, discussion, etc., can proceed smoothly). In this embodiment, the robot R2 speaks with the keyword ("Umeda" in this example), so the robot R1 picks it up and says, for example, "Umeda is convenient, isn't it?" H has the effect of not giving a sense of incongruity to Umeda's discussion.

Other assumed interactions include example 1 and example 2 below.
<Example 1>
Robot R1: Where would you like to go on your day off? (Step S7)
Person H: First is Umeda Kana (recognized as "Umeda") (step S11)
Robot R2: Umeda after all (step S17)
Robot R1: Umeda is convenient (step S19)
<Example 2>
Robot R1: Where would you like to go on your day off? (Step S7)
Person H: It's Ome (recognized as "O, Umeda") (step S11)
Robot R2: Umeda after all (step S17)
Robot R1: Umeda is convenient (step S19)
In this way, the utterance of the pre-phrase "As expected, Umeda kana" by the robot R2 becomes the context of the reply uttered by the robot R1 in the next step S19.

Here, in step S17, the robot R2, which is different from the robot R1 that uttered the question sentence in step S7, is made to utter a preliminary phrase, and in step S19, the following recognition response utterance (after the robot R2 has recognized the preliminary phrase) is made to utter a preliminary phrase. ) is made to be uttered by another robot R1. In other words, the two robots R1 and R2 alternately made the question sentence, the preliminary phrase, and the recognition response utterance, but the order may be reversed. Furthermore, the same robot R1 or R2 may be made to utter all of the question sentence, preliminary phrase, and recognition response utterance. Alternatively, another robot (R3), not shown, may be made to perform the recognition response utterance of step S19.

It should be noted that when the robot R2 is made to utter a pre-phrase such as "Expoland" in step S17, a dialogue may be considered in which the robot R1 (or R3) is made to make a recognition response utterance such as "Expo is popular" in step S19. be done.

Further, when the robot R2 is made to utter a preliminary phrase, for example, "Yes, there was Hokkaido" in step S17, the recognition response utterance, for example, "Hokkaido recommends crab" is made to the robot R1 (or R3) in step S19. It is also conceivable to have a dialog that

"Umeda", "Expoland", or "Hokkaido" included in such preliminary phrases are considered to be further subtopics of the subtopic "where to go" of the main topic "travel story".

Note that the CPU 20a that executes step S17 functions as a pre-phrase utterance unit.

After having the robot R2 make the recognition response utterance in step S19, the CPU 20a determines whether or not to end the dialogue in the next step S21. Here, conditions for terminating the dialogue are determined, for example, that a certain period of time (for example, 15 minutes) has passed since the start of the dialogue, or that the person H is absent according to the image of the camera 16 (FIG. 1).

If it is to be terminated, the process is terminated after termination processing is performed in step S23. Termination processing includes, for example, processing such as saving a log of dialogue.

If not, the process returns to step S3.

If "NO" in the previous step S9, that is, if the response utterance of the person H could not be obtained through the microphone 14, the CPU 20a selects from the non-response dialogue pool 23b the main topic currently in progress in the next step S23. Randomly select a keyword with . Then, in step S25, the robot R2, which is different from the robot R1 that uttered the question sentence, is caused to utter a proxy response sentence (for example, "I am Umeda kana after all") according to the selected keyword. Here, unlike the pre-phrasing in step S17, for example, by uttering a word representing the subject of the utterance, such as "I am", it is clarified that it is a subjective substitute response sentence of the robot R2. The CPU 20a that executes step S25 (which may include step S23) functions as a proxy response sentence utterance unit.

After that, in step S29, the CPU 20a causes the robot R1 to make a recognition response utterance similar to that in step S19. However, as described above, steps S7, S27 and S29 may all be performed by the same robot R1 or R2.

By causing the robot R2 to utter a substitute response sentence in step S27, even if there is no response utterance from the person H, the dialogue will not be broken for the time being, and the recognition response utterance in step S29 will increase the willingness of the person H to have a dialogue. recovery can be expected. For example, even if the person H cannot immediately reply to the question in step S7, there is a possibility that the person H will come up with a response sentence triggered by the proxy response sentence in step S27. In this case, person H can continue the dialogue on the main topic at the time, for example, "talk about travel." In that sense, the substitute response sentence in step S27 can have the same effect as the preliminary phrase in step S17 (creating a dialogue context).

It should be noted that the main topic may be changed in steps S3 and S5 when "YES" is determined in step S13 and the number of times the keyword has been changed in step S25 exceeds a predetermined number of times.

After step S29, the process advances to the previous step S21 to determine whether or not the process is completed.

When "NO" is determined in step S11, the CPU 20a, in step S31, causes the robot R2 to utter an ambiguous response sentence (eg, "I want to go somewhere else") selected from the non-response dialog pool 23b. This ambiguous response has the meaning that "the robot R2 answers the question of the robot R1 on behalf of the person H" in step S27, whereas "the robot R1 or R2 answers the utterance of the person H". It has the meaning of "answer". In other words, when the response utterance from the person H is detected in step S9 but the response utterance sentence cannot be recognized in step S11, an ambiguous response sentence may be generated, although the dialogue may collapse if the response sentence is not recognized in step S11. By making the robot speak, it is possible to draw out the next speech of the person H, thereby avoiding the breakdown of the dialogue.

The CPU 20a that executes this step S31 functions as an ambiguous response sentence utterance unit, and after step S31 proceeds to step S21.

In the above description, it was explained that the main topic can be changed according to the passage of time, the person H speaking or not speaking, etc. However, the subtopic can be changed in the following cases.

If there is an unspoken dialogue (scenario) highly relevant to the previous utterance within the same subtopic as the current subtopic, the utterance of person H moves to that highly relevant subtopic. For example, it is determined whether or not each scenario includes a word (word) that is close to a keyword that has been registered in advance. The distance is evaluated using a technique such as Word2Vec. However, even if such a word is included, if another word similarly similar to other keywords is included, it will be discounted and evaluated.

The subtopic may be changed when no keyword is hit in step S11 and the number of times of proceeding to step S31 reaches a predetermined number N or more.

In the above-described embodiment, each robot R1 and dialog database 22 are stored in advance. However, instead of the database 22, the necessary dialogs (script data) may be sequentially supplied to the group manager 20 from the net, for example.

Furthermore, the above-described embodiment is a dialogue system using a robot that is a physical agent, but the present invention can be applied not only to such a physical agent, but also to an avatar or robot displayed on the screen of a display. It is also possible to use agents such as characters. In this case, the robot controller 24 and the robot sensor 26 of FIG. 1 are replaced by a display controller (not shown) for displaying such avatars and characters, and the dialogue location is near the display displaying the agent. can be assumed.

Furthermore, in place of the robot agent or the CG agent described above, an agent that only speaks can also be employed as an agent for dialogue with people. For example, if car navigation speakers are installed on the left and right sides of the vehicle, the voice heard from the left side is R1 (corresponding to the robot R1 in the embodiment), and the voice from the right side is R2 (the robot R2 in the embodiment). equivalent to ). In this case, the place of interaction is in a car, and the robot controller 24 of FIG. 1 is replaced by an audio controller (not shown) that controls the speech of such a voice agent.

In other words, the present invention is a dialogue system with humans using arbitrary agents.

10 ... dialogue system 12 ... dialogue place R1, R2 ... robot 18 ... sensor manager 20 ... group manager 22 ... dialogue database 24 ... robot controller

Claims (4)

A dialogue system comprising at least one agent at a dialogue location, wherein the agent at the dialogue location interacts with a person according to a dialogue,
a question sentence utterance unit that causes the agent to utter a question sentence;
A first determination unit that determines whether or not the person has uttered a response sentence to the question sentence,
a second determination unit that determines whether the response sentence hits a predetermined keyword when the first determination unit determines that the response sentence is uttered;
When the second determination unit determines that the response sentence hits the specific keyword, the agent utters a pre-phrase that invites the agent to next utterance after the response from the person. A dialogue system, comprising: a pre-phrase utterance unit; and a recognition response sentence utterance unit that, following utterance of the pre-phrase by the pre-phrase utterance unit, utters a recognition response sentence that creates a dialogue context for the pre-phrase. A dialogue system comprising at least one agent at a dialogue location, wherein the agent at the dialogue location interacts with a person according to a dialogue,
a question sentence utterance unit that causes the agent to utter a question sentence;
A first determination unit that determines whether or not the person has uttered a response sentence to the question sentence,
proxy response, wherein the first determination unit causes the agent to utter a proxy response sentence for responding to the question sentence on behalf of the person when the first determination unit does not determine that the response sentence has been uttered. A dialogue system comprising: a sentence utterance unit; and a recognition response sentence utterance unit that, following utterance of a proxy response sentence by the proxy response sentence utterance unit, utters a recognition response sentence that creates a dialogue context for the proxy response sentence. a computer of a dialogue system comprising at least one agent at a dialogue place, where the agent interacts with a person according to a dialog at the dialogue place;
a question sentence utterance unit that causes the agent to utter a question sentence;
A first determination unit that determines whether or not the person has uttered a response sentence to the question sentence,
a second determination unit that determines whether the response sentence hits a predetermined keyword when the first determination unit determines that the response sentence is uttered;
When the second determination unit determines that the response sentence hits the specific keyword, the agent utters a pre-phrase that invites the agent to next utterance after the response from the person. A dialogue system program that functions as a pre-phrase utterance unit, and a recognition response sentence utterance unit that utters a recognition response sentence that creates a dialogue context for the pre-phrase following the utterance of the pre-phrase by the pre-phrase utterance unit. . a computer of a dialogue system comprising at least one agent at a dialogue place, where the agent interacts with a person according to a dialog at the dialogue place;
a question sentence utterance unit that causes the agent to utter a question sentence;
A first determination unit that determines whether or not the person has uttered a response sentence to the question sentence,
proxy response, wherein the first determination unit causes the agent to utter a proxy response sentence for responding to the question sentence on behalf of the person when the first determination unit does not determine that the response sentence has been uttered. A dialogue system program that functions as a sentence utterance unit, and a recognition response sentence utterance unit that, following the utterance of the proxy response sentence by the proxy response sentence utterance unit, utters a recognition response sentence that creates a dialogue context for the proxy response sentence. .

Images