KR20200031245A - Adaptive robot communication system and method of adaptive robot communication using the same - Google Patents

Abstract

In an adaptive robot dialogue system and an adaptive robot dialogue method using the same capable of improving the satisfaction in interaction with a user, the adaptive robot dialogue system includes an initial dialogue model management unit, a dialogue generation unit, and a personal dialogue model management unit. The initial dialogue model management unit stores and updates an initial dialogue model and a dialogue model for each personality group. The dialogue generation unit initializes a personal dialogue model of a user by receiving an initial dialogue model from the initial dialogue model management unit based on personality information of the user and selects a plurality of utterances based on the personal dialogue model. The personal dialogue model management unit evaluates the personal dialogue model by compensating for each utterance, learns an evaluation result, and provides the evaluation result to the initial dialogue model management unit.

Description

ADAPTIVE ROBOT COMMUNICATION SYSTEM AND METHOD OF ADAPTIVE ROBOT COMMUNICATION USING THE SAME}

The present invention relates to an adaptive robot dialogue system and an adaptive robot dialogue method using the same, and more particularly, as a part of the interaction between the robot and the human, in the dialogue between the robot and the human, a utterance method in consideration of a user's personality By selecting, it relates to an adaptive robot dialogue system based on reinforcement learning that considers a user's personality to improve satisfaction in interaction with a user, and an adaptive robot dialogue method using the same.

As the service robot market expands and related research increases, the importance of effective use of robotic intelligent elements such as recognition, manipulation, and human-robot interaction is increasing. In particular, in the case of a dialogue system, which is a field of natural language processing, in the field of human-robot interaction, the recognition that the user's state is needed to improve the efficiency of communication with the robot and user satisfaction is also increasing.

However, it is not easy to understand the user's state, and understanding the user's state or changing the conversation system already designed to reflect the state of various users requires an additional cost in addition to the cost used for the pre-designed conversation system. It is difficult to implement easily.

However, Korean Patent Registration No. 10-0580619 discloses a technique for managing a conversation between a user and an agent by referring to a recent conversation history with a user, but in the case of a conversation system utilizing such so-called learning information Edo, there is a problem that a lot of time for learning, or the user's state is not accurately reflected.

Furthermore, as an important factor in determining a user's state, a technology related to a conversation system or a conversation model reflecting the user's personality is not yet developed.

Republic of Korea Registered Patent No. 10-0580619

Accordingly, the technical problem of the present invention has been devised in this regard, and the object of the present invention is to select a utterance method in consideration of the user's personality, and update and learn a personal conversation model based on the satisfaction level in interaction with the user. It is related to the adaptive robot dialogue system based on reinforcement learning considering the user's personality that can improve.

In addition, another object of the present invention relates to an adaptive robot dialog method using the adaptive robot dialog system.

The adaptive robot dialogue system according to an embodiment for realizing the object of the present invention includes an initial dialogue model management unit, a conversation generation unit, and a personal conversation model management unit. The initial dialogue model management unit stores and updates the initial dialogue model and the dialogue model for each personality group. The conversation generating unit receives the initial conversation model from the initial conversation model management unit based on the user's personality information, initializes the user's personal conversation model, and selects a plurality of speech sentences based on the personal conversation model. The personal conversation model management unit evaluates the personal conversation model by compensating for each of the utterances, learns the evaluation results, and provides the initial conversation model management unit.

In one embodiment, the initialized personal conversation model is a conversation model initialized according to a personality group to which the user belongs, and may include information on a speech that has high satisfaction in the personality group to which the user belongs.

In one embodiment, the personality group is divided into four personality groups based on the degree of openness and extroversion of the user, and in the initialized personal conversation model, speech sentences with high user satisfaction for each of the four personality groups are in order. Can be saved.

In one embodiment, the conversation generating unit may include a speech sentence selection unit for selecting the speech sentences based on the information on the speech with high satisfaction and the evaluation result.

In one embodiment, compensation for each of the utterances may be performed directly by the user.

In one embodiment, the initial conversation model management unit, the initial conversation model personalized for each conversation system, and a database unit for storing a conversation model distinguished by a user's personality group, based on evaluation results provided from the personal conversation model management unit It may include an initial conversation model update unit for updating the initial conversation model, and an initial conversation model transmission unit for transmitting the initial conversation model according to the request of the conversation generation unit.

In one embodiment, the conversation generating unit may further include a personality information requesting unit for requesting information to a conversation model classified for each personality group of the user, based on the acquired personality information of the user.

In one embodiment, the personal conversation model management unit includes a personal conversation model learning unit that learns the evaluation results of the personal conversation model, and a personal conversation model transmission unit that provides the learned evaluation results to the initial conversation model update unit. can do.

An initial dialogue model is designed based on the initial dialogue model stored in the adaptive robot dialogue method according to an embodiment for realizing another object of the present invention. In response to the request for the initial conversation model, the initial conversation model is transmitted to the conversation generation unit. Upon receiving the initial conversation model, the user's personal conversation model is initialized based on the user's personality information. A plurality of speech sentences are selected based on the personal conversation model. The personal conversation model is evaluated by compensating for each of the speeches. The evaluation result of the personal conversation model is learned. The initial conversation model is updated with the learned personal conversation model.

In one embodiment, selecting the utterances may include selecting each of the utterances in the personal conversation model, and sequentially speaking the selected utterances through a robot.

In one embodiment, evaluating the personal conversation model may include requesting an evaluation of the personal conversation model from a user, and sequentially evaluating each speech statement uttered through the robot. You can.

In one embodiment, in the step of initializing the personal conversation model, information on a utterance sentence having high satisfaction in a personality group to which the user belongs may be provided.

In one embodiment, in the step of selecting the plurality of speeches, the speeches may be selected based on information on the speech with high satisfaction and evaluation results of the personal conversation model.

According to embodiments of the present invention, in selecting a utterance method or a utterance sentence in a conversation between a robot and a human, satisfaction of interaction with a user may be improved by considering a user's personality.

In particular, the user can implement a more user-friendly adaptive robot dialogue system by performing an evaluation on whether the robot's utterance is satisfied, and learning the result of the evaluation by the robot to update the dialogue model.

In this case, in order to reflect the personality of the user, the personal conversation model is initialized based on the information on the preferred speech in the personality group to which the user belongs, by utilizing the pre-stored dialogue model for the personality group of the user. In order to further consider the personality factors of, a certain exploration is performed in the information on the preferred speech statement to select a speech statement and utter it to the user, so a speech statement that is more suitable for the user although not preferred by the personality group Alternatively, it is possible to select a ignition method. Through this, the adaptability between the robot and the user can be further improved.

In particular, in performing the predetermined exploration, the evaluation result for each speech statement is reflected and the personal conversation model is updated, so that it is possible to quickly select the optimal speech statement or speech method suitable for the user. .

In addition, since information on the preferred utterances for each personality group is utilized, in the initial state, utterances in a form completely independent of the user's personality are provided, thereby minimizing the time to select the optimal utterance or utterance method. Of course, it is possible to minimize dissatisfaction in communication without wasting unnecessary time to the user.

1 is a block diagram showing an adaptive robot dialog system according to an embodiment of the present invention.
2 is a flowchart illustrating an adaptive robot dialog method using the adaptive robot dialog system of FIG. 1.
3 is a flowchart illustrating a step of selecting a robot speech statement and evaluating a personal dialogue model in the adaptive robot dialogue method of FIG. 2.
4 is a graph illustrating an example of distinguishing a user's personality group in the robot dialogue system of FIG. 1.
FIG. 5 is a schematic diagram illustrating an example of an initialized personal conversation model applied according to a user's personality group in FIG. 4.
FIG. 6 is a schematic diagram illustrating an example of selecting a utterance statement by reflecting a user's personality factor in the robot dialogue system of FIG. 1 and updating a personal dialogue model through evaluation of the dialogue.

The present invention can be applied to various changes and can have various forms, and the embodiments are described in detail in the text. However, this is not intended to limit the present invention to a specific disclosure form, it should be understood to include all modifications, equivalents, and substitutes included in the spirit and scope of the present invention. In describing each drawing, similar reference numerals are used for similar components. Terms such as first and second may be used to describe various components, but the components should not be limited by the terms.

The terms are used only for the purpose of distinguishing one component from other components. The terms used in this application are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise.

In this application, terms such as “comprises” or “consisting of” are intended to indicate the presence of features, numbers, steps, operations, components, parts or combinations thereof described in the specification, one or more other features. It should be understood that the existence or addition possibilities of fields or numbers, steps, actions, components, parts or combinations thereof are not excluded in advance.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by a person skilled in the art to which the present invention pertains. Terms such as those defined in a commonly used dictionary should be interpreted as having meanings consistent with meanings in the context of related technologies, and should not be interpreted as ideal or excessively formal meanings unless explicitly defined in the present application. Does not.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram showing an adaptive robot dialog system according to an embodiment of the present invention. 2 is a flowchart illustrating an adaptive robot dialog method using the adaptive robot dialog system of FIG. 1. 3 is a flowchart illustrating a step of selecting a robot speech statement and evaluating a personal dialogue model in the adaptive robot dialogue method of FIG. 2. 4 is a graph illustrating an example of distinguishing a user's personality group in the robot dialogue system of FIG. 1. FIG. 5 is a schematic diagram illustrating an example of an initialized personal conversation model applied according to a user's personality group in FIG. 4. FIG. 6 is a schematic diagram illustrating an example of selecting a utterance statement by reflecting a user's personality factor in the robot dialogue system of FIG. 1 and updating a personal dialogue model through evaluation on this.

Hereinafter, the adaptive robot dialog system 10 and the adaptive robot dialog method using the same will be described with reference to FIGS. 1 to 6.

The adaptive robot dialog system 10 is a part of interaction between a robot and a human, and in performing conversation with each other, the robot provides a dialogue in a pattern or manner that is more familiar to the user or preferred by the user. For the sake of clarity, it will be described later, but it is a technology for implementing providing a conversation through appropriate utterance in consideration of a personality factor among various characteristics of a corresponding user.

First, referring to FIG. 1, the adaptive robot conversation system 10 includes an initial conversation model management unit 100, a conversation generation unit 200, and a personal conversation model management unit 300.

In addition, the initial conversation model management unit 100 includes a database unit 110, an initial conversation model update unit 120 and an initial conversation model transmission unit 130, and the conversation generation unit 200 is a personality information request unit (210), the initial conversation model request unit 220, a personal conversation model initialization unit 230, and a speech selection unit 240, wherein the personal conversation model management unit 300 is a personal conversation model evaluation unit 310 , A personal conversation model learning unit 320 and a personal conversation model transmission unit 330.

In the adaptive robot dialogue method using the adaptive robot dialogue system 10, as shown in FIG. 2, first, an initial dialogue model is designed based on the stored initial dialogue model (step S100).

The database unit 110 includes an initial dialogue model database 111 and a dialogue model database 112 for each personality group. In this case, the initial conversation model database 111 stores various types of conversation models that can be used in conducting conversations with a user, and specifically, the gender of the environment in which the conversation is performed and the personality of the user in which the conversation is performed. Various types of dialogue models according to characteristics such as age, age, and topic of dialogue are stored in the form of a database.

Therefore, in the adaptive robot conversation method in this embodiment, among the various types of conversation models stored in this way, an initial conversation model is designed in consideration of a conversation environment, a conversation topic, and user characteristics.

However, in the prior art, a dialogue system has been implemented in which the dialogue is performed by providing the initial dialogue model designed as such to the user. It is designed and dialogue is carried out.

That is, after the initial conversation model is designed as described above, and when a request for transmission of the initial conversation model is received from the initial conversation model request unit 220, the initial conversation model transmission unit 130 receives the initial conversation model request unit 220 ), The initial dialogue model is transmitted (step S200).

Meanwhile, the personal conversation model initialization unit 230 initializes the personal conversation model of the corresponding user based on the transmitted initial conversation model (step S300).

However, in the present embodiment, when initializing the personal conversation model, in addition to simply applying the pre-designed initial conversation model, the personal conversation model is initialized based on the personality information of the user.

In this case, in order to reflect the personality information of the user, the personality information request unit 210 is information about the conversation model distinguished according to the personality group of the user from the dialogue model database 112 for each personality group of the database unit 110. By requesting, a conversation model suitable for a user's personality is provided to the personal conversation model initialization unit 230.

Thus, the personal conversation model initialization unit 230 is based on the initial conversation model and the conversation model for each personality group reflecting the characteristics of the conversation model for each personality group according to the personality information of the user. Will initialize.

More specifically, in the conversation model database 112 for each personality group, the personality group to which the corresponding user belongs is classified according to the personality characteristics of various users, and the user's preferred personality group has a preferred or satisfactory speech or speech method. Information about the conversation model is stored.

In this case, the stored conversation model may not be said to be a speech or speech method that is preferred or highly satisfactory only to the corresponding user, and is generally referred to as a speech or speech method that is generally preferred or highly satisfactory to users belonging to the corresponding personality group. can do.

For example, referring to FIG. 4, among the Big5 factors, which are generally five factors that determine a user's personality, the user is based on the factors of openness (P _O , X axis) and extroversion (P _E , Y axis). The personality of can be divided into four groups (Group1 to Group4).

That is, if both openness and extroversion are high (> 0.5), it can be said to be a personality group belonging to Group1, and if openness is high (> 0.5) but low extroversion (<0.5), it can be said to be a personality group belonging to Group2. Similarly, if both openness and extroversion are low (<0.5), it can be said to be a personality group belonging to Group4, and if extroversion is high (> 0.5), but openness is low (<0.5), it can be said to be a personality group belonging to Group3.

As described above, according to the personality group to which the user's personality belongs, as shown in FIG. 5, the personal conversation models 231, 232, 233, and 234 may be initialized and set to match the personality group.

Thus, based on the personality of the user, through the personality information requesting unit 210, the conversation model database 112 for each personality group can request a conversation model for the personality group to which the user's personality belongs. The conversation model is initialized as a personal conversation model by the personal conversation model initialization unit 230 together with the initial conversation model.

The personal conversation model initialized in this way is illustrated in FIG. 6.

That is, if a conversation model corresponding to a personality group that matches a user's personality is selected from among the conversation models (231, 232, 233, 234) matched to the personality group, and is initialized as a personal conversation model together with the initial conversation model , As shown in detail in Figure 6, for each user, for each conversation state (S1, S2, ..., Sm), utterances or speech methods (A1, A2, ..., An) A dialogue model is defined, which includes information about preferences for (V1, V2, ...).

For example, if a user wants to inquire about the conversation state S1 called schedule registration, the utterance method A2 in which time and place information is simultaneously inquired rather than the utterance method A1 in which time and place information is separately inquired separately , If preferred by the personality group to which the user belongs, A2 section (V1) is provided with information having a higher priority than A1 section (V2) (meaning that V1 has higher priority than V2), so that the conversation model is defined. Can be.

However, as described above, even if the conversation model for the user is initialized by the personal conversation model initialization unit 230, the conversation model stored in the conversation model database 112 for each personality group is necessarily the most preferred or satisfactory to the user. It cannot be judged to be a high utterance or a utterance.

Therefore, in the present embodiment, a so-called predetermined exploration is performed in consideration of this in the step of selecting a utterance door in the utterance door selection unit 240 (step S400).

In this case, the exploration in the step of selecting the utterance statement is that the utterance or utterance method proposed through the dialogue model in the personality group to which the user's personality belongs is not the most preferred or satisfactory to the user, To find the most preferred or highly satisfactory speech or speech method for the user, a speech sentence or speech method that is relatively unfavorable or has low satisfaction in the speech or speech method proposed through the dialogue model in the personality group (hereinafter It means that the user's satisfaction is evaluated by performing utterance on the utterance statement.

That is, in the step of selecting the spoken sentence (step S400), the spoken sentence selection unit 240 selects a plurality of spoken sentences based on the initialized personal conversation model.

More specifically, in the speech selection unit 240, as shown in FIG. 6, preferences (V1, V2) based on preferences or satisfaction, etc. in the personality group to which the user belongs by the personal conversation model initialization unit 230 With respect to the set of utterance statements, ...), the utterance statements are sequentially selected in order of preference or satisfaction (steps S410 to S430), and each of the selected utterance statements is uttered through the robot to the user (step) S440).

In this case, the number of selected utterances may be variously selected, and may be sequentially selected for utterances having preference or satisfaction belonging to the top 30% or 50%.

On the other hand, it is not necessarily selected in the order of utterances having high preference or satisfaction, and may be selected in a predetermined order for utterances belonging to the upper level.

Furthermore, the criteria for the selected speech sentences may be set in various ways in addition to the above-described criteria.

Meanwhile, as described above, when a plurality of utterances selected by the utterance statement selection unit 240 are uttered through a robot, the personal conversation model evaluation unit 310 performs evaluation on each of the utterance utterances. It is made (step S500).

In this case, evaluation of each spoken sentence is performed by performing compensation for each spoken sentence by the user, and through this, evaluation of the initialized personal conversation model is performed.

The reward means that the user evaluates satisfaction or preference with respect to the ignition of the robot as a numerical value. When the satisfaction with the utterance of the robot is high, a relatively high score is obtained. The low score is compensated for, but the reward score can be compensated with an absolute value or with a relative value for other speeches.

That is, referring to FIGS. 3 and 6, in the step of selecting the utterance statement (step S400) and the step of evaluating the personal conversation model (step S500), the personal conversation model initialization unit 230 is initialized. From among the utterances defined in the personal conversation model, when the first utterance is selected in order of preference or satisfaction (step S410), and the utterance of the selected first utterance through the robot (step S440), the user After requesting the evaluation of the dialogue model (step S510), evaluation is performed by the user through compensation for the first speech statement (step S520).

As described above, when the evaluation is performed through the reward for the first utterance, the preference or satisfaction of the initialized personal conversation model is updated by the user's evaluation, and the conversation model for the user is updated.

Likewise, when the second utterance statement is selected (step S420), when the second utterance statement is uttered through the robot (step S440), a user is requested to evaluate (step S510), and the second utterance by the user The door is evaluated (step S530).

Similarly, if there are a total of n utterances selected as utterance targets, the same procedure is performed for the nth utterance statement and evaluated by the user (steps S430, S440, S510, S540), through which The step of selecting (step S400) and the step of evaluating the personal conversation model (step S500) are terminated.

As described above, when the evaluation of the personal conversation model ends, the personal conversation model is learned (step S600).

That is, in the personal conversation model learning unit 320, when evaluation is performed by the user on each of the selected utterances, the evaluation result for each of the utterances is the update data for the corresponding item as shown in FIG. Since it is input, learning is performed for each of these evaluation results.

In this case, the learning may be performed at the same time as the evaluation of each speech sentence is performed, or alternatively, the learning may be performed at once after the evaluation of all speech sentences is completed.

In addition, the user's evaluation result for each of the utterances is provided to the initial conversation model update unit 120 through the personal conversation model transmission unit 330, and the initial conversation model update unit 120 As a result of the same learning, the initial dialogue model is updated (step S700).

Thus, the initial dialogue model reflecting only the characteristics of the general group to which the user's personality belongs is updated by the dialogue model more suitable for the user by reflecting the evaluation result of the user and the learning result about it, and thus the updated dialogue model is used. By replacing or updating the initial dialogue model, a dialogue model more suitable for the user can be provided as the initial dialogue model.

Accordingly, the process of selecting a utterance in the utterance selection unit 240 described with reference to FIG. 6 is more easily performed, so that a utterance suitable for the user can be selected more easily.

Furthermore, the dialogue model database 112 for each personality group belonging to the group to which the user's personality belongs may be updated based on the updated initial dialogue model, and thus the updated dialogue model database 112 for each personality group is used. Therefore, it can be used for other users in the future.

According to the embodiments of the present invention as described above, in selecting a utterance method or a utterance sentence in a conversation between a robot and a human, satisfaction in interaction with a user can be improved by considering a user's personality. .

In particular, the user can implement a more user-friendly adaptive robot dialogue system by performing an evaluation on whether the robot's utterance is satisfied, and learning the result of the evaluation by the robot to update the dialogue model.

In this case, in order to reflect the personality of the user, the personal conversation model is initialized based on the information on the preferred speech in the personality group to which the user belongs, by utilizing the pre-stored dialogue model for the personality group of the user. In order to further consider the personality factors of, a certain exploration is performed in the information on the preferred speech statement to select a speech statement and utter it to the user, so a speech statement that is more suitable for the user although not preferred by the personality group Alternatively, it is possible to select a ignition method. Through this, the adaptability between the robot and the user can be further improved.

In particular, in performing the predetermined exploration, the evaluation result for each speech statement is reflected and the personal conversation model is updated, so that it is possible to quickly select the optimal speech statement or speech method suitable for the user. .

In addition, since information on the preferred utterances for each personality group is utilized, in the initial state, utterances in a form completely independent of the user's personality are provided, thereby minimizing the time to select the optimal utterance or utterance method. Of course, it is possible to minimize dissatisfaction in communication without wasting unnecessary time to the user.

Although described above with reference to the preferred embodiment of the present invention, those skilled in the art may variously modify and change the present invention without departing from the spirit and scope of the present invention as set forth in the claims below. You will understand that you can.

The adaptive robot dialogue system and the adaptive robot dialogue method using the same according to the present invention have the potential for industrial use that can be used in an adaptive robot interacting with a human.

10: Adaptive robot dialogue system
100: initial conversation model management unit 110: database unit
111: initial conversation model DB 112: conversation model DB by personality group
120: initial conversation model update unit 130: initial conversation model transmission unit
200: conversation generating unit 210: personality information request unit
220: initial conversation model request unit 230: personal conversation model initialization unit
240: speech door selection unit 300: personal conversation model management unit
310: personal conversation model evaluation unit 320: personal conversation model learning unit
330: personal conversation model transmission unit

Claims (13)

An initial dialogue model management unit that stores and updates the initial dialogue model and dialogue models for each personality group;
A conversation generating unit receiving the initial conversation model from the initial conversation model management unit based on the personality information of the user, initializing the user's personal conversation model, and selecting a plurality of speech sentences based on the personal conversation model; And
An adaptive robot dialogue system comprising a personal dialogue model management unit that evaluates the personal dialogue model by compensating for each of the spoken sentences, learns the evaluation results, and provides the initial dialogue model management unit. According to claim 1,
The initialized personal conversation model is a conversation model initialized according to a personality group to which the user belongs, and includes an adaptive robot dialog system characterized by including information on a speech with high satisfaction in the personality group to which the user belongs. According to claim 2,
The personality group is divided into four personality groups based on the degree of openness and extroversion of the user,
In the initialized personal conversation model, an adaptive robot conversation system characterized in that speech sentences with high user satisfaction are stored in order for each of the four personality groups. According to claim 2, The conversation generating unit,
An adaptive robot dialog system, comprising: a spoken sentence selection unit for selecting the spoken sentences based on the information on the highly spoken sentence and the evaluation result. According to claim 4,
Adaptive robot dialogue system, characterized in that the compensation for each of the utterances is performed directly by the user. According to claim 1, The initial conversation model management unit,
A database unit for storing an initial conversation model personalized for each conversation system and a conversation model distinguished for each user's personality group;
An initial conversation model update unit that updates the initial conversation model based on the evaluation result provided from the personal conversation model management unit; And
And an initial dialogue model transmission unit transmitting the initial dialogue model according to the request of the dialogue generation unit. The method of claim 6, wherein the conversation generating unit,
Based on the acquired personality information of the user, the adaptive robot dialog system further comprises a personality information requesting unit for requesting information to the conversation model classified for each personality group of the user. The method of claim 6, wherein the personal conversation model management unit,
A personal conversation model learning unit for learning evaluation and results of the personal conversation model; And
And a personal conversation model transmission unit that provides the learned evaluation result to the initial conversation model update unit. Designing an initial dialogue model based on the stored initial dialogue model;
Transmitting the initial conversation model to a conversation generator in response to a request for the initial conversation model;
Receiving the initial conversation model and initializing the user's personal conversation model based on the user's personality information;
Selecting a plurality of speech sentences based on the personal conversation model;
Evaluating the personal conversation model by compensating for each of the spoken sentences;
Learning an evaluation result of the personal conversation model; And
And updating an initial conversation model with the learned personal conversation model. The method of claim 9, wherein the step of selecting the utterance sentences,
Selecting each of the spoken sentences in the personal conversation model; And
And sequentially speaking the selected speech sentences through a robot. The method of claim 10, wherein evaluating the personal conversation model comprises:
Requesting a user to evaluate the personal conversation model; And
And sequentially evaluating each utterance statement uttered through the robot. 10. The method of claim 9, In the step of initializing the personal conversation model,
Adaptive robot communication method characterized in that it provides information on the speech with high satisfaction in the personality group to which the user belongs. The method of claim 12, In the step of selecting the plurality of speech sentences,
An adaptive robot communication method, characterized in that the speech sentences are selected based on the information on the speech with high satisfaction and the evaluation result of the personal conversation model.

Images