KR101964300B1 - Apparatus and method for implementing motion of robot using knowledge model - Google Patents

Abstract

Embodiments include a morpheme pattern generation unit for generating one or more morpheme patterns by dividing an utterance sentence into morpheme patterns including at least one morpheme, a gesture type corresponding to each morpheme pattern classified based on the knowledge model, And an operation command generation unit for generating an operation command for operating the robot in accordance with the applied gesture type, and relates to an apparatus and method for modeling and operating the robot using the knowledge model.

Description

[0001] APPARATUS AND METHOD FOR IMPLEMENTING MOTION OF ROBOT USING KNOWLEDGE MODEL [0002]

BACKGROUND OF THE INVENTION 1. Field of the Invention [ More specifically, the present invention relates to an apparatus and method for implementing an action on a social behavior expression of a robot to which a morphological pattern and a knowledge model are applied.

Various robot systems are being developed for busy modern man such as personal secretary robot that searches the user 's schedule and necessary information, helper robot that takes care of the elderly, home agent robot which performs on behalf of the user on behalf of errands or cleaning work. One of the parts to be considered in this service robot system is the interaction between human and robot. Recently, much attention has been paid to research for maximizing intimacy and bond between human and robot. In this regard, especially, gesture generation technology for expressing how human beings and robots communicate with each other and communicate and understand meaning can be considered as a major research field.

The issue in gesture generation technology is to efficiently store gesture information, to create various gestures, and to automatically analyze gesture sentences to generate appropriate gestures. By applying these three techniques, It is not easy to generate a gesture of a robot automatically in consideration of a spoken sentence and a context. Conventionally, a method of creating a gesture by generating emotions of a robot based on a change in emotion of a recognized user and combining facial expressions and actions stored therein is proposed. However, such a study can repeatedly generate the same gesture, and there is a limit to generating a finite number of gestures, which may degrade the user's concentration and interest. In addition, a conventional method of generating a type of gesture using a simple rule-based morphological pattern and applying a deformation based on manually input emotion information to a basic gesture connected to the keyword has been proposed. However, this study does not fully reflect the methodology for conveying meaning and feelings together with the contents of spoken sentences during the conversation with human beings, and it is not clear how the rules are defined, low. In addition, if the morpheme pattern is not defined as a rule, it is impossible to create a gesture type, and the gesture information associated with the keyword must be stored in all, which results in poor memory efficiency. Also, there is a problem that various gestures can not be generated unless the atmosphere of the sentence is manually specified.

Korean Patent Application No. 10-2013-0029395

 Synchronized gesture and speech production for humanoid robots (2010 IEEE / RSJ International Conference on Intelligent Robots and Systems October 18-22, 2010, Taipei, Taiwan)

In order to solve the above problems, it is necessary to understand not only the existing method of creating a motion by connecting the keyword and the gesture in a one-to-one manner but also the gesture type and keyword by using the morphological characteristics of the sentence, Model and apparatus and method using the same are required.

An apparatus for realizing an operation of a robot using a knowledge model according to an embodiment of the present invention includes a morpheme pattern generating unit for generating one or more morpheme patterns by dividing an utterance sentence into morpheme patterns including at least one morpheme, A gesture type application unit for applying a gesture type corresponding to each of the divided morpheme patterns to each morpheme pattern, and an operation command generation unit for generating an operation command for operating the robot according to the applied gesture type.

In addition, in the apparatus for implementing an operation of a robot using the knowledge model according to an embodiment, the utterance sentence may include a sentence input by a user or a response sentence to a question sentence input to the outside.

Also, in the apparatus for implementing an operation of a robot using a knowledge model according to an embodiment, the morphological pattern generator may generate one or more morphological patterns based on the knowledge model.

In addition, the gesture type of the robot using the knowledge model according to an exemplary embodiment may include at least one of a description of the shape and size of the object, a description of the position and state of the object, Describing a rhythmic element, having meaning without being generated with speech, and not performing a gesture can be included.

Further, in the apparatus for implementing an operation of a robot using a knowledge model according to an embodiment, the gesture type application unit may apply a gesture type only to a part of a morpheme pattern of an utterance sentence based on a query sent to the outside .

In an apparatus for implementing a robot using a knowledge model according to an embodiment,

The knowledge model may be generated based on data for one or more sample sentences in which a user generates a morphological pattern and applies a gesture type.

The gesture type may include at least one unit operation having one or more corresponding keywords, and the operation command generation unit may generate a morpheme pattern of the utterance sentence using the knowledge model according to one embodiment. The unit operation may be determined in the corresponding gesture type based on the included keyword, and the operation command may be generated based on the determined unit operation.

In an apparatus for implementing a robot using a knowledge model according to an embodiment,

Wherein the action command generation unit arbitrarily or predetermined unit action is determined when the keyword corresponding to the keyword included in the morpheme pattern of the utterance sentence does not exist in the gesture type of the morpheme pattern, and based on the determined gesture, Can be generated.

The apparatus further includes an emotion information generation unit that generates emotion information based on the sensing information received from the sensor, wherein the operation command generation unit generates the emotion information using the knowledge model, The motion command can be modified on the basis of the motion command.

In an apparatus for implementing a robot using a knowledge model according to an embodiment,

The operation command generation unit may modify at least one of a size of an operation and a position where a speed operation for performing an operation is implemented based on the emotion information.

A method of operating a robot using a knowledge model according to an embodiment of the present invention includes generating at least one morpheme pattern by dividing an utterance sentence into morpheme patterns including at least one morpheme, Applying a gesture type corresponding to each morpheme pattern to each morpheme pattern, and generating an action command to operate the robot according to the applied gesture type.

Also, in the method of implementing a robot operation using a knowledge model according to an embodiment, the utterance sentence may include a sentence input by a user or a response sentence to a question sentence input to the outside.

Also, in the method of operating a robot using a knowledge model according to an embodiment, the step of generating the morpheme pattern may generate one or more morpheme patterns based on the knowledge model.

The gesture type may be one or more of describing the shape and size of the object, describing the position and the state of the object, describing the operation of the instruction, May include at least one of depicting a rhythmic element, having a symbolic meaning, or not performing a gesture.

Further, in the method of implementing a robot operation using a knowledge model according to an exemplary embodiment, the step of applying the gesture type may include: applying a gesture type only to a part of morpheme patterns of the utterance sentence, Can be applied.

In another embodiment, the knowledge model may be generated based on data on one or more sample sentences in which a user generates a morphological pattern and applies a gesture type.

According to another embodiment of the present invention, there is provided a method of operating a robot using a knowledge model, wherein the gesture type includes at least one unit operation having a corresponding one or more keywords, The unit operation may be determined in the corresponding gesture type based on the keyword included in the morpheme pattern, and the operation command may be generated based on the determined unit operation.

According to another aspect of the present invention, there is provided a method of operating a robot using a knowledge model, the method comprising the steps of: generating a motion command, , It is possible to arbitrarily or to determine a predetermined unit operation and to generate an action command based on the determined gesture.

According to another aspect of the present invention, there is provided a method of operating a robot using a knowledge model, the method including generating emotion information based on sensed information input from a sensor, The operation command can be modified based on the operation command.

According to another aspect of the present invention, there is provided a method of operating a robot using a knowledge model, the method comprising: generating an action command based on the emotion information; At least one can be modified.

According to an embodiment of the present invention, various types of dialogue-actions are generated to maintain long-term interest and bond with the user. Unlike the existing technique of generating an arbitrary gesture when there is no keyword information, By using the gesture type based on the dialogue-action that fits the spoken sentence will be created to allow the blur of the conversation to continue naturally.

In addition, the present invention can be applied to a robot system that provides a service to a person, and is a device for generating various types of conversation-actions reflecting the gesture type matching the spoken sentence and the feelings of the robot. Unlike existing techniques that generate arbitrary gestures in the absence of keyword information, the proposed technique has the advantage of generating dialogue-actions that match the spoken sentence by using the gesture type based on the morpheme. That is, various types of gestures corresponding to the flow of conversation can be continuously generated, thereby giving the user a natural feeling and intimacy. In addition, it is possible to maximize the memory efficiency by including a method of separately managing the gestures having the same form without storing them in duplicate.

1 is a block diagram of an apparatus for implementing an operation of a robot using a knowledge model according to an embodiment of the present invention.
2 is a view for explaining the operation of the morphological pattern generating unit according to an embodiment.
3 is a view for explaining the operation of the knowledge model M used in an embodiment of the present invention.
4 is a diagram for explaining a gesture type applied to a spoken sentence.
5 is a diagram for explaining the relationship between the gesture type stored in the database 400 and the unit operation.
6 shows an example in which an operation command is modified based on emotion information.
7 is a flowchart of a method of implementing an operation of a robot using a knowledge model according to an embodiment of the present invention.

Embodiments described herein may be wholly hardware, partially hardware, partially software, or entirely software. A "unit," "module," "device," or "system" or the like in this specification refers to a computer-related entity such as a hardware, a combination of hardware and software, or software. A processor, an object, an executable, a thread of execution, a program, and / or a computer, for example, a computer, but is not limited to, a computer. For example, both an application running on a computer and a computer may correspond to a part, module, device or system of the present specification.

Embodiments have been described with reference to the flowcharts shown in the drawings. While the above method has been shown and described as a series of blocks for purposes of simplicity, it is to be understood that the invention is not limited to the order of the blocks, and that some blocks may be present in different orders and in different orders from that shown and described herein And various other branches, flow paths, and sequences of blocks that achieve the same or similar results may be implemented. Also, not all illustrated blocks may be required for implementation of the methods described herein. Furthermore, the method according to an embodiment of the present invention may be implemented in the form of a computer program for performing a series of processes, and the computer program may be recorded on a computer-readable recording medium.

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

1 is a block diagram of an apparatus for implementing an operation of a robot using a knowledge model according to an embodiment of the present invention. Referring to FIG. 1, an apparatus 1000 for implementing an operation of a robot using a knowledge model includes a morphological pattern generator 100, a gesture type application unit 200, and an operation command generator 300.

The morpheme pattern generation unit 100 may generate one or more morpheme patterns by dividing the utterance sentence into morpheme patterns including one or more morpheme. That is, the utterance sentence can be divided into one or more morpheme patterns. Here, the utterance sentence may be a sentence input by a user or a response sentence to an externally inputted question sentence, but is not limited thereto. For example, the sentence 'I am here', which is a response generated for a user inputting a typed or spoken phrase 'I am here' or an externally inputted question 'Where are you?' May be a spoken sentence. The utterance sentence can include all the sentences that the robot outputs. To this end, the device may further comprise a microphone.

2 is a view for explaining the operation of the morphological pattern generating unit according to an embodiment. Referring to FIG. 2, an exemplary spoken sentence 10 appears. The morpheme pattern generator 100 divides the utterance sentence 10 into morpheme units (1,2,3,4,5,6,7), and then generates morpheme patterns 11,12,13 , ≪ / RTI > 14). In one example, the morpheme pattern generation unit 100 may generate a morpheme pattern by user input. In another embodiment, the morpheme pattern generation unit 100 may generate morpheme patterns based on the knowledge model.

3 is a view for explaining the operation of the knowledge model M used in an embodiment of the present invention. The knowledge model M can generate morpheme patterns by analyzing the input sentence and dividing it into one or more morpheme patterns. To create an algorithm that distinguishes morpheme patterns, the user can preliminarily sort the input sentences by morpheme patterns and input the results into the knowledge model. Based on the inputted input sentence and the morpheme patterns, the knowledge model M can form an algorithm that divides input sentences into morpheme pattern units. Therefore, the more morpheme patterns the user inputs to the knowledge model M, the more accurate the morpheme patterns generated by the knowledge model can be. That is, the knowledge model can be gradually improved in function depending on use.

The gesture type application unit 200 may apply a gesture type corresponding to each of the divided morpheme patterns to each morpheme pattern based on the knowledge model. The gesture type application unit 200 can deduce and apply gesture types that can be generated in the utterance sentence based on the morpheme of each morpheme pattern.

The gesture type can be described by describing the shape and size of the object, depicting the location and state of the object, describing the dictation action, depicting the rhythmic element, depicting the symbolic meaning But does not perform the gesture), and does not perform the gesture. However, the present invention is not limited thereto.

For example, the gesture type can be divided into six types as shown in Table 1 below. In the table below, the gesture type shows a gesture type defined based on the Iconic, Metaphoric, Deictic, and Beat type gestures defined by McNeil in 2006 and the Emblem type gesture defined by Kendon in 1995.

Gesture type Justice Iconic Describes the shape and size of the object Metaphoric A description of the location and state of the object Deictic Indicating behavior Beat Rhythmic factor Emblem Describe symbolic meaning (meaning that it is not created with speech) Nothing Not to be expressed as a gesture

4 is a diagram for explaining a gesture type applied to a spoken sentence. The utterance sentence 10 is divided into one or more morpheme patterns 11, 12, 13, and 14 by the morpheme pattern generation unit 100. Then, the gesture type application unit 200 applies a corresponding gesture type to each morph pattern 11, 12, 13, and 14. Referring to Fig. 4, the morphological pattern 11 is applied to the gesture type (D). The uppercase alphabet shown in FIG. 4 represents the first alphabet of the gesture type described in Table 1 above. That is, in FIG. 4, the gesture type of the morpheme pattern 11 is applied as Deictic (D).

Tables 2 and 3 show exemplary output values of the values input to the gesture type application unit 200. [

Contents input In the room Fragrance full in here Print Iconic Nothing Metaphoric

Contents input Over there Big box one I have it Print Deictic Iconic Emblem Metaphoric

In addition, the gesture type application unit 200 may select and apply a gesture type that is important for meaning transfer considering the context and the gesture playback time. For example, the gesture type application unit 200 may be based on the theory that when creating a gesture for an utterance sentence, it recalls an image of the object and the object.

Tables 4 and 5 below are exemplary sentences showing that the selected gesture type is different for each context. Table 6 also shows a case in which a gesture to be emphasized is applied in a flexible manner even when the order of words in a spoken sentence is changed.

Question Where is the box? input Over there Big box one I have it Print Deictic - Emblem -

Question You need to pack a big thing? input Over there Big box one I have it Print - Iconic Emblem -

Question You need to pack a big thing? input Big box Over there one I have it Print Iconic - Emblem -

The gesture type applied to the utterance characters as described above can also be implemented by the knowledge model M described with reference to Fig. As described above, the knowledge model M can generate and evolve an algorithm that accumulates information on gesture types applied to each morphological pattern by a user and applies a gesture type. This knowledge model may also be referred to as a teacher learning model.

Also, as shown in Tables 5 and 6, the gesture type application unit 200 may apply the gesture type only to a part of the morpheme patterns of the utterance sentence based on the externally inputted question sentence. This is because, when a gesture type is applied to all the morpheme patterns constituting the utterance sentence and the operation is performed, the operation may not be performed smoothly, and on the other hand, it is advantageous to communicate only that the desired operation is performed.

Also, in one embodiment, the apparatus 1000 for implementing operations of a robot using a knowledge model may further include a database 400. [ The database 400 may store algorithm information and various other data information required for execution of the knowledge model M. [ Such a database 400 may also be located remotely.

5 is a diagram for explaining the relationship between the gesture type stored in the database 400 and the unit operation. Referring to FIG. 5, a gesture type and unit operation information may be stored in a portion 410 of a database. One gesture type may include one or more unit actions. Each unit operation may be, for example, a motion for drawing a square, an action for drawing a triangle, or an operation for pointing a far point. Each unit operation may also be matched to one or more keywords and stored in the database 400.

For example, the unitary action I ₂ is to draw a square, and the matched keyword may be 'tv', 'box', 'monitor' etc. related to the square. Accordingly, the unit operation is not matched for each keyword, and a plurality of keywords are connected to one unit operation, thereby reducing the burden of data amount.

The operation command generation unit 300 may generate an operation command for operating the robot according to the applied gesture type. The motion command generation unit 300 may generate an operation command for moving an arm, a leg, a neck, a waist, a hand, a foot, and the like of the robot. The driving unit 500 can operate a body part of the robot based on an operation command. The driving unit 500 may be configured to include a motor, an encoder, and the like.

As described in FIG. 5, the gesture type may include one or more unit actions having a corresponding one or more keywords. In this case, the action command generating unit 300 can determine the unit action in the corresponding gesture type based on the keyword included in the morpheme pattern of the utterance sentence, and generate the action command based on the determined unit action. For example, if the keyword 'box' exists in the morpheme pattern of the utterance sentence, the action command generating unit 300 can determine the matched unit operation of the keyword 'box'. The association between the 'gesture type-unit operation-keyword' and the word data can be stored in the database 400.

That is, one gesture type can be applied to one morph pattern, and the gesture type has a corresponding plurality of unit motions. Each unit operation has a plurality of corresponding keywords. Therefore, the operation command generation unit 300 can generate an operation command by selecting the unit operation corresponding to the keyword included in the morpheme pattern.

 For example, the gesture type I may be associated with a unitary action I1, I2, I3, etc., and each unit action I1 may have one or more keywords (e.g., tv, monitor, have.

In another embodiment, the keyword corresponding to the keyword included in the morpheme pattern may not be included in the action command associated with the gesture type. In this case, the operation command generation unit 300 may arbitrarily or determine a predetermined unit operation, and generate an operation command based on the determined gesture.

For example, a morphological pattern including a keyword 'box' is applied to a gesture type I, but a unit operation in which a keyword 'box' is matched may not exist. In this case, the operation command generation unit 300 can arbitrarily determine the unit operation in the gesture type applied to the morpheme pattern. In another example, the operation command generation unit 300 may set a basic unit operation of the gesture type and generate an operation command for the set basic unit operation when there is no matched unit operation.

In the above example, when the gesture type of the morphological pattern including the keyword " box " is M, the operation of touching the square or circle shape may be set as a basic unit operation for the gesture type M.

In addition, the apparatus 1000 for implementing the robot using the knowledge model may further include an emotion information generating unit for generating emotion information based on the sensing information input from the sensor. The sensor 600 is installed in a part of the robot and can sense a sensory signal such as temperature, sound, touch, temperature, touch from the outside. For example, when the touch sensor is attached to the head of the robot and the user strokes the head of the robot, the emotion information generating unit 700 can generate emotion information of satisfaction.

In one embodiment, the operation command generation unit 300 may modify the operation command based on the emotion information. 6 shows an example in which an operation command is modified based on emotion information. Referring to FIG. 6, before the emotion information is inputted, the motion command is that the robot draws a square 71 using two hands. However, when the emotion information generating unit 700 generates emotion information such as a feeling of humming based on the sensing information acquired by the sensor, the motion command generating unit 72 generates the motion information Can be reduced. Although the size of the operation is exemplified in the above description, the type of operation modified in other embodiments may include the size of the operation, the position where the operation is implemented, the speed, and the like, but the present invention is not limited thereto.

In addition, the operation command generation unit 300 can distinguish concepts and elements constituting the conversation behavior as shown in Table 7 below in order to generate an entirely natural operation for the utterance sentence.

Element Contents Start gesture - from the beginning of the firing sentence to the main posture standby gesture
Creating Meaningless Actions
- Create a natural gesture until the main gesture is created Main Attitude Waiting Gesture - Create the required standby action to the starting point for the main gesture presentation Main gesture - Keyword expression of selected type
- Actions with the main meaning of spoken sentences Main Posture Gesture - Create motion between the main gesture and the finish gesture Finishing gesture - Behavior returning to initial position

To this end, the operation command generation unit 300 may allocate the gesture reproduction time based on the utterance sentence, and may load the corresponding unit operation command from the gesture type and the keyword. And the speed of the unit operations can be adjusted according to the assigned time.

The above-described robot may be a humanoid robot having a limb similar to a human being, but the present invention is not limited thereto, and the robot may be included in the knowledge modeling and operation apparatus of the present invention.

7 is a flowchart of a method of implementing an operation of a robot using a knowledge model according to an embodiment of the present invention. The method for implementing the robot using the knowledge model can be implemented by the elements of the apparatus for implementing the robot using the knowledge model described above.

Referring to FIG. 7, a method of operating a robot using a knowledge model includes generating one or more morpheme patterns by dividing an utterance sentence into morpheme patterns including one or more morpheme (S100) A step S200 of applying a gesture type corresponding to each morpheme pattern to each morpheme pattern and a step S300 of generating an action command for operating the robot according to the applied gesture type. Here, the utterance sentence may include a sentence input by a user or a response sentence to an externally inputted question sentence.

In addition, the step of generating morpheme patterns (S100) may generate one or more morpheme patterns based on the knowledge model. Gesture types can also be used to describe the shape and size of an object, to describe the position and state of an object, to describe an instructional action, to describe a rhythmic element, , Or not performing a gesture, but the present invention is not limited thereto.

 In the step S200 of applying the gesture type, the gesture type can be applied only to a part of the morpheme pattern of the utterance sentence based on the externally inputted question sentence. The knowledge model may be generated based on data for one or more sample sentences in which the user generates a morphological pattern and applies a gesture type and the gesture type includes one or more unit actions having corresponding one or more keywords .

In operation S300, an operation command is generated by determining a unit operation in the corresponding gesture type based on the keyword included in the morpheme pattern of the utterance sentence, and generating the operation command based on the determined unit operation .

In yet another embodiment, the step of generating an operation command (S300) includes the steps of, if the keyword corresponding to the keyword included in the morpheme pattern of the utterance sentence does not exist in the gesture type of the morpheme pattern, And generate an action command based on the determined gesture.

The method for implementing the robot using the knowledge model according to an embodiment may further include generating emotion information based on the sensing information input from the sensor. In this case, the step of generating an action command (S300) may modify the action command based on the emotion information.

The step of generating an operation command (S300) may also modify at least one of the size of the operation, the speed at which the operation is performed, and the position where the operation is implemented based on the emotion information.

While the invention has been shown and described with reference to certain embodiments thereof, it will be understood by those skilled in the art that various changes and modifications may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. However, it should be understood that such modifications are within the technical scope of the present invention. Accordingly, the true scope of the present invention should be determined by the technical idea of the appended claims.

1 to 7: morpheme unit
10: spoken sentence
11 ~ 14: Morphological pattern
100: morpheme pattern generator
200: Gesture type application part
300: Operation command generation unit
400: Database
500:
600: Sensor
700: Emotion information generating unit
1000: Modeling and operation device
M: Knowledge model

Claims (20)

A morpheme pattern generation unit for generating one or more morpheme patterns by dividing the utterance sentence into morpheme patterns including two or more morpheme;
A gesture type applying unit for inferring a gesture type corresponding to each morpheme pattern in consideration of a context of each morpheme pattern or a gesture reproduction time based on a knowledge model; And
And an operation command generator for generating an operation command for operating the robot according to the inferred gesture type,
The knowledge model is a teacher learning model that forms an algorithm for classifying sentences pre-inputted from a user into morpheme pattern units. The more morpheme patterns input from a user, the more accurate the morphological patterns generated by the knowledge model, The functions of the knowledge model are gradually improved by accumulating information on the gesture types applied to the morpheme patterns,
The gesture type application unit applies the gesture type only to a part of the morpheme patterns of the utterance sentence based on the externally inputted question sentence and omits the application of the gesture type to the remaining morpheme patterns, Thereby improving the user's attainment of the doctor,
Wherein the operation command generation unit generates an operation command based on the inferred gesture in consideration of the context in the utterance sentence of each morpheme pattern or the gesture playback time when the gesture type does not exist in the morpheme pattern of the utterance sentence A device for realizing the motion of a robot using a knowledge model.
The method according to claim 1,
Wherein the spoken sentence includes a sentence input by a user or a response sentence to a question sentence input to the outside.
The method according to claim 1,
Wherein the morpheme pattern generator comprises:
And generates at least one morpheme pattern based on the knowledge model.
The method according to claim 1,
The gesture type includes:
At least one of depicting the shape and size of an object, describing the location and state of an object, describing an instructional action, describing a rhythmic element, having a symbolic meaning, or not performing a gesture Wherein the robot includes a robot and a robot.
delete The method according to claim 1,
The knowledge model,
Wherein the user generates a morphological pattern and is generated based on data on one or more sample sentences to which the gesture type is applied.
The method according to claim 1,
Wherein the gesture type comprises one or more unit actions having a corresponding one or more keywords,
Wherein the operation command generation unit comprises:
Determining a unit operation in the corresponding gesture type based on a keyword included in a morpheme pattern of the utterance sentence, and generating the operation command based on the determined unit operation.
delete The method according to claim 1,
A sensor for sensing a sensory signal according to a user's motion; And
And an emotion information generation unit for generating emotion information based on a signal according to an operation of the user,
Wherein the motion command generation unit modifies the motion command based on the emotion information.
10. The method of claim 9,
Wherein the operation command generation unit comprises:
Wherein at least one of a magnitude of an operation, a speed at which an operation is performed, and a position at which an operation is implemented is modified based on the emotion information.
Dividing the utterance sentence into morpheme patterns including two or more morphemes to generate at least one morpheme pattern;
Inferring a gesture type corresponding to each morpheme pattern in consideration of a context of each morpheme pattern or a gesture reproduction time based on a knowledge model; And
Generating an action command to operate the robot according to the inferred gesture type,
The knowledge model is a teacher learning model that forms an algorithm for classifying sentences pre-inputted from a user into morpheme pattern units. The more morpheme patterns input from a user, the more accurate the morphological patterns generated by the knowledge model, The functions of the knowledge model are gradually improved by accumulating information on the gesture types applied to the morpheme patterns,
The step of applying the gesture type may include applying a gesture type only to a part of a morpheme pattern of the utterance sentence based on an externally input question sentence and omitting application of a gesture type to the remaining morpheme patterns, To facilitate the operation of the user,
Generating the action command based on the inferred gesture in consideration of the context in the utterance sentence of each morpheme pattern or the gesture playback time when the gesture type for the morpheme pattern of the utterance sentence does not exist Wherein the method comprises the steps of:
12. The method of claim 11,
Wherein the spoken sentence includes a sentence input by a user or a response sentence to a question sentence input to the outside.
12. The method of claim 11,
Wherein the generating the morpheme pattern comprises:
And generating at least one morpheme pattern based on the knowledge model.
12. The method of claim 11,
The gesture type includes:
At least one of depicting the shape and size of an object, describing the location and state of an object, describing an instructional action, describing a rhythmic element, having a symbolic meaning, or not performing a gesture Wherein the step of generating the knowledge model comprises the steps of:
delete 12. The method of claim 11,
The knowledge model,
Wherein a user generates a morphological pattern and is generated based on data on one or more sample sentences to which a gesture type is applied.
12. The method of claim 11,
Wherein the gesture type comprises one or more unit actions having a corresponding one or more keywords,
Wherein generating the action command comprises:
Determining a unit operation in a corresponding gesture type based on a keyword included in a morpheme pattern of a spoken sentence, and generating the operation command based on the determined unit operation.
delete 12. The method of claim 11,
Sensing a sensory signal according to an operation of a user; And
Further comprising generating emotion information based on a signal according to an operation of the user,
Wherein the step of generating the action command modifies the action command based on the emotion information.
20. The method of claim 19,
Wherein generating the action command comprises:
Wherein at least one of a magnitude of an operation, a speed at which an operation is performed, and a position at which an operation is implemented is modified based on the emotion information.

Images