KR102191160B1 - System for modelling artificial intelligent data of robot data and modelling method thereof - Google Patents

Abstract

An objective of the present invention is to provide a method for modeling and processing artificial intelligent data of robot data to convert various elements such as image data through a robot, surrounding environment data, movement and driving data of a robot into data, transmit data to an artificial intelligence cloud server, model data into artificial intelligence data, analyze a surround environment of a robot, and assign a decision-making analysis function of a robot utilizing existing learned data. In order to achieve the above objective, a system for modeling artificial intelligent data of robot data according to the present invention includes a cloud server with an artificial intelligence modeling unit modeling situation information data collected by a situation information collection unit or previously learned situation data so that a service robot can use the information for decision-making.

Description

Artificial intelligent data modeling system of robot data and its modeling method {SYSTEM FOR MODELLING ARTIFICIAL INTELLIGENT DATA OF ROBOT DATA AND MODELLING METHOD THEREOF}

The present invention relates to an artificial intelligent data modeling and data processing method of robot data, and more particularly, a robot that processes data acquired using a server having an artificial neural network model generation function or an analysis function using an artificial neural network model It relates to an artificial intelligent data modeling and data processing method of data.

The artificial neural network, developed by simulating the human neural structure, is an artificial intelligence model that played a leading role in the development of artificial intelligence technology as a model that forms the basis of deep learning in the field of artificial intelligence.

However, the existing logic/rule-based model has limitations in its scope of application because it is difficult to clearly define all problems in the real world and cannot inform all situations and knowledge in the world.

Therefore, recently, artificial neural network technology has been applied to a field where it is difficult to apply the existing logic/rule-based model, and the performance and scalability of the artificial intelligence-based system are greatly improved.

On the other hand, artificial neural networks use artificial neural network models in a mobile device, application, or computer environment with limited power, because many operations are performed to process input data, many weights are required, and many intermediate result traffic is generated. There is a problem that it is difficult.

Republic of Korea Patent Publication No. 10-2017-0101455 (published on September 6, 2017)

An object of the present invention for solving the conventional problems as described above is to convert not only image data through a robot, but also various elements such as surrounding environment data, movement and driving data of the robot, etc., and transmit the data to an artificial intelligence cloud server to provide artificial intelligence. It provides an artificial intelligent data modeling and data processing method of robot data that analyzes the environment around the robot after modeling with data, and gives the robot's own decision-making analysis function by using the existing learned data.

Another object of the present invention is to provide a large number of resources required for artificial intelligence through a cloud service based on wireless communication in order to minimize the resources required for state recognition such as object recognition and face recognition around the robot, thereby making the robot more economical. It is to provide an artificial intelligent data modeling and data processing method of robot data that can perform and expand service functions.

In order to achieve the above object, the artificial intelligent data modeling system of robot data according to the present invention is provided with context information data or previously learned context data collected by the context information collection unit so that the service robot can use it for decision making. It characterized in that it comprises a cloud server having an artificial intelligence modeling unit for modeling.

In addition, in the artificial intelligent data modeling system of robot data according to the present invention, the context information collection unit includes a visual sensor that recognizes an object and recognizes a distance between the object and the service robot.

In addition, in the artificial intelligent data modeling system of robot data according to the present invention, the context information collection unit recognizes the compressive force of the service robot hand for gripping the object, and a tactile sensor that recognizes the friction force between the object and the service robot hand. It characterized in that it comprises a.

In addition, in the artificial intelligent data modeling system of robot data according to the present invention, the context information collection unit recognizes the hardness, texture, and weight of the object to be gripped, and the movement direction for gripping the object, and the object. It characterized in that it comprises a tactile sensor for recognizing a state of vibration generated during the grip.

In addition, in the artificial intelligent data modeling system of robot data according to the present invention, the service robot includes a response data output unit that outputs response data for context information from the cloud server.

In addition, in the artificial intelligent data modeling system of robot data according to the present invention, the cloud server comprises: an artificial intelligence data storage unit for storing pre-learned context data; And a context recognition result derivation unit for providing the modeled context information model to the service robot.

In addition, in the artificial intelligent data modeling system of robot data according to the present invention, when modeling by the artificial intelligence modeling unit, a consistency verification step (S10) of performing a theoretical verification whether the collected context information is appropriate data; includes; Characterized in that.

In addition, in the artificial intelligent data modeling system of robot data according to the present invention, the collected situation information and the situation data previously learned from the artificial intelligence data storage unit are combined, or each of a position and a position and a position and a value according to the passage of time are individually , A modeling step (S20) of classifying according to size and image, analyzing and analyzing the classified context information according to user requirements, learning by a data learning unit, and converting it into a data industry standard language (PMML); It characterized in that it comprises a.

In addition, in the artificial intelligent data modeling system of robot data according to the present invention, the context information model meeting the requirements among the plurality of modeled context information models is recommended by the artificial intelligence modeling unit and provided to the service robot. It characterized in that it comprises a data application step (S30).

In addition, in the artificial intelligent data modeling system of robot data according to the present invention, the cloud server is characterized in that it further comprises a context information processing unit for processing the context information data.

In addition, in order to achieve the above object, the artificial intelligent data modeling method of robot data according to the present invention is an artificial intelligent data modeling method of robot data for modeling collected context information data or pre-learned context data. It characterized in that it comprises a; consistency verification step (S10) of performing a theoretical verification whether the context information is appropriate data.

In addition, in the artificial intelligence data modeling method of robot data according to the present invention, the collected situation information and the situation data previously learned from the artificial intelligence data storage unit are combined, or each individual position and , A modeling step (S20) of classifying according to size and image, analyzing and analyzing the classified context information according to user requirements, learning by a data learning unit, and converting it into a data industry standard language (PMML); It characterized in that it comprises a.

In addition, in the artificial intelligent data modeling method of robot data according to the present invention, the context information model meeting the requirement among the plurality of modeled context information models is recommended by an artificial intelligence modeling unit and provided to the service robot. It characterized in that it includes; data application step (S30).

Details of other embodiments are included in "Specific Contents for Carrying out the Invention" and the attached "Drawings".

Advantages and/or features of the present invention, and a method of achieving them will become apparent with reference to various embodiments described below in detail together with the accompanying drawings.

However, the present invention is not limited to the configuration of each embodiment disclosed below, but may also be implemented in various different forms, and each embodiment disclosed in the present specification makes the disclosure of the present invention complete, and the present invention It should be understood that the present invention is provided to completely inform the scope of the present invention to those of ordinary skill in the art to which it belongs, and that the present invention is only defined by the scope of each claim in the claims.

According to the present invention, not only image data through the robot, but also various elements such as surrounding environment data, robot movement and driving data, etc., are converted into data and transmitted to an artificial intelligence cloud server, modeled as artificial intelligence data, and then the surrounding environment of the robot is analyzed. It has the effect of giving the robot's own decision-making analysis function by using the existing learned data.

In addition, according to the present invention, in order to minimize the resources required for state recognition such as object recognition and face recognition around the robot, many resources required for artificial intelligence are provided through a cloud service based on wireless communication. It has the effect of performing and expanding the robot service function.

1 is a system configuration diagram showing the overall configuration of an artificial intelligent data modeling system for robot data according to the present invention.
Figure 2 is a block diagram showing the configuration of a service robot in the artificial intelligent data modeling system of robot data according to the present invention.
3 is a block diagram showing the configuration of a cloud server in the artificial intelligent data modeling system for robot data according to the present invention.
Figure 4 is a flow chart showing the flow of the artificial intelligent data modeling method of robot data according to the present invention.

Before describing the present invention in detail, terms or words used in the present specification should not be interpreted as being unconditionally limited in a conventional or dictionary meaning, and in order for the inventor of the present invention to describe his invention in the best way It should be understood that the concepts of various terms can be appropriately defined and used, and furthermore, these terms or words should be interpreted as meanings and concepts consistent with the technical idea of the present invention.

That is, the terms used in this specification are only used to describe a preferred embodiment of the present invention, and are not intended to specifically limit the content of the present invention, and these terms represent various possibilities of the present invention. It should be noted that this is a term defined in consideration.

In addition, in this specification, it should be understood that the singular expression may include a plurality of expressions unless clearly indicated in a different meaning in the context, and may include the singular meaning even if similarly expressed in the plural. .

Throughout the present specification, when a component is described as "including" another component, it does not exclude any other component, but further includes any other component unless otherwise indicated. It could mean you can do it.

Furthermore, when a component is described as "existing inside or connected to and installed" of another component, this component may be directly connected to or installed in contact with other components, It may be installed spaced apart by a distance, and in the case of installation spaced apart by a certain distance, a third component or means may exist for fixing or connecting the component to other components. It should be noted that a description of the elements or means of 3 may be omitted.

On the other hand, when a component is described as being "directly connected" to another component or "directly connected", it should be understood that there is no third component or means.

Likewise, other expressions describing the relationship between each component, such as "between" and "directly between", or "neighbor to" and "directly neighbor to" have the same effect. Should be interpreted as.

In addition, in this specification, terms such as "one side", "the other side", "one side", "the other side", "first", "second", etc., if used, this one component for one component It is used in order to be clearly distinguishable from other components, and it should be noted that the meaning of the component is not limited by such terms.

In addition, terms related to positions such as "upper", "lower", "left", and "right" in the present specification, if used, should be understood as indicating a relative position in the drawing with respect to the corresponding component, These position-related terms should not be understood as referring to absolute positions unless absolute positions are specified for their positions.

In addition, in the present specification, in specifying the reference numerals for each component of each drawing, the same reference numerals for the same components, even if the components are indicated in different drawings, that is, the same reference throughout the specification. The symbols indicate the same components.

In the drawings attached to the present specification, the size, position, and coupling relationship of each component constituting the present invention are partially exaggerated, reduced, or omitted in order to sufficiently clearly convey the spirit of the present invention or for convenience of description. It may have been described, and therefore its proportion or scale may not be exact.

Further, in the following description of the present invention, a detailed description of a configuration that is determined to unnecessarily obscure the subject matter of the present invention, for example, a known technology including the prior art, may be omitted.

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

1 is a system configuration diagram showing the overall configuration of an artificial intelligent data modeling system for robot data according to the present invention.

Referring to FIG. 1, an artificial intelligent data modeling system 1000 of robot data according to the present invention includes a service robot 100 and a cloud server 200.

Since the cloud server 200 can implement an artificial neural network model, it is possible to provide a system in which the learning and classification functions of artificial intelligence are extended by providing the result of performing the artificial neural network model even in a device that is difficult to have an artificial intelligence function on its own. have.

The cloud server 200 may include software that provides functions such as artificial intelligence modeling technology and a user interface (UI).

Specifically, the software may perform data preprocessing, machine learning, deep learning, etc. through the cloud server 200, visualize it, and adjust various variables and perform computing.

That is, the cloud server 200 and software may provide artificial intelligence modeling, programming, compilation, result analysis, interaction design, or computing function for this, along with a database role.

In addition, the cloud server 200 and software may provide functions such as a storage device and a computing workstation, as well as a function of analyzing various programs, web-based real-time compilation, and the like.

Meanwhile, the service robot 100 may provide data acquired from various sensors provided to the cloud server 200 as data for classification of an already learned artificial intelligence model.

The service robot 100 may provide the acquired data to the cloud server 200 as training data of an artificial neural network model.

The cloud server 200 may receive classification data and apply it as an input value to an already learned artificial intelligence model.

In addition, the cloud server 200 may receive training data and use it as training data for artificial intelligence modeling.

At this time, the cloud server 200 may build a new learning database using the received learning data.

In addition, the service robot 100 may receive and output an analysis result or the like from the server 200.

Accordingly, it is possible to provide a user with an artificial intelligent data modeling system 1000 of robot data linking hardware and software capable of generating and outputting an artificial neural network model and designing an interaction.

2 is a block diagram showing the configuration of a service robot in the artificial intelligent data modeling system of robot data according to the present invention, and FIG. 3 is a block diagram showing the configuration of a cloud server in the artificial intelligent data modeling system of robot data according to the present invention. to be.

2, in the artificial intelligent data modeling system 1000 of robot data according to the present invention, the service robot 100 includes a context information collection unit 110, a control unit 120, a communication unit 130, and It includes a response data output unit 140.

Here, the context information collection unit 110 serves to collect information on the surrounding environment or surrounding situation of the service robot 100 using various sensors.

The controller 120 plays a role of controlling the movement or movement of the service robot 100.

The communication unit 130 serves to transmit the context information collected by the context information collection unit 130 to the cloud server 200.

In addition, the cloud server 200 serves to receive a model for the context information 10 requested by the customer for which data modeling has been completed.

The response data output unit 140 serves to output the context information model received by the communication unit 130.

In addition, referring to FIG. 3, in the artificial intelligent data modeling system 1000 of robot data according to the present invention, the cloud server 200 includes an artificial intelligence modeling unit 210, an artificial intelligence data storage unit 220, and It includes a context information processing unit 230, a data learning unit 240, and a context recognition result derivation unit 250.

The artificial intelligence modeling unit 210 receives the context information 10 data collected by the context information collection unit 110 or the previously learned context data, and performs a modeling so that the service robot 100 can use it for decision making. Perform.

The artificial intelligence data storage unit 220 serves to store previously learned context data.

The context information processing unit 230 serves to receive the context information 10 collected by the context information collection unit 130 and process the service robot 100 to use it for decision making.

Here, gripping based on a learning algorithm using only visual information is not easy to obtain a force relationship model due to contact with an object, and is simplified and utilized through several assumptions such as a rigid body model, but to increase the usability, the tactile sense that can grasp the contact state You need informational learning skills.

Accordingly, in the artificial intelligent data modeling system 1000 of robot data according to the present invention, the received situation information 10 is processed by the situation information processing unit 230 in a total of three steps.

Specifically, in step A, correlation matching is performed by measuring a distance between two observation values of the observation values of the context information 10, respectively, to measure a degree of association-based dissimilarity.

Here, correlation matching is performed by Equation 1 below.

[Equation 1]

는 변수값들의 비유사도이고, 모든 i,t∈{1,2,...,p}이고, 모든 j,k∈{1,2,...,qⁱ}이며,

은 확률 분포의 거리 계산 함수이다.here,

Is the dissimilarity of the variable values, all i,t∈{1,2,...,p}, all j,k∈{1,2,...,q ⁱ },

Is the distance calculation function of the probability distribution.

Next, in step B, a value having a high cross probability is obtained by measuring the frequency of crossing between values measured by association matching.

Finally, in step C, cluster analysis is performed using the distance between values measured by association matching.

Here, in the cluster analysis, after forming a cluster for the measured value, one of the actual value and the measured value included in the cluster close to the actual value is used as a representative value.

At this time, it is very important to minimize the sum of the distances between the measured value and the representative value.

In more detail, Rand Index is used as a measure to compare the performance of cluster analysis.

The Rand Index is a method of comparing whether the results of two cluster analysis are similar.The numerator of the Rand Index is the frequency of whether two identical pairs of observations belong to the same cluster or different clusters, and the denominator is the number of all observation pairs. It's a number.

As the volume of data collected with the advancement of IT technology becomes very large and complex, analysis methods other than traditional statistical analysis methods are required.

Accordingly, data mining and machine learning analysis methods are desirable to extract useful information from a large amount of complex data and find meaningful rules or patterns therefrom.

Unlike in the past, observations measured with categorical variables or qualitative variables, rather than observations measured with continuous variables, are increasingly being collected.

In the present invention, an Association-Based Dissimilarity method is used, which is a method for measuring distances of categorical variables in consideration of associations between variables.

First, the first step measures the distance between two observations made up of categorical variables.

To do this, it is first necessary to measure the dissimilarity between these values when various values of a categorical variable are given.

This is a method of calculating the dissimilarity between the values of two variables by using the distances of the conditional probability distribution of the values of other variables. When a categorical variable is composed, cluster analysis is performed by measuring the distance between observations considering the association between the variables. .

A data set consisting of n observations is X={x ₁ ,x ₂ ,...x _n } , and each observed observation is composed of a p-dimensional categorical variable.

Categorical variable A _i has q _i variable values, and A _i ={a _i1 ,a _i2 ,...,a _iq ⁱ } .

For the distance between two observations composed of categorical variables, the dissimilarity of the variable values of the categorical variable is first calculated, and the distance between the two observations is calculated using this dissimilarity.

In the present invention, the dissimilarity of the variable values of the categorical variable is d(a _ij ,a _ik ) , and the distance between the two observed values is d(x _g ,x _h ) .

[Equation 1]

는 변수값들의 비유사도이고, 모든 i,t∈{1,2,...,p}이고, 모든 j,k∈{1,2,...,qⁱ}이며,

은 확률 분포의 거리 계산 함수임.here,

Is the dissimilarity of the variable values, all i,t∈{1,2,...,p}, all j,k∈{1,2,...,q ⁱ },

Is the distance calculation function of the probability distribution.

In the second step, cluster analysis is performed based on the distance between two observations measured based on the association between variables.

The categorical variable handled in the present invention cannot be expressed in numerical coordinates, and only distances between observations can be calculated.

That is, since the association-based dissimilarity method calculates the distance between observations, the PAM method is used as a cluster analysis method in the present invention.

The PAM method uses medoids that are less sensitive to outliers and missing values as representative values of the cluster.

Here, medoid means one of the observations belonging to the cluster.

The PAM method forms a cluster by finding a medoid that minimizes the sum of the distances between the observed value and the nearest medoid.

The data learning unit 240 serves to learn the context information 10 from the data learning unit 260.

The context recognition result derivation unit 250 serves to provide the modeled context information model to the service robot 100.

In addition, the cloud server 200 may further include a data receiving unit that receives the context information 10 from the service robot 100 or a data storage unit that stores the context information 10.

Meanwhile, in the artificial intelligent data modeling system 1000 of robot data according to the present invention, the situation information 10 is preferably at least one of image information, driving information, location information, gripping information, distance information, and voice information. It is not limited and may include all information obtained by various sensors.

With the configuration as described above, the artificial intelligent data modeling system 1000 of robot data according to the present invention receives the context information 10 data collected by the context information collection unit 130 or the previously learned context data. It includes a cloud server 200 having an artificial intelligence modeling unit 210 that models the service robot 100 to be used for decision making.

Here, the context information collection unit 110 includes various sensors such as a visual sensor, a tactile sensor, a tactile sensor, and a gyro sensor.

The visual sensor is a sensor that recognizes a distance between an object or an object and the service robot 100.

In addition, the tactile sensor is a sensor that recognizes the compressive force of the hand of the service robot 100 for gripping the object or object, and recognizes the friction force between the object or object and the hand of the service robot 100.

On the other hand, the tactile sensor is a tactile sensor that recognizes the hardness, texture, and weight of an object or object to be gripped, the direction of movement to grip the object or object, and a vibration state that occurs when the object or object is gripped. .

In this way, the data collected by the visual sensor is learned by the data learning unit 260.

In addition, the data collected by the tactile sensor and the data collected by the tactile sensor are learned together by the data learning unit 240.

The gripping point is detected through the data obtained by the learned visual sensor and the learned data by the tactile sensor and the tactile sensor.

Considering that the main function of the service robot 100 is to physically interact through contact with the outside, unlike the existing industrial robot, the cooperative robot or service robot 100 that requires performing multiple tasks in an unstructured environment is a robot Hand gripping and manipulation techniques are more important.

Recently, artificial intelligence technology is being used in various fields, and combined with robot technology (especially vision technology), it is being used as a core technology for part gripping and assembly handling in changes in manufacturing environments such as smart factories.

In order to handle parts gripping and assembly handling, object recognition and distance information recognition are performed by a visual sensor, and a Convolutional Neural Network (CNN) algorithm may be used.

However, in the gripping and manipulation functions through physical interaction, there is a difficulty in stable gripping due to nonlinear features (contact, friction, deformation, etc.) that cannot be easily modeled and sensor uncertainty. The information by the tactile sensor can be considered important.

Human skin has a high-density tactile sensor (tactile receptor) capable of detecting the hardness, texture, weight, movement direction and vibration of an object, and the role of the tactile sensor is classified according to the stimulation.

Not only is the function of the robot limited by visual recognition alone, it is difficult to create a robot that can respond quickly to a dynamic environment.

To this end, it is meaningless to simply transmit all tactile signals, and a tactile recognition technology that can secure necessary information using tactile signals is required.

In the existing robot control based on mechanics, it is possible to develop a gripping and manipulation technology using an equation representing the force relationship between the object and the robot based on the exact kinematics and dynamic model of the object to be gripped and the robot. In real situations where there are nonlinear forces such as points that are not accurately known and friction forces that cannot be accurately modeled, a gripping technique using a learning algorithm is required.

In addition, gripping based on learning algorithms using only visual information is not easy to obtain a force relationship model due to contact with an object, and it is simplified and used through several assumptions such as a rigid body model, but to increase the usability, it is possible to grasp the contact state. You need informational learning skills.

Accordingly, in the artificial intelligent data modeling system 1000 of robot data according to the present invention, the situation information 10 collected by the artificial intelligence modeling unit 210 and the previously learned situation data are modeled in a total of three steps. Made by

In the first step, the consistency verification step (S10), a theoretical verification is performed whether the collected context information 10 is appropriate data.

In addition, in the second step, the modeling step (S20), the collected situation information 10 and the situation data previously learned from the artificial intelligence data storage unit 220 are combined, or the location and the position and the position and each individually , Size and image, the classified context information 10 is analyzed and analyzed according to the user's requirements, is learned by the data learning unit 240, and then converted into a data industry standard language (PMML).

Finally, in the third step, the data application step (S30), a context information model that meets the requirements among the plurality of modeled context information models is recommended by the artificial intelligence modeling unit 210 and provided to the service robot 100 do.

Specifically, in the consistency verification step (S10), context information data is selected, the validity of the selected context information data is analyzed, and the analyzed context information data is classified.

That is, with the result of applying the algorithm and data collected after driving the service robot 100 to the service robot, a theoretical verification of whether the data is appropriate to improve the movement/manipulation/social performance of the service robot 10 Perform.

In the modeling step S20, after pre-processing the context information data for which the theoretical verification has been completed, the user's requirements are analyzed.

After that, data transformation and learning are performed.

Here, in the data conversion, raw data is patched, CSV conversion is performed, and data industry standard language conversion is performed.

At this time, the collected data is classified by time, location, size, image, etc. according to characteristics over time, and converted into a standard format that robots can use in common.

In the data application step (S30), the modeled data is searched, and the most effective artificial intelligence data is recommended by the artificial intelligence modeling unit 210, and then the recommended useful data is downloaded and applied to the service robot 100. .

That is, in the present invention, the basic work for converting context information data into artificial neural network data and utilizing it as standard-based data is performed using PMML, which is a data industry standard language used to represent data mining.

Predictive Analytic Models and Data Mining Models are terms that refer to mathematical models for learning patterns hidden in large amounts of historical data by using statistical techniques.

The predictive analysis model predicts the existence of perceived patterns in new data using knowledge acquired during education.

PMML, the data industry standard language, makes it easy to share predictive analytics models between different applications.

Therefore, a user can learn one model in the system, can express its contents in PMML, which is a data industry standard language, and can transfer the contents of a model used in another system and use it for prediction such as machine failure.

Therefore, most of the current data mining solutions provide input or output functions in PMLL, which is a standard language for the data industry.

4 is a flow chart showing the flow of the artificial intelligent data modeling method of robot data according to the present invention.

Referring to FIG. 4, the method for modeling artificial intelligent data of robot data according to the present invention includes a total of three steps.

In the consistency verification step (S10), as an artificial intelligent data modeling method of the collected context information 10 data or robot data modeling pre-learned context data, a theoretical verification of whether the collected context information 10 is appropriate data is performed. And a conformity verification step (S10) to be performed.

In addition, in the modeling step (S20), the collected context information 10 and the context data previously learned from the artificial intelligence data storage unit 210 are combined, or individually, the location, size, and Classified according to images and analyzed and analyzed according to the user's requirements of the classified context information 10, is learned by the data learning unit 240, and then converted into a data industry standard language (PMML).

Finally, in the data application step (S30), a context information model meeting a requirement among a plurality of modeled context information models is recommended by the artificial intelligence modeling unit 210 and provided to the service robot 100.

Accordingly, according to the present invention, various elements such as image data through the robot as well as surrounding environment data, movement and driving data of the robot are converted into data and transmitted to the artificial intelligence cloud server, modeled as artificial intelligence data, and then the environment around the robot is analyzed And, it has the effect of giving the robot's own decision-making and analysis function by using the existing learned data.

In addition, according to the present invention, in order to minimize the resources required for state recognition such as object recognition and face recognition around the robot, many resources required for artificial intelligence are provided through a cloud service based on wireless communication. It has the effect of performing and expanding the robot service function.

In the above, several preferred embodiments of the present invention have been described with some examples, but the description of the various various embodiments described in the "Specific Contents for Carrying out the Invention" section is merely exemplary, and the present invention is Those of ordinary skill in the art to which it belongs will be well understood from the above description that the present invention can be variously modified and implemented or equivalent to the present invention.

In addition, since the present invention can be implemented in a variety of other forms, the present invention is not limited by the above description, and the above description is intended to complete the disclosure of the present invention, and is generally used in the technical field to which the present invention pertains. It should be understood that it is provided only to completely inform the scope of the present invention to those skilled in the art, and that the present invention is only defined by each claim of the claims.

10: situation information
100: service robot
110: situation information collection unit
120: control unit
130: communication department
140: response data output unit
200: cloud server
210: artificial intelligence modeling unit
220: artificial intelligence data storage unit
230: situation information processing unit
240: data learning department
250: situation recognition result derivation unit
1000: artificial intelligent data modeling system of robot data

Claims (13)

A service robot that provides the context information data collected by the context information collection unit as learning data for artificial intelligence modeling; And
Includes; a cloud server that receives the learning data provided from the service robot and uses it as learning data for artificial intelligence modeling,
The service robot,
A context information collection unit that collects information on the surrounding environment or surrounding situation of the service robot using various sensors,
And a communication unit for transmitting the context information collected by the context information collecting unit to the cloud server and receiving a result of the artificial intelligence modeling completed by the cloud server,
The cloud server,
An artificial intelligence data storage unit that stores pre-learned context information data,
An artificial intelligence modeling unit that receives the context information data from the service robot and the pre-learned context information data from the artificial intelligence data storage unit and performs the artificial intelligence modeling so that the service robot can use it for decision making;
A context information processing unit that processes the context information data provided so that the service robot can make a decision,
And a data learning unit for learning the context information data,
The situation information collection unit,
A visual sensor that recognizes an object and recognizes a distance between the object and the service robot;
A tactile sensor that recognizes a compressive force of a service robot hand for gripping the object, and recognizes a friction force between the object and the service robot hand; And
A tactile sensor that recognizes the hardness, texture, and weight of an object to be gripped, a moving direction to grip the object, and a vibration state generated when the object is gripped,
The data collected by the visual sensor is learned by the data learning unit,
Learning the data collected by the tactile sensor and the data collected by the tactile sensor together by the data learning unit,
A gripping point is detected through the learned data by the visual sensor and the learned data by the tactile sensor and the tactile sensor,
When the received situation information data is processed by the situation information processing unit, correlation matching is performed to measure the dissimilarity based on the association by measuring the distance between two observed values of the observed values of the situation information data, and performing correlation matching. The crossover frequency between the measured values is measured to obtain a value with a high probability of crossing, and a cluster analysis is performed using the distance between the values measured by association matching, and the cluster analysis is a cluster of the measured values. After formation, the actual value and one of the measured values included in the cluster close to the actual value are used as a representative value, and the sum of the distances between the measured value and the representative value is minimized,
The association matching is characterized in that it is performed by Equation 1 below,
Artificial intelligent data modeling system of robot data.
[Equation 1]

- here,

Is the dissimilarity of the variable values, all i,t∈{1,2,...,p}, all j,k∈{1,2,...,q ⁱ },

Is the distance calculation function of the probability distribution-
delete delete delete The method of claim 1,
The service robot,
It characterized in that it comprises a response data output unit for outputting response data to the context information from the cloud server,
Artificial intelligent data modeling system of robot data.
The method of claim 1,
The cloud server,
Characterized in that it comprises a; context recognition result derivation unit for providing the modeled context information model to the service robot,
Artificial intelligent data modeling system of robot data.
The method of claim 6,
When modeling by the artificial intelligence modeling unit,
Conformity verification step (S10) of selecting the context information data, analyzing the validity of the selected context information data, classifying the analyzed context information data, and performing a theoretical verification whether the collected context information data is appropriate data (S10 ); characterized in that it comprises,
Artificial intelligent data modeling system of robot data.
The method of claim 7,
The collected situation information and the situation information data previously learned from the artificial intelligence data storage unit are combined, or each individually classified according to the location, size, and image according to time flow, and the classified situation information Characterized in that it comprises a; analyzing and analyzing according to the user's requirements, after learning by the data learning unit, and then transforming into a data industry standard language (PMML) (S20);
Artificial intelligent data modeling system of robot data.
The method of claim 8,
It characterized in that it comprises a data application step (S30) of receiving the context information model that satisfies the requirement among the plurality of modeled context information models, which are recommended by an artificial intelligence modeling unit and provided to the service robot,
Artificial intelligent data modeling system of robot data.
delete delete delete delete

Images