KR20230151593A - Artificial Intelligence Education System using Robot - Google Patents

Abstract

The present invention provides a learning robot that has at least one sensor that detects the surrounding environment, wirelessly transmits an acquired detection signal, and performs an operation according to a wirelessly transmitted control signal corresponding to the detection signal, and a learning robot is disposed. A play ground that is implemented as a two-dimensional plate with a predetermined size and pattern to limit the activity area of the learning robot, each representing different pre-designated objects to create a changing surrounding environment for the learning robot, and random objects on the play ground. Receives a detection signal by performing wireless communication with at least one environmental object tool and a learning robot placed at a location in a predetermined manner, and learns in response to the detection signal using an artificial neural network that is learned in a manner according to the learner's learning settings. An artificial intelligence capable of performing artificial intelligence training while inspiring learner interest at a low cost, including an artificial intelligence control device that generates and wirelessly transmits control signals for the robot to interact with at least one environmental object tool on the playground. Provides an education system.

Description

Artificial Intelligence Education System using Robot}

The present invention relates to an artificial intelligence education system using a robot that can effectively learn artificial intelligence at low cost.

Recently, artificial intelligence is not only being applied to a wide variety of fields, but also is very closely related to people's lives in various forms as its scope of application continues to expand more and more rapidly. Accordingly, the number of students wishing to learn the concepts and principles of artificial intelligence is increasing, and talent development through early education is becoming an important social issue at the national level for the development of an advanced information and communication society.

However, because artificial intelligence education is difficult to understand its principles and operations, there is a problem that it is difficult to produce teaching aids suitable for learning. Additionally, as artificial intelligence requires high-performance computing devices, the cost of manufacturing teaching aids is very high. Due to these problems, early education on artificial intelligence can only be provided to some people, and it is difficult to spread widely.

Korea Registered Patent No. 10-1926640 (registered on December 3, 2018)

The purpose of the present invention is to provide an artificial intelligence education system that can effectively perform artificial intelligence education at low cost.

Another purpose of the present invention is to provide an artificial intelligence education system that can arouse and sustain interest by allowing learners to easily create various environmental factors in which learned artificial intelligence is learned and check the learned results.

The artificial intelligence education system according to an embodiment of the present invention for achieving the above purpose is provided with at least one sensor that detects the surrounding environment and wirelessly transmits the obtained detection signal, and wirelessly transmits the detection signal corresponding to the detection signal. A learning robot that performs actions according to given control signals; A play ground implemented as a two-dimensional flat plate with a predetermined size and pattern on which the learning robot is placed to limit the activity area of the learning robot; at least one environment object tool each representing a different pre-designated object to create a changing surrounding environment of the learning robot and disposed at a random location on the play ground; And the detection signal is received by performing wireless communication with the learning robot in a predetermined manner, and the learning robot moves on the play ground in response to the detection signal using an artificial neural network that is learned in a manner according to the learner's learning settings. and an artificial intelligence control device generating and wirelessly transmitting the control signal to interact with the at least one environmental object tool.

The learning robot includes a sensor unit that includes at least one sensor including an image sensor and detects the surroundings in a predetermined manner to obtain the detection signal; a robot communication unit that performs wireless communication with the artificial intelligence control device in a predetermined manner, transmits the detection signal to the artificial intelligence control device, and receives the control signal corresponding to the detection signal from the artificial intelligence control device; a first driving unit having a driving structure for moving the learning robot; a second driving unit having a driving structure that allows the learning robot to grasp the at least one environmental object tool; and a robot control unit that receives the control signal and generates first and second drive signals for driving the first and second drive units.

The first driving unit includes two first actuators each driven in response to the first driving signal; And it may include a plurality of legs that move in conjunction with each other when the corresponding first actuator of the two first actuators is driven.

The second driving unit includes a second actuator driven in response to a second driving signal; And it may include a tong that performs a picking or placing operation on the second actuator.

The artificial intelligence control device includes a communication unit that performs wireless communication with the learning robot to receive the detection signal and transmits the control signal to the learning robot; It is implemented as an artificial neural network that performs learning, and when the detection signal is transmitted, the image transmitted as the detection signal is analyzed according to a learned method to determine the area of the play ground and the location and type of the at least one environmental object tool. an environmental identification unit that identifies each; and is implemented as an artificial neural network through which learning is performed, wherein the learning robot determines an action to interact with and controls the learning robot according to a learned method according to the type and location of the at least one environmental object tool identified within the identified play ground area. It may include an operation decision unit that generates a signal.

The at least one environmental object tool may be implemented as a two-dimensional image card displaying an object image for creating the surrounding environment of the learning robot.

During learning, the environment identification unit may identify the play ground and the at least one environmental object tool according to a weight specified by the learner for each area of the image transmitted as the detection signal.

The environment identification unit is pre-trained to identify each of the play ground and the at least one environment object tool manufactured according to a preset method, and identifies the location of each of the play ground and the at least one environment object tool according to the preset method. and types can be identified.

During learning, the operation determination unit determines an operation for the learning robot to approach or avoid the identified environmental object tool according to a weight specified by the learner for each of the at least one environmental object tool identified in the environment identification unit. Thus, the control signal can be generated.

When the at least one environmental object tool manufactured according to a preset method is identified, the operation determination unit determines an operation for the learning robot to approach or avoid the identified environmental object tool according to a preset method and sends the control signal. can be created.

The at least one environmental object tool may be implemented as a three-dimensional tool that the learning robot can grasp using the second driving unit.

When the pattern of the play ground identified by the environment identification unit is a stadium pattern, the operation determination unit may determine the operation of the learning robot so that the environmental object tool is moved to a predetermined feature position or is not moved to another specific position. there is.

The artificial intelligence control device may be implemented by installing a software application provided for artificial intelligence learning on a user terminal owned by the learner.

Therefore, the artificial intelligence education system according to an embodiment of the present invention is easy to spread artificial intelligence education because the artificial intelligence that controls the robot's movements is implemented in software on a user terminal owned by the learner separately from the robot and can be implemented at low cost. In addition, learners can easily set the surrounding environment of the learning robot that is the subject of learning, and learn and observe the robot's movements to perform movements corresponding to the set surrounding environment, thereby increasing interest in artificial intelligence.

1 to 3 show a schematic configuration of an artificial intelligence education system according to an embodiment of the present invention.
Figure 4 shows a schematic structure of the learning robot and artificial intelligence control device of Figure 1.
FIG. 5 is a diagram for explaining the detailed configuration of the robot of FIG. 1.
Figure 6 shows an example of utilization of the artificial intelligence education system according to this embodiment.

In order to fully understand the present invention, its operational advantages, and the objectives achieved by practicing the present invention, reference should be made to the accompanying drawings illustrating preferred embodiments of the present invention and the contents described in the accompanying drawings.

Hereinafter, the present invention will be described in detail by explaining preferred embodiments of the present invention with reference to the accompanying drawings. However, the present invention may be implemented in many different forms and is not limited to the described embodiments. In order to clearly explain the present invention, parts that are not relevant to the description are omitted, and the same reference numerals in the drawings indicate the same members.

Throughout the specification, when a part "includes" a certain element, this does not mean excluding other elements, unless specifically stated to the contrary, but rather means that it may further include other elements. In addition, terms such as "... unit", "... unit", "module", and "block" used in the specification refer to a unit that processes at least one function or operation, which is hardware, software, or hardware. and software.

1 to 3 show a schematic configuration of an artificial intelligence education system according to an embodiment of the present invention.

The artificial intelligence education system according to this embodiment includes the learning robot 100 and the artificial intelligence control device 200 shown in FIG. 1, the play ground shown in FIG. 2, and at least one environmental object tool shown in FIG. 3. can do.

In this embodiment, the learning robot 100 may be implemented as a multi-legged learning robot simulating an earwig (or stag beetle), as shown in FIG. 1. The learning robot 100 is equipped with at least one sensor to detect the surrounding environment to obtain a detection signal, and transmits the obtained detection signal to the artificial intelligence control device 200. Then, the learning robot 100 receives a control signal from the artificial intelligence control device 200 and performs an operation according to the applied control signal.

The learning robot 100 may be implemented as a movable robot equipped with wheels, but since a robot that moves on wheels has a very simple movement method, it is difficult to arouse learners' interest in the movements of the learning robot. Additionally, the learning robot 100 can be implemented as a bipedal learning robot similar to a human. However, when implemented as a two-legged learning robot, the robot must be able to maintain balance while moving, so not only are a large number of different types of sensors required, but also difficult to control motion, requiring complex manipulation control. Therefore, there is a problem that it is difficult to manufacture at low cost.

Accordingly, in this embodiment, it is assumed that the learning robot 100 is implemented as a multi-legged learning robot simulating an earwig so that it can be manufactured at low cost and operate stably while encouraging learner interest.

The artificial intelligence control device 200 receives the detection signal transmitted from the learning robot 100, recognizes the surrounding environment of the learning robot 100, and generates a control signal corresponding to the recognized surrounding environment to control the learning robot 100. send to Here, the artificial intelligence control device 200 can learn how to recognize the surrounding environment from a detection signal transmitted from the learning robot 100 by the learner and the actions to be performed by the learning robot 100 according to the recognized surrounding environment. . Then, depending on the learned method, the learner can perform the required actions under various environmental conditions.

For example, the learning robot 100 and the artificial intelligence control device 200 can transmit detection signals and control signals to each other by performing wireless communication according to a predefined wireless communication technique such as Direct Wifi or Bluetooth. there is. In this embodiment, the learning robot 100 and the artificial intelligence control device 200 are implemented as separate devices and perform wireless communication in order to reduce the cost required to build an artificial intelligence education system.

Artificial intelligence for controlling the operation of the learning robot 100 must recognize the surrounding environment on its own according to a learned method and perform actions corresponding to the recognized surrounding environment, and requires a large amount of high-speed computation in this process. And the learning process of artificial intelligence requires several times more calculations than that. Therefore, the artificial intelligence control device 200 must be implemented as a high-performance computing device. However, in the case of existing artificial intelligence learning aids, the manufacturing cost of the learning robot was very high as the computing device was implemented to be directly embedded in the learning robot that senses the surrounding environment and performs actual movements. This increase in manufacturing costs is acting as the biggest obstacle to the spread of artificial intelligence learning. In addition, when a computing device is built into a learning robot, the performance of the computing device is limited due to issues such as installation space, power, and cost, so the learning robot 100 recognizes only a very simple surrounding environment, and even in the recognized environment, the performance is limited. There was a limitation that it could only be learned to perform monotonous movements. And these limitations have the problem of lowering learners' interest in artificial intelligence, resulting in reduced learning effectiveness.

In order to overcome this limitation, in this embodiment, a learning robot 100 corresponding to the body and an artificial intelligence control device 200 to act as a brain are separately provided, and the robot 100 and the artificial intelligence control device 200 are provided separately. to communicate with each other wirelessly. Here, the artificial intelligence control device 200 may be implemented with a user terminal that the learner already owns, such as a desktop computer, laptop computer, tablet, and smart phone on which an artificial intelligence learning application prepared in advance for artificial intelligence learning is installed. In general, user terminals not only exhibit a level of computational performance that is incomparable to the computational processors installed in existing artificial intelligence learning robots, but also do not incur additional cost burden because most learners already own them.

Therefore, if each learner uses the existing user terminal as the artificial intelligence control device 200 that acts as the brain of the learning robot 100, not only can a high-performance artificial intelligence learning system be built without additional cost. , it has the advantage of being able to learn and use artificial intelligence better, thereby encouraging learners' interest in learning artificial intelligence.

A detailed description of the learning robot 100 and the artificial intelligence control device 200 will be described later.

The play grounds 301 and 302 shown in FIG. 2 provide an activity space where the learning robot 100 can be active. Basically, the learning robot 100 driven under the control of the artificial intelligence control device 200 performs wireless communication with the artificial intelligence control device 200, so the activity space does not need to be restricted. However, if the activity area of the learning robot 100 can change infinitely like the real world, there will be too many things that the learner must consider when learning educational artificial intelligence, which may cause a problem that reduces motivation to learn. There will be.

In addition, the user terminal implemented as the artificial intelligence control device 200 can exhibit very excellent performance compared to the processor built into the learning robot 100, but has limitations in computing power compared to expensive artificial intelligence-specific computing devices. exist. Therefore, if there are no restrictions on the activity space of the learning robot 100, the artificial intelligence may not be properly learned or the learning robot 100 may not be operated stably due to the limited computing power of the artificial intelligence control device 200.

Accordingly, the artificial intelligence learning system of this embodiment provides a play ground (301, 302) that specifies and limits the activity space of the learning robot (100) so that the learner can easily learn artificial intelligence, and the artificial intelligence control device (200) ), so that interest in learning can be maintained by ensuring that the learning robot 100 can be operated stably even with limited computing power.

Figures 2 (a) and (b) show examples of various types of playgrounds. As shown in (a) of FIG. 2, the play ground 301 may be implemented as a flat board with a predetermined size and a basic pattern such as a predetermined color. In this way, when the play ground 301 is implemented in the form of a simple two-dimensional board with a specified pattern and size, the learner can easily limit the area of the surrounding environment in which the artificial learning robot 100 can recognize and operate, thereby facilitating artificial intelligence learning. can be simplified. And, as shown in (b) of FIG. 2, additional patterns for specific purposes may be formed in the play ground 302. In (b) of FIG. 2, an additional pattern simulating a soccer stadium is further formed, and in this case, the learner can additionally learn the activity rules of the learning robot 100 according to the additional pattern through artificial intelligence. That is, artificial intelligence is applied when the learning robot 100 is placed on the play ground 302 in the form of a stadium as shown in (b) of FIG. It can be learned to behave differently.

In this way, the play grounds 301 and 302 not only simplify artificial intelligence learning by providing an activity area for the learning robot 100, thereby improving the ease of artificial intelligence learning, but also provide a plurality of different play grounds 301 and 302. ) can be used to learn the diversity of artificial intelligence so that the learning robot 100 performs different designated actions according to the surrounding environment set by ).

The environmental object tool is placed on the play grounds 301 and 302 to provide the surrounding environment of the learning robot 100 together with the play grounds 301 and 302. While the play grounds (301, 302) are used to provide a pre-designated static surrounding environment, the environment object tool is an object that provides an additional surrounding environment to the learning robot (100) and can be used to create any object on the play grounds (301, 302). It is freely placed in a location so that the surrounding environment of the learning robot 100 can be varied in various ways even in a limited area called the play ground 301 and 302. In other words, the environmental object tool allows the surrounding environment of the learning robot 100 to be easily changed, so that artificial intelligence for the learning robot 100 can be learned in various environments, and the learned artificial intelligence can learn in various environments. It is possible to verify whether the robot 100 can be operated normally.

As shown in FIG. 3, the environmental object tool may be implemented as a two-dimensional image card such as (a) to (c), or a three-dimensional tool such as (d). The environmental object tool implemented as a two-dimensional image card as shown in (a) to (c) of FIG. 3 can be used to create the learning robot 100 at low cost by simply placing the environmental object tool at an arbitrary location on the play ground 301 and 302. Various changes can be made to the surrounding environment. In addition, since pre-designated images are used, the type or number of objects that artificial intelligence must recognize is reduced compared to the case of using real objects, allowing artificial intelligence to learn more easily.

In FIG. 3, (a) represents an example of a symbol object among environmental object tools, (b) represents an example of a surviving entity, and (c) represents an example of an obstacle object. The learner can train the artificial intelligence so that the learning robot 100 prefers one of the banana image shown on the left and the apple image shown on the right in the symbol object of (a), while not preferring the other one. Additionally, the learner may train artificial intelligence to allow the learning robot 100 to avoid its natural enemy, a bird, shown on the right, while searching for the same species of bugs shown on the left in the survival tool in (b). Additionally, when the learning robot 100 encounters an obstacle such as (c) while moving on the play ground 301, artificial intelligence may be trained to avoid the obstacle.

Meanwhile, in (d), a ball is shown as an example of a 3D object. In this case, the learner can teach the learning robot 100, which simulates an earwig, to use a 3D object in a predetermined manner. As an example, as shown in (b) of FIG. 2, when the play ground 302 is provided in the form of a stadium, the learner recognizes the ball placed on the play ground 302 by the learning robot 100 and plays a predetermined ball. It can be taught to move the ball to a set location, that is, to the goal, or to interfere when another learning robot carries the ball to the goal.

In this way, the environmental object tool can be placed on the play ground 301, 302 on a simple two-dimensional plane to provide additional surrounding environment objects for learning artificial intelligence for operating the learning robot 100 under various conditions. , It can be used to create an environment to verify whether the learned artificial intelligence operates the learning robot 100 normally.

FIG. 4 shows the schematic structure of the learning robot and the artificial intelligence control device of FIG. 1, and FIG. 5 is a diagram for explaining the detailed structure of the robot of FIG. 1.

4 and 5, the learning robot 100 may include a sensor unit 110, a robot communication unit 120, a robot control unit 130, a first drive unit 140, and a second drive unit 150. there is. The sensor unit 110 includes at least one sensor disposed at a predetermined position on the main body of the learning robot 100 and acquires a detection signal. Here, the sensor unit 110 is disposed in front of the learning robot 100, as shown in FIG. 5, and includes at least one image sensor 111 for detecting the play grounds 301 and 302 and environmental objects and tools. can do. In this case, the sensor unit 110 may capture images of the surrounding environment of the learning robot 100 and transmit them to the robot communication unit 120 as a detection signal.

The robot communication unit 120 receives a detection signal from the sensor unit 110 and wirelessly transmits it to the artificial intelligence control device 200 through the antenna 121. Then, the control signal transmitted from the artificial intelligence control device 200 is received and transmitted to the robot control unit 130. The robot communication unit 120 may be implemented as a communication module that communicates with the artificial intelligence control device 200 using a pre-designated wireless communication method.

Meanwhile, the robot control unit 130 may be provided on the body of the learning robot 100, and in response to a control signal applied from the artificial intelligence control device 200 through the robot communication unit 120, the first and second driving units At least one of (140, 150) is driven to operate the learning robot (100). The robot control unit 130 may generate and output first and second drive signals for driving the first and second drive units 140 and 150, respectively, in response to the applied control signal.

The first driver 140 is driven in response to the first drive signal, and the second driver 150 is driven in response to the second drive signal.

The first and second driving units 140 and 150 may each include at least one actuator driven in response to a corresponding driving signal. In this embodiment, the first driving unit 140 includes a plurality of actuators for moving the six legs 141 of the learning robot 100, which is a multi-legged walking robot in the form of an earwig, as shown in (a) of FIG. 5, The second driving unit 150 may include one actuator for moving the tongs 151 of the learning robot 100. The first driving unit 140 may be equipped with six actuators depending on the number of legs 141 of the learning robot 100, but in this case, not only does the manufacturing cost of the learning robot 100 increase, but it also controls multiple actuators. This increases control complexity, making it difficult to learn artificial intelligence. In other words, there is a risk that the learner's learning achievement may decrease. Accordingly, in this embodiment, as shown in (b) of FIG. 5, the three legs located on one side of the learning robot 100 are moved in conjunction with each other by one actuator, so that the first driving unit 140 Even with only two actuators, various movements of the learning robot 100 can be realized.

The two actuators of the first driving unit 140 operate individually according to the corresponding first driving signal, allowing the learning robot 100 to move forward, backward, left and right. In addition, one actuator of the second driving unit 150 operates in response to the second driving signal, allowing the tongs 151 to perform a movement of grasping or placing a three-dimensional object among environmental object tools.

Meanwhile, as described above, the artificial intelligence control device 200 may be implemented as a user terminal owned by a learner with an artificial intelligence application for artificial intelligence learning pre-installed. The artificial intelligence control device 200 may include a communication unit 210, an environment identification unit 220, and an operation determination unit 230. The communication unit 210, an environment identification unit 220, and an operation determination unit 230 Each can be implemented as a software module included in an artificial intelligence application.

The communication unit 210 performs wireless communication with the learning robot 100 to receive a detection signal and transmits it to the environment identification unit 220. Then, the control signal obtained from the operation determination unit 230 is received and transmitted to the learning robot 100. The environment identification unit 220 is implemented with an artificial neural network and recognizes the surrounding environment of the learning robot 100 by analyzing the detection signal transmitted from the learning robot 100 according to a pre-learned method. As an example, the environment identification unit 220 recognizes the area of the play grounds 301 and 302 from the image acquired by the sensor unit 110 of the learning robot 100 and transmitted as a detection signal, and detects the areas of the play grounds 301 and 302. It is possible to identify whether an environment object tool is placed on the screen and the type and location of the placed environment object tool. And the operation determination unit 230 determines the operation to be performed by the learning robot 100 in response to the surrounding environment identified by the environment identification unit 220 and generates a control signal. Like the environment identification unit 220, the motion determination unit 230 is also implemented with an artificial neural network and determines an operation corresponding to the surrounding environment according to a previously learned method.

Artificial intelligence, that is, the environment identification unit 220 and the motion determination unit 230 implemented with an artificial neural network, must be trained before actually operating the learning robot 100, and the learner receiving artificial intelligence training must identify the environment. By learning the unit 220 and the operation decision unit 230, the principles and concepts of artificial intelligence can be understood.

When learning the environment identification unit 220, the learner designates a play ground area in the image acquired as a detection signal and specifies the placement location and type of the environment object tool, so that the environment identification unit 220 recognizes the surrounding environment. You can learn to do it. However, in the present invention, since the play grounds 301 and 302 and various environmental object tools are manufactured and provided in advance, the environment identification unit 220 learns in advance to identify the provided play grounds 301 and 302 and various environmental object tools. It may also be provided as is.

And when learning the motion determination unit 230, the learner moves the learning robot 100 to the location of the environmental object tool according to the type and location of the environmental object tool recognized by the environment identification unit 220 on the play ground 301, 302. The motion decision unit 230 can be trained by providing weights depending on whether to move or avoid. Additionally, if the environmental object tool is recognized as a three-dimensional object such as a ball, the learning robot 100 may learn by providing weights so that it can pick up the environmental object tool with the pincers and move it to a designated location.

Figure 6 shows an example of utilization of the artificial intelligence education system according to this embodiment.

In FIG. 6 (a), a case where the learning robot 100 is taught to move across the play ground 301 according to a symbol object among environmental object tools. As shown in (a), the learner places two symbolic objects, a banana and an apple, at random positions on the play ground 301 and places the learning robot 100 on one side of the play ground 301. , the path along which the learning robot 100 moves on the play ground 301 is confirmed. The learner sets a weight (or penalty) according to the path along which the learning robot 100 moves, and then artificial intelligence prefers both symbol objects and sets a path for the learning robot 100 to approach the two care objects. Alternatively, the learning robot 100 can be trained to set a path that avoids the two symbol objects by disfavoring both symbol objects. Additionally, as shown in (a), artificial intelligence can be trained to set a path for the learning robot 100 to approach apples while avoiding bananas.

Meanwhile, (b) shows a case where the learning robot 100 is taught to move across the play ground 301 according to survival objects among environmental object tools. In (b), the learner can place two symbolic objects, a natural enemy such as a bird and a homogeneous insect such as a bug, at an arbitrary location on the play ground 301, and can additionally place an obstacle object such as a tree. In the case of symbol objects, artificial intelligence can be individually trained to prefer or disfavor each symbol object, but in the case of survival objects, artificial intelligence can be trained in advance to prefer or disfavor each symbol object. Additionally, artificial intelligence can learn to avoid obstacles. In this case, the learner observes whether the learning robot 100 randomly placed on the play ground 301 moves in a way that avoids obstacles and areas where natural enemies are located among survival objects and approaches its own kind, and moves along a normal path. By setting weights depending on whether the robot moves along an abnormal path, the learning robot 100 can be taught to move along a normal path.

In (c), the play ground 302 is implemented in the pattern of a stadium, and artificial intelligence can also be set to have weights corresponding to the stadium. As an example, artificial intelligence is learned in such a way that a weight is given when the learning robot 100 moves a goal, which is a three-dimensional object, to a position designated as an opponent's goal, and a penalty is added when the goal is located in a position designated as its own goal. , while the learning robot 100 scores a goal in the opponent's goal, it is possible to prevent the opponent's learning robot from moving the ball to its own goal.

That is, in the artificial intelligence learning system according to this embodiment, the learner trains artificial intelligence implemented in the form of a software application in the artificial intelligence control device 200 in a manner according to the learning settings desired by the learner, and the artificial intelligence is used to create a learning robot (100) ) can be controlled. Artificial intelligence can be learned according to a supervised learning method in a way that weights are given to the movements of the learning robot 100 by the learner, and the learner understands how to learn artificial intelligence, thereby understanding the concept of artificial intelligence and I can understand the principle.

As such, the artificial intelligence learning system according to this embodiment is separately equipped with a learning robot 100 that performs actual movements and an artificial intelligence control device 200 that performs artificial intelligence learning, and performs wireless communication with each other, and the artificial intelligence As the control device 200 is implemented by installing a pre-designated artificial intelligence education application on the user terminal, it is possible to provide a system that can educate learners about artificial intelligence at low cost. In addition, the learning robot 100 is implemented in the form of a multi-legged walking robot that simulates an earwig, and by driving multiple legs with two actuators, it can stably perform various movements at low cost, and the learning robot 100 It provides a play ground that limits the activity area and a variety of environmental object tools to easily implement various environmental conditions, allowing learners to easily understand the concepts and principles of learning and operating artificial intelligence.

The present invention has been described with reference to the embodiments shown in the drawings, but these are merely illustrative, and those skilled in the art will understand that various modifications and other equivalent embodiments are possible therefrom.

Therefore, the true technical protection scope of the present invention should be determined by the technical spirit of the attached claims.

100: Learning robot 110: Sensor unit
120: Robot communication unit 130; robot control unit
140: first driving unit 150: second driving unit
200: artificial intelligence control device 301, 302: play ground

Claims (13)

A learning robot that has at least one sensor that detects the surrounding environment, wirelessly transmits an acquired detection signal, and performs an operation according to a wirelessly transmitted control signal corresponding to the detection signal;
A play ground implemented as a two-dimensional flat plate with a predetermined size and pattern on which the learning robot is placed to limit the activity area of the learning robot;
at least one environment object tool each representing a different pre-designated object to create a changing surrounding environment of the learning robot and disposed at a random location on the play ground; and
The detection signal is received by performing wireless communication with the learning robot in a predetermined manner, and in response to the detection signal using an artificial neural network that is learned in a manner according to the learner's learning settings, the learning robot moves to the above on the play ground. An artificial intelligence training system comprising an artificial intelligence control device generating and wirelessly transmitting said control signal to interact with at least one environmental object tool. The method of claim 1, wherein the learning robot
A sensor unit that includes at least one sensor including an image sensor and detects the surroundings in a predetermined manner to obtain the detection signal;
a robot communication unit that performs wireless communication with the artificial intelligence control device in a predetermined manner, transmits the detection signal to the artificial intelligence control device, and receives the control signal corresponding to the detection signal from the artificial intelligence control device;
a first driving unit having a driving structure for moving the learning robot;
a second driving unit having a driving structure that allows the learning robot to grasp the at least one environmental object tool; and
An artificial intelligence education system comprising a robot control unit that receives the control signal and generates first and second drive signals for driving the first and second drive units. The method of claim 2, wherein the first driving unit
two first actuators each driven in response to the first driving signal; and
An artificial intelligence education system including a plurality of legs that move in conjunction with each other when the corresponding first actuator of the two first actuators is driven. The method of claim 3, wherein the second driving unit
a second actuator driven in response to a second drive signal; and
An artificial intelligence education system including a tong that performs a picking or placing operation on the second actuator. The method of claim 2, wherein the artificial intelligence control device
a communication unit that performs wireless communication with the learning robot to receive the detection signal and transmits the control signal to the learning robot;
It is implemented as an artificial neural network that performs learning, and when the detection signal is transmitted, the image transmitted as the detection signal is analyzed according to a learned method to determine the area of the play ground and the location and type of the at least one environmental object tool. an environmental identification unit that identifies each; and
Implemented as an artificial neural network through which learning is performed, the learning robot determines an action to interact with according to the type and location of the at least one environmental object tool identified within the identified play ground area according to a learned method, thereby generating the control signal. An artificial intelligence training system that includes a motion decision unit that generates. The method of claim 5, wherein the at least one environment object tool
An artificial intelligence education system implemented with a two-dimensional image card displaying object images to create the surrounding environment of the learning robot. The method of claim 6, wherein the environmental identification unit
An artificial intelligence education system that, upon learning, identifies the play ground and the at least one environmental object tool according to a weight specified by the learner for each region of the image transmitted as the detection signal. The method of claim 6, wherein the environmental identification unit
is pre-trained to identify each of the play ground and the at least one environmental object tool manufactured according to a preset method, and identifies the location and type of each of the play ground and the at least one environmental object tool according to a preset method; artificial intelligence education system. The method of claim 6, wherein the operation determination unit
During learning, the learning robot determines an operation to approach or avoid the identified environmental object tool according to the weight specified by the learner for each of the at least one environmental object tool identified in the environment identification unit, thereby sending the control signal. An artificial intelligence education system that generates. The method of claim 6, wherein the operation determination unit
When the at least one environmental object tool manufactured according to a preset method is identified, artificial intelligence determines an operation so that the learning robot approaches or avoids the identified environmental object tool according to a preset method and generates the control signal. Education system. The method of claim 5, wherein the at least one environment object tool
An artificial intelligence education system in which the learning robot is implemented as a three-dimensional tool that can be grasped using the second driving unit. The method of claim 11, wherein the operation determination unit
If the pattern of the play ground identified in the environment identification unit is a stadium pattern, an artificial intelligence education system that determines the operation of the learning robot so that the environment object tool is moved to a predetermined feature position or is not moved to another specific position. . The method of claim 5, wherein the artificial intelligence control device
An artificial intelligence education system implemented by installing a software application provided for artificial intelligence learning on a user terminal owned by the learner.