KR100503652B1 - Artificial creature system and educational software system using this - Google Patents

Abstract

The present invention relates to a virtual living system and an educational software system using the same. The virtual living organism according to the present invention has the 'emotion', 'motivation', 'alwaysness' and decides its own behavior, adapts to the surrounding environment, and has the ability to learn a behavior suitable for various user's commands. In particular, the educational software system using the virtual life system of the present invention is a system for fostering virtual life and learning the characteristics of living things and inanimate objects, a system for fostering virtual life according to the learning amount of the curriculum, and a virtual life teacher of another nature. It is composed of an educational system that can be learned.

Description

Virtual creature system and educational software system using this

The present invention relates to the use of virtual life in educational software systems. To this end, we first created a new virtual creature structure and propose a learning structure of a virtual creature that can learn various commands of users. In addition, we constructed an educational software system that enables learners to learn using the proposed virtual life and the learning structure of virtual life. The educational software system consists of a system for fostering virtual creatures, a system for learning the characteristics of living things and inanimate objects, a system for fostering virtual creatures according to the amount of learning, and a system for learning from virtual creature teachers of different characteristics.

As shown in FIG. 1, a conceptual diagram of a general virtual creature uses a recognition unit 11 that recognizes the external environment 10, an internal state unit 12 that represents the emotions and motivations of the virtual creature, and external information and an internal state. Thus, it consists of a behavior determination unit 13 for determining the behavior of the virtual creature, a learning unit 14 for adapting the virtual creature according to the external state, and a behavior implementation unit 15 for actually implementing the behavior.

In the case of a virtual creature or an entertainment robot, it is important to give a feeling that the virtual creature is a real creature by interaction with the user. For this to happen, the virtual creature must be able to select the behavior appropriate for the situation by itself and be able to learn to perform the proper behavior through repetitive training for any instruction of the user.

However, most voice command learning of the conventional virtual life has a structure in which the virtual life responds by simply recognizing the user's voice in a predetermined command-action pair. 2 shows a structure that depends on such general speech recognition. In this case, only the instruction in the predetermined command 22-action 21 is learned, where learning simply refers to the process of speech recognition (23). Learning within these rules prevents the implementation of various behaviors of virtual creatures or entertainment robots, and is not only simple speech recognition but also difficult to be seen as growth through real robot learning. In addition, learning how to implement learning that connects behavior to a given command, by searching for dozens of behaviors, one by one, was not only a matter of time, but also a very low probability of learning.

There were also various problems with educational software that use them for educational purposes. In general, educational software has simply used the classic method of asking students to submit and solve problems. Recently, as a training assistant, a lot of methods are used to help education using avatars or simple graphic characters, but since the avatars or graphic characters are simply shown as pre-programmed, these methods create various educational scenes. It was hard and forced to show and educate all the same. As a result, there was a limit to diverse and more interactive education.

Therefore, the present invention is to solve the conventional problems as described above, the present invention proposes a new virtual life structure and proposes an efficient structure that can learn the user's command, as well as artificial emotion and virtual personality in educational software It is an object of the present invention to use the virtual life of the present invention.

According to the present invention as a technical idea for achieving the above object of the present invention, 1) virtual creatures can perform a variety of actions according to their emotions, motives, homeostasis, unlike ordinary avatars or simple graphic characters, each emotion By controlling the traits of consciousness, motivation, and homeostasis, they can also create their own virtual personality. In addition, the method of probabilistically selecting behaviors by internal state and external environment and the method of decisively selecting behaviors by mimicking the instincts of animals can implement various behaviors. In addition, it is possible to efficiently learn the desired behavior among dozens of behaviors for any command of the user by constructing a set of similar behaviors among similar behaviors and reinforcing the result of the behavior.

2) Also, the educational software system using the proposed virtual life consists of the following. Even if a learner nurtures a virtual creature having the virtual personality of each creature or inanimate object and does not directly touch the object, the learner can learn the characteristics of living creatures and inanimate creatures that are difficult to meet.

3) In addition, the virtual creature can be nurtured only when the predetermined amount of learning is satisfied in parallel with the learner's curriculum, thereby increasing the learner's motivation to learn. And, unlike the unilateral educational helper, the character can be trained by the educational virtual teacher of various personalities, personalities through virtual life having their own virtual personality.

Hereinafter, with reference to the accompanying drawings, the configuration and operation of the embodiment of the present invention will be described in detail.

The structure of the virtual creature proposed in the present invention is configured as shown in FIG. 3. The virtual life structure is largely composed of a perception system 32a, 32b, 32c, an internal status system 33, a behavior system 34a, 34b, 34c, a learning system. 35a, 35b, and a motor system 36. The functions of each part are as follows.

1) recognition system

The recognition system is a part 32b that receives and preprocesses sensor information 32a used for changing the internal state of a living creature, selecting a behavior, and learning. In addition, it selects an object of interest (32c) from among a lot of information currently input according to the sensor priority and is used to select an action. The double preprocessing part is called a symbolization system (32b). It converts sensor information input from outside and inside of virtual life into meaningful information and is used to change internal state and behavior. The input information is converted into information expressed as a boolean vector as in Equation 1 and used.

In addition, the concept of 'compliance' was introduced to simulate a more lifelike encoding system. Adaptation refers to a phenomenon in which sensitivity is slowed down when a strong stimulus persists. A good example can be seen from the perspective. If you stay in a bright place for a long time, your vision will deteriorate. If you go to a dark place, you will not see it well. This adaptation was implemented by changing the sensitivity of the same stimulus in succession.

When the stimulus 41 continuously enters as shown in FIG. 4, the sensitivity 42 of the sensor stimulus gradually decreases, and the magnitude thereof changes from 1 to 0, which is the maximum value. If the stimulus does not enter for a period of time, it gradually returns to its original sensitivity. Sensitivity is set separately for each stimulus, and changing values can be set differently for each stimulus. As shown in the algorithm of the virtual creature that adopts the acclimation of FIG. It is not recovered. If there is no stimulus for more than T_IGNORE time, the recovery degree is gradually restored by DT_RECOVER (42b). The sensitivity of each stimulus may be represented by a sensitivity vector as shown in Equation 2. Each element has a value between 0 and 1.

A coding vector S of Equation 3 may be generated using the stimulus vector equation X and the sensitivity vector equation P. This vector is used as the input to the internal state system as the final output of the coding system.

The interest selector 32c is a part that selects a target that is currently paying attention from information coming into the artificial life, and creates a signal for excluding an action that is not suitable for the predetermined target. The object of interest may be all values of the encoding vector used in the encoding system. Priority was assigned to the coding vector S so that the target of interest was determined and the target of interest was not easily changed even if an unimportant input came in. If the priority of the values s _i of the encoding vector S is pri ( s _i ), the selected interest s _i is obtained as shown in Equation 4, and the interest having the highest priority among the inputs is selected.

2) internal state machine

The internal state system 33 is composed of three subsystems: 'synchronous' 33a, 'always' 33b, and 'emotion' 33c.

Motivation system is the driving force that determines the main behavior of artificial life. For example, curiosity, friendliness, boredom, avoidance, domination, possessive desire. i is the number of sync states, m _i (t) is the sync value at time t, Is a synchronous change gain, S is an encoding vector, Is the connection strength between the S and the i-th synchronization m _i (t), the update expression of the synchronization state is shown in Equation (5).

here Is defined as the mean value of motives. If there is no input, the average value of the synchronization indicates a value where the corresponding synchronization state value converges, and the larger the average value, the greater the desire to solve the synchronization is. Indicates how long this desire lasts for external stimuli. These two values represent the characteristics of the artificial creature with respect to the corresponding motives and form part of the character of the artificial creature. Connection strength with coding vector S Indicates how the sensor input value affects the internal state of artificial life. That is, the weight values indicate whether the input satisfies the synchronization, the amplification, or the input. At this time, by using the preference learning portion to be described in 4) By changing the value of, we can change and learn the signs of artificial life. Each sync state value constituting the sync system is arranged in a vector as shown in Equation (6).

Homeostasis system is the desire to maintain the survival of artificial life, for example, may be tired or sleepy. Where j is the number of homeostatic states, Is the homeostasis value at time t, Is the homeostatic change gain, Is the coding vector S and the j th homeostasis Suppose the connection strength with, the update equation of the homeostatic state is as follows.

Coupling Strength with Coding Vector S Indicates how the sensor input value affects the internal state of artificial life. That is, the weight values indicate whether the input amplifies or decreases the homeostasis or is independent of the input. This homeostatic state system is not trained like the other two internal states. In addition, the output values of the homeostasis system are arranged in a vector as follows.

Finally, the emotion (Emotion system) expresses the emotional state of the virtual life, there may be joy, sadness, anger, fear. The emotional state indirectly influences the choice of the behavior of the virtual creature and is used directly to express facial expressions. When selecting an action in the action selector described in 3), after selecting the action primarily by motivation and homeostasis, the action is finally selected in consideration of emotion. Emotions are also used to represent facial expressions of artificial life. The renewal expression of the emotional model is defined like the other two internal states. k is the number of emotional states, Is the emotional value at time t, To change the emotion gain, S is the encoding vector, Is the vector S and the kth emotion The connection strength with and can be summarized as follows.

here Is defined as the mean value of emotion. The average value of emotion refers to the value that the emotional state value converges when there is no input. It indicates how long this feeling lasts for external stimuli. These two values represent the characteristics of artificial life for the corresponding emotion and form part of the personality. Connection strength with S Indicates how the sensor input value affects the internal state of artificial life. That is, the weight values determine whether the input increases or decreases the emotion. In this case, we will use the preference learning Changing the value of can change the sign of artificial life and control the effects of changing emotions. Each emotion of the changing emotional state system is arranged in a vector as shown in Equation (10).

However, when modeling emotions by dividing them into states, emotions may suddenly change. As shown in FIG. 5, when the difference between the two emotion values is small and the change in the emotion value is large, the occupied emotion may vibrate between two corresponding emotions. In the left 51 of FIG. 5, since the value of m ₁ is 100 and the value of m ₃ is 95, the occupancy emotion is m ₁ . However, if m ₁ is changed to 90 in a short time, as in the right (52) situation, the occupied emotion is changed to m ₃ , and if this situation is repeated, the occupied emotion will vibrate between the two emotions.

In addition, there is a problem in that even when the difference between emotion values is subtle, it is a problem to see the emotional state as high as a little. For example, when the value of 'joy' is 90 and the remaining emotion values are 0, the occupied emotion is 'joy'. However, if the value of 'joy' is 100 and the value of the other emotions is 110, the problem of not being occupied emotion occurs even though 'joy', which was an occupied emotion at 90, has a higher value.

To solve this problem, the 'occupation emotion determination threshold', in which the difference in emotion values is defined, is defined. Less than does not find occupied emotion in E Assume that the 'normal' state is indicated by. Figure 6 illustrates this algorithm. By extending the emotional state to this structure, as shown in FIG. 7, a situation 71 in which the emotion suddenly changes from e ₁ to e ₂ is solved. Can be changed to In other words, it does not suddenly turn into 'joy' → 'anger', but instead of 'joy' → 'normality' → 'anger', the emotion can be changed more naturally. By modifying the emotion structure, the emotion vector in Equation 10 Is added and expanded as shown in Equation (11).

3) Behavior System

There are two ways in which the behavior system produces behavior. There is a method of selecting an action probabilistically based on the input from a recognition system and an internal state system, and in some cases, there is a method of deciding an action to select an appropriate action when a desired condition is satisfied. These probabilistic and deterministic methods make action choices in a complementary relationship. The behavior system consists of a subsystem of a behavior selector 34a and an inherent behavior logic 34c.

The behavior selector probabilistically selects a behavior using values of motivation M and homeostasis H obtained from an internal state system. The selection is probabilistic and consists of an algorithm as shown in FIG. First, the provisional vote value 81 for each behavior is obtained as shown in Equation 12.

As in Equation 12, the values of the synchronization M, the homeostasis H, and the weight values of the synchronization, the homeostasis, affect the behavior. The vote value for which the action will be selected using the value of. Calculate first. If the probabilistic behavior is performed using only the state values of the internal state system, the virtual creature acts only its own state irrespective of its surroundings and thus makes sense. Various masking 34b is used to select an action by considering both internal and external inputs. There are four kinds of blocking, respectively, 'Block by Interest (82)', 'Block by Voice Command (83)', 'Block by Emotion (84)', and 'Block by Instinct Behavior (85)' Is implemented.

For example, if the virtual creature is not currently aware of the ball, the behavior related to the ball should be blocked by the 'blocking method by the object of interest' so that it cannot be expressed.If the voice command is recognized, 'blocking by the voice command' How to block actions other than voice commands. In addition, if the current emotional state is sad, blocking the behaviors that are happy when using the 'blocking method according to the emotional state' to prevent the artificial life to behave in an emotional state.

Interest blocking matrix Is the object of interest obtained from the interest selector. Determined by Each object of interest f has a blocking value for actions, and the interest blocking matrix is expressed by Equation 13.

Voice command blocking matrix Is defined as a voice command in voice learning. Determined by X voice commands each Has a blocking value that affects the vote value of the actions, allowing the action to be selected as a result of the voice command.

Blocking matrix based on emotional state Blocks actions that cannot be performed according to current emotions. Current occupancy feelings Has a blocking value that affects the ballot value of actions, which reduces the ballot value of actions that do not fit into the current emotional state or should not be performed.

Blocking matrices according to instinct behaviors block actions that cannot be performed according to current instinct behaviors. The current instinct behavior has a blocking value that affects the vote value of the actions, so that when the instinct behavior is manifested and affects the performance of the current action, it controls the vote value.

Finally, the voting value V is calculated by considering four blocking matrices.

Once the vote vector V for each action is calculated, it is finally used to determine the probability that the action will be performed.

Finally act Is the action selection rate obtained from Equation 16 The action is chosen accordingly. The probability of action selection and the conditions of action are as follows.

Changes in behavior are caused by four causes: end-signal, urgent flag, higher priority interest discovery, and probabilistic transition.

First, when the execution end signal comes from the motor system, that is, when the action is completed, another action is selected and performed. The second is when an emergency call comes in from the instinct behavior selector. If you find a wall or find an obstacle, even if you are currently performing an action, the sensor makes an emergency call to change the behavior so that the robot responds to the urgent environment change. Third, if a higher priority interest is found, then the appropriate action is converted according to the perceived interest. Finally there is a stochastic shift. If there is an action that does not have an end time such as a 'rest' action, you should switch to another action according to the probability. Otherwise, if there is no external stimulus, it will continue to show the same action. I can't show it.

There is a limit to the behavior implemented by the behavior selector that selects the behavior by considering both external input and internal state. Thus, the instinct behavior logic 34c, which selects behavior decisively only by external input, independent of the internal state, complements the behavior selector. Instinct behavior selection logic models the animal's instincts and expresses immediate avoiding behavior when an obstacle such as a wall or cliff is found.

Instinct behavioral logic expresses the behavior that virtual life must take according to the conditions of sensor information and internal state. This function can be used for reflection behavior of artificial life. For example, you may be surprised when the distance sensor changes suddenly, or you may be surprised when it suddenly darkens. When a dangerous situation such as a cliff occurs, the action selector is urgently called to stop the current action and take appropriate response.

4) Learning System

When praised or scolded, the virtual creature can perform the actions desired by the user in response to the preference learner 35a and the arbitrary voice command that change the tendency of identifying and storing the interest of interest and evading or approaching according to the interest of interest. It consists of a voice learner 35b.

The preference learner is the part that learns the likes and dislikes of artificial life about an object. For example, complimenting artificial creatures adjusts the strength of the connection between the coding vector and its internal state so that the faces and fair objects you've seen recently raise the joyous emotional state of the creature and reduce its motivation to evade. . Preference learning is the strength of the connection between motivation and emotional state By adjusting the value of, the effect of the corresponding object on the internal state is controlled. At time t Wow The equation is as follows.

The voice learner can learn the behaviors that are appropriate to any command among all behaviors by reducing the behaviors that are the targets of learning by constructing similar behaviors among similar behaviors among similar behaviors, and reinforcing the results of the behaviors for each similar behavior set. It was configured to be. The contents of the voice learner are in FIG.

First, as shown in FIG. 9, a subset of the entire behavior set B 91 is used as an analogous behavior set. Configure 92. Where a is the number of similar actions. N actions as elements of the full set of actions May be one or more similar behavior elements, and may not be elements of any similar behavior set. For example, if we reconstruct behaviors into behavior sets, we could do the following: In FIG. 10, C 101 is a set of instructions, and B is a set of behavior sets. For example, elements of instruction set B, in turn, stand up, sit down, come here, stop, come forward, go back, go left, go right , And "dancing," and the elements of behavior set B are the same, If the set is a collection of the remaining behaviors except these, it is possible to create a similar set of behaviors as in the set of the equation (21).

Here the actions in the 'sit' and 'come' behavior sets can be summarized as follows, for example: 'Sit' Similar behavior set consists of 'Sit', 'Squat', and 'Lying', and 'Come here' is composed of 'Follow', 'Access', 'Kick' and 'Touch' have.

At this time, if the desired behavior comes out from the command to reward the corresponding set of similar behaviors at the same time, if the wrong behavior comes out penalty for the similar set of behaviors (115). In this way, the strength 103 of the pair of command behaviors reinforces learning 114 and configures the behavior according to the strength. Is the i th input command, Is the jth set of similar behaviors, Assume the connection strength between the two having a value between 0 and 1 at time t. Command When a desired action comes up for, the set of similar actions the action belongs to Associated with Is calculated as in Equation 23 and updated.

In (23), artificial organisms As many phases or If you get a punishment, this value is called the emotion parameter (113) is multiplied Will change. This emotional parameter is the value of the emotional state that affects learning efficiency. This is a factor that considers that learning efficiency is good when we are happy, and learning efficiency is poor when we are in a bad mood. Therefore, according to the emotional state, the learning rate, that is, the speed of learning is controlled.

j-th set of similar behaviors to which the i-th command and the performed action b belong Connection strength After updating to adjust the strength of the command-action pair, the voice command blocking matrix of Equation 14 is updated so that the action can be executed after the command is entered. The following equation (25) illustrates this.

The voice command blocking matrix In proportion to, the difference between the action just performed and the rest affects the vote. The behavior you just performed is a behavior that directly affects the award, so give more weight and give a smaller amount of weight to the other similar behaviors to learn similar behaviors at once. That is, α (112) and β (111) in the equation (25) have a relationship of α>β> 0.

After this reinforcement learning, each action's vote By increasing, the behavior in the similar behavior set increases the probability of being selected when the behavior is selected in the post-learning behavior selector, especially those that were directly involved in learning.

5) Application to educational software system

The educational software system using the invented virtual life and the new learning structure consists of the following three types.

First, even if learners cultivate virtual creatures that have unique virtual characteristics of each creature or inanimate object and do not directly contact them, they can indirectly learn the characteristics of living creatures and inanimate creatures that are difficult to meet by nurturing the virtual creatures. Educational software is generally configured as shown in FIG. The input device consisting of a keyboard, a mouse, and a speaker is used to stimulate the virtual creature, and the virtual device receives a response from the output device, such as a monitor and a speaker. At this time, the educational content provider provides virtual characteristics of various living and non-living things as shown in FIG. 13, and the educator interacts with the virtual living things through the education system as shown in FIG. Even without contact with the virtual creatures provided, it is possible to indirectly learn about various living things.

In addition, the description for the educational software that can increase the motivation to learn virtual creatures can be raised only when the learner satisfies a predetermined amount of learning in parallel with the curriculum is shown in FIG. The learner performs the educational program provided by the educator through the educational software (142), and is assessed whether the specified learning amount or goal has been achieved (143). At this time, when the amount of learning or the goal is achieved, the right to nurture the virtual creature is given a prize, and when the goal is not achieved, the virtual life is gradually worsened (144). Educators can use this process to motivate learners.

In addition, educational software that allows learners to be educated by educational virtual teachers of various personalities and personalities using virtual creatures having their own virtual personalities, unlike the unilateral educational helpers and characters, is configured as shown in FIG. 15. . Virtual teachers composed of virtual creatures have their own virtual personalities. This allows the learner to select the desired personality of the virtual teacher to learn in the desired learning atmosphere or to change the virtual personality of the virtual teacher according to the learner's learning attitude, leading to the learner in the appropriate learning atmosphere. .

As described above, according to the present invention, a more realistic virtual life may be realized by modeling an internal state having motivation, homeostasis, and emotion, and using all of them to select an action. In addition, in the behavior selection method, various behaviors can be created by using the new probabilistic and definite behavior selection methods.

In addition, any command of the user will be able to efficiently learn the desired behavior of dozens of virtual creatures. It is expected that more effective learning methods will allow each user to have a virtual creature that can recognize his or her own unique commands and be able to apply this method to educational programs.

When the invention is applied to the educational software system by applying the invented virtual life system structure and the learning method of the virtual life, the following invention effects can be expected.

First, learners can expect similar educational effects to learners through interactive indirect experiences, without direct experience with target organisms or inanimate objects, by fostering virtual creatures with virtual personalities that mimic the characteristics of living and inanimate objects.

In addition, only when the learner completes a predetermined amount of learning through the educational program process, the virtual life can be nurtured, thereby increasing the motivation for learning.

In addition, the learner can learn in the learning atmosphere desired by the virtual living teacher having a unique virtual personality, or the virtual personal teacher's virtual personality is changed according to the learning attitude of the learner, thereby providing a suitable learning atmosphere for the learner. Can change. Therefore, learners can learn in a variety of learning atmospheres.

1 is a conceptual diagram of a configuration of a general virtual life.

2 is a diagram illustrating a conventional voice learning portion of a virtual creature or entertainment robot.

3 is a system configuration diagram of a virtual creature according to the present invention.

FIG. 4 is a diagram illustrating the concept of compliance in the recognition system of FIG. 3.

5 is a view showing a problem of the existing emotion when configured as a state.

FIG. 6 is an algorithm for solving the problem of the existing emotion shown in FIG. 5.

7 is a schematic diagram of emotion change when an existing emotion model and an emotion model of the present invention are used.

8 is a behavior selection algorithm of the behavior selector according to the present invention.

9 is a block diagram of a similar behavior set according to the present invention.

10 is a relationship diagram of a command set and a similar behavior set according to the present invention.

11 is a block diagram of a voice learner according to the present invention.

12 is a block diagram of an educational software system through interaction with a virtual living organism according to the present invention.

Figure 13 is a schematic diagram of a software system for education through the virtual living things having various virtual characteristics according to the present invention.

14 is an algorithm of an educational software system capable of cultivating virtual life according to the learning amount according to the present invention.

15 is a schematic diagram of a software system for education through a virtual creature teacher having various virtual personalities according to the present invention.

<Description of Symbols for Main Parts of Drawings>

10: external environment

11: recognizer recognizes the external environment

12: Internal state part representing emotion, motivation, and homeostasis of virtual life

13: Behavior decision unit that uses external information and internal state to determine behavior

14: Learning unit that virtual life adapts to external state

15: Behavior implementation unit where virtual creatures implement behavior

21: Acting on a traditional voice learner

22: Voice command from existing voice learner

23: speech recognition unit in the existing speech learner

31: external environment

32a: sensor in recognition system

32b: encoding system of recognition system

32c: Interest selector in recognition system

33: Internal state system 33a: Synchronous

33b: homeostasis 33c: emotion

34a: behavior selector in behavior system

34b: Various shielding of behavior selectors in the behavior system

34c: Instinct Behavioral Logic of the Behavioral System

35a: Learning system preference learner

35b: Voice Command Learner for Learning System

36: motor system

41: Sensor stimulation of recognition system

42: Sensitivity of recognition system

42a: decrease in sensitivity

42b: Recovery of sensitivity

43: Time to measure incoming sensor stimulus

43a: Time without sensor stimulation

51: Emotional state 1 for example

52: Emotional state 2 for example

71: Emotional change in existing emotional models expressed as states

72: Emotional change in the emotion model according to the improved invention expressed in state

81: Calculation of Provisional Voting Value Vector V

82: mask according to interest

83: mask according to voice command

84: mask according to the emotional state

85: mask according to instinct behavior

86: Calculation of Probability P (b) of Each Action Using Temporary Voting Value Vector V

87: Probabilistic selection of actions based on probability P (b)

91: full set of actions

92: similar behavior set

101: complete command set

102: individual commands

103: Strength of association with command and similar behavior set

111: Influence of link strength on a set of similar behaviors, β

112: Influence of link strength on a set of similar behaviors, α

113: Emotional parameter ρ

114: Reinforcement Learning Algorithm

115: Awards / penalties to virtual creatures after users have determined their actions on orders.

141: Start the learning program to learn according to the amount of learning

142: Learning according to the amount of learning

143: Evaluation of the amount of learning of the learning program to learn according to the learning amount

144: foster virtual life of learning program to learn according to learning quantity

145: End of the learning program to learn according to the learning amount

Claims (12)

In the system for implementing virtual (artificial) life, Recognition means having a coding system, for receiving the sensor information used for the change of the internal state of the virtual life and the behavior selection and learning; Motivation for determining the main behavior of the virtual creature, and lower life structures such as homeostasis for maintaining survival, emotion selection for expressing facial expression, and the like, and based on output information of the recognition means, Internal state means for affecting the internal state; Learning means for adapting the learning of the virtual creature according to an external state based on the output information of the recognition means and the internal state means; A behavior selector having a behavior selector and an instinctive behavior logic, the behavior selection means for determining the behavior of the virtual creature based on the output information of the recognition means, the internal state means and the learning means; And And a movement means for driving the action of the virtual creature based on the output information of the action selection means. The virtual living system according to claim 1, wherein after encoding the sensor information in the coding system of the recognizing means to use the inside of the virtual living body, the adaptation concept is introduced to change the sensitivity when the same sensor stimulus is continuously input. . The virtual living system according to claim 2, wherein the sensitivity is set for each sensor stimulus, and changing values are set differently for each sensor stimulus. The method according to claim 1, wherein when the emotion of the state having the highest value in the 'emotion' made of the state of the internal state means the occupied emotion, the occupied emotion is changed or has a low emotion value by a small change between the two emotions. In addition, the virtual life system, characterized in that the normal emotional state is added to the emotional state to prevent the phenomenon of occupied emotion. The method according to claim 4, wherein after checking the magnitude of the highest emotion value, if the result value is smaller than the 'minimum occupancy emotion determination value', the occupancy emotion is set as' intermediate feeling ', which is an intermediate emotion, and the result value is' minimum occupancy emotion. Virtual life system, characterized by comparing the difference between the largest emotional value and the second emotional value. The method according to claim 5, wherein when the difference between the largest emotion value and the second emotion value is greater than the 'comparison occupancy emotion determination value', the emotion is suddenly changed or each emotion state with the current occupied emotion as the largest value. A virtual living system, characterized in that even if the value is low, emotions are prevented from being determined. The method of claim 1, wherein the action selection means includes a probabilistic action selection method that probabilistically determines an action based on information input from the recognition means and the internal state means, and decisively determines the action in an instinct action by directly receiving a sensor input. A virtual living system, characterized in that the deterministic behavior selection method is applied. The method of claim 1, wherein when the learning means determines a behavior of a virtual creature according to a user's command to determine a connection pair of actions for a corresponding command, the total behavior set is similar to the entire set of actions. Virtual life system, characterized in that the group of 'action set' and then reward or punish the set according to the reward or punishment of the user. The virtual creature of claim 8, wherein the learning rate of the virtual creature is adjusted according to the internal state value in consideration of a process of rewarding or punishing the internal state value of the virtual creature in the 'like behavior set' in the learning process. system. In the educational software system for education and learning using virtual (artificial) life, Virtual and nonliving virtual life, each having a unique virtual character; An input means having a keyboard and a mouse and indirectly instructing the virtual living body using educational software that allows the learner to indirectly learn the characteristics of living and inanimate objects; And And a monitor and a speaker, and output means for outputting a response of the virtual creature to the learning stimulus of the input means. The educational software system of claim 10, wherein the learner performs an educational program provided by the educator through the educational software, evaluates whether the learner achieves a predetermined amount of learning or a goal, and rewards the student with the reward. The educational software system of claim 10, wherein the learner is able to receive learning through a virtual living being acting as a virtual teacher having a virtual personality.