JP2002273674A - Control device for robot, method therefor, program code and storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize a learning function and, especially, characterization via a simple constitution. SOLUTION: When a robot is hit, a panel 2 is stored as a call destination in an empty box of a link table No.1, and motion data corresponding to a painful action are reproduced. If the robot next detects something ahead, one is selected randomly from among boxes of the link table No.1, and a panel is called for according to a call destination stored in the selected box. If the panel 2 is called for, the robot backs at high speed, and the target of processing is moved to the unit next to the unit having called for the panel 2. Upon a pat, the oldest box is deleted from the boxes stored in the link table No.1, and motion data corresponding to a joyful action are reproduced.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a robot control apparatus and method, a program code, and a storage medium.

[0002]

2. Description of the Related Art In the field of robot technology in recent years, there is a robot (communication robot) for realizing virtual communication with a user. There are a plurality of types of communication robots that simply differ in appearance, reality as a creature, and the like.

[0003] As a simple appearance, there are many familiar pet types such as a cat type and a dog type. The reality as the creature depends on the number of action patterns taken by the robot. In other words, the more actions the robot takes depending on the situation, the more the robot will look like a creature.

[0004] For this purpose, some action patterns, such as movements of limbs and body, facial and eye expressions, and sounds such as squeals, are registered in advance in the communication robot. Depending on the user's action on the robot (for example, stroking or hitting the robot) or the situation of the robot (for example, the robot hits something or loses its posture), the robot's behavior pattern to be taken is determined. The robot is selected from the previously stored action patterns, and the robot takes an action according to the selected action pattern.

[0005]

However, in order to make a communication robot more creature-like, it is necessary to register more behavior patterns in the robot.
In addition, creatures have "learning ability" and try to spontaneously change their behavior patterns with reference to past situations.
When the communication robot is provided with this "learning ability", it is necessary to provide a device for realizing the function in addition to the function for selecting the above-mentioned behavior pattern, and to improve the complexity and cost of the communication robot device. There was a problem.

[0006] The present invention has been made in view of the above problems, and has as its object to realize a learning function, particularly a personality rating, with a simple configuration.

[0007]

In order to achieve the object of the present invention, for example, a robot controller of the present invention has the following arrangement.

That is, a panel storage means for storing a plurality of panels storing at least one predetermined condition and at least one process associated with each of the predetermined conditions, and one of the plurality of panels. And a control means for controlling the robot using the panel selected by the panel selection means, wherein the selection means selects a panel candidate to be used next. Candidate information storage means for storing a predetermined number of candidate information to be represented, and panel determination means for determining a panel to be used next based on the storage contents of the candidate information storage means, wherein the candidate information is stored in the candidate information storage means. The memory of
This is performed by the processing stored in the panel.

In order to achieve the object of the present invention, for example, a robot control device of the present invention has the following configuration.

That is, the behavior pattern of the robot and / or
Or, a panel storage unit that stores a plurality of panels that describe a state of the robot, conditions for the robot to take the action, and / or conditions for the robot to enter the state, and a storage unit that stores a plurality of panels. And when determining the behavior pattern of the robot and / or determining the state of the robot, selects one of the panels stored in the storage means, and based on the selected panel, the behavior pattern and / or Or a determining means for determining the state.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail according to preferred embodiments with reference to the accompanying drawings.

[First Embodiment] In this embodiment, a device (hereinafter simply referred to as a control) that controls a communication robot (hereinafter, sometimes simply referred to as a robot) that acts in accordance with the external situation and / or its own situation. And the method will be described.

<Structure of Control Device> FIG. 1 shows the structure of a control device according to the present embodiment.

Reference numeral 101 denotes a physical sensor unit for detecting the external situation of the robot and its own situation. The physical sensor unit 101 as a sensor for detecting the situation of the outside world of the robot includes an optical sensor (for example, a cds cell) for detecting brightness, a sound sensor for detecting sound, and detecting the situation when the robot hits something. Touch sensor, an infrared objective sensor that detects whether or not an object (for example, an obstacle) exists around the robot with infrared rays. Further, the physical sensor unit 101 as a sensor for sensing the situation of the robot itself is configured by an attitude sensor for detecting the attitude of the robot. The above sensor is an example in the present embodiment, and is not limited to this. or,
Since the mechanism of each sensor and its operation rule are not in the category described in the present embodiment, the description is omitted here.

Reference numeral 102 denotes an internal sensor unit for detecting a situation inside the robot. The internal sensor unit 102 includes an internal clock that functions as a clock inside the robot, a counter, and the like.

Reference numeral 103 denotes a sensor processing unit, which is a physical sensor unit 1
01 and the detection result from the internal sensor unit 102,
Specifically, results such as “collision” and “brightness” are output.

An action calculation unit 104 determines an action pattern corresponding to an input from the sensor processing unit 103 by using a panel 105a stored in a data storage unit 105 described later. This determination method will be described later. Also,
A work area 104a for reading a panel to be used from the data storage unit 105 is provided inside.

A data storage unit 105 includes a panel 105a (which will be hereinafter referred to as a character of the robot) which indicates a causal relationship of the behavior of the robot (this behavior will be hereinafter referred to as a character of the robot), and various types of data taken by the robot. Motion data 10
5b, pattern data (eye pattern data) 105c displayed on the eyes of the robot (configured by LEDs and the like, and capable of displaying various patterns), and data (sounds) such as barks and sound effects emitted by the robot Data) 105d and a link table 105f (details will be described later).

Reference numeral 106 denotes an action processing unit which controls various parts (hands, feet, necks, eyes, etc.) of the robot based on the action pattern determined by the action calculation processing unit 104. The action processing unit 106 includes a motion reproduction unit (constituted by an actuator and a traveling wheel) for reproducing the motion of the robot, the above-described eye display unit, and a sound generation unit (speaker or D / A converter) and an operation unit for performing various operations.

<Panel and Unit> FIG. 2 shows an example of the panel. Units are provided on the panel, and the types are roughly classified into a sensor unit 201 and an action unit 202. The sensor unit 201 indicates conditions for the robot to take the action indicated by the action unit on the same line in the panel. For example, if the touch sensor corresponding to the right hand of the robot detects that something has hit the right hand of the robot, the action unit corresponding to the sensor unit that indicates "hit the right hand", that is, "Expression pain!", "Sound surprise" ”,“ Walk back 1 second,
The four action units of “interruption prohibited”, “left-hand rotation 1 second, interrupt prohibited” are executed in this order. That is, the action unit is always executed from the left.

When the action unit indicating “expression is painful” is executed, the eye pattern indicating the expression “painful” is converted from the eye pattern data 105 c stored in the data storage unit 105 to the action calculation processing unit 104. Is selected, and the selected eye pattern is displayed on the action processing unit 106 (eye display unit).

When the action unit indicating “surprise” is executed, the sound data indicating “surprise” is selected from the sound data 105 d stored in the data storage unit 105 by the action calculation processing unit 104, and is selected. The processed sound data is reproduced as a sound by the action processing unit 106 (sound generating unit).

When the action unit indicating “retreat 1 second, interrupt prohibited” is executed, the data storage 10
5 is selected by the action calculation processing unit 104 from the motion data 105b stored in No. 5 and the selected motion data is only for one second (timed by the internal clock of the internal sensor 102). At 106 (motion reproducing unit), the motion is reproduced as a motion. That is, the robot takes a walking backward action. At this time, since there is an instruction of “interruption prohibition”, interruption of other processing is prohibited during walking retreat. But,
It is also possible to provide an action unit that has this interrupt, and in that case, an interrupt of a process with a higher priority is permitted.

When an action unit indicating “left walking rotation for 1 second, interrupt prohibition” is executed, the data storage unit 1
The motion data indicating “left walking rotation” is selected from the motion data 105b stored in the memory unit 05 by the action calculation processing unit 104, and the selected motion data is processed for one second (measured by the internal clock of the internal sensor 102). Unit 106 (motion playback unit)
Is played back as a motion. In other words, the robot takes a left walking rotation action. At this time, since there is an instruction of “prohibition of interruption”, interruption of other processing is prohibited during the left walking rotation.

As described above, the action unit corresponding to the sensor unit satisfying the condition is executed.
Then, the action processing unit 106 performs a process according to the action unit to be executed. In practice, the action calculation processing unit 104 repeats each sensor unit provided on the panel read into the work area 104a from top to bottom and scans periodically (for example, from the top unit to the bottom unit of the panel). Scan repetition cycle up to 0.1 seconds). Then, it is checked whether there is any sensor unit corresponding to the sensor result input from the sensor processing unit 103. If there is a corresponding sensor unit, the action unit corresponding to the corresponding sensor unit is executed as described above.

In the present embodiment, the priority order of the units executed on the panel (when the input from the sensor processing unit 103 corresponds to a plurality of sensor units) is lower in the present embodiment. , The lower the unit, the higher the priority, but the present invention is not limited to this.

In FIG. 2, there is an action unit indicating “panel jump panel No. 2”, which is a unique number assigned to each panel. 2 indicates that the panel corresponding to 2 is to be loaded into the work area 104a. Specifically, from the panel 105a stored in the data storage unit 105, the No. The second panel is selected, and the work area 104a is rewritten to the selected panel. By doing so, the behavior of the robot can be varied, and by having more panels, it is possible to have more characters.

<Link Table> FIG. 3 shows a link table. The link table stores data (for example, a pointer that points to an address where the panel is stored) that specifies the call destination panel.
5f is stored in the data storage unit 105. In this embodiment, one link table can store up to 16 call destinations (16 boxes are provided).
However, the present invention is not limited to this.

FIG. 4 shows an action unit relating to the link table.

Reference numeral 401 denotes one of 16 boxes of the designated link table (link table No. 1 in the example of FIG. 3).
This is an action unit indicating that one is selected at random. The call destination stored in the selected box is, for example, panel No. If it is 5, the panel No. 5
Is selected from the data storage unit 105 and the work area 10
4a. If no call destination is stored in the selected box (for example, NUL
If the L code is stored), this action unit is ignored and the next action unit is executed.

An action unit 402 indicates that the designated panel (panel No. 5 in the figure) is stored in the designated link table (link table No. 1 in the figure). When storing panels in the designated link table, as shown in FIG. 5, the panels are sequentially stored in empty boxes. However, if all the boxes included in the designated link table are not empty, they are always pushed out of the old box (in FIG. 6, the lower box is older) and disappear.

An action unit 403 indicates that the specified link table (table No. 1 in the figure) is to be deleted from the box in the oldest order.

An action unit 404 indicates that all the boxes of the designated link table (table No. 1 in the figure) are to be deleted.

A method of learning a robot will be described below using the action units and the link table relating to the link table.

<Learning method using link table> In this learning example, if the user hits repeatedly, he or she will be hit before it is hit (when his / her hand comes in front of him) and escape behind. Will be described.

FIG. 7 shows panels 1 and 2 used in this learning example. Panel 1 includes sensor units 711, 713, 71
6,718 and action units 712,714,715,717,71 corresponding to the respective sensor units.
9, 720 are provided. On the other hand, the panel 2 is provided with a sensor unit 751 and corresponding action units 752 and 753. In the following, it is assumed that the panel 1 has been loaded in the work area 104a in advance.

In the above state, the robot basically has nothing (no sensor is responding, or the value detected by the sensor is a predetermined value (the sensor processing unit 103 detects that the sensor has detected something). And the action unit 712 corresponding to the sensor unit 711 is executed, and the action calculation processing unit 104 specifies the motion data of “sitting” from the data storage unit 105. The action processing unit 106 reproduces the specified “sitting” motion data.

If the robot is hit (the brightness detected by the optical sensor becomes dark within a predetermined time) within a predetermined time (for example, within one second), the robot is hit once (the internal sensor 1).
02 (count by the counter of 02) sound (physical sensor unit 1)
01 is detected), the action units 714 and 71 corresponding to the sensor unit 713 are detected.
5 are executed in this order. First, action unit 71
4 is executed, and the link table No. 1 (Since the link table uses only link table No. 1 in the description of this learning example, it may be simply called a link table)
One piece of data (for example, a pointer indicating the address where the panel 2 is stored) for making the panel 2 a call destination is stored in the empty box. Then, the action unit 715 is executed, the motion data corresponding to the painful action is specified from the data storage unit 105, and the action processing unit 106 reproduces the motion data corresponding to the specified “painful action”. By doing this,
Each time the robot is hit, the link table stores data with the call destination as panel 2 (of course, if the user hits 16 times, the data with the panel 2 as the call destination is stored in all the boxes of the link table). Will be.)

Next, the robot (the physical sensor unit 101)
If something is detected before (for example, by an infrared objective sensor), the action unit 717 corresponding to the sensor unit 716 is executed. Specifically, one is randomly selected from the boxes (16 boxes) in the link table, and the panel is called based on the call destination stored in the selected box. If the box is empty, the process proceeds to the next without doing anything. However, according to the above-described processing, if the robot is hit many times, the probability of selecting a box to call the panel 2 increases. Therefore, if the robot is hit many times, the probability of calling the panel 2 is increased by the execution of the action unit 717 just because there is something in front of the robot.

The panel 2 is provided with a sensor unit 751 and action units 752 and 753. When the panel 2 is called by the execution of the action unit 717, the panel 2 is loaded in an area different from the panel 1 in the work area 104a. When the panel 2 is called, there is nothing in the robot, and the action units 752 and 753 corresponding to the sensor unit 751 are executed in this order. First, when the action unit 752 is executed, a fast retreating motion is stored in the data storage unit 10.
5, and the action processing unit 106 reproduces the fast-retreating motion. Then, the action unit 753 is executed, and the panel return, that is,
The processing target is moved to the unit next to the unit that called panel 2.

In this way, the frequency of the processing in the panel 2, that is, the processing of retreating at a high speed, increases as described above in accordance with the frequency of the number of times the robot head is hit.

Further, in the panel 1, when a stroke is made (a sudden brightening or darkening is detected five or more times within a predetermined time (for example, within 5 seconds) five times or more (can be detected by the above-described sensors)), the sensor unit Action units 719 and 720 corresponding to 718 are executed. First, by executing the action unit 719, the oldest box is deleted from the boxes stored in the link table. In other words, the number of boxes for storing “data for making panel 2 a call destination” is reduced by one in the link table (one or more pieces of data that make panel 2 a call destination are stored in the link table). Is assumed). In other words, when the action unit 717 is executed, the probability of calling the panel 2 is reduced. As a result, the probability of retreating at high speed is reduced, and it can be seen that "a little better mood" has been seen in personification. Then, by executing the action unit 720, motion data corresponding to the action of joy is selected from the data storage unit 105, and the selected motion data is reproduced by the action processing unit 106.

FIG. 8 is a diagram for explaining a learning method using the panels 1 and 2 and the link table. The figure shows the change in the number of panels 2 stored in the same one link table.
(In FIG. 1, only panel 2) is stored. Therefore, by executing the action unit 810, another piece of the same data is added, and the state becomes the state 801. Here, action unit 8
When step 20 is executed, one more piece of the same data is added, and a state 802 is set. On the other hand, the action unit 83
When 0 is executed, the stored data is deleted from the oldest one, the same data is deleted from the state of 801, and 803 is deleted.
State.

Also, in the above-described processing, the mood of the robot is improved little by little every time it is stroked, but, for example, by using the action unit 404 shown in FIG. If it is stroked, all the boxes in the link table will be deleted, so the data (panel 2) identifying the call destination panel will be deleted,
It becomes a character that makes you feel good at once.

Also, the action unit 403 shown in FIG. 4 is used instead of the action unit 719, and the action unit 403 is executed every predetermined time (for example, one minute) (time measurement by the internal clock of the internal sensor 102). Since the boxes in the link table are deleted one by one every time the predetermined time elapses, the robot is in a good mood every time the predetermined time elapses.

As described above, the learning object is expressed as a panel (panel 2 in the above example), the learning is performed by storing the panel in the link table, and the learning degree is expressed as the number of panels stored in the link table. By doing so, it is possible to realize a learning function (including a personality rating) by extending the action program of the robot using the panel and the unit.

FIG. 9 is a flowchart of the learning process described above.
Will be described. First, the panel 1 is loaded into the work area 104a (step S901). Then, an input from the sensor processing unit 103 is received (step S90).
2). Then, in the input from the sensor processing unit 103,
It is determined whether there is an input corresponding to the sensor unit (step S903), and if so, the process proceeds to step S904. FIG. 10 shows a flowchart of the specific processing in step S904.

When the input from the sensor processing unit 103 indicates that the player has been hit (step S100).
1) The process proceeds to step S1002, and panel 2 is stored in the link table as a call destination (step S1002).
1002). Then, the motion data based on the "painful action" is reproduced (step S1003).

When the input from the sensor processing unit 103 indicates that "there is something before" (step S10).
04), the process proceeds to step S1005, and one is randomly selected from the boxes of the link table (step S1005). If the selected box is empty, the process proceeds to step S1009. If the selected box stores data for calling the panel 2, the process proceeds to step S 1007, calls the panel 2 (step S 1007), and performs fast retreat (step S 1008).

When the input from the sensor processing unit 103 indicates "patched" (step S100)
9), the oldest box is deleted from the boxes of the link table (step S1010), and the “joy action”
(Step S1)
011).

When there is no input from the sensor processing unit 103 or when the input is equal to or less than a predetermined value (step S
1012), the motion data of "sitting" is reproduced (step S1013).

Then, the process returns to step S905 in FIG.

In the present embodiment, if no panel is registered as a call destination in the link table (in this learning example, the probability that the panel 2 will be the call destination is 0), the robot will be sitting all the time. Although the motion data is reproduced, the present invention is not limited to this. At least one of the call destinations (panel 2 in this learning example) may be registered.

[Second Embodiment] In the present embodiment, a neutral sensor (a sensor that makes it difficult for the robot to express emotions) is replaced by a sensor that is meaningful (for easy expression of emotions) to the robot by learning by the robot. Here is an example of learning.

For example, in a pet as a general creature, inputs such as “hit” and “hit” are not good inputs for the pet. Then, the emotions become unpleasant emotions, and the emotions are output as facial expressions and actions.
However, for example, "sound the handclap once"
For such inputs, the emotions of the pet are not unique.
Therefore, three examples of learning an emotional expression for such a neutral input will be described in the present embodiment. The method of using the panel (including the sensor unit and the action unit) and the link table for performing the learning is as described in the first embodiment.

<Learning example 1> For example, when a hand clap is sounded once and then the robot is praised (for example, when the robot is stroked), a panel (arranged in the panel) is made to cause the robot to reproduce a motion "pleasant with stiffness". (Including a sensor unit and an action unit).
As a result, each time the handclap sounds, the robot learns to play a “slippery” motion.

In order to realize the above-mentioned learning, the panel 1 detects the sound of the hand clap once and detects the sound of the hand clap once.
Is stored in the link table, and when praised, one box is selected at random from the link table and the panel specified by the call destination stored in the box is read and executed. Deploy.
On the other hand, the sensor unit and the action unit are arranged on the panel 2 so as to reproduce a motion of “pleasant with stiffness” when there is nothing.

<Learning Example 2> For example, when there is something in front of the robot (when it is detected that there is something in front of the robot), hitting the robot causes the robot to reproduce a "painful" motion. And a panel (including a sensor unit and an action unit arranged in the panel). As a result, the robot learns to play a "painful" motion each time it detects something in front.

In order to realize the above-mentioned learning, when it detects that something is in front of the panel 1, one panel 2 described later is stored in the link table, and when it is hit, one box is randomly selected from the link table. Select and arrange the sensor unit and action unit to read and execute the panel specified by the call destination stored in the box. On the other hand, a sensor unit and an action unit are arranged on the panel 2 so as to reproduce a motion that "aches" when nothing is done.

<Learning Example 3> For example, if the robot suddenly becomes bright (when it is detected that the robot suddenly becomes bright), the throat is gurgling (a touch sensor or the like is provided on the throat of the robot, and " (When it is possible to detect ")", a panel (including a sensor unit and an action unit arranged in the panel) is configured to cause the robot to reproduce a "pleasant" motion. As a result, the robot learns to play a “pleasant” motion each time it detects that it suddenly becomes bright.

In order to realize the above-mentioned learning, when it is detected that the panel 1 suddenly becomes bright, one panel 2 described later is stored in the link table. Select one box and arrange the sensor unit and action unit to read and execute the panel specified by the call destination stored in the box. On the other hand, a sensor unit and an action unit are arranged on the panel 2 so as to reproduce a “pleasant” motion if there is nothing.

As a result of the above learning, the robot can learn emotions according to neutral inputs, and can learn actions and facial expressions according to the learned emotions.

[Third Embodiment] The causal relationship of the behavior of the robot (that is, the character of the robot) in the first and second embodiments is realized by a combination of a panel and units provided therein. This unit combination is externally programmable. For example, it is possible to arrange units freely in a panel by using a GUI that operates on a personal computer. as a result,
Since an original panel is created, the created panel is transferred to the robot. The transfer is performed by connecting a personal computer and a robot to terminals of a predetermined standard (for example, UBS).
Terminals, etc.). Of course, it is obvious that the above-mentioned personal computer is installed with the above-mentioned GUI program code and data, and a driver for transferring data to the robot.

[0064]

As described above, according to the present invention, it is possible to realize a learning function, particularly, a personality rating with a simple configuration.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a configuration of a control device according to a first embodiment of the present invention.

FIG. 2 is a diagram illustrating an example of a panel.

FIG. 3 is a diagram showing a link table.

FIG. 4 is a diagram showing an action unit relating to a link table.

FIG. 5 is a diagram illustrating a case where a panel is stored in a specified link table.

FIG. 6 is a diagram illustrating a case where all boxes included in the link table are not empty when a panel is stored in a designated link table.

FIG. 7 is a diagram showing panels 1 and 2 used in a learning example.

8 is a diagram illustrating a learning method using panels 1 and 2 shown in FIG.

FIG. 9 is a flowchart of a learning process.

FIG. 10 is a flowchart of a specific process in step S904.

 ──────────────────────────────────────────────────の Continued on the front page F term (reference) 2C150 CA01 CA02 CA04 DA05 DA24 DA25 DA26 DA27 DA28 DF03 DF04 DF06 DF33 ED42 ED47 ED52 EF03 EF07 EF13 EF16 EF23 EF28 EF29 EF33 EF36 3C007 AS36 KS10 KS12 KS12 KS20 KS12 KS12 KS20

Claims (10)

[Claims] 1. A panel storage means for storing a plurality of panels storing at least one predetermined condition and at least one process associated with each of the predetermined conditions, and one of the plurality of panels. And a control means for controlling the robot using the panel selected by the panel selection means, wherein the selection means selects a panel candidate to be used next. Candidate information storage means for storing a predetermined number of candidate information to be represented; and panel determination means for determining a panel to be used next based on the storage content of the candidate information storage means. Is stored by the processing stored in the panel. 2. The robot according to claim 1, wherein the panel determination unit makes the determination in accordance with the number of pieces of candidate information stored in the candidate information storage unit that represent the same panel. Control device. 3. When the candidate information storage means receives new candidate information while the predetermined number of candidate information is stored, the candidate information storage means stores the new candidate information in place of the oldest candidate information. The control device for a robot according to claim 1 or 2, wherein: 4. The apparatus according to claim 1, wherein said candidate information storage means stores at least one piece of candidate information.
The control device for a robot according to any one of claims 3 to 3. 5. When each of the predetermined conditions included in the panel has a priority, and the control unit determines that a plurality of the predetermined conditions included in the panel are satisfied.
The robot control device according to any one of claims 1 to 4, wherein a process associated with the condition having the higher priority is executed. 6. A panel storage for storing a plurality of panels describing a behavior pattern of a robot and / or a state of the robot and a condition for the robot to perform the behavior and / or a condition for the robot to enter the state. Means, a storage means for storing a plurality of panels, and when determining an action pattern of the robot and / or determining a state of the robot, one of the panels stored in the storage means is selected, A control device for a robot, comprising: determination means for determining the behavior pattern and / or the state based on the selected panel. 7. A plurality of panels each storing at least one predetermined condition and at least one process associated with each of the predetermined conditions are stored in a panel storage means, A control method for a robot, comprising: a panel selection step of selecting one; and a control step of controlling the robot using the panel selected in the panel selection step. A predetermined number of candidate information representing candidates is stored in the candidate information storage means, a panel to be next used is determined based on the storage content of the candidate information storage means, and the candidate information is stored in the candidate information storage means. Is performed by a process stored in the panel. 8. A panel storing means for storing a panel describing a behavior pattern of a robot and / or a state of the robot and a condition for the robot to perform the behavior and / or a condition for the robot to enter the state. A storage step of storing a plurality of panels in a storage unit; and determining a behavior pattern of the robot and / or a state of the robot, by selecting one of the panels stored in the storage unit. And selecting the behavior pattern and / or the state based on the selected panel. 9. A program code for executing the method of controlling a robot according to claim 7 on a computer. 10. A computer-readable storage medium storing the program code according to claim 9.