JP2001157981A - Robot device and control method thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a robot device and a control method thereof capable of extraordinarily improving amusement performance. SOLUTION: In this robot device and this control method for that for generating an action based on an action model, the action model is selected among plural action models based on at least one of input information from the external and self-action and/or growth history. As a result, in this control method for the robot device, the next action that is the optimum can be continuously performed in accordance with a current situation, thereby a robot device and a control method thereof capable of extraordinarily improving amusement performance can be provided.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a robot apparatus and a control method therefor, and is suitably applied to, for example, a pet robot.

[0002]

2. Description of the Related Art In recent years, a pet robot of a four-legged walking type that performs a predetermined operation in accordance with a command from a user or a surrounding environment has been proposed and developed by the present applicant. Such a pet robot has a shape similar to a dog or cat bred in a general household, and operates according to a command from a user, the surrounding environment, and the like.

[0003]

However, in such a pet robot, functions such as real dogs and cats for performing the next optimal action and operation according to the current situation and the past experience are used. If a function to change the next action and action can be provided, it is considered that the user can be provided with a further sense of intimacy and satisfaction, and the amusement as a pet robot can be further improved.

The present invention has been made in view of the above points, and an object of the present invention is to propose a robot apparatus and a control method thereof that can significantly improve amusement.

[0005]

According to the present invention, there is provided a robot apparatus and a control method for the same, the apparatus having a plurality of types of action models, and using an action selection means to input externally input information. The output of one behavior model is selected from the output of each behavior model based on at least one of the self behavior history and / or the growth history. As a result, according to the robot apparatus and the control method thereof, it is possible to continuously perform the optimal next action according to the current situation.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the drawings.

(1) Configuration of Pet Robot According to the Present Embodiment In FIG. 1, reference numeral 1 denotes a pet robot according to the present embodiment as a whole, and leg units 3A to 3D are connected to the front, rear, left and right of body unit 2, respectively. The head unit 4 and the tail unit 5 are connected to the front end and the rear end of the body unit 2, respectively.

A cooling fan (not shown) is provided inside the body unit 2, and an exhaust port 2AX and an intake port (not shown) are formed on the upper surface 2A and the lower surface 2B via the cooling fan. Have been. Accordingly, in the pet robot 1, in response to the driving of the cooling fan, the air sucked from the intake port is discharged to the outside from the exhaust port 2AX through the inside of the body unit 2, so that the inside of the body unit 2 is The temperature can be reduced.

(2) Internal Configuration of Pet Robot System Here, in the pet robot 1 shown in FIG. 2, the body unit 2 includes a controller 10 for controlling the operation of the pet robot 1 as a whole, and the power of the pet robot 1 An information reading unit 1 to which a battery 11 serving as a power source, an internal sensor unit 14 including a battery sensor 12 and a heat sensor 13, and an external storage memory 15 and an internal storage memory 16 are connected.
7, an acceleration sensor 18 and an angular velocity sensor 19 are housed.

The head unit 4 has a microphone 20 corresponding to an "ear" and a CCD (Char) corresponding to an "eye".
ge Coupled Device) Camera 21 and Touch Sensor 22
, A speaker 23 corresponding to the "mouth", and a distance sensor 24 such as an infrared distance sensor are provided at predetermined positions.

Further, the joints of the leg units 3A to 3D, the leg units 3A to 3D and the trunk unit 2
Each coupling portion of the head unit 4 and the connecting portion of the body unit 2, and the tail unit 5 and the body unit, respectively, etc. The coupling portion actuator 3AA ₁ of 2
３3AA _K , 3BA ₁ ３3BA _K , 3CA ₁ Ｃ3C
_{A K, 3DA 1 ~3DA K,} 4A 1 ~4A L, 5A 1 ~
5A _M is disposed. The microphone 20 of the head unit 4 is provided with a sound commander (a commander that generates a sound of a different scale according to the operation content) from the user as a musical scale such as “walk”, “down” or “chase the ball”. Is collected, and the obtained audio signal S1 is sent to the controller 10. Also C
The CD camera 21 captures an image of the situation in the forward direction, sends the obtained image signal S2 to the controller 10, and the distance sensor 24 measures the distance to the object in front, and sends the measurement result to a distance measurement signal. It is sent to the controller 10 as S3.

Further, the touch sensor 22 is provided on the upper part of the head unit 4 as apparent from FIG. 1, and detects a pressure received by a physical action such as "stroke" or "slap" from the user. The detection result is sent to the controller 10 as a pressure detection signal S4.

Further, the battery sensor 12 of the body unit 2 detects the remaining amount of the battery 11 in a plurality of stages, and outputs the detection result of each stage to a battery remaining amount detection signal S5.
To the controller 10 sequentially.

Further, the heat sensor 13 of the body unit 2
Detects the internal temperature of the pet robot 1 and sends the detection result to the controller 10 as a heat detection signal S6.

Further, the acceleration sensor 1 of the body unit 2
8 detects acceleration in three axes (X-axis, Y-axis, Z-axis) in units of several tens of milliseconds, and sends the detection result to the controller 10 as an acceleration detection signal S7. In addition, the angular velocity sensor 19 has three (R angle, P angle,
The rotation angular velocity in the (Y angle) direction is detected, and the detection result is sent to the controller 10 as an angular velocity detection signal S8.

The controller 10 includes a microphone 20,
CCD camera 21, distance sensor 24, touch sensor 2
2, audio signal S1 given from battery sensor 12, heat sensor 13, acceleration sensor 18, angular velocity sensor 19,
Image signal S2, distance measurement signal S3, pressure detection signal S4,
Based on the remaining battery level detection signal S5, the heat detection signal S6, the acceleration detection signal S7, the angular velocity detection signal S8, and the like, it is determined whether there is a surrounding situation, a command from the user, an action from the user, and the like.

The controller 10 determines a subsequent action based on the result of the determination and a control program input from the external storage memory 15 via the information reading unit 17,
Actuators 3AA _{1 to} 3 required based on determination result
_{AA K, 3BA 1 ~3BA K,} 3CA 1 ~3CA K, 3
_{DA 1 ~3DA K, 4A 1 ~4A} L, by driving the 5A ₁ to 5 A _M, or to shake the head unit 4 up and down and right and left, or to move the tail unit 5, the leg units 3A~3D To perform an action such as walking.

At this time, the controller 10 outputs a sound based on the sound signal S9 to the outside by giving a predetermined sound signal S9 to the speaker 23, if necessary, or the position of the "eye" of the pet robot 1. LED (Light Emitting Diode) (not shown) provided in the LED is turned on, off, or blinks.

As described above, the pet robot 1 can behave autonomously based on the surrounding conditions, control programs, and the like.

The external storage memory 15 in the body unit 2 is a recording medium such as a memory stick that can be inserted from the outside, and the above-described control program for the entire pet robot 1 is stored in the recording medium in advance. I have. Further, in the internal storage memory 16, hardware information, calibration information, an error operation execution program, learning information, personality information, and the like are stored in a rewritable state. Various programs and information stored in the external storage memory 15 and the internal storage memory 16 can be read and written by the information reading unit 17 under the control of the controller 10.

Accordingly, the controller 10 determines the behavior of the pet robot 1 based on various programs and information read from the external storage memory 15 and the internal storage memory 16 by the information reading unit 17,
Corresponding to the determined driving the address 3AA ₁ to 5 A _M required, and to output the sound from the speaker 23 as required.

In this way, in the pet robot 1, the autonomous behavior of the pet robot 1 can be individually changed according to various programs and information stored in the external storage memory 15 and the internal storage memory 16. It has been made like that.

(3) Behavioral Expression of Pet Robot In this pet robot 1, in addition to acting autonomously based on surrounding conditions and a control program, etc., the pet robot 1 responds to the current situation as if it were a real animal. In addition to performing the optimal next action and action, the next action and action are changed based on past experience.

Here, the processing of the controller 10 regarding the action generation of the pet robot 1 will be described.

As shown in FIG. 3, when the content of the processing of the controller 10 relating to the action generation of the pet robot 1 is functionally classified, a recognition object 30 for recognizing a specific external state and a recognition result of the recognition object 30 are input. A semantic conversion object 31 for converting into a character string called semantics, an emotion / instinct generation object 32 expressing an emotion and an instinct state based on input semantics obtained from the semantic conversion object 31, and the semantic conversion object 31 Semantics and emotion / instinct generation object 3 obtained from
2. An action generation object 33 that enumerates action models that can be selected based on the expression result of 2, and an action selection object 3 that determines an optimum action model for a subsequent action among the action models enumerated by the action generation object 33.
4, can be divided into a behavioral expression object 35 to express the behavior corresponding to the behavior model determined the pet robot 1 by the action selection object 34 by controlling driving the actuators 3AA ₁ to 5 A _M.

In this case, the recognition object 30 is composed of an audio signal S1 supplied from the microphone 20, an image signal S2 supplied from the CCD camera 21, a distance measurement signal S3 supplied from the distance sensor 24, and a pressure detection signal S4 supplied from the touch sensor 22. Based on the acceleration detection signal S7 given from the acceleration sensor 18 and the angular velocity detection signal S8 given from the angular velocity sensor 19, the surrounding state of the pet robot 1, the instruction from the user, the presence or absence of the user's action, etc. are recognized. At the same time, based on the remaining battery level detection signal S4 and the heat detection signal S5 provided from the battery sensor 22 and the heat sensor 23 constituting the internal sensor 24, the remaining state of the battery 21 and the pet robot 1
After recognizing the internal temperature of, the recognition result is notified to the meaning conversion object 31.

After converting the recognition result supplied from the recognition object 30 into input semantics, the semantic conversion object 31 converts the input semantics into emotion / semantics.
Instinct generation object 32 and action generation object 3
Supply 3

The emotion / instinct generation object 32 is based on the supplied input semantics and the personal information 36 read from the internal storage memory 16 in the torso user 2, and the state of the emotion and instinct of the pet robot 1. Is determined, and only when the state of the determined emotion and instinct exceeds a predetermined level, the state of the emotion and the instinct is notified to the action generation object 33. In this personal information, the relationship between the input semantics and the state of emotion and instinct is parameterized.

The emotion / instinct generation object 32
Holds a parameter indicating the strength of the emotion for each of the six emotions of "joy,""sadness,""anger,""surprise,""fear," and "disgust," , "Search desire", "exercise desire", and "appetite", each of which holds a parameter indicating the strength of the desire. Parameter values of these emotions and desires are Each of them is sequentially updated based on a specific recognition result (“struck” and “stroke”) given from the input semantics, the passage of time, and the like.

Specifically, the emotion / instinct generation object 32
Indicates that the input semantics supplied are pet robot 1
Of the pet robot 1, the degree of the parameter values of each emotion and each desire held at that time on the current action of the pet robot 1, and the suppression and stimulation received from other emotions. , The parameter values of the emotion and the desire are updated based on the behavior output of the pet robot 1 and the elapsed time.

The action generation object 33 has input semantics supplied from the meaning conversion object 31 and:
Based on the parameter values of each emotion and each desire supplied from the emotion / instinct generation object 32 and the state transition table 37 read from the internal storage memory 16 in the body user 2, the pet robot 1 is expressed next. Choose an action to take.

The behavior generation object 33, such as shown in FIG. 4, the action model sequence then expressible behavior patterns the pet robot 1 according to various sensor outputs described above as each behavior model 40A ₁ ~40A _N may be set as a library 40, it is among the behavioral model library 40, adapted to enumerate selectable behavior model 40A ₁ ~40A _N of the supplied sensor output.

The behavior model library 40 includes a "ball response behavior model", an "autonomous search behavior model", an "emotional expression behavior model", an "obstacle avoidance behavior model",.
..., "remote control mode behavior model", ..., "falling recovery behavior model", "battery management behavior model", etc. corresponding to several pre-selected condition items,
Independent behavior model 40A ₁ ~40A _N, respectively.

[0034] In the case of this embodiment, the behavior model 40A ₁ ~40A _N is, as a method to determine the next action, one node, as shown in FIG. 5 (state) NODE ₀
~ NODE _n to any other node NODE ₀ ~ NODE
_The probability of stochastically determining whether to transition to _n based on the transition probabilities P _{1 to} P _{n + 1} set for the arcs ARC ₁ to ARC _{n + 1} connecting the nodes NODE _{0 to} NODE _n. An algorithm called automaton is used.

Specifically, each of the behavior models 40A _{1 to} 40A
_N is each node N forming its own behavior model
Respectively in correspondence to the ODE ₀ ~NODE _n, and a state transition table 37 as shown in FIG. 6 of each of these nodes NODE ₀ ~NODE _n.

In this state transition table 37, the node NO
Input events (recognition results) as transition conditions in DE _{0 to} NODE _n are listed in order of priority in a row of “input event name”, and further conditions for the transition condition are rows of “data name” and “data range”. In the corresponding column.

Therefore, in the node NODE ₁₀₀ represented by the state transition table 37 in FIG. 6, "ball detected (BALL)"
When the recognition result is given, the condition for transitioning to another node is that the "size (SIZE)" of the ball given together with the recognition result is in the range of "50 to 100". I have. Also this node NODE
_{In 100} , when a recognition result of “obstacle detected (OBSTACLE)” is given, the “distance (DISTANCE)” to the obstacle given together with the recognition result is given.
E) is in the range of “40 to 80” or “JOY” held in the emotion / instinct generation object 32
Parameter values are "50 to 60", "Surprised (SURPRI
SE) "or the parameter value of" Sadness "is" 90 to 100 ".
Is a condition for transitioning to another node.

Further in the state transition table 37, with the node name that can transition from the node NODE ₀ ~NODE _n in the column of "transition destination node" in the column of "probability of transition to another node" is listed, the "input Other nodes NODE _{0 to} N that can transition when all the conditions described in the rows of “event name”, “data value”, and “data range” are met
The transition probability to ODE _n is described in a corresponding part in the column of “transition probability to another node”, and the node N
The action to be output when transitioning from ODE _{0 to} NODE _n is “output action” in the column of “transition probability to another node”.
Is described in the line. Note that the sum of the probabilities of the respective rows in the column of “transition probability to another node” is 100 [%].

Therefore, in the node NODE ₁₀₀ represented by the state transition table 37 in FIG.
L) "and the" size "of the ball is" 50 ".
Is given in the range of “node NODE” with a probability of “30 [%]”.
₁₂₀ (node 120) ”, and then“ ACTIO ”
The action of "N1" will be output.

Each of the behavior models 40A ₁ to 40 _N is
Is configured as such a state node NODE ₀ ~NODE _n written as transition table, each several leads, such as when a given input semantics recognition result from the corresponding node NODE ₀ ~NOD
Stochastically determine the next action by utilizing a state transition table of E _n, and to output the determination result to the action selection object 34.

The action selection object 34 selects an action output from a predetermined high-priority action model among actions output from the action models 40A ₁ to 40 _N of the action model library 40, respectively. The command to execute the action is converted into a character string having output semantics and output. In this embodiment, the behavior model 4 shown in the lower part of FIG.
0A ₁ priority as ~40 _N is set to be higher.

Further, the action selection object 34 gives feedback by notifying the emotion / instinct generation object 32 of the output semantics as the action history after the action is completed. Such behavior history information is information in which behaviors of the pet robot 1 are accumulated in a cumulative manner, and the order of the behavior manifestation (for example, “kick”, “anger”)
And the order of “joy”), the content of the history information is determined.

As a result, the emotion / instinct generation object 3
2 means “joy”, “sadness”,
For a total of six emotions of "anger", "surprise", "disgust" and "fear", a parameter indicating the strength of the emotion is stored for each emotion. And the parameter values of each of these emotions are "hit" by the input semantics.
And a specific recognition result such as “stroke” and the output semantics given from the time selection and the action selection object 34, and the like, and are sequentially updated.

Specifically, in the action selection object 34, the detection result of the battery sensor 12 is
If the toner status of indicating that it is below a predetermined level, regardless of the detection result of the other various sensors, selecting preferentially "battery management behavior model" 40A _N.

When the remaining state of the battery 11 is equal to or higher than a predetermined level and the detection result of the acceleration sensor 18 indicates that the pet robot 1 has fallen, the detection results of the other various sensors are used. Regardless of the above, the “falling return behavior model” 40A _N-1 is preferentially selected.

Further, "Remote control mode behavior model" 40
When the detection result of each sensor corresponding to the behavior model having a higher priority than A _K does not determine the corresponding behavior model, and the detection result of the microphone 20 indicates that the user has given an instruction, preferentially select the "remote control mode behavior model" 40A _K.

Further, the “obstacle avoidance behavior model” 40A ₄
The detection result of each sensor corresponding to the behavior model having a higher priority than that of the pet model does not determine the corresponding behavior model, and the detection result of the CCD camera 21 indicates that there is an obstacle other than the ball in front of the pet robot 1. when representing the preferentially "obstacle avoidance behavior model" 40A ₄ is selected.

Further, among the behavior models 40A _{1 to} 40A _N arranged in accordance with the above-described priorities, the behavior model up to the “obstacle avoidance behavior model” 40A ₄ is not determined, and the detection result of the touch sensor 22 is not determined. If the emotion of the pet robot 1 represents that it has changed more than a predetermined level, selects preferentially "emotional expression behavior model" 40A _3.

[0049] Without further "emotional behavior model" behavior model to 40A ₃ is determined among the behavioral model 40A ₁ ~40A _N arranged according to the priority described above, and a microphone 20, CCD camera 21, touch detection result of the sensor 22 and or the battery sensor 22 when the pet robot 1 indicates that in the boring and or fasted state, selects preferentially "autonomous exploratory behavior model" 40A _2.

[0050] The "Autonomous exploratory behavior model" 40A without action that the model is determined to ₂ of each behavior model 40A ₁ ~40A _N arranged according to the priority described above, and the distance sensor 24 is referred to as a CCD camera 21 the detection result when the pet robot 1 has recognized the existence of the ball (such as color and shape) is preferentially select the "ball corresponding behavior model" 40A _1.

Thus, the action selection object 34
After selecting the higher priority behavior model 40A ₁ ~40A _N, the selected behavior model 40A ₁ ~40A _N
Causes the pet robot 1 to express an action corresponding to
The expression behavior is converted into output semantics and fed back to the emotion / instinct generation object 32.

Thus, the emotion / instinct generation object 32
Based on the output semantics fed back from the action selection object 34, the parameter values of each emotion and each desire according to the emotion and instinct state of the pet robot 1 determined based on the input semantics and the personal information 36 described above. By updating, the next action of the pet robot 1 can be expressed so that the action content represented by the output semantics reflects the action.

More specifically, as a first example, as shown in FIG. 7 in which the same reference numerals are given to the parts corresponding to those in FIG. Each time the touch sensor 22 of the head is pressed, the emotion / instinct generation object 32 sets a parameter such as “pleasant when you see a favorite color” to “a favorite color”. I'm not very happy or hate to see it. ''
By changing the value as described above, the action generating object 33 and the action expressing object 35 initially express the action of "joy", but gradually express the action of "anger".

As a result, even when the pet robot 1 shows the pink ball next time, the pet robot 1 may ignore the pink ball less than the previous time and ignore it. Therefore, depending on how the user treats the pet, the preference of the pet robot 1 can be changed to various types of walking.

As a second example, when the pet robot 1 is stroked while watching (or chasing) the ball, the parameter value of the emotional variation with respect to the ball is set so as to “like the ball more”. By changing the character, the ball can be changed to a very personality.

As a third example, the pet robot 1
If the user tries to make the pet robot express his / her hand using a remote commander (not shown) when his / her desire to search, exercise, and affection falls, the user will act on the hand. Parameter value in the state to be
When the parameter values representing the emotion and the instinct have not been reached, the pet robot 1 may not take any action.

Still, if the user attempts to express the "hand" many times, the pet robot 1 is repeatedly input with the expression command of the "hand" as input semantics. Since the behavior expression such as “sleep” or “yawn” is fed back to the robot 1 as output semantics, as a result, the pet robot 1 decreases the parameter value of the desire for “hand” and performs the action. The motivation is lost, and thus the state can be changed to a state in which “there are no difficulties” that does not suit the user's will.

As a fourth example, when the pet robot 1 finds a pink ball, as shown in FIG. 8 in which the same reference numerals are given to portions corresponding to FIG. 3, the pet robot 1 approaches the ball and tries to kick it. At this time, in the emotion / instinct generation object 32, the “ball (large)”
Is input as input semantics, and then the action expression of "kick" is fed back as output semantics. At that time, if the pet robot 1 succeeds in kicking the ball, the ball moves away, so that the recognition result of “ball (small)” is input as input semantics.

At this time, the emotion / instinct generation object 32
Then, based on the transition of the input semantics from “kick” to “ball (small)”, the pet robot 1 determines that “succeeded in kicking the ball”, and responds to “ball” in accordance with the determination result. By changing the parameter value based on the emotion of "joy" and the state of the instinct of "exercise desire" to be large, the pet robot 1 increases the joy and desire for exercise when looking at the ball, so that the degree of playing the ball is increased. You can change your personality to be higher.

Conversely, if the pet robot 1 fails to kick the ball, the emotion / instinct generation object 3
In 2, the pet robot 1 can be changed to a character in which the pet robot 1 does not play a ball much by changing the parameter value based on the emotional state such as “sadness” or “dislike” so as to increase.

(4) Actions and Effects According to the Present Embodiment In the above configuration, in the pet robot 1, action patterns that can be expressed next from the current state are set as action models, respectively, and the plurality of actions are set. after extracting each behavior model 40A ₁ ~40A _N corresponding to the urging from the surrounding and internal situations and user from the model, these highest priority behavior model 40A ₁ ~40A
_N to select the selected behavior model 40A ₁ -40
By expressing an action corresponding to A _N, it may be performed continuously optimal next action in accordance with the current situation.

Then, after the pet robot 1 actually expresses an action according to the action model, the expressed action result (ie, action history) is fed back, so that the pet robot 1 responds to the emotion and instinct state of the pet robot 1. By changing the degree of emotions and desires according to the result of the action, the result of the current action can be reflected in the next action, thus associating the transition from the current action to the next action. As a result, the pet robot 1 with high environmental adaptability can be realized.

[0063] The previously set a plurality of action models 40A ₁ ~40A _N behavioral model library 40, by which from the each behavior model so as to expand the possible transitions action respectively, the concept of action model library 40 The number of transitionable actions can be significantly increased as compared with the case where a transitionable action is developed from a single behavior model without providing the behavior pet robot 1, thus allowing the behavior pet robot 1 to exhibit a more complicated behavior. be able to.

Further, a plurality of behavior models 40A _{1 to} 40A
By providing _N in the behavior model library 40,
Compared to the case of a single behavior model, the settings such as the transition probability and the corresponding action can be adjusted for each behavior model, and the pet robot 1 can behave as if it were a real animal. Can be set relatively easily (maintenance, version upgrade, etc.).

According to the above configuration, in the pet robot 1, the action having the next highest priority is expressed in a state where the action selected according to the surrounding and internal conditions and the action from the user is expressed. By doing,
The optimal next action according to the current situation can be continuously performed, and thus the pet robot 1 that can significantly improve the amusement property can be realized.

Then, after the pet robot 1 actually expresses the action selected as described above, the degree of emotion and desire according to the emotion and instinct state of the pet robot 1 is determined based on the expressed action result. By making the change, the current action result can be reflected in the next action to enhance the environmental adaptability, and thus the pet robot 1 that can significantly improve the amusement property can be realized.

(5) Other Embodiments In the above-described embodiment, a case has been described in which the invention is applied to the four-legged walking type pet robot 1 configured as shown in FIG. The present invention is not limited to this, and can be widely applied to robots having various other configurations.

In this case, in the above-described embodiment,
The controller 10, the actuators 3AA ₁ to 5 A _M
And a case in which the action is generated based on the output of the action model by applying the speaker 23 and the like, but other various configurations may be applied depending on the form of the robot apparatus to which the present invention is applied. Can be.

In this case, in the above-described embodiment, a plurality of action models 40A _{1 to} 40A _N provided in the action model library 40 as shown in FIG. Has been described, but the present invention is not limited to this, and various other actions may be provided in the action model library as action models.

[0070] As shown in FIG. 9 that the same reference numerals are added to corresponding parts in FIG. 4, for example, a behavioral model library 50 is provided on the behavior generation object 33, of each behavior model 50A _1X ~50A _N categorized by purpose , "Ball-response behavior model 1" 50A _1X and "Ball-response behavior model 2" that match the "ball-response behavior model" with the growth stage
In addition to the two types of 50A _1Y , the “autonomous search action model 1” 50A _2X and the “autonomous search action model 2” 50A _2Y that match the “autonomous search action model” to the respective growth stages.
May be classified into two types.

In this case, in FIG. 3 described above, the behavior generation object 33 of the controller 10 determines the growth stage (for example, “birth period”, “childhood period”) of the pet robot 1 based on the output semantics fed back from the behavior expression object 35. Period), "childhood", "adolescence" and "adult". As a result, the behavior result of the behavior model selected in accordance with the growth stage is fed back to the emotion / instinct generation object 32 as the growth history, so that not only the behavior history up to now but also the growth history is reflected in the next behavior. As a result, the environmental adaptability can be improved, and thus the pet robot 1 that can significantly improve the amusement property can be realized.

[0072] Furthermore, in the above-described embodiment, among the outputs of the behavioral model 40A ₁ ~40A _N of a plurality of types provided in behavioral model library 40, input Iman Genetics externally applied (input information) with its own Although the case where the action selection object 34 of the controller 10 is applied as the action selection means for selecting one action model based on at least one of the action and / or the growth history has been described, the present invention is not limited to this. As the selection means, various other configurations can be applied according to the form of the robot device to which the present invention is applied.

Further, in the above-described embodiment, the degree of emotion corresponding to the emotion model and / or the instinct model is determined based on the behavior model which is fed back after expressing the behavior according to the selected behavior model. As described above, the emotion / instinct generation object 32 of the controller 10 is applied as the emotion desire changing means for changing the degree of desire. However, the present invention is not limited to this. If it can be reflected in the above-mentioned action, various other configurations can be applied as the emotion desire changing means.

Further, in the above-described embodiment, FIG.
Each behavior model 4 has a higher priority as described in the lower side as shown in the behavior model library 40 of FIG.
It has dealt with the case of setting the priority of 0A ₁ ~40A _N, the present invention is not limited to this, in any other order may be set priorities.

Further, in the above embodiment, the emotions according to the emotion model include “joy”, “sadness”,
In addition to providing six emotions of “anger”, “surprise”, “disgust” and “fear”, as desires according to the instinct model,
4 for "Motivation", "Love", "Appetite" and "Curiosity"
Although the case where one desire is provided has been described, the present invention is not limited to this, and various other numbers and types can be widely applied as the number and types of emotions and desires.

[0076]

As described above, according to the present invention, there are a plurality of types of behavior models, and the output of each behavior model is based on at least one of the input information from the outside and the own behavior and / or growth history. By providing the action selecting means for selecting the output of one action model from among the above, it is possible to continuously perform the optimal next action according to the current situation, thus greatly improving the amusement property. A robot device that can be improved can be realized.

According to the present invention, a plurality of types of behavior models are provided, and one of the outputs of each behavior model is selected based on at least one of input information from the outside and own behavior and / or growth history. The behavior model output is now selected. As a result, according to the control method of this robot device,
An optimal next action according to the current situation can be continuously performed, and thus a control method of a robot apparatus that can significantly improve amusement can be realized.

[Brief description of the drawings]

FIG. 1 is a perspective view illustrating an external configuration of a pet robot according to an embodiment.

FIG. 2 is a block diagram illustrating a circuit configuration of the pet robot.

FIG. 3 is a conceptual diagram showing a software configuration in a controller.

FIG. 4 is a conceptual diagram for explaining an action model library;

FIG. 5 is a schematic diagram illustrating a stochastic automaton.

FIG. 6 is a diagram showing a state transition table.

FIG. 7 is a schematic diagram for explaining a change in the degree of emotion and desire due to feedback of an action history.

FIG. 8 is a schematic diagram for explaining a change in the degree of emotion and desire due to the feedback of the action history.

FIG. 9 is a conceptual diagram explaining a behavior model library according to another embodiment.

[Explanation of symbols]

1 ... pet robot, 2 ... body unit, 3A ~
3D ... leg unit, 4 ... head unit, 5 ... tail unit, 10 ... controller, 30 ... recognition object, 31 ... meaning conversion object, 32 ...
Emotion / instinct generation object, 33: action generation object, 34: action selection object, 35: action expression object, 36: personal information, 37: state transition table, 40, 50: action model library, 40A ₁
４０40A _N , 50A _1X ５０50A _N ... Behavior model.

Continued on the front page F-term (reference)

Claims (8)

[Claims] The present invention has a plurality of types of behavior models, and based on at least one of externally input information and a self behavior history and / or a growth history, one of the behavior models is selected from the outputs of the behavior models. A robot apparatus comprising an action selecting means for selecting an output of a robot. 2. An emotion desire changing means for changing a degree of emotion according to an emotion model and / or a degree of desire according to an instinct model based on an output of the behavior model selected by the behavior selection means. The robot device according to claim 1, wherein: 3. A priority order is set in advance for each of said behavior models, and said behavior selecting means includes: an output unit for each of said behavior models based on at least one of said input information and said behavior and / or growth history. After extracting at least one or more outputs of the behavior model from
2. The robot apparatus according to claim 1, wherein an output of the behavior model having the highest priority is selected from the extracted outputs of the behavior models. 4. An emotion desire changing means for changing the degree of emotion according to the emotion model and / or the degree of desire according to the instinct model based on the output of the behavior model selected by the behavior selection means. The robot device according to claim 3, wherein: 5. A behavior model having a plurality of types of behavior models, wherein one of the behavior models is selected from outputs of the behavior models based on at least one of input information from the outside and the behavior history and / or growth history of the user. A method for controlling a robot device, comprising selecting an output of a robot. 6. The method according to claim 5, wherein the degree of emotion according to the emotion model and the degree of desire according to the instinct model are changed based on the output of the selected behavior model. A method for controlling a robot device. 7. A plurality of types of behavior models are set in advance with priorities, and based on at least one of the input information and the behavior and / or growth history, at least one or more of outputs of each of the behavior models. After extracting the output of the above behavior model,
The method according to claim 5, wherein an output of the behavior model having the highest priority is selected from the outputs of the extracted behavior models. 8. The robot according to claim 7, wherein the degree of emotion according to the emotion model and the degree of desire according to the instinct model are changed based on the output of the selected behavior model. How to control the device.