JP2001179666A - Pet type robot and recording medium readable by computer with program recorded - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a curious pet type robot and recording medium readable by computer with program recorded so as to realize its movement. SOLUTION: This robot comprises an input means 1 adapted to detect data relating to an object and an output means 3 adapted to move according to the data obtained by the input means 1. The robot, has CPU 21 that calculates degree of interest on the basis of the data obtained by the input means 1 for deciding to make it an interest object or not by the result compared of the degree of interest calculated newly with that calculated previously to decide movement from curiosity about the object.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a curious pet-type robot and a storage medium.

[0002]

2. Description of the Related Art An electronic pet type robot that performs various actions based on internal state parameters including data on emotions has been proposed (Japanese Patent Laid-Open No. 11-12601).
No. 7).

[0003]

However, conventional pet-type robots have not been able to take curiosity behaviors as do actual pets and infants.

The present invention has been made based on such a technical background. That is, an object of the present invention is to provide a curious pet-type robot and a computer-readable recording medium on which a program for realizing such an operation is recorded.

[0005]

Means for Solving the Problems In order to achieve the above object, a first aspect of the present invention provides an input unit for detecting data related to an object, and an output unit for performing an action based on data obtained from the input unit. In a pet-type robot comprising: an interest degree calculating means for calculating an interest degree for an object based on data obtained from the input means; an interest degree newly calculated by the interest degree calculating means; It is characterized by having target setting means for determining whether or not to be an object of interest based on the result of comparison with the degree of interest, and action determining means for determining curiosity behavior with respect to the target set by the target setting means.

According to a second aspect, in the pet type robot according to the first aspect, a reflex action inducing means for determining whether data obtained from the input means is a reflex action factor, and the reflex action inducing means And a reflex action executing means for executing the selected reflex action by selecting the reflex action based on the judgment of (1).

According to a third aspect, in the pet-type robot according to the second aspect, the output means includes a priority action means for controlling the execution of the reflex action prior to the interest action. .

A fourth invention is the pet-type robot according to the first invention, the pet-type robot according to the second invention, or the pet-type robot according to the third invention, wherein the interest degree calculating means becomes more interested in time. The calculation is performed so as to reduce the degree.

According to a fifth aspect, in the pet-type robot according to the first to fourth aspects, the action determining means performs a curious action based on the degree of interest and the distance to the object of interest. It is characterized in that it is determined.

A sixth invention is a computer-readable recording medium in which an input function for detecting data related to an object and an action function for acting based on data obtained by the input function are recorded, Interest level calculation function that calculates the degree of interest for the target based on the data obtained from the input function, and the interest level calculated by the interest level calculation function and the interest level calculated before that. A computer-readable recording medium in which a program for causing a computer to perform a target setting function for determining a target setting function and an action determining function for determining curiosity behavior with respect to a target set by the target setting function is stored.

[0011]

Hereinafter, a pet-type robot according to the present embodiment will be described.

The pet-type robot according to the present embodiment shows an interest in a moving object, takes a curiosity action such as approaching the object, touching with a limb, hitting its own body, or suddenly making a sound. It is a robot that can take reflex actions that scare you when it sounds.

FIG. 1 is a block diagram illustrating an example of a hardware configuration of a pet robot according to the present embodiment.

The pet-type robot has an input unit 1 for detecting data related to the target, and a recognition / judgment unit 2 for recognizing the target based on the data from the input unit 1 and determining an action.
And output means 3 for executing the action determined by the recognition / judgment unit 2.

The input means 1 detects data related to the object as described above, and the distance measuring sensor 11
, Camera 12, microphone 13, touch sensor 14
And A / D converters 15 and 16.

The distance measuring sensor 11 is a distance sensor that measures a distance from an object, and includes, for example, a light emitting unit that emits light from an infrared light emitting diode like a small high-performance distance measuring sensor (GP2D02: manufactured by Sharp), An LED drive circuit that controls the driving of the light emitting unit, a lens unit, a lens unit that emits light from a light receiving unit having a different output voltage from each light receiving element, and collects light reflected from an object, and a lens unit. A light receiving unit for receiving the reflected light, and a signal processing unit for processing output signals from the light receiving unit to output 8-bit serial data corresponding to the distance to the target. Such ranging data is used to determine which curiosity behavior to perform.

The camera 12 is, for example, red (hereinafter, red)
This is input means for reading out three primary color signals of frame (hereinafter simply referred to as R), green (hereinafter simply referred to as G), and blue (hereinafter simply referred to as B). A signal from the camera 12 is used for optical flow or tracking of a specific color area in order to recognize the position of a moving object in a frame (corresponding to a visual field).

The microphone 13 is an input unit for obtaining an audio signal, and although only one is shown in the figure, actually two microphones are provided, and these microphones collect sound from a sound source. These audio signals are used, for example, to specify the direction of a sound source.

The touch switch 14 is, for example, a pressure-sensitive sensor, and is buried in the body or limbs of the pet-type robot. The signal from the touch switch 14 is used to detect a portion that has been hit or touched. Thereby, for example, reflex behavior is induced.

The recognition / judgment unit 2 performs recognition processing and action determination processing based on information from the input unit 1 as described above. The recognition processing includes visual processing including optical flow and tracking, auditory processing for recognizing the direction of a sound source, and tactile processing for specifying a contacted part. The action determination process gives priority to the reflex action, takes interest in what moves when the reflex action is not performed, and takes various curiosity actions.

The recognition / judgment unit 2 stores a peripheral control unit that performs interrupt control with the CPU 21, a main memory 23 that is a random access memory, and programs such as an operating system that manages a recognition system and an action determination program. ROM 22 and interfaces (hereinafter abbreviated as I / F) 24 and 25. The CPU 21 controls the entire pet-type robot by executing an operating system stored in the ROM 22. Here, the main memory means a memory on which a program can be executed.

The I / F 24 is an interface for inputting / outputting data by serial communication with the input unit 1. The I / F 25 is an interface for communicating data with the output unit 3 by serial communication.

The output unit 3 performs the action determined by the recognition / judgment unit 2. For example, the output unit 3 generates a cry such as a surprised voice, a D / A converter 33, and a speaker 31.
It comprises a drive mechanism 32 for driving each joint for generating various operations, and a motor 34 for driving the same. Since the output means 3 is provided with the above-described configuration, the action of shrinking the body as a reflex action, the action of pulling away from a reflection factor such as a sound or an object, or directing a gaze in the direction of a reflection factor such as a sound or an object ,
It can generate a surprised voice, and can perform actions such as chasing, approaching, and touching the subject as a curious action.

The hardware configuration of the pet robot according to the present embodiment has been described above.

Next, the outline of the process of recognizing and determining the behavior of the pet-type robot in the present embodiment will be described.

FIG. 2 is a flowchart showing the execution of the curiosity behavior from the setting of the degree of interest I. The program shown in the flowchart shown in FIG. 2 is stored in the ROM 22, and is executed by the CPU 21.

When the program shown in FIG. 2 is started, the CPU 21 determines whether or not a reset factor of interest has occurred (step 1). Here, the reset factor is a factor that triggers a change from the current target to a new target, and includes, for example, a pressure-sensitive signal from the touch switch 14, an audio signal from the microphone 13, and image data from the camera 12. . The pressure-sensitive signal is a signal for detecting a hit or touched state. The audio signal is used to specify the direction of the sound source (target).

If the CPU 21 determines in step 1 that a pressure-sensitive signal or an audio signal, which is a reset factor, has been generated, the motor 34 and the drive mechanism 32 move a joint such as a neck to change the line of sight (step 2). As a result, the content of the frame captured from the camera 12 is changed.

The CPU 21 starts an alerting routine (step 3). This alerting routine is a routine for setting a new object of interest based on a signal from the camera 12. Specifically, the CPU 21 detects a movement vector in a frame captured from the camera 12, and specifies a lattice-like image corresponding to an optical flow (hereinafter, also referred to as a movement vector) indicating a maximum value. This is a candidate of interest. Next, the color data having the maximum area in the candidate of interest is specified, and the color data is stored. By repeating this process for a certain period of time, the color data having the highest frequency in the candidate of interest is set as the specific color.

The CPU 21 calculates the value of the degree of interest I by substituting the movement vector of the candidate of interest into a predetermined equation (step 4). Therefore, the CPU 21 implements the interest calculation function.

The CPU 21 determines whether or not the degree of interest I for the candidate of interest obtained in step 4> the degree of interest I for the current object of interest (step 5). This is because the pet-type robot of the present embodiment takes a curiosity behavior with respect to a target having a high degree of interest I.

If the CPU 21 determines in step 5 that the degree of interest I for the candidate of interest> the degree of interest I for the current target of interest, the CPU 21 resets the candidate to the candidate of interest (step 6). That is, the CPU 21 sets the interest target candidate identified in step 4 as an interest target. Thus, the CPU 21 implements the target setting function.

If the CPU 21 determines in step 5 that the interest level I for the interest target candidate <the interest level I for the current interest target, the CPU 21 does not reset the interest target (step 7).

Steps 1 to 7 described above correspond to the reset process of the object of interest. In addition, the pet-type robot of the present embodiment changes the factor that determines the degree of interest I from “the magnitude of the movement” to “irregularity of the movement”, “color”, etc. Can be changed.

The CPU 21 determines in step 1 that no resetting factor of interest is generated,
After finishing the reset process of the object of interest consisting of step 7, the following routine is started (step 8).

Here, the follow-up routine acquires position data corresponding to an object of interest in the frame by tracking processing using color data corresponding to the specific color set in the alerting routine, and acquires interest based on the position data. Each joint is moved by the motor 34 and the drive mechanism 32 so as to capture the target at the center of the frame, and in parallel with this processing, distance measurement data from the distance measurement sensor 11 to the object of interest is acquired.

The CPU 21 determines a curiosity behavior from the distance measurement data acquired in step 8 and the value of the degree of interest I (step 9).

The CPU 21 performs a curious action by driving the motor 34 and the driving mechanism 32 so as to correspond to the action determined in step 9 (step 1).
0).

Steps 8 to 10 described above include:
This is tracking processing of an object of interest by visual processing, determination of curiosity behavior, and execution processing of curiosity behavior. The pet robot according to the present embodiment can take various curiosity behaviors corresponding to the degree of interest I and the distance D to the object of interest.
As described above, the CPU 21 implements the action determination function.

FIG. 3 is a flowchart showing a reflex action routine.

The reflex action routine is an interrupt routine for the routine shown in FIG. The CPU 21
By detecting the reflex action factor, the processing of the routine shown in FIG. 2 is interrupted and the reflex action routine is executed with priority. Therefore, the CPU 21 corresponds to priority action means.

The CPU 21 detects that an object having a speed higher than a predetermined value approaches from the distance measurement data from the distance measurement sensor 11, detects a sound exceeding a predetermined value from an audio signal from the microphone 13, If it is detected from the pressure-sensitive signal from the switch 14 that the object has been hit more than a predetermined value, or if it is detected that the object has moved closer than a predetermined distance from the distance measuring sensor 11 and the camera 13, this routine is executed. to start. Thus, the CPU 21 realizes the reflex action induction function.

The CPU 21 determines a reflex behavior corresponding to the reflex behavior factor detected as described above (step 21).

The CPU 21 executes the reflex action determined in step 21 (step 22). For example, a pet-type robot shrinks, pulls away from a sound source or object that is a reflex action factor, generates a surprised voice, or turns its gaze on the direction of a sound source or object that is a reflex factor. . As described above, the CPU 21 corresponds to a reflex action execution unit.

The reflex action routine shown in FIG.
The pet-type robot according to the present embodiment executes the reflex action first and executes the reflex action first when the reflex action factor is not detected. I was able to take curiosity.

Next, an optical flow constituting a part of the visual recognition processing according to the present embodiment will be described. This processing is a technique used in the alerting routine described with reference to FIG.

The optical flow is a technique for detecting the movement of each area on a frame read from the camera 12, and is a technique for detecting a movement vector (composed of a speed and a direction) of a moving object on the frame. . This processing is for obtaining, for example, a movement vector corresponding to the direction and speed corresponding to the motion of the arm of the person.

The optical flow is a process of dividing a frame into a grid as shown in FIG. 4 and tracking whether each grid has moved from the previous frame to the next frame, and a grid-like image on the previous frame and the next frame. Calculating a movement vector from the upper grid image.

First, the tracking processing is described with reference to FIGS.
This will be described with reference to FIG. FIG. 4 is a schematic diagram in which an image to be searched is extracted and a previous frame and a next frame are displayed in a superimposed manner. FIG. 5 is a conceptual diagram showing tracking.

In FIG. 4, an image A ₁ is a lattice-like image stored in the previous frame, and an image A ₂ is an image in a state where the image A ₁ has moved in the next frame. Tracking, if the lattice-like image A ₁ stored in the previous frame,
It is a process for searching an image A ₂ resembling most the image A ₁ in the next frame.

An evaluation function G (x, y) shown in Expression (1) is used to evaluate whether or not images are similar.

(Equation 1)

[0052] the evaluation function G is _{A 1} in the search area B to _{A 1}
This is to calculate the sum of squares of the error of the luminance value of each dot for the grid-like area of the same size as that of. Image A ₁ stored within the search area B as shown in FIG. 5 moves one dot at a time in the x or y direction, obtaining the evaluation function G with the image stored as the image of the next frame. An image with the smallest evaluation function G is searched for. This is a specific tracking process. Such processing is performed for all frames or a specific area of the frame.

If there is a grid-like image that completely matches A ₁ in the search area B, the evaluation function G becomes zero. However, for example, taking the human palm example, and gradually open the gripping state, assuming that gradually moves, will not completely exist same image A ₁ in the next frame . Therefore, x and y at which the evaluation function G is minimized in the search area B are set as the movement positions in the next frame. here,
(X, y) is the centroid of the coordinates of the image _{A 1.} A movement vector is calculated from the coordinate data of the center of gravity. Track the position of the image A ₁ to be searched from such motion vectors. Such a process is an optical flow process. A dedicated chip for calculating such an optical flow in real time over the previous frame may be used.

In the present embodiment, the position of a moving object (search image) is recognized using the optical flow in the visual recognition process by alerting, and at the same time, the process of extracting the color of the search image is also performed. This enables tracking using a specific color to recognize the position of a stationary object (search image) as a visual recognition process that induces curious behavior because optical flow can always search only a moving object. That's why.

Next, a tracking process using a specific color will be described. This process is a technique employed in the following routine described with reference to FIG. The tracking process using a specific color is a process of cutting out a specific color region on a frame and tracking the center of gravity of the region when the color of an object to be tracked is specified. If there is a similar color area in the frame, it will be confused. Therefore, the moving direction is predicted from the tracking result up to the previous frame, and the correct color area is selected.

The entire frame is scanned to determine whether there is a specific color area to be tracked, and binarized. In particular,
If a specific color exists, “1” is set. If no specific color exists, “0” is set.

This determination processing is performed according to the following equation (2).

(Equation 2)

Next, the area determined to include the specific color will be described.
Area center of gravity coordinates (X_G, Y _G) Is obtained by the following equation (3).
You.

(Equation 3)

The above is the tracking processing using the specific color used in the present embodiment.

Next, the degree of interest I used to determine curiosity behavior will be described with reference to FIG. FIG. 6 is a graph showing an attenuation function used for calculating interest.

The degree of interest I indicates the degree of interest. In the present embodiment, it is calculated by the following equation (4). Interest level I = _Fmax × D (t) Equation (4)

Here, F _max is the maximum value of the movement vector calculated at the time of executing the alert routine, and D max
(T) indicates that immediately after the start of tracking as shown in FIG.
This is an attenuation function that is set to 0. As can be seen from the above equation, the degree of interest I increases as the movement of the object increases or immediately after the reset factor occurs. No matter how big the movement, the subject gets tired over time.

In the present embodiment, the degree of interest I is calculated from the magnitude of the movement of the target, but is not limited to this. May be calculated. Furthermore, these may be calculated in combination.

FIG. 7 is a schematic diagram showing an algorithm for determining curiosity behavior.

FIG. 7 shows four types of actions based on the distance D to the target and the degree of interest I, “turn gaze, but immediately shift gaze to another,” “turn gaze at face,” and “approach. It indicates that the action is selected from "go" and "stretch and touch". In addition, the pet-type robot according to the present embodiment changes its behavior from "reaching and touching." Further, the behavior of the pet-type robot according to the present embodiment changes from "approaching" to "gazing at the face" with the passage of time.
This is because the degree of interest I attenuates with time, as can be seen from equation (2).

The state in which the pet-type robot according to the present embodiment turns his / her gaze, but immediately takes an action to shift the gaze to another
The target is far away, and the movement of the target is small. There is no change from this state to a new curiosity behavior. This is because the degree of interest I attenuates with time, and the pet-type robot does not take an action of reducing the distance to the target in the schematic diagram shown in FIG.

The case where the computer program having the above-described contents is recorded on, for example, a recording medium and distributed is also included in the technical scope of the present invention. The recording medium is a magnetic tape, flexible disk, CD-ROM or DVD
, And a magneto-optical disk such as an MO.

As described above, the pet robot according to the present embodiment can be realized as a curious robot.

[0069]

The present invention was able to provide a pet-type robot with an inquisitive curiosity and a computer-readable recording medium on which a program for realizing such an operation is recorded.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a hardware configuration example of a pet robot according to the present embodiment.

FIG. 2 is a flowchart showing execution of curiosity behavior from setting of degree of interest I;

FIG. 3 is a flowchart showing a reflex action routine.

FIG. 4 is a schematic diagram in which an image to be searched is extracted and a previous frame and a next frame are superimposed and displayed.

FIG. 5 is a conceptual diagram illustrating tracking.

FIG. 6 is a graph showing a damping function used for calculating interest.

FIG. 7 is a schematic diagram showing a curious behavior determination algorithm.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 input means 2 recognition / judgment unit 3 output means 11 distance measuring sensor 12 camera 13 microphone 14 touch switch 21 CPU 22 ROM 23 RAM 31 speaker 32 drive mechanism 34 motor

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G06F 17/00 G06F 15/20 Z F-term (Reference) 2C150 BA06 CA02 DA05 DA24 DA26 DA27 DA28 DF03 DF04 DF06 DF33 ED42 ED52 EF16 EF23 EF29 EF33 3F059 AA00 BA00 BB06 DA01 DB05 DB09 DC00 DC01 DC08 DD11 FC00 3F060 AA00 BA10 5B049 AA02 BB07 DD00 DD03 EE00 EE07 EE12

Claims (6)

[Claims] 1. A pet-type robot comprising: input means for detecting data related to an object; and output means for performing an action based on data obtained from the input means, based on data obtained from the input means. Means for calculating an interest level for an object, and an object setting for determining whether or not the interest level newly calculated by the interest level calculation means and the interest level calculated before the interest level are set as an interest level. And a behavior determining means for determining curiosity behavior with respect to the target set by the target setting means. 2. A reflex action inducing means for judging whether data obtained from the input means is a reflex action factor,
2. The pet-type robot according to claim 1, further comprising: a reflex action executing means for selecting a reflex action based on the judgment of the reflex action inducing means, and executing the selected reflex action. 3. The pet-type robot according to claim 2, wherein the output means includes priority action means for controlling the reflex action prior to the interest action. 4. The pet-type robot according to claim 1, wherein said interest degree calculating means calculates the degree of interest so as to decrease with time. 5. The pet-type robot according to claim 1, wherein the action determining means determines a curiosity action based on the degree of interest and a distance to the object of interest. 6. A computer-readable recording medium in which an input function for detecting data related to an object and an action function for acting based on data obtained by the input function are recorded, the computer-readable recording medium being provided with the input function. Interest level calculation function that calculates the degree of interest for the target based on the data obtained, and the interest level newly calculated by the interest level calculation function and the interest level calculated before that are determined as the interest level. A computer-readable recording medium that stores a program for causing a computer to execute a target setting function to perform and a behavior determination function that determines curiosity behavior with respect to the target set by the target setting function.