WO2009107185A1

WO2009107185A1 - On-vehicle robot

Info

Publication number: WO2009107185A1
Application number: PCT/JP2008/053147
Authority: WO
Inventors: 裕昭柴▲崎▼
Original assignee: パイオニア株式会社
Priority date: 2008-02-25
Filing date: 2008-02-25
Publication date: 2009-09-03
Also published as: JPWO2009107185A1

Abstract

An on-vehicle response robot (1) is provided with an adaptation index calculating part (11) and an operation control part (13) which decide an operation to be reacted in accordance with an adaptation degree with an operator, an operation mechanism (14) performing an operation corresponding to decision by the adaptation index calculating part (11) and the operation control part (13), and an evaluating part (12) correcting the adaptation degree with the operator based on reaction of the operator who is brought into contact with the operation that the operation mechanism (14) performs and changing the operation that the operation mechanism (14)is to perform.

Description

In-vehicle robot

The present invention relates to a vehicle-mounted robot that reacts to a driver.

In recent years, not only industrial robots but also various humanoid robots that perform operations close to humans have been developed. Conventionally, as such a humanoid robot, there is a robot interaction device that recognizes a speech content of an operator and analyzes the speech content and responds to the operator with a corresponding content ( Patent Document 1). If this robot interaction device is mounted on a vehicle, this robot interaction device can perform simple conversations with the driver or the like according to the utterance content of the driver or the like.
JP 2004-291176 A (paragraph number 0025, FIG. 1)

In the above-mentioned conventional technology, this robot interaction device simply reacts and exchanges simple conversations regardless of who the driver is, and what is the trigger to make the driver to drive safely by such simple conversation alone? I couldn't.

The problems to be solved by the present invention include the above-mentioned problems as an example.

In order to solve the above-mentioned problem, the invention according to claim 1 performs an operation according to the determination by the operation determination unit, an operation determination unit that determines an operation to be reacted according to the degree of familiarity with the driver. An operation mechanism; and a control unit that corrects a degree of familiarity with the driver based on a reaction of the driver in contact with an operation performed by the operation mechanism, and changes an operation to be performed by the operation mechanism.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a perspective view illustrating a configuration example of a part of a vehicle 100 according to the present embodiment.
The vehicle 100 includes left and

right seats

6 a and 6 b, a side brake 7, a door 5, a handle 4, a meter 7, a device panel 21, and a dashboard 9 in the vehicle.

The equipment panel 21 has a display unit 15. The display unit 15 is, for example, a display device for a car navigation device or a television receiver. The door 5 is an opening and closing device for opening and closing when a driver or a passenger gets on or off the vehicle. The left and

right seats

6a and 6b are seats for the driver and passengers to sit on. The handle 4 is an operation member that is operated by the driver for driving. The side brake 7 is a kind of braking device for fixing the vehicle to a place so that the vehicle does not move any more when the vehicle is stopped. The meter 7 is a display unit that displays the speed of the vehicle 100, the engine speed, the water temperature, and the like. The dashboard 9 represents a member in front of the passenger seat side among the left and

right seats

6a and 6b.

On the dashboard 9, an in-vehicle robot 1 corresponding to a robot is disposed particularly at a position where the driver enters the field of view while driving. In the present embodiment, the dashboard 9 is arranged in the vicinity of the center of the driver's seat and the

passenger seat

6a, 6b. The in-vehicle robot 1 may be placed on the dashboard 9 so as to be detachable or may be fixed. This in-vehicle robot 1 has a photographing function, a recognition function, an individual reaction function, and the like, which will be described later, determines a familiar person, identifies a familiar driver and a passenger who frequently rides, and gives greetings and reactions respectively. Has the function to perform separately. In the following description, this vehicle-mounted robot 1 is exemplified as a reaction to the driver as an example.

FIGS. 2 and 3 are enlarged perspective views showing an example of an appearance in which the vehicle-mounted robot 1 shown in FIG. 1 is operating.
As shown in FIG. 2, the in-vehicle robot 1 includes a head cover 1C, a head 1H, a main body 1B, a left arm 1L, a right arm 1R, and a leg 1FT. The head cover 1C is a transparent member provided on the upper part of the main body 1B and covers the head 1H. The left arm portion 1L and the right arm portion 1R can be moved relative to the main body 1B.

The in-vehicle robot 1 has a head 1H that moves relative to the main body 1B, and a camera 5 is provided at a position corresponding to the human eye in the head 1H. In the head 1H, the position of the camera 5 corresponding to the human eye has an eye shape, and the mouth has a shape corresponding to the human mouth. Therefore, the head 1H represents the expression 1F as if it were the shape of the eyes and mouth. These eyes and mouth can be deformed by an operating mechanism described later.

In FIG. 2, the in-vehicle robot 1 shows that the facial expression 1F creates a mild atmosphere and is intimately contacting the driver. On the other hand, in FIG. 3, the vehicle-mounted robot 1 expresses a state in which the expression 1F creates an unpleasant atmosphere and touches the driver in an unfriendly atmosphere. The facial expression 1F of the in-vehicle robot 1 is not limited to these, and for example, it may be a scary and uneasy atmosphere.

FIG. 4 is a block diagram showing an example of functions of the in-vehicle robot 1 shown in FIGS.
The in-vehicle robot 1 includes a camera 5, a subject recognition unit 6, a familiarity index calculation unit 11, a memory 19, a factor acquisition unit 18, an operation control unit 13, an evaluation unit 12, and an operation mechanism 14.

Of these, the familiarity index calculation unit 11 and the operation control unit 13 correspond to an operation determination unit, and determine an operation to react according to the familiarity with the driver. Further, the familiarity index calculation unit 11 and the operation control unit 13 may determine an operation to be reacted according to the ride history of the driver in addition to the familiarity with the driver.

Specifically, the familiarity index calculation unit 11 corresponds to a familiarity index calculation means, and calculates a familiarity index in which the degree of familiarity with the driver is quantified. The operation control unit 13 corresponds to an operation control unit, and controls the operation mechanism 14 by determining an operation to be reacted according to the familiarity index.

The operation mechanism 14 corresponds to an operation mechanism, and performs an operation according to the determination by the familiarity index calculation unit 11 and the operation control unit 13 as operation determining means. This operation mechanism 14 is a generic term for actuators for moving the above-described head 1H, right arm 1R, and left arm 1L relative to the main body 1B.

The evaluation unit 12 corresponds to control means, and corrects the degree of familiarity with the driver based on the reaction of the driver in contact with the operation performed by the operation mechanism 14, and changes the operation to be performed by the operation mechanism 14. Hereinafter, the function of the vehicle-mounted robot 1 will be specifically described.

The camera 5 is a photographing device for photographing a landscape when a driver or a passenger enters the field of view from the head 1H of the in-vehicle robot 1 disposed in the vehicle 100. This landscape includes at least one of a driver and a passenger. The camera 5 starts to operate when a driver (or one of the passengers) gets into the vehicle 100. The camera 5 may start operating when the power is turned on to start the engine.

The target person recognition unit 6 corresponds to a recognition means and recognizes each of the plurality of drivers. The target person recognition unit 6 has a function of recognizing at least one of a driver and a passenger as a target person by at least one of voice recognition and face recognition, for example. The target person recognition unit 6 includes, for example, an analog / digital conversion unit, converts an analog signal photographed by the camera 5 into a digital signal, and captures the image of the driver photographed based on the digital image data. Identify the face.

The memory 19 is a storage medium capable of storing information, and stores a factor for calculating a familiarity index, which will be described later, for expressing the degree of familiarity between the in-vehicle robot 1 and the driver as a numerical value. Specifically, the memory 19 has a quantitative factor 17 and a qualitative factor 16.

Among these, one example of the quantitative factor 17 is boarding history information representing past boarding history. The quantitative factor 17 will be described in detail later. For example, the cumulative number of times the vehicle 100 has been boarded, the cumulative boarding time of the vehicle 100, and driving determination results regarding safe driving or rough driving can be given. . The quantitative factor 17 includes the detection result of the driving state by the G sensor as an example of the factor acquisition unit 18 in view of the “method of calculating the familiarity index by including the operational determination as familiar factor” described later. Also good. The evaluation by the evaluation unit 12 may be one item of the quantitative factor 17. The part may be updated independently of other items, and among some items of the quantitative factor 17, there are some items for which the evaluation of the evaluation unit 12 is fed back. Yes, and only the contents may be updated.

On the other hand, as an example of the qualitative factor 16, the details will be described later. For example, the emotion determination result for the robot based on the tone of the opponent's call voice, the determination result based on the meaning content of the call voice of the opponent recognized by the voice, Judgment result based on whether or not the person who has called the person including the name of 1 and the result of judging whether the person's gesture, reaction, facial expression, etc. were gentle rather than the tone of the voice, etc. Can be mentioned. The evaluation by the evaluation unit 12 may be one item of the qualitative factor 16. Then, the part may be updated independently of other items, and among some items of the qualitative factor 16, there are some items to which the evaluation of the evaluation unit 12 is fed back. Yes, and only the contents may be updated.

The familiarity index calculation unit 11 calculates a familiarity index, which will be described later, based on at least one of the quantitative factor 17 and the qualitative factor 16 read from the memory 19 in correspondence with the driver recognized by the subject recognition unit 6. . This familiarity index is a numerical index that represents the degree of familiarity between the vehicle-mounted robot 1 showing various facial expressions and the driver.

The motion control unit 13 has a function of determining an operation to be performed by the in-vehicle robot 1 based on the familiarity index calculated by the familiarity index calculation unit 11. The operation control unit 13 controls the operation mechanism 14 based on the determination items to perform an operation to be reacted. As described above, the operation mechanism 14 operates to move any one of the head 1H, the right arm 1R, and the left arm 1L of the in-vehicle robot 1.

The driver who is in contact with any one of the movements of the head 1H, the right arm 1R, and the left arm 1L, for example, from the content of the movement, for example, the vehicle-mounted robot 1 is unique and is familiar to him. For example, it feels as if the driver is scared.

The evaluation unit 12 evaluates how the driver's attitude has changed due to the operation of the operation mechanism 14. As described above, the evaluation unit 12 determines how the driver's attitude toward the vehicle-mounted robot 1 is based on the recognition result of the driver's image captured by the camera 5 and then recognized by the subject recognition unit 6. It has a function to evaluate whether it has changed.

The in-vehicle robot 1 has one configuration example as described above. Next, an example of a procedure included in the operation control process based on the familiarity factor according to the one configuration example will be described with reference to FIGS. 1 to 4.

FIG. 5 is a flowchart illustrating an example of a procedure included in the operation control process based on the familiarity factor.
In step S <b> 1, the target person recognition unit 6 recognizes a driver to be a target from an image in the vehicle 100 captured by the camera 5. Next, in step S <b> 2, the familiarity index calculation unit 11 acquires a factor from the memory 19. Specifically, the memory 19 includes a quantitative factor 17 and a qualitative factor 16 as factors. The familiarity index calculation unit 11 acquires at least one of the quantitative factor 17 and the qualitative factor 16 from the memory 19.

Next, in step S3, the familiar index calculation unit 11 calculates a familiar index based on the acquired factor. Variations of the method of calculating the familiarity index by the familiarity index calculating unit 11 will be described later. Next, in step S4, the operation control unit 13 selects a determination method. Variations of this determination method will also be described later.

In step S4, when a determination method is selected by the operation control unit 13, the familiarity with the driver is determined using the determination method, and the operation mechanism 14 is operated based on the familiarity index described above. Here, for example, when the driver is driving violently and the familiarity index is low, the operating mechanism 14 moves the left arm 1L downward relative to the main body 1B as shown in FIG. Let 1H be an expression 1F that is not intimate.

Then, the driver notices that the in-vehicle robot 1 that was felt to be a unique presence and a close presence to him has such an intimate expression, and his own driving behavior is He notices that the vehicle-mounted robot 1 has some unfavorable influence, and he has to change his driving attitude if he has to drive carefully.

Next, in step S5, the evaluation unit 12 recognizes the driver's image taken by the camera 5 after the subject recognition unit 6 has passed for a while and recognizes the driver who is the subject from the facial expression (or voice). Get the reaction. Next, in step S6, the evaluation part 12 evaluates and determines the driver's reaction acquired in this way. Next, in step S7, the familiarity index calculation unit 11 calculates the familiarity index again based on the determination result of the driver's reaction by the evaluation unit 12. At this time, the familiarity index calculation unit 11 may recalculate the familiarity index based on at least one of the quantitative factor 17 and the qualitative factor 16 described above as well as such a determination result.

Next, in step S8, the motion control unit 13 determines whether or not the new familiarity index is higher than the previous familiarity index. If the new familiarity index is higher, step S9 described later is executed. On the other hand, if the new familiarity index is not high, step S10 described later is executed.

In step S8, the operation control unit 13 controls the operation mechanism 14 to execute an operation with a higher familiarity than the previous time. Specifically, the motion control unit 13 controls the motion mechanism 14 to move the left arm 1L upward relative to the main body while making the head 1H an intimate facial expression 1F as shown in FIG. To make it work.

Then, the vehicle-mounted robot 1 expresses an attitude that makes the driver feel intimate, so that the driver feels that the vehicle-mounted robot 1 is more unique and closer to him / her. The driver tries to continue driving safely. Next, it returns to step S5 mentioned above and is performed.

On the other hand, when the motion control unit 13 determines in step S8 that the new familiarity index is not higher than the previous familiarity index, the operation mechanism 14 is controlled to be the same as the previous familiarity or the previous familiarity. Perform low action.

Then, the in-vehicle robot 1 expresses an attitude that is not intimate with respect to the driver, and the driver is estranged from the in-vehicle robot 1 that has felt that it is a unique existence and is closer to him / her. In order to make the driver feel that there is a factor to act, the driver changes his / her driving and becomes aware that if he / she does not drive safely, he / she will not be intimate. Therefore, the driver is careful to drive carefully so that the in-vehicle robot 1 that he / she feels is intimate.

The following methods can also be adopted in each procedure in the above flow. In the following example, the in-vehicle robot 1 is mounted in the vehicle 100 as described above.

<Example of factors that determine the familiarity index>
The familiar index calculation unit 11 may calculate the familiar index using at least one of the following quantitative factor 17 and qualitative factor 16.
1. Quantitative factor 17
(1) Cumulative number of rides on the vehicle 100 (2) Cumulative boarding time of the vehicle 100-Frequency of boarding the vehicle 100-It is determined whether or not a person who often rides as a normal driver. That is, the determination is made according to the ratio of the number of times / period / frequency of boarding as a driver to the total number of times / period / frequency.
・ Even if it is the same person, it is determined whether the boarding at that time is a boarding as a driver or not.
・ Determine the relationship with the driver based on whether the passenger seat is often in the passenger seat or the back seat. An example of such a relationship is the familiarity between the in-vehicle robot 1 and the driver.
(3) Driving judgment result of safe driving or rough driving Qualitative factor 16
-You may make it determine the emotion with respect to the vehicle-mounted robot 1 from the tone of the other party's calling voice.
-You may judge from the meaning content of the other party's calling voice.
The determination may be made according to whether or not the person has made a call including a predetermined word such as the name of the robot.
・ If there is a word that is not normally spoken by someone who is not the first, such as “I met you again,” you may make a decision equivalent to your name.
-It may be determined by image recognition whether the person's gesture, reaction, facial expression, etc. are gentle rather than the tone of the voice.
-The state of driving may be determined by an image captured by a camera, or may be determined based on engine sound, road noise, presence or absence of sleepiness by recognition of a driver's facial expression, and the like.

<Example of determination method>
The familiarity index calculation unit 11 described above uses the following factors singly or in combination. The familiarity index calculation unit 11 calculates a familiarity index for each driver, and the operation control unit 13 determines an action to be taken based on the familiarity index. The quantitative factor 17 as accumulated data may be accumulated for each driver.

1. Quantitative factor 17
(1) Cumulative number of rides on the vehicle 100 (1-1) Determination method that a driver or the like “gets on” the vehicle 100 ・ Even if the determination is based on the recognition of the person's face and voice by the robot good.
-Data of a personal authentication system other than the robot (navigation, vehicle fingerprint sensor, car key detection, etc.) may be used.
-It may be used in combination with a seating sensor or door opening / closing sensor (not shown).
(1-2) Entrance / exit other than during traveling may be excluded from the count.
(1-3) Boarding and entering / leaving less than a predetermined time may be excluded.
For example, when a driver stops at a convenience store or gasoline stunt while driving and then gets off the vehicle again, the driver makes a decision for convenience at such a predetermined time, and the driver goes in and out of the vehicle. You may make it handle with what was not.
(1-4) Exit / exit that can be determined from the stop-off point when it is a stop on the way may be excluded.
The determination may be made when the point is a place where a temporary boarding / exiting is generally performed such as a convenience store, a gas stunt, a restaurant, a shopping center, or a parking lot in the town.
(1-5) It is possible to exclude entry / exit that can be determined as a stopover from the set destination or set route.
-Entering / exiting at a point other than the destination-Entering / exiting on the way of the set route (1-6) One day (or during a predetermined period such as 24 hours) any number of times may be counted as one time.

(2) Cumulative boarding time of vehicle 100 (2-1) Cumulative boarding time is accumulated.

(3) Frequency of boarding the vehicle 100 (3-1) The determination may be made based on the elapsed period from the previous boarding of the driver to the current boarding.
(3-2) The determination may be made based on the number of boarding times or boarding time in a predetermined period.
(3-3) The number of times of boarding may be obtained, for example, by calculating the ratio of the number of days actually boarded to the number of days in a predetermined period.
(3-4) The frequency may be calculated in units of a predetermined coefficient period, and it may be determined whether the frequency is increasing.

(4) It may be based on whether the person who rides normally as a normal driver or not. That is, the number of rides as a driver, the period, and the total number of times, the period, and the ratio to the frequency may be used as a reference.
・ For example, recognize the face at the driver's seat. The driver's seat position may be registered first, or the vehicle-mounted robot 1 may recognize the handle 4 as an image.

(5) Even if it is the same person, is the driver at that time or not? ・ For example, recognize the face at the location of the driver's seat. The position of the driver's seat 6a may be registered first, or the vehicle-mounted robot 1 may recognize the handle 4 as an image.

(6) The relationship (intimacy) with the driver is determined and taken into account based on whether the passenger seat 6b is often used or the back seat is often used.
The seat position may be registered first, or the vehicle-mounted robot 1 may recognize the presence or absence of the handle 4 or the presence or absence of the front seat.
・ For example, the passenger seat 6b may be determined to have higher intimacy, or may be determined for each age group based on the passenger's appearance (children are determined to have higher intimacy even in the rear seat), and driving This determination may be performed only when there are a plurality of passengers in addition to the passenger.

In the present embodiment, the number of times, the period, the frequency, etc. described above may be accumulated from the first time, or may be accumulated for a predetermined period such as the last one month, for example.

2. Qualitative factor 16
(1) The result of judging the feelings of the robot from the tone of the opponent's calling voice ・ If the tone shows affection, the familiarity index increases as a positive judgment, and if it is angry, the negativeness is judged as negative. Go down.
-Accumulated and saved each time this judgment is called, it may be reflected in the index by accumulating data.
-The numerical value of each determination may be accumulated plus or minus, and the determination may be made with the total plus or minus degree.
・ Determine based on the tendency to increase or decrease plus.
-When accumulating, abnormal values such as especially low values in the positive trend and vice versa may be omitted from the judgment.
-Graph each judgment value in chronological order or in order of occurrence, and judge the friendliness by the pattern of this graph.
For example, the familiarity index generally shows a certain positive value, but even if a negative value sometimes shows a particularly low judgment and the cumulative total is low, the underlying trend is positive because the emotional undulation of the person is intense from the graph pattern judge.

(2) Result of judgment from the meaning of the caller's call voice that has been recognized ・ If the utterance content of the driver, etc., is a meaning content that indicates dearness, then the familiarity increases as a positive judgment and the meaning content is angry If there is, it can be judged as negative and the familiarity index can be lowered.
-You may determine from the meaning content only of a call word, and you may determine the intention and meaning content of the call word from the context before and behind.
-Accumulated and saved each time this judgment is called, it may be reflected in the index by accumulating data.

(3) Result of determination based on whether or not the person who has called with the name of the vehicle-mounted robot 1 ・ In other words, determine whether or not he / she knows or remembers himself / herself (vehicle-mounted robot 1) .
・ If you have called once in the past, you can make a positive decision, or you can make a decision based on whether you have called at least once in a predetermined period such as the last month. You may judge whether you gave me.
・ If there is a word that is not normally spoken by someone who is not the first, such as “I met you again,” you may make a decision equivalent to your name.

(4) Results of image recognition to determine whether the person's gestures, reactions, facial expressions, etc., rather than the tone of the voice, were friendly. • Determined when communication with the robot occurred, such as when called. May be. The determination may be made by accumulating the determination.
-A reaction of a person may be determined when the in-vehicle robot 1 performs some operation.
Such determination by image recognition may be used in combination with determination of voice tone.

In the present embodiment, the total may be a total from the first time, or may be a total for a predetermined period such as the last month. Further, in the present embodiment, it may be accumulated only during the drive, and may be reset at the next drive, or only data of a predetermined drive such as only the previous drive may be stored, and the next time may be started from there. As a specific example, the evaluation unit 12 may make the determination by adding the current total to the previous total, for example.

<Quantification of familiarity>
1. Quantitative factor 17
FIG. 6 is a diagram showing a first example in which the quantitative factor 17 is quantified. In FIG. 6, the vertical axis indicates the evaluation, and the horizontal axis indicates the number of times, the frequency, and the like.
For example, the evaluation unit 12 can employ a pattern in which the quantitative factor 17 is evaluated only in the positive direction by a stacking method. That is, when the evaluation unit 12 performs numerical evaluation of the quantitative factor 17 and calculates a familiarity index to perform the evaluation, the evaluation increases to the right as illustrated in accordance with an increase in the number of times, frequency, and the like.

FIG. 7 is a diagram showing a second example in which the quantitative factor 17 is digitized. In FIG. 7, the vertical axis indicates the evaluation, and the horizontal axis indicates the number of times, the frequency, and the like.
For example, the evaluation unit 12 can employ a pattern that evaluates positively or negatively depending on the numerical value at that time. That is, when the evaluation unit 12 converts the quantitative factor 17 into a numerical value, calculates a familiarity index, and performs the evaluation, the evaluation increases or decreases with the evaluation value 0 as a boundary according to an increase in the number of times, frequency, and the like.

2. Qualitative factor 16
(1) Binarization and qualitative judgment criteria are provided and stepwise digitization is performed.
(2) The determination may be changed according to the situation or individual cumulative data.
(3) For example, as to whether or not to call the name of the in-vehicle robot 1, it is positive if it is called without calling minus even if it is not yet called at the stage where the number of boarding is small, but when many appear, You may make it minus if you don't call it.

<Change in familiarity index>
1. It is good also as the operation which accumulates only a positive factor and acclimatizes according to it, and an index increases.
・ For example, the acclimatization index may increase as the number of boarding increases simply by accumulating the number of boarding.
2. On the other hand, the familiarity index may be determined including a negative factor, and may be changed in a direction in which the familiarity decreases.
-You may decide whether to change the index to plus or minus depending on whether the value of the judgment factor at that time is plus or minus.
-For example, if the elapsed period from the previous boarding to the current boarding becomes equal to or greater than a predetermined value, a negative determination of a numerical value corresponding to the period may be made (see a graph application example described later).
-It is possible to accumulate both positive and negative values of a factor (predetermined period) and determine whether the familiarity index is changed to positive or negative depending on whether the cumulative value is positive or negative.
・ For example, according to the elapsed time from the previous boarding to the current boarding, the numerical value obtained by the graph as in the graph application example described later is accumulated over a predetermined period such as the last month, and the result value is positive or negative It may be determined whether the familiarity index is changed to plus or minus.
-It is also possible to accumulate both positive and negative values of a plurality of factors for a predetermined period, and determine whether the acclimatization index changes to positive or negative depending on whether the cumulative value is positive or negative. A factor determined only as a positive factor and a factor determined only as a negative factor may be mixed by multiplying a factor determined for both by a predetermined coefficient.
Here, as an example of calculating the familiarity index k, the following equation can be given.
Familiarity index k = p [cumulative ride count evaluation] + q [previous elapsed period evaluation] + r [name evaluation]

<Application example of graph>
FIG. 8 is a diagram illustrating a first example of how the index is reduced according to the elapsed period from when the driver got on the last time.
The evaluation unit 12 may adopt a pattern in which the index is adapted according to the elapsed period from when the driver got on the last time.

FIG. 9 is a diagram illustrating a second example in which the habituation index decreases according to the elapsed period from when the driver got on the last time.
The evaluation unit 12 may adopt a pattern in which the acclimatization index is comparatively long and the qualitative factor 16 or the quantitative factor 17 is held in the memory 19 even after an elapsed period from when the driver got on the last ride. .

FIG. 10 is a diagram illustrating a third example in which the habituation index decreases according to the elapsed period from when the driver got on the previous ride.
The evaluation unit 12 may adopt a pattern in which the habituation index decreases immediately after the elapsed period from when the driver boarded the vehicle last time.

8 to 10, a plurality of graphs (formulas) may be prepared, and the graph to be applied may be changed according to the character, growth, and mood of the vehicle-mounted robot 1. The character of the vehicle-mounted robot 1 here represents an individual difference.

The in-vehicle robot 1 in the above embodiment is equivalent to the operation determining means 11 and 13 (corresponding to the familiarity index calculation unit and the operation control unit) that determine the operation to be reacted according to the familiarity with the driver (corresponding to a qualitative factor). ), An operation mechanism 14 that performs an operation according to the determination by the operation determination means 11 and 13 (corresponding to the operation mechanism), and the driver based on the reaction of the driver in contact with the operation performed by the operation mechanism 14 And a control unit 12 (corresponding to an evaluation unit) that changes the operation to be performed by the operation mechanism 14.

In this way, for example, when the driver performs a safe driving, the in-vehicle robot 1
In response to such safe driving, a friendly gesture is performed so as to show familiarity to the driver. Then, the driver comes into contact with the intimate reaction of the in-vehicle robot 1 and tries to continue such safe driving.

On the other hand, for example, when the driver performs rough driving, the vehicle-mounted robot 1 makes a frightening response in response to such rough driving. Then, the driver comes into contact with the sparse response of the in-vehicle robot, notices that the in-vehicle robot 1 that he / she feels important is feared, and tries to change such a rough operation. As a result, the driver's own reflection on driving can be promoted, and the driver can be led to safer driving. This effect can be obtained as the vehicle-mounted robot 1 becomes a target of feelings or emotion transfer more than just a machine for the driver.

More specifically, the in-vehicle robot 1 performs a terrible operation, an uneasy behavior, and a voice utterance / pronunciation when a rough operation such as a sudden handle or a sudden brake is detected by a G sensor or the like. The vehicle-mounted robot 1 gradually shows a friendly gesture to a familiar person, but if the driver continues rough driving, the driver will become accustomed and improve his friendship. The usual behavior for, will shift to something strange.

When the vehicle-mounted robot 1 recognizes a familiar driver when riding and starts to show a more pleasant gesture than before, the driver will show a scary gesture if the driver drives rough. become. Then, if the driver continues the driving, it will eliminate the accumulated effort that has been accustomed until now, and in order to make the vehicle-mounted robot 1 behave amiably again, familiarize from the beginning and do not accumulate safe driving. It will be recognized that it should not, and psychological restraint to refrain from rough driving will work.

At this time, the vehicle-mounted robot 1 changes its operation and reaction according to such a familiarity level (corresponding to the familiarity index). That is, when the vehicle-mounted robot 1 detects a rough driving, the vehicle-mounted robot 1 changes the operation and reaction according to the degree of familiarity with the driver at that time. At this time, the vehicle-mounted robot 1 may be operated more exaggerated than usual when, for example, a highly accustomed person performs rough driving. As a betrayal of the trust that the driver has built for the in-vehicle robot 1 until now, this in-vehicle robot 1 performs a more fearful operation, and if the driving is immediately changed, the relationship of trust becomes stronger. For example, you may make it make a gesture more amiable than before. This is because a driver with a high familiarity index is considered to be a person with a high degree of emotional transfer, so even if the degree of fear and reaction is slightly exaggerated and a high psychological suppression effect is aimed at, it is difficult to impair emotional transfer It is because it becomes. In addition, for example, for a person who is not so familiar, if the emotional transfer is not enough and the fear level is increased too much, the in-vehicle robot 1 may be adversely affected because the emotional transfer is impaired. It is also good to have a slow reaction.

In the vehicle-mounted robot 1 in the above embodiment, in addition to the above-described configuration, the motion determination means 11 and 13 (corresponding to the familiarity index calculation unit and the operation control unit) not only include the degree of familiarity with the driver, but also the driving The operation to be reacted is determined according to the rider's boarding history 17 (corresponding to a quantitative factor).

If it does in this way, the vehicle-mounted robot 1 will perform the operation | movement according to the driver's familiarity, considering the ride history of the driver, and the driver | operator's vehicle-mounted robot 1 which the driver has. It is possible to perform an operation that more closely reflects the familiarity.

The in-vehicle robot 1 in the embodiment further includes a recognition unit 6 (corresponding to a target person recognition unit) that identifies and recognizes each of the plurality of drivers, in addition to the above-described configuration. No. 13 is characterized in that the familiarity degree is managed according to the reaction for each driver identified by the recognition means 6.

In this way, the vehicle-mounted robot 1 can operate to perform different reactions depending on each driver even when any of a plurality of drivers is randomly driving.

In the vehicle-mounted robot 1 in the above embodiment, in addition to the above-described configuration, the motion determination unit further includes a familiarity index calculation unit 11 (into the familiarity index calculation unit) that calculates a familiarity index obtained by quantifying the familiarity with the driver. And a motion control means 13 (corresponding to a motion control unit) for determining a motion to be reacted according to the familiarity index.

In this way, the vehicle-mounted robot 1 objectively determines the familiarity of the driver with the driver using the familiarity index that quantifies the relationship with the driver, and according to the familiarity index, the driver If an intimate relationship is not established, for example, an unfriendly operation can be performed, and if an intimate relationship is established, a favorable operation can be performed, for example.

In the vehicle-mounted robot 1 in the above embodiment, in addition to the above-described configuration, the motion determination means 11 and 13 are further based on the familiarity level according to the driver's speech content recognized by the recognition means 6. The operation mechanism 14 is controlled by determining the operation to be reacted.

In this way, the in-vehicle robot 1 can determine the degree of familiarity according to the speech of the driver who has recognized the voice, and can perform an action that should react accurately according to the degree of familiarity with the driver.

In the vehicle-mounted robot 1 in the above embodiment, in addition to the above-described configuration, the control unit 12 (corresponding to the evaluation unit) further includes the familiarity index according to the utterance content of the driver recognized by the recognition unit 6. It is characterized by correcting.

In the vehicle-mounted robot 1 in the above embodiment, in addition to the above-described configuration, the familiarity index calculation means 11 is further adapted to the familiarity depending on whether or not a predetermined word is included in the speech content that has been voice-recognized. Calculate the index. As such a predetermined wording, the name given to the vehicle-mounted robot 1 can be mentioned, for example.

In this way, the vehicle-mounted robot 1 can easily determine the degree of familiarity with the driver based on whether or not a predetermined word is present in the utterance content of the driver.

In the vehicle-mounted robot 1 in the above embodiment, in addition to the above-described configuration, the motion determination means 11 and 13 react based on the degree of familiarity according to the driver's facial expression image-recognized by the recognition means 6. It is characterized by determining an operation to be performed.

In this way, the vehicle-mounted robot 1 can determine the degree of familiarity with the driver according to the driver's facial expression that has been image-recognized, and can perform an action that should react according to the degree of familiarity with the driver.

In the vehicle-mounted robot 1 in the above embodiment, in addition to the above-described configuration, the control unit 12 further corrects the familiarity index according to the driver's facial expression image-recognized by the recognition unit 6. To do.

As described above, in this embodiment, in addition to the following familiarity calculation method, one of two types of familiarity index calculation methods such as the latter familiarity calculation method can be selectively employed. .

1. Method for calculating familiarity index by including driving judgment in familiarity factor (the above embodiment)
The vehicle-mounted robot 1 gets used to the driver when the driver gets on the vehicle many times. However, even if the driver gets used to it, if the driver performs rough driving, the familiarity index is subtracted and the familiarity is lowered. May return to its original unfamiliar state. Therefore, the driver comes to drive safely to restore the familiarity of the in-vehicle robot 1.

2. How to calculate the familiarity index without including the driving judgment as a familiar factor (application example)
The in-vehicle robot 1 gets used to the driver when the driver gets on the vehicle many times. However, when the driver makes a rough driving, the familiarity index does not change, so the familiarity does not change, but an uneasy expression is expressed. When you return to safe driving, you will be happy with the familiar appearance. For this reason, the driver will think that it would be good to reflect on the familiar presence of the in-vehicle robot 1 showing an uneasy expression and return to safe driving. This uses the utility that a psychological effect can be obtained if the relationship built over time is based on the relationship rather than the familiar one from the beginning.

In addition, this embodiment is not restricted above, Various deformation | transformation are possible. Hereinafter, such modifications will be described in order.

In the above embodiment, the motion control unit 13 and the familiar index calculation unit 11 may be integrated.

In the said embodiment, although the vehicle-mounted robot 1 illustrated in the inside of the vehicle 100 as an example, it is not restricted to this, For example, even if it applies to the robot or toy put in a house, such as a front door and a room good. The in-vehicle robot 1 may be applied to a robot intended to contribute to other purposes instead of the robot intended to contribute to safe driving as described above. A robot having substantially the same function as the in-vehicle robot 1 is equivalent to the above-described driving determination. Instead, for example, the room is overlooked, and if the room is messed up by image recognition, the robot will be displeased and organized. The function may be demonstrated. For example, the state of being in a bad mood represents a state in which habituation is reduced.

In addition to those already described above, the methods according to the above-described embodiments and modifications may be used in appropriate combination.

It is a perspective view which shows the example of a part of structure in the vehicle of the vehicle in this embodiment. It is the perspective view which expanded and represented an example of the external appearance in which the robot shown in FIG. 1 is operate | moving. It is the perspective view which expanded and represented an example of the external appearance in which the robot shown in FIG. 1 is operate | moving. FIG. 4 is a block diagram illustrating an example of functions of the robot illustrated in FIGS. 1 to 3. It is a flowchart which shows an example of the procedure contained in the operation control process based on a familiarity factor. It is a figure which shows the 1st example which digitized the quantitative factor. It is a figure which shows the 2nd example which digitized the quantitative factor. It is a figure which shows the 1st example of a mode that an acclimatization index falls according to the elapsed period from the time when a driver | operator got on last time. It is a figure which shows the 2nd example of a mode that a habituation index falls according to the elapsed period from the time when a driver | operator got on last time. It is a figure which shows the 3rd example of a mode that a habituation index falls according to the elapsed period from the time when a driver | operator got on last time.

Explanation of symbols

1 In-vehicle robot 6 Target person recognition unit (equivalent to recognition means)
11 Familiarity index calculation unit (equivalent to familiarity index calculation means, action determination means)
12 Evaluation part (equivalent to control means)
13 Operation control unit (equivalent to operation control means and operation determination means)
14 Operating mechanism (equivalent to operating mechanism)
16 Qualitative factor 17 Quantitative factor 100 Image display control device

Claims

An action determining means for determining an action to be reacted according to the familiarity with the driver;
An operation mechanism for performing an operation according to the determination by the operation determination means;
Control means for correcting the degree of familiarity with the driver based on the reaction of the driver in contact with the operation performed by the operation mechanism, and changing the operation to be performed by the operation mechanism;
A vehicle-mounted robot comprising:
The in-vehicle robot according to claim 1,
The in-vehicle robot characterized in that the operation determining means determines not only the degree of familiarity with the driver but also the operation to be reacted according to the ride history of the driver.
In the in-vehicle robot according to claim 1 or 2,
Recognizing means for recognizing each of the plurality of drivers,
The operation determining means includes
The vehicle-mounted robot characterized by managing the degree of familiarity according to the reaction of each driver recognized by the recognition means.
The in-vehicle robot according to claim 3,
The operation determining means includes
Familiarity index calculating means for calculating a familiarity index that quantifies the familiarity degree with the driver,
An operation control means for controlling the operation mechanism by determining an operation to react according to the familiarity index;
A vehicle-mounted robot comprising:
The in-vehicle robot according to claim 3,
The operation determining means includes
An in-vehicle robot characterized in that an action to be reacted is determined based on the degree of familiarity according to the utterance content of the driver that is voice-recognized by the recognition means.
The in-vehicle robot according to claim 4,
The control means includes
An in-vehicle robot, wherein the familiarity index is corrected according to the utterance content of the driver that is voice-recognized by the recognition means.
The in-vehicle robot according to claim 4,
The familiarity index calculating means is:
An in-vehicle robot characterized in that the familiarity index is calculated according to whether or not a predetermined wording is included in the speech content that has been voice-recognized.
The in-vehicle robot according to claim 3,
The operation determining means includes
An in-vehicle robot characterized in that an action to be performed is determined based on the degree of familiarity according to a driver's facial expression image-recognized by the recognition means.
The in-vehicle robot according to claim 4,
The control means includes
An in-vehicle robot, wherein the familiarity index is corrected according to the facial expression of the driver whose image is recognized by the recognition means.