JP2003217055A - Method and system for preventing accident - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for preventing an accident such as a crime by using an interactive device. <P>SOLUTION: This method prevents the accident by using the interactive device capable of interacting with a person, and includes a step for the interactive device to detect the person, a step for the interactive device to interact with the person, a step for the interactive device to determine whether or not abnormality is caused on the basis of a result of interaction, and a step for the interactive device to report when determining that the abnormality is caused. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for preventing an accident such as a crime by using an interactive device capable of interacting with a person, and a system having a plurality of interactive devices.

[0002]

2. Description of the Related Art Conventionally, a device called an immobilizer is known as a theft prevention device for protecting a vehicle from theft. This device prevents the engine from starting if the ID code transmitted from the key does not match the ID code registered in the vehicle. This device prevents the engine from being started by an illegally duplicated key. This is because an illegally duplicated key cannot transmit an ID code that matches the ID code registered in the vehicle. Therefore, this device is effective for preventing the theft of a vehicle using an illegally duplicated key.

The G used in car navigation systems
Vehicle security systems using PS have also been developed. In this system, a vehicle is equipped with a specific device in advance so that the position of the vehicle can be searched when the vehicle is stolen. Some security companies that use this system also provide services to expedite security personnel to the scene where a stolen vehicle is found.

A system called a home security system is also known. In this system, when a suspicious person invades a house, the security sensor detects an abnormality and the home controller sends an abnormality signal to the control center. The control center is configured to issue an instruction to urge a security guard to the scene where the security sensor detects an abnormality.

Further, there is disclosed a technique for preventing a serious accident by notifying a third party when an abnormality of the human body is detected by using a sensor attached to the human body (for example,
See Patent Document 1).

[0006]

[Patent Document 1] Japanese Patent Laid-Open No. 10-155749

[0007]

Although the above-described immobilizer is effective in preventing the theft of a vehicle using an illegally duplicated key, it prevents the theft of a vehicle using a real key. I can't. Therefore, if the real key is lost, the vehicle may be stolen.

Further, in the above-mentioned vehicle crime prevention system, when a thief destroys the vehicle crime prevention system, it becomes impossible to search the position of the stolen vehicle, so that it becomes difficult to find the stolen vehicle. .

Further, in the above-mentioned home security system, if the thief cannot impersonate the house resident and the security sensor cannot detect the abnormality,
It will be damaged.

Further, in the technique described in Patent Document 1 mentioned above, it is necessary to attach a sensor to the subject and monitor the health condition of the subject. For this reason, the subject has been forced to inconvenience such as wearing the sensor and being hard to move.

The present inventor has thought that crimes can be prevented by identifying a thief or an intruder into a house by using an interactive device. In addition, the present inventor operates a plurality of interactive devices in cooperation with each other,
I thought that crime could be prevented. further,
The present inventor thought that an accident can be prevented in advance by detecting the occurrence of an abnormality using an interactive device.

An object of the present invention is to provide a method for preventing accidents such as crimes using an interactive device and a system equipped with a plurality of interactive devices.

[0013]

The method of the present invention is a method for preventing an accident by using an interactive device capable of interacting with a person, wherein the interactive device detects a person. The step of performing the dialogue with the person, the interactive apparatus determining whether or not an abnormality has occurred based on the result of the dialogue, and the abnormality has occurred. If so, the interactive device makes a notification, which achieves the object of the invention.

The abnormality may be that the person is not the user of the interactive apparatus, or that the person is the user of the interactive apparatus and the state of the user is not normal.

Another method of the present invention is a method of preventing an accident using an interactive device capable of interacting with a person, the interactive device detecting the person. Said interactive device having an interaction with said person and a user of said interactive device; said interactive device determining whether said person is said user based on the result of said interaction; When it is determined that the person is the user, the interactive device determines whether the state of the user is normal, and when it is determined that the state of the user is not normal Includes the step of making a notification by the interactive device, whereby the above object is achieved.

The interactive apparatus may determine whether or not the state of the user is normal by speaking to the user and confirming the response to the speech.

Another method of the present invention is a method for preventing crimes by using an interactive device capable of interacting with a person, the interactive device detecting the person, Said interactive device receiving whereabouts information indicating the whereabouts of the user of said interactive device; said interactive device determining whether said person is said user based on said whereabouts information; If it is determined that the person is not the user, the interactive device makes a notification, thereby achieving the above object.

The interactive device may receive the location information from another interactive device via a communication line.

The step of the interactive device determining whether the person is the user based on the location information is a step of determining whether the interactive device is absent based on the location information. Determining that the user is absent, the interactive device performs a dialogue regarding the person and the user, and the interactive device performs the dialogue based on a result of the dialogue. Determining whether the person is the user.

[0020] When it is determined that the person is the user, the interactive device determines whether the state of the user is normal, and the state of the user is not normal. If determined, the interactive device may further comprise making a notification.

The interactive device may refer to the dialog history from a dialog history database of another interactive device.

The interactive device may be provided in a vehicle.

The interactive device may be provided in a house.

The system of the present invention is a system including a plurality of interactive devices connected to each other via a communication network, and each of the plurality of interactive devices can interact with a person. Each of the plurality of interactive devices includes a detection unit that detects a person, a location information memory that stores location information indicating the location of the user of the interactive device, and another interactive device. The receiving unit that receives the location information from the location information memory via the communication network, and the person detected by the detecting unit based on the location information received from the other interactive apparatus is the user. And a notification unit that makes a notification when the person is determined not to be the user, thereby achieving the above object.

The determination unit determines whether or not the user is absent based on the location information, and when it is determined that the user is absent, the person detected by the detection unit is detected. And a dialogue regarding the user may be performed, and whether or not the person is the user may be determined based on the result of the dialogue.

When it is determined that the person is the user, the determination unit further determines whether or not the state of the user is normal, and determines that the state of the user is not normal. In the case of being notified, the notification unit may make a notification.

[0027]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

(First Embodiment) FIG. 1A shows a configuration example of an interactive agent 1 as an example of an interactive apparatus.

The interactive agent 1 includes an image recognition unit 10
The voice recognition unit 20, the language processing unit 30, the voice synthesis unit 40, the voice output unit 50, the communication unit 60, and the database unit 70 are included.

The image recognition section 10 is connected to an image input section 12 (for example, a camera), and performs image recognition processing on the image input by the image input section 12 to detect a person from the image. To do. The voice recognition unit 20
It is connected to the voice input unit 22 (for example, a microphone),
The voice input by the voice input unit 22 is recognized.

The language processing unit 30 understands the content of the dialogue based on the voice recognition result output from the voice recognition unit 20, searches the database unit 70, and responds to the personal information of the person and the situation of the dialogue. Generate a sentence.

The response sentence generated by the language processing unit 30 is synthesized into a voice by the voice synthesizing unit 40. The voice synthesized in this way is output to the voice output unit 50 (for example,
Speaker).

The communication unit 60 is used to notify a security company or police via a communication line. The communication line may be a wireless communication line or a wire communication line. For example, when a wireless communication line is used, the communication unit 60 transmits / receives data via the antenna 62.

The database section 70 is a dialogue database 7 in which dialogue patterns and response sentence generation rules are stored.
1, a dialogue history database 72 in which history of past dialogues is stored, information for identifying the person (for example, information such as gender, age, name, occupation, personality, hobby, date of birth of the person), It includes a personal database 73 that stores information that only the person can know and an information database 74 that stores information such as today's weather and news. Information such as today's weather and news is transmitted from outside the interactive agent 1 from the communication unit 60 and the language processing unit 30, for example.
And is stored in the information database 74.

Here, the "person" means the user (owner) of the interactive agent 1. The number of users may be one or more. The user of the interactive agent 1 can be registered in the personal database 73 in advance, for example. In this case, the "person" is
A person who is registered in advance in the personal database 73 as a user of the interactive agent 1.

As described above, the interactive agent 1 is constructed so that it can interact with a person. It is preferable that the interactive agent 1 has a function (software agent function) that moves around in an information space such as the Internet and performs information processing such as information retrieval, filtering, and schedule adjustment on behalf of a human. Since the interactive agent 1 interacts with a person as if it were a human being, it is sometimes called an anthropomorphic agent.

The interactive agent 1 is a kind of computer. Components 10 to 7 of the interactive agent 1
The function of 4 can be realized by, for example, a CPU (not shown) in the computer executing various programs stored in a memory (not shown) in the computer. However, the components 10 to 7 of the interactive agent 1
The functions of 4 are not limited to being realized by software. It is also possible to realize some or all of the functions of the components 10 to 74 of the interactive agent 1 by hardware.

For details of research on interactive agents, for example, the websites of the Ministry of International Trade and Industry, the Agency of Industrial Science and Technology, the Institute of Electronics Technology, and the Information Integration Dialogue Lab (http://www.etl.go.jp/ ~ 72
33 /).

FIG. 1B schematically shows how the driver 80 and the interactive agent 1 are having a conversation. The interactive agent 1 may be installed anywhere in the vehicle such as a dashboard.

FIG. 2 shows an example of the procedure of the crime prevention program executed by the interactive agent 1 shown in FIG. 1A.

The steps of the crime prevention program shown in FIG. 2 will be described below by taking the vehicle theft prevention as an example. The interactive agent 1 is assumed to be installed in the vehicle.

In ST1, the image recognition section 10 detects a person from the image input by the image input section 12. S
The person detected at T1 may be the principal or may not be the principal (eg, a thief).

In ST2, the image recognition section 10 obtains the certainty that the detected person is the person himself. The probability that the detected person is the person is calculated by, for example, comparing the characteristic amount indicating the detected person with the characteristic amount indicating the person stored in advance in the personal database 73. The certainty that the detected person is the person is represented by a numerical value in the range of 0% to 100%, for example.

In ST2, the image recognition unit 10 determines whether or not the probability that the detected person is the person himself / herself satisfies a predetermined criterion (for example, 95% or more). When it is determined that the certainty satisfies the predetermined criterion, the image recognition unit 10 determines that the detected person is the person (that is, the user). As a result, the process proceeds to ST6, and the normal mode dialogue is performed between the interactive agent 1 and the user. This dialogue is controlled by the language processing unit 30. Here, the dialogue in the normal mode is, for example, everyday conversation. The dialogue in the normal mode may be started by the utterance of the interactive agent 1 or the utterance of the user.

The following is an example of normal mode interaction. Here, S represents the utterance of the interactive agent 1, and U1 represents the utterance of the user.

S: Good morning, Mr. U.

U1: Good morning. Is the weather fine today? S: It's a nice day today.

U1: That was good. Tell me the news today.

S: Director Nagashima has resigned. The successor is Mr. Hara.

In ST2, when it is determined that the probability that the detected person is the person does not satisfy the predetermined criterion (for example, 95% or more), the image recognition unit 10
It is determined that the detected person is not the person (that is, not the user). As a result, the process proceeds to ST3, and the dialogue in the doubt mode is performed between the detected person and the interactive agent 1. This dialogue is controlled by the language processing unit 30. Here, in the suspicion mode dialogue, the interactive agent 1 presents a question about the person himself / herself for the purpose of confirming that the detected person is not the person himself, and the detected person is presented by the interactive agent 1. It is done in the form of answering the questions asked.

The following is an example of a suspicious mode interaction. Here, S represents the utterance of the interactive agent 1, and U2 represents the utterance of a person who is not the user (for example, a thief).

S: Good morning, U. When was your birthday U2: It's April.

S: It was disappointing. It's October.

In ST4, the language processing unit 30 finally determines whether or not the detected person is the person based on the result of the dialogue in the doubt mode in ST3. In ST4, if it is determined that the detected person is the person himself / herself, the process proceeds to ST6, and shifts from the doubt mode dialogue to the normal mode dialogue. If it is determined in ST4 that the detected person is not the person himself / herself, the process proceeds to ST5.

In ST5, the language processing section 30 instructs the communication section 60 to notify the security company (or the police) via the communication line. The communication unit 60 may transmit such position information and the position information of the interactive agent 1 to the security company (or the police).

As described above, according to the crime prevention program shown in FIG. 2, when it is determined that the person detected in ST2 is less likely to be the person in question, the suspicious mode dialogue is performed in ST3. Based on the result of the dialogue, whether or not the person detected is the person is finally determined. The notification is given only when it is determined in ST2 and ST4 that the person is not the person himself.
By thus making the double determination that the person is “not the person”, it is possible to more reliably determine whether or not the detected person is the person.

Note that ST2 shown in FIG. 2 may be omitted. In this case, when a person is detected in ST1, a suspicious mode dialogue is started in ST3. S
By omitting T2, the processing load on the image recognition unit 10 is reduced.

As described above, according to the present embodiment, it is possible to determine whether or not the person detected by using the interactive agent 1 is the user. If it is determined that the detected person is not a user (for example, a thief),
Be notified to the security company (or police). This makes it possible to prevent crime in advance.

If the interactive agent 1 is mounted on a vehicle and the vehicle is to be checked by an attendant to park the vehicle, the interactive agent 1
It is preferable to turn off the power. This is to prevent the interactive agent 1 from interacting with the staff.
Alternatively, the interactive agent 1 is used when the current position of the vehicle is a parking lot, a hotel, etc. by using the vehicle position information.
May not enter suspicion mode.

It should be noted that the number of persons who should be judged to be the true person by the interactive agent 1 may be only one person or plural persons. For example, when one vehicle is used by four people (Mr. A, Mr. B, Mr. C, and Mr. D), information for identifying each of the four people (for example, sex, age, Information such as name / occupation / personality / hobbies / date of birth) and information that only the person can know may be stored in the personal database 73.

The present embodiment can also be applied to prevent a suspicious person from entering a house. For example, if the interactive agent 1 is provided on the intercom, it is possible to prevent a suspicious person from entering the house through a dialogue with the interactive agent 1. Further, if the interactive agent 1 is installed in a house, it is possible to prevent theft even if a suspicious person intrudes into the house.

FIG. 5 shows an example of the detailed flow of ST6 shown in FIG.

In ST51, since it is confirmed that the person detected in ST1 shown in FIG. 2 is the person (that is, the user), the interactive agent 1 and the user interact in the normal mode. Is done. This dialogue is controlled by the language processing unit 30, for example.

In ST52, it is determined whether or not the user's state is normal based on the result of the dialogue performed in ST51. Such a determination is performed by the voice recognition unit 20 and the language processing unit 30, for example. For example, the voice recognition unit 20 extracts a keyword from the voice of the user input by the voice input unit 22. The language processing unit 30 determines whether or not the keyword extracted by the voice recognition unit 20 matches a predetermined keyword such as “painful” or “help” (that is, a keyword indicating that the user is not normal). . The predetermined keyword is stored in advance in the dialogue database 71, for example. The language processing unit 30 determines that the user is not normal (abnormal) when the keywords match, and otherwise determines that the user is normal. Alternatively, when the voice recognition unit 20 detects in the voice of the user input by the voice input unit 22 that a silent period has continued for a predetermined period or longer, the voice recognition unit 20 outputs a detection signal to the language processing unit 30. You may The language processing unit 30
When the detection signal from the voice recognition unit 20 is received,
It may be determined that the user is not normal (abnormal), and in other cases, it may be determined that the user is normal.

If the determination result in ST52 is "YES", the process proceeds to ST56, and if the determination result in ST52 is "NO", the process proceeds to ST53.

In ST53, the user is uttered for abnormality confirmation. This utterance is, for example, the language processing unit 3
0, the voice synthesis unit 40, and the voice output unit 50. For example, the language processing unit 30 generates a response sentence such as “Is it okay?” And outputs it to the voice synthesis unit 40. The speech synthesis unit 40 synthesizes the response sentence output from the language processing unit 30 into speech. The synthesized voice is output from the voice output unit 50. As a result, an utterance for confirming an abnormality, such as "Is it okay?", Is issued to the user. In ST54, S
A response from the user to the utterance made in T53 (for example, the presence or absence of a response from the user, and / or
Based on the content of the response from the user), it is finally determined whether the user's state is normal. Such determination is performed by, for example, the voice recognition unit 20 and the language processing unit 3
Performed by 0. Here, the determination in ST54 is performed, for example, according to the same method as the determination method described in ST52.

If the determination result in ST54 is "YES", the process proceeds to ST56, and if the determination result in ST54 is "NO", the process proceeds to ST55.

In ST55, the language processing section 30 instructs the communication section 60 to notify the emergency center via the communication line. The communication unit 60 makes such a notification and, at the same time, the user's personal information (for example, the personal database 73).
The age, sex, medical history, etc. of the user stored in the above may be transmitted to the emergency center, or the position information of the interactive agent 1 may be transmitted to the emergency center.

At ST56, the normal mode dialogue between the interactive agent 1 and the user is continued.

As described above, according to the detailed flow shown in FIG. 5, whether or not the user is normal is determined through the dialogue between the interactive agent 1 and the user. It can be said that such a determination method is more user-friendly than a conventional determination method for determining whether or not the user is normal by using a sensor or an image attached to the human body. This is because the sensor is not bothered to be worn on the human body and the interactive agent 1 does not feel that it is being monitored.

Further, according to the detailed flow shown in FIG. 5, if it is determined in ST52 that the user is not normal, an utterance for abnormality confirmation is issued in ST53, and ST5
In 4, it is finally determined whether or not the user is normal, depending on the presence or absence of a response to the utterance (or the content of the response). In both ST52 and ST54, the notification is made only when it is determined that "the user is not normal". By thus making the double determination that "the user is not normal", it is possible to more reliably determine whether or not the user is normal.

Further, according to the detailed flow shown in FIG. 5, in the notification to the emergency center, it is confirmed that the person detected in ST1 shown in FIG. 2 is the person (that is, the user). Will be done later. Therefore, even in a situation where the user cannot speak, the personal information of the user (for example, age, sex, medical history, etc.) can be transmitted to the emergency center simultaneously with the notification. Thereby, the emergency center can obtain the personal information of the user before the user is transported to the emergency center. As a result, it becomes possible to promptly perform appropriate treatment for the user in the emergency center.

Incidentally, ST53 and ST shown in FIG.
54 may be omitted. In this case, if it is determined in ST52 that the user is not normal, the emergency center will be immediately notified.

As described above, according to this embodiment, it is possible to determine whether or not the person detected by using the interactive agent 1 is the user. When it is determined that the detected person is not the user, the security company (or the police) is notified. This makes it possible to prevent crime in advance. Further, when it is determined that the detected person is the user, it is further determined whether the user is normal (or abnormal). If it is determined that the user is not normal (abnormal), the emergency center is notified. This makes it possible to prevent accidents.

Further, according to the present embodiment, a dialogue is conducted between the interactive agent 1 and the detected person, and it is determined whether or not an abnormality has occurred based on the result of the dialogue. If it is determined that an abnormality has occurred, a notification is sent. Here, “abnormal” means that the detected person is not the user of the interactive agent 1. Alternatively,
“Abnormal” means that the detected person is the interactive agent 1
It means that the user is a user and the state of the user is not normal.

(Second Embodiment) FIG. 3 shows an example of a procedure of a crime prevention program executed by the interactive agent 1 shown in FIG. 1A.

Each step of the crime prevention program shown in FIG. 3 will be described below by taking an example of preventing a suspicious person from entering a house. It is assumed that the interactive agent 1 is provided in a house, for example, a living room.

In ST11, the image recognition section 10 detects a person from the image input by the image input section 12.
The person detected in ST11 may be the person or may not be the person (for example, a thief).

In ST12, the language processing section 30 receives location information indicating the location of the person. The language processing unit 30 may receive the location information input to the interactive agent 1 by the person himself / herself via an input unit (not shown), or the location information from another interactive agent via the communication unit 60. You may receive

In ST13, the language processing section 30 determines in ST1.
Based on the location information received in 2, it is determined whether the person is absent.

In ST13, when it is determined that the person is absent (for example, he is out), the process proceeds to ST.
Proceed to 14. When it is determined in ST13 that the person is not absent (for example, at home), the process proceeds to ST.
Proceed to 17.

The processing of ST14 to ST17 is the same as the processing of ST3 to ST6 shown in FIG.
The description is omitted here. As in the first embodiment, the detailed flow of ST17 shown in FIG.
The flow chart of may be applied. In this case, the same effect as that described above can be obtained.

As described above, according to the crime prevention program shown in FIG. 3, when it is determined that the person is absent in ST13, the suspicious mode dialogue is performed in ST14,
Based on the result of the dialogue in the doubt mode, it is finally determined whether or not the person detected in ST11 is the user. The notification is given only when it is determined in ST13 that the person is absent, and when it is determined in ST15 that the person is not the person. As described above, the determination of "the person is absent" and the determination of "the person is not the person" are performed repeatedly, so that it is possible to more reliably determine whether the detected person is the person.

Incidentally, ST14 and ST15 shown in FIG.
May be omitted. In this case, when it is determined that the person is absent in ST13, the security company (or the police) is immediately notified.

As described above, according to the present embodiment, it is possible to determine whether or not the person detected by using the interactive agent 1 is the user. If it is determined that the detected person is not a user (for example, a thief),
Be notified to the security company (or police). This makes it possible to prevent crime in advance.

The procedure of the crime prevention program shown in FIG.
For example, it is suitable for the multi-agent environment shown in FIG.

FIG. 4 shows an example of a multi-agent environment. In the example shown in FIG. 4, the home agent 81
, A vehicle agent 82, and a portable agent 83
And are connected to the communication network 84. The multi-agent environment shown in FIG. 4 is an example of a system including a plurality of interactive agents connected to each other via a communication network. The number of agents in a multi-agent environment is 3
Of course, it is not limited to. The number of agents in a multi-agent environment can be any integer greater than one.

As these three agents 81 to 83, the interactive agent 1 shown in FIG. 1A can be used. However, in order to use the agents 81-83 in a multi-agent environment, the agents 81-83
Each of 83 must further include a location information memory 90 (see FIG. 1A). Details of the location information memory 90 will be described later. Agents 81-83
By communicating with each other via the communication network 84, they can operate in cooperation with each other.

The home agent 81 is provided in the house. The vehicle agent 82 is provided in the vehicle. The portable agent 83 is provided so that the person can carry it. The portable agent 83 is preferably a wrist watch type in terms of excellent portability.

For example, assume that the person is in a car and is out. In this case, the vehicle agent 82 detects that the person himself / herself is in the vehicle, and informs the home agent 81 of the information that the person himself / herself is in the vehicle as location information indicating the location of the person himself / herself. Contact agent for vehicle 8
The home agent 81 received from 2 can understand that the person is out. Therefore, the home agent 81 can determine that the person is absent in ST13 shown in FIG.

Furthermore, by operating a plurality of interactive agents in cooperation with each other, it is possible not to repeat past dialogs. For example, when a person gets up in the morning and has a dialogue with the home agent 81 about the weather today, when the person goes to work by riding in a car, the vehicle agent 82 mounted on the car is It is preferable to have a dialogue with the person on another topic (eg sports today) without repeating the dialogue on. To control such a dialogue, for example, a dialogue between the home agent 81 and the person is accumulated in the dialogue history database 72 of the home agent 81, and the car agent 82 makes the dialogue history of the home agent 81. This can be accomplished by looking up the database 72 and avoiding dialogues that duplicate past dialogues.

The multi-agent interactive agent executes various agent functions (including reading of electronic mail) in various life scenes of the user to interact with humans through dialogue. Such interactive agents are friendly and useful to human beings as they are closely related to life.
Therefore, the user will utilize the interactive agent in various life scenes.

The procedure of the crime prevention program shown in FIG.
It is assumed that the number of users (owners) of the interactive agent 1 is one. Here, the procedure of the security program assuming a case where there are a plurality of users in a multi-agent environment will be described.

FIG. 6 shows an example of a procedure of a crime prevention program executed by each of the home agent 81, the vehicle agent 82 and the portable agent 83. Here, it is assumed that each of the home agent 81, the vehicle agent 82, and the portable agent 83 has the same configuration as that of the interactive agent 1 shown in FIG. 1A. It is also assumed that the multi-agent environment is set so that the home agent 81, the vehicle agent 82, and the portable agent 83 are connected via the communication network 84 as shown in FIG.

Here, it is assumed that there are two users, "user A" and "user B". In this case, the user name of “user A” and the user name of “user B” are registered in the personal database 73 of each agent. Of course, as in the first embodiment, the personal database 73 may store, for each user, information that identifies the user (image, audio feature amount, etc.) or information that only the user can know.

Each step of the crime prevention program shown in FIG. 6 will be described below by taking an example of preventing a suspicious person from entering a house.

The processing of ST81 is the same as ST1 shown in FIG.
The process is the same as that of. Therefore, the description thereof is omitted here. For example, the image recognition unit 10 functions as a detection unit that detects a person.

In ST82, the communication section 60 receives location information indicating the location of the user from another agent. Here, the "other agent" refers to an agent other than itself among a plurality of agents in a multi-agent environment. For example, the communication unit 60 of the home agent 81 receives the location information indicating the location of the user A from the vehicle agent 82 and the location information indicating the location of the user B from the portable agent 83.

As described above, the communication unit 60 functions as a receiving unit that receives the location information from the location information memory 90 of another interactive agent via the communication network 84.

FIG. 7 schematically shows how the home agent 81 receives location information indicating the location of the user A from the vehicle agent 82 and location information indicating the location of the user B from the portable agent 83. The location information is stored in the location information memory 9 of the home agent 81.
Stored in 0.

FIG. 8 shows an example of the contents of the location information memory 90 of the home agent 81. The example shown in FIG. 8 is
The home agent 81 is not used by either user A or user B (that is, unused or absent),
The vehicle agent 82 is used by the user A, and the portable agent 83 is used by the user B. The initial state of the location information memory 90 is unused for any agent.

In ST83, the language processing section 30 refers to the location information memory 90 to determine whether all of the plurality of users are using other agents.

If the determination result in ST83 is "YES", the process proceeds to ST87, and if the determination result in ST83 is "NO", the process proceeds to ST84.

For example, when the contents of the location information memory 90 of the home agent 81 are as shown in FIG. 8, both the users A and B have other agents (that is,
This means that the vehicle agent 82 and the portable agent 83) are used. In this case, the process is ST87.
Proceed to.

The processing of ST84, ST85 and ST86 is as follows.
This is the same as the processing of ST2, ST3, and ST4 shown in FIG. Therefore, the description thereof is omitted here.

As described above, the language processing unit 30 functions as a determination unit that determines whether or not the detected person is the user based on the location information received from another agent.

In ST87, the language processing section 30 instructs the communication section 60 to notify the security company (or the police) via the communication line.

As described above, the language processing unit 30 and the communication unit 60 function as a reporting unit that reports when it is determined that the detected person is not the person.

It is determined that at least one of the plurality of users is not using another agent (ST8
2) And when it is determined that the detected person is the user (ST84, ST86), the process proceeds to ST88.

In ST88, the language processing section 30 records the user name in the location information memory 90 of the home agent 81.

If the location information from another agent received in ST82 is information indicating that one of the plurality of users (for example, user A) is using the other agent, The authenticity of the user in ST84 may be determined for a user excluding the user name using another agent from the plurality of users. Thereby, the target persons of the user to be judged can be limited. For example, the language processing unit 30 of the home agent 81
However, if it is determined in ST84 that the person detected is the user (the user name is also determined), the user name may be stored in the home agent field of the location information memory 90 in ST88.

The processing of ST89 is the same as ST6 shown in FIG.
The process is the same as that of. Therefore, the description thereof is omitted here. Note that, as in the case of the first embodiment, S shown in FIG.
The flowchart of FIG. 5 may be applied as the detailed flow of T89. In this case, the same effect as that described above can be obtained.

(Third Embodiment) Portable Agent 83
(FIG. 4), when detecting that the response from the user in the normal mode is different from usual, the normal mode is switched to the suspicious mode, and information identifying the person (for example, the date of birth of the person). It is preferable to have a question-type conversation with the user asking for. If the user is unable to answer the question asking for information identifying the person, there is a high possibility that the person has been harmed or that the portable agent 83 has been stolen. Therefore, it is preferable that the portable agent 83 notify the security company (or the police) of that fact. At the same time as such notification, the location information of the portable agent 83 may be transmitted to the security company (or the police).

Alternatively, a keyword is decided in advance between the portable agent 83 and the person, and when the person is at risk of danger, the person himself or herself will emit the keyword, and the portable agent 83 will output the keyword. If detected, a security company (or police) may be notified. The keyword is preferably a phrase that cannot be captured by a third party (for example, a phrase such as “Happa Fumi Fumi”).

In the above-described first to third embodiments, the interactive agent has been described as an example of the interactive device, but the interactive device is not limited to this. Interactive device
It can be any device configured to be able to interact with a person. The interactive device may be, for example, an interactive toy.

[0116]

According to the method of the present invention, a thief or an intruder into a house can be identified by using an interactive agent. This makes it possible to prevent crime in advance. The present invention can determine an abnormal state of the body using an interactive agent. Thus, it is possible to prevent an accident by notifying the preset addressee.

[Brief description of drawings]

FIG. 1A is a block diagram showing a configuration example of an interactive agent 1.

FIG. 1B is a diagram schematically showing how the driver 80 and the interactive agent 1 are having a conversation.

FIG. 2 is a flowchart showing an example of a procedure of a crime prevention program executed by the interactive agent 1 shown in FIG. 1A.

FIG. 3 is a flowchart showing an example of a procedure of a crime prevention program executed by the interactive agent 1 shown in FIG. 1A.

FIG. 4 is a diagram showing an example of a multi-agent environment.

5 is a flowchart showing an example of a detailed flow of ST6 shown in FIG.

FIG. 6 is a household agent 81 and a vehicle agent 82
And a flowchart showing an example of a procedure of a crime prevention program executed by each of the portable agents 83.

FIG. 7 is a diagram schematically showing how the home agent 81 receives the location information indicating the location of the user A from the vehicle agent 82 and the location information indicating the location of the user B from the portable agent 83.

FIG. 8 is a diagram showing an example of contents of a location information memory 90 of a home agent 81.

[Explanation of symbols]

1 Interactive agent 10 Image recognition unit 20 Speech recognition unit 30 Language Processing Department 40 Speech synthesizer 50 voice output section 60 Communication unit 70 Database Department 90 location information memory

   ─────────────────────────────────────────────────── ─── Continued front page    F term (reference) 5C087 AA02 AA03 BB02 BB19 BB74                       BB75 DD03 DD13 DD24 EE05                       EE08 FF01 FF04 GG02 GG19                       GG20 GG66 GG70 GG83

Claims (14)

[Claims] 1. A method of preventing an accident using an interactive device capable of interacting with a person, the method comprising: detecting the person by the interactive device; A step of interacting with a person, a step of determining whether or not the interactive device has an abnormality based on a result of the interaction, and, if it is determined that the abnormality has occurred, the interactive type A device making a notification. 2. The abnormality is that the person is not a user of the interactive apparatus, or that the person is a user of the interactive apparatus and the state of the user is not normal. The method according to Item 1. 3. A method of preventing an accident using an interactive device capable of interacting with a person, the method comprising: detecting the person by the interactive device; Performing a dialogue about a person and a user of the interactive device; determining the person is the user based on the result of the dialogue by the interactive device; and the person is the user. If it is determined that the interactive device determines whether the state of the user is a normal state, and if it is determined that the user state is not normal, the interactive device A device making a notification. 4. The interactive device according to claim 3, wherein the interactive device speaks to the user and confirms a response to the utterance to determine whether or not the user is in a normal state. Method. 5. A method for preventing crimes using an interactive device capable of interacting with a person, the method comprising: detecting the person by the interactive device; Receiving location information indicating the location of the user of the interactive device, determining whether the interactive device is the user based on the location information, the interactive device if the person is not the user If determined, the interactive device makes a notification. 6. The method of claim 5, wherein the interactive device receives the location information from another interactive device via a communication line. 7. The step of the interactive device determining whether the person is the user based on the location information, the interactive device determining whether the user is absent based on the location information. Determining whether or not the user is absent, the interactive device performs a dialog regarding the person and the user, the interactive device based on the result of the dialog. Determining whether the person is the user. 8. If the person is determined to be the user, the interactive device determines whether the state of the user is normal, and the state of the user is normal. The method of claim 5, further comprising the step of: if the interactive device determines not to do so. 9. The interactive device refers to a dialog history from a dialog history database of another interactive device.
The method described in. 10. The method according to claim 1, wherein the interactive device is provided in a vehicle. 11. The method according to claim 1, wherein the interactive device is provided in a house. 12. A system comprising a plurality of interactive devices connected to each other via a communication network, each of the plurality of interactive devices being configured to be able to interact with a person. Each of the plurality of interactive devices includes a detection unit that detects a person, a location information memory that stores location information indicating the location of the user of the interactive device, and the location information of another interactive device. Whether the person detected by the detecting unit is the user based on the receiving unit that receives the location information from the memory via the communication network and the location information received from the other interactive device. A system that includes a determining unit that determines whether or not the person, and a reporting unit that reports if the person is determined not to be the user. 13. The determination unit determines whether or not the user is absent based on the location information, and when it is determined that the user is absent, it is detected by the detection unit. 13. The system according to claim 12, wherein a dialogue is performed regarding the person and the user, and whether the person is the user is determined based on a result of the dialogue. 14. When it is determined that the person is the user, the determination unit further determines whether or not the state of the user is normal, and the state of the user is not normal. The system according to claim 12, wherein when it is determined that the notification unit makes a notification.