JP2015109040A - Emergency call device and emergency call system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an emergency call device that preferably operates.SOLUTION: An emergency call device 10 used for watching a person to be watched comprises: a voice unit 17 for obtaining voice; an object information storage unit 24 that has stored therein in advance constant feature point information that does not depend on a health status of the person to be watched or is low in dependency; a voice information comparison unit 23 that extracts constant feature point information from the voice obtained by the voice unit 17, compares the extracted constant feature point information with the constant feature point information on the person to be watched stored in the object information storage unit 24, and thereby authenticates whether the voice obtained by the voice unit 17 is voice of the person to be watched or not; and a communication unit 16 for outputting information on the person to be watched on the basis of the authentication result of the voice information comparison unit 23.

Description

  The present invention relates to an emergency call device and an emergency call system used in, for example, a monitoring service system for elderly people and the like.

  2. Description of the Related Art Conventionally, an emergency call system for reporting an emergency such as a crime or a sudden illness to a specific destination is known. In the conventional emergency call system, the emergency button of the emergency call device or the wireless device (pendant) worn by the user is pressed in the event of an emergency, and the emergency status of the person is reported to a preset destination. Met. However, in recent years, monitoring services that actively monitor users of emergency call devices and avoid accidents in advance through daily monitoring services have become widespread.

  In such a watch-type emergency call system, there is a problem of avoiding false detection of the motion of the target person by watching the motion of an object other than the watch target (for example, a small animal such as a dog or a cat). For example, Patent Document 1 discloses a system in which a sensor is arranged on an indoor wall surface so that the detection area of the sensor is in a range in which the movement of an object other than a person is not detected.

Japanese Patent Laid-Open No. 11-9559

  However, the technique described in Patent Document 1 has a problem in operation when there are a plurality of people. That is, even if the person being watched over is not active, if another person is within the detection range of the sensor, it is recognized that the person being watched over is active due to the movement of the other person, and an emergency call is made May not be done.

  This invention is made | formed in view of such a condition, The objective is to provide the emergency call apparatus and emergency call system which operate | move suitably.

  In order to solve the above-described problems, an emergency call device according to an aspect of the present invention stores in advance a voice unit that acquires voice and permanent feature point information that does not depend on the health condition of the person being watched over or has low dependence. Permanent feature point information is extracted from the voice acquired by the target person information storage unit and the voice unit, and the extracted permanent feature point information is stored in the target person information storage unit. By comparing with the voice information comparison unit for authenticating whether or not the voice acquired by the voice unit is that of the person to be watched, and information on the person to be watched based on the authentication result of the voice information comparison part And a communication unit that outputs to the outside.

  It should be noted that any combination of the above-described constituent elements and a representation of the present invention converted between an apparatus, a method, a system, a program, a recording medium storing the program, etc. are also effective as an aspect of the present invention.

  ADVANTAGE OF THE INVENTION According to this invention, the emergency call apparatus and emergency call system which operate | move suitably can be provided.

It is a figure for demonstrating the emergency call system which concerns on embodiment of this invention. It is a figure for demonstrating the operation example of an emergency call system. It is a figure for demonstrating another operation example of an emergency call system. It is a figure for demonstrating the structure of an emergency call apparatus. It is a figure for demonstrating the detail of the storage information of a subject information storage part. It is a figure for demonstrating the detailed structure of an audio | voice information comparison part. It is a figure which shows the detail of the speaker's feature point information. It is a figure which shows an example of the status information of a monitoring subject stored in a subject information storage part. It is a figure which shows another example of the monitoring subject's state information stored in a subject information storage part. It is a figure which shows another example of the monitoring subject's state information stored in a subject information storage part. It is a figure which shows an example of the state continuation record information of a monitoring subject stored in a subject information storage part. It is a figure for demonstrating the structure of an emergency call center apparatus.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a diagram for explaining an emergency call system 100 according to an embodiment of the present invention. The emergency call system 100 is used for a monitoring service for an elderly person or the like. As shown in FIG. 1, in the emergency call system 100, the emergency call device 10 installed in the monitoring subject's house 101 and the emergency call center device 106 installed in the emergency call center 102 are connected via a communication network 103. It is connected.

  The emergency call system 100 according to the present embodiment makes a fixed call to the watch target person 110 in the watch target person's house 101 via the emergency call device 10, and the watch target person 110 responds to the call. Based on the above, personal authentication is performed, and the physical condition of the watching target person 110 is inferred from the tone of the voice, so that an appropriate watching service is implemented. The conventional monitoring service using the emergency call device is mainly targeted for the elderly living alone, but the emergency reporting system 100 is intended for not only the elderly living alone but also a plurality of elderly people and the like. In order to enable management of a plurality of persons to be watched over, the emergency call system 100 employs personal authentication technology.

  The emergency call device 10 shown in FIG. 1 shows the configuration of the front panel of the emergency call device 10. As shown in FIG. 1, an emergency call button 11, a consultation button 12, a cancel button 13, a speaker 14, and a microphone 15 are provided on the front panel of the emergency call device 10. The person 110 to be watched can make a hands-free call with an operator or the like in the emergency call center 102 by using the speaker 14 and the microphone 15 by pressing the emergency call button 11 or the consultation button 12. If the emergency call button 11 or the consultation button 12 is pressed by mistake, the notification is canceled by pressing the cancel button 13.

  As shown in FIG. 1, an external telephone 105 is connected to the emergency call device 10 in order to enable a call using a normal handset and a call to a destination other than a registered call destination. The emergency call device 10 is connected to a sensor 104 that monitors a fire or the like via wireless or wired communication. When any abnormality is detected by the sensor 104, a voice message and data report are made to an appropriate report destination.

  An example of the operation of the emergency call system 100 will be described. The emergency call device 10 is set to output from the speaker 14 “Welcome” in the morning when the person being watched over is at home. When the watching target person 110 answers "Good morning" to the output "Someday" from the speaker 14, as will be described later, authentication of the watching target person by voice, keyword recognition, and estimation of the physical condition are performed. When the response is from the person being watched over and the physical condition is determined to be good, the state of the person being watched over is determined to be “normal” and no emergency call is made to the emergency call center 102.

  On the other hand, if the response is the person being watched over and the physical condition is determined to be poor, the emergency call device 10 determines the state of the person being watched as “abnormal” and sends an emergency call to the emergency call center 102 registered in advance. To communicate the user's emergency situation. When there is no response or when it is determined that the response is not the person to be watched over, a notification to the effect that there is no response or is not the person to be watched is sent to the emergency call center 102.

  Also, for example, for those who are required to take medicine every three meals, a message such as “Did you eat?” Or “Did you take your medicine?” Is output from the emergency call device 10 If it is set and configured to recognize keywords such as “not yet”, “spent”, “drinked”, it is possible to perform watching according to the state of the person being watched over.

  FIG. 2 is a diagram for explaining another example of the operation of the emergency call system 100. In FIG. 2, there are a plurality of watching target persons (watching target persons 110 </ b> A and 110 </ b> B) in the watching target person's house 101. According to the emergency call system 100 according to the present embodiment, the monitoring service can be appropriately implemented even when there are a plurality of monitoring target persons in this way. As shown in FIG. 2, when the watching target person 110A answers “Good morning” and the watching target person 110B answers “Good morning” to the output from the speaker 14 simultaneously or successively, as will be described later. The voice information is separated for each speaker, and the target person is verified and the physical condition is estimated for each speaker. Therefore, according to this embodiment, even if there are a plurality of persons to be watched over, it is possible to suitably perform the watch service.

  FIG. 3 is a diagram for explaining still another example of the operation of the emergency call system 100. In the watching service, if the person being watched over cannot be confirmed for a certain period of time, an error may be reported. In this case, it is necessary not to report as abnormal when going out for a long time. As shown in FIG. 3, when a predetermined keyword (for example, “I will go”, “I will travel”, etc.) is voice input to the emergency call device 10, the emergency call device 10 recognizes that the person being watched has gone out. To do. As shown in FIG. 3, when the emergency call device 10 responds with “Taro-san, Hanako-san, I welcome you” when two couples talk to the emergency call device, they will go to the emergency call device. When the two people go out, it can be confirmed that the emergency terminal has recognized. In the emergency call system 100, even in the case of elderly couples who are not living alone, the home and outing states of each target person are recorded, and a message for confirming safety is not output while the watching target person is out.

  FIG. 4 is a diagram for explaining the configuration of the emergency call device 10. As shown in FIG. 4, the emergency call device 10 includes a communication unit 16, a voice unit 17, a control unit 18, an operation unit 19, a connection unit 20, and a display unit 21.

  The operation unit 19 is an interface of buttons that can be operated by the watching target person 110, and is connected to the emergency call button 11, the consultation button 12, the cancel button 13, and the like described in FIG. When the emergency call button 11 or the consultation button 12 is pressed, the operation unit 19 notifies the control unit 18 of the information. The control unit 18 controls the communication unit 16 so as to transmit a report content corresponding to a button operated to a previously registered report destination.

  The voice unit 17 is an interface for controlling a voice-related device, and is connected to the speaker 14 and the microphone 15 described in FIG. The audio unit 17 includes an A / D (analog / digital) conversion unit 22. The A / D converter 22 digitizes the voice information collected by the microphone 15, converts the digitized voice information into an analog signal, and outputs it through the speaker 14. The voice unit 17 manages output voice information (message) that is set in advance or can be arbitrarily registered. The voice unit 17 is configured to periodically call a predetermined message (for example, “Good morning”, “How is your physical condition”) to the person to be watched over.

  The connection unit 20 is an interface with an external connection device, and the sensor 104 described with reference to FIG. 1 is connected in the present embodiment. When the sensor 104 detects some abnormality such as a fire, the connection unit 20 notifies the control unit 18 of the information. The control unit 18 controls the communication unit 16 to transmit a voice message report, a data report, and the like to an appropriate report destination according to the reason that the sensor 104 is activated.

  The display unit 21 is an interface of the display device, and is connected to, for example, an emergency button lamp (not shown).

  The communication unit 16 communicates with a preset report destination and the emergency call center 102 via the communication network 103. The communication network 103 includes not only a telephone line but also an Internet protocol line. In connection with the Internet protocol line, in order to protect the emergency call device 10 from unintentional connection, protective measures are taken so that connection by the Internet protocol is not possible except for connection with a specific emergency call center 102.

  As shown in FIG. 4, the control unit 18 includes an audio information comparison unit 23 and a target person information storage unit 24.

  The subject information storage unit 24 depends on the sound feature point information (hereinafter referred to as “constant feature point information”) that does not depend on or is less dependent on the health condition of the person being watched over and on the health condition of the person being watched over. Voice feature point information (hereinafter referred to as “health feature point information”) to be stored. The permanent feature point information and the health feature point information are set when the emergency call device 10 is watched and installed in the subject person's home, for example.

  The voice information comparison unit 23 extracts the permanent feature point information from the voice acquired by the voice unit 17 via the microphone 15, and the watched target stored in the target person information storage unit 24 with the extracted permanent feature point information. By comparing with the permanent feature point information (template) of the person, it is authenticated whether or not the voice acquired by the voice unit 17 belongs to the person being watched over.

  The voice information comparison unit 23 extracts health feature point information from the voice acquired by the voice unit 17 via the microphone 15, and the extracted health feature point information is stored in the target person information storage unit 24. The physical condition of the person being watched over is estimated by comparing with the health feature point information (template) of the person. Extraction of feature point information from speech can be performed by a method such as spectrum analysis.

  Based on the authentication result of the voice information comparison unit 23 and the estimation result of the voice information comparison unit 23, the communication unit 16 outputs information related to the watching target person to the emergency call center device 106 of the emergency call center 102.

  FIG. 5 is a diagram for explaining the details of the stored information in the target person information storage unit 24. In FIG. 5, as an example, stored information in the case where there are two watching subjects (watching subjects A and B) is shown. As shown in FIG. 5, the target person information storage unit 24 stores information for each target person to watch. Here, the watching target person A will be described as an example, but the same information is also stored for the watching target person B.

  The target person information storage unit 24 stores the constant feature point information A1 as the first information regarding the watching target person A. As described above, the “constant feature point information” is voice feature point information that does not depend on the health condition of the person being watched over or has a low degree of dependency. Status). The permanent feature point information A1 includes feature point information that does not depend on a specific keyword. By comparing the permanent feature point information A1 with the voice permanent feature point information acquired by the voice unit 17, it is possible to authenticate the person being watched at home or absent.

  In the present embodiment, the permanent feature point information A1 further includes permanent feature point information related to a specific keyword (for example, keywords X, Y, Z...). It is possible to perform identity authentication without depending on a specific keyword. However, in the present embodiment, in order to appropriately recognize keywords (for example, “good morning”, “we will go”, etc.) for controlling the emergency call device 10, constant feature point information which does not depend on a specific keyword. In addition, permanent feature point information for each keyword is stored.

  Moreover, the subject information storage unit 24 stores health feature point information A2 when the health status of the watching target person A is normal as the second information regarding the watching target person A. As described above, the “health feature point information” is voice feature point information that changes in accordance with the health condition of the person being watched over, and includes, for example, the pitch of the voice, the interval between phonemes, the degree of voice blur. By comparing the health feature point information A2 with the voice health feature point information acquired by the voice unit 17, the health condition of the person being watched over can be estimated.

  In addition, the target person information storage unit 24 stores the latest state information A3 regarding the monitoring target person A such as the health condition, staying at home, absence, etc. as the third information regarding the watching target person A.

  In addition, the target person information storage unit 24 stores time-series history information of the state information of the watching target person A (hereinafter referred to as “state continuation record information A4”) as the fourth information regarding the watching target person A. To do. The state continuation record information A4 is stored in a certain amount (for example, for one week) together with the information acquisition time.

  FIG. 6 is a diagram for explaining a detailed configuration of the voice information comparison unit 23. The input voice information converted into a digital signal by the A / D conversion unit 22 of the voice unit 17 is delivered to the voice information comparison unit 23 of the control unit 18.

  As shown in FIG. 6, the speech information comparison unit 23 includes a speaker separation unit 40, a feature point separation unit 41, an acquired speech information feature point storage unit 42, and a feature point comparison unit 43.

  The speaker separation unit 40 analyzes whether or not a plurality of speakers are included in the voice information input from the A / D conversion unit 22. As a result of the analysis, when it is determined that the voice information of a plurality of speakers is included, the speaker separation unit 40 separates the input voice information for each speaker and delivers it to the feature point separation unit 41. The feature point separation unit 41 extracts feature points from the speech information separated for each speaker, and delivers the extracted feature points for each speaker to the acquired speech information feature point storage unit 42. The acquired voice information feature point storage unit 42 stores the feature point information for each speaker delivered from the feature point separation unit 41. FIG. 6 illustrates, as an example, a state in which the feature point information of the speaker α and the feature point information of the speaker β are stored.

  FIG. 7 shows details of the feature point information of the speaker α. As shown in FIG. 7, the acquired voice information feature point storage unit 42 includes permanent feature point information α1 of the speaker α and health feature point information α2 of the speaker α as the feature point information of the speaker α. Stored separately.

  Returning to FIG. 6, the feature point comparison unit 43 stores the constant feature point information stored in the acquired voice information feature point storage unit 42 for each speaker by the watching target person stored in the target person information storage unit 24. Compared with the permanent feature point information, it is authenticated whether or not the acquired voice belongs to the person being watched over, and a predetermined keyword is recognized. Also, the feature point comparison unit 43 monitors the health feature point information stored in the acquired voice information feature point storage unit 42 for each speaker, and stores the health feature point information for each watch target stored in the subject information storage unit 24. And estimate the health status of the acquired speech speaker. Then, the feature point comparison unit 43 records the authentication result and the estimation result in the state information (A3, B3, etc.) of the watching target person and the state continuation record information (A4, B4, etc.) in the subject person information storage unit 24.

  FIG. 8 shows an example (example 1) of the status information of the watching target person stored in the target person information storage unit 24. In Example 1, the current time is May 16, 2013, 10:00 am. If safety confirmation is carried out every day, it is time for the voice unit 17 of the emergency call device 10 to call for safety confirmation, but the keyword “I will go” is confirmed at 10:15 on May 15th. Since it is recorded that the person X to be watched is out of the office and the health condition is normal, it is not necessary to confirm the watch by calling. The control unit 18 of the emergency call device 10 uses the state information of the watching target person X stored in the target person information storage unit 24 to control the voice unit 17 so as not to make a call. At the time when the outing state is changed to the home state, the control unit 18 resumes watching confirmation by calling.

  FIG. 9 shows another example (example 2) of the status information of the watching target person stored in the target person information storage unit 24. In this example 2, since the watching target person Y is at home, the voice unit 17 of the emergency call device 10 performs a call for watching (for example, “Good morning”). In this example 2, three calls are made and unconfirmed at 10:00 is recorded in the status information of the watching target person Y. In this case, since it is presumed that some kind of abnormality has occurred in the watching target person Y, the control unit 18 notifies the predetermined reporting destination registered in advance that the watching target person Y is in an abnormal state. The communication unit 16 is controlled.

  FIG. 10 shows still another example (example 3) of the status information of the watching target person stored in the target person information storage unit 24. In this example 3, since the watching target person Z is at home, the voice unit 17 of the emergency call device 10 performs a call for watching (for example, “Good morning”). When the watching target person Z responds to the call of the emergency call device 10, the feature point comparing unit 43 of the voice information comparing unit 23 is a constant feature point of the plurality of watching target persons stored in the target person information storing unit 24. The information (see FIG. 5) is compared with the permanent feature point information (see FIG. 7) stored in the acquired voice information feature point storage unit 42, and it is confirmed that the responder to the challenge is the watching target person Z. Certify. The feature point comparison unit 43 of the voice information comparison unit 23 is stored in the health feature point information (see FIG. 5) stored in the target person information storage unit 24 and the acquired voice information feature point storage unit 42. The health feature point information (see FIG. 7) is compared, and the physical condition of the watching target person Z is estimated. Based on the result of the authentication and estimation, the feature point comparison unit 43 updates the identity confirmation and health status of the status information. In Example 3 shown in FIG. 10, the person to be watched over is confirmed by the keyword “good morning”, is home, and the health state is estimated to be abnormal. In this case, when the health state is registered as abnormal, the control unit 18 of the emergency call device 10 controls the communication unit 16 so as to report an abnormal situation to a preset report destination.

  FIG. 11 shows an example of the state continuation record information of the watching target person stored in the target person information storage unit 24. In the status continuation record information of the watching target person, the monitoring target person authentication result and the comparison result of the parameters for health condition confirmation are recorded every time the watching target person authentication is performed, and a certain amount is stored in the emergency call device 10. The parameters recorded in this example are the average pitch of speech, the phoneme interval, the degree of blur. These parameters are health feature point information that depends on the health condition of the person being watched over, and the values compared with those of the person being watched in the normal state (that is, the health state) are recorded. Is guessed. In this example, the judgment of the parameters when going out on May 14, 2013 and when returning home is normal, and the next morning, at 9:30 on May 15, 2013, the keyword “Good morning” is used to identify the person being monitored. This indicates that the judgment by the whole parameter is abnormal. By simply recording the comparison values of time-series parameters in addition to simply determining whether the abnormality is normal or not, an expert (for example, an operator of the emergency call center 102) can perform a determination based on time-series information. .

  FIG. 12 is a diagram for explaining the configuration of the emergency call center device 106. As shown in FIG. 12, the emergency call center device 106 includes a center side control unit 120 and a center side communication unit 121. The center side control unit 120 includes a center side target person information storage unit 122 and a center side audio information comparison unit 123.

  In the emergency notification system 100 according to the present embodiment, not only the monitoring target person authentication in the emergency notification apparatus 10 but also the emergency notification center 102 analyzes the voice of the monitoring target person who is responding to the emergency notification apparatus 10, and the emergency notification system 102 Compared with the data of the person to be watched stored in the reporting center device 106, it is possible to carry out appropriate watching.

  In the emergency call system 100, the call from the emergency call device 10 to the person to be watched is a call by a voice message built in the voice unit 17 of the emergency call device 10 and an operator from the emergency call center 102. Thus, both voice calls output from the emergency call device 10 are possible. As the watch target authentication by the response to the challenge, both the authentication at the emergency call device 10 and the authentication at the emergency call center device 106 can be used, but generally the CPU equipped in the emergency call device 10 is an emergency call. It is assumed that the capability is lower than that of the CPU of the center device 106. Therefore, in the present embodiment, voice information is sent from the emergency call device 10 to the emergency call center device 106, and the emergency call center device 106 can also perform voice analysis. For example, when it is more difficult to analyze the response of the person being watched than usual due to a serious cold, etc., and when a high CPU capacity is required for the person being watched for authentication, the emergency call center device 106 keeps track of it. Target person authentication is implemented.

  In order to enable voice analysis at the emergency call center device 106, the communication unit 16 of the emergency call device 10 is configured to transmit the voice information acquired by the voice unit 17 to the emergency call center device 106. The voice information transmitted from the communication unit 16 of the emergency call device 10 is received by the center side communication unit 121 of the emergency call center device 106 via the communication network 103.

  As illustrated in FIG. 12, the center-side target person information storage unit 122 of the center-side control unit 120 stores in advance permanent feature point information and health feature point information of the watching target person. Since the monitoring target person is limited for each emergency call device, the center-side target person information storage unit 122 manages the permanent feature point information and health feature point information of the watch target person for each emergency call device. Each emergency call device is assigned an identification number (for example, 0001, 0002, etc.), and when communication with the emergency call device is started, watching target person information stored in the identification number of the emergency call device in communication Is referenced.

  The center-side voice information comparison unit 123 extracts the permanent feature point information from the voice acquired from the emergency call device 10, and the watched person stored in the center-side target person information storage unit 122 with the extracted permanent feature point information. By comparing with the permanent feature point information, it is authenticated whether or not the voice acquired from the emergency call device 10 belongs to the person being watched over.

  Further, the center side audio information comparison unit 123 extracts health feature point information from the voice acquired from the emergency call device 10, and the extracted health feature point information is stored in the center side target person information storage unit 122. The physical condition of the person being watched over is estimated by comparing with the health feature point information.

  In the present embodiment, the conversion of the voice data from analog to digital is executed by the emergency call device 10 (because the frequency band that can be used for transferring voice information through a telephone line is limited). However, if feature points must be extracted through the operator, such as advanced voice data processing and extraction of the difference between the response of one subject and multiple responses, such as extracting feature points of multiple speakers Extracts the voice feature points of the person being watched over by the emergency call center device 106, and the extracted result is stored in the center-side target person information storage unit 122. Alternatively, the extracted result may be stored in the acquired voice information feature point storage unit 42 of the voice information comparison unit 23 of the emergency call device 10.

  Further, the emergency call center device 106 provides more detailed watching service by acquiring the status continuation record information of the watching target person from the emergency calling device 10 installed in the watching target person's house 101 as necessary. Can do.

  The embodiment has been described in detail above. With the technique described in the above-mentioned Patent Document 1, it is possible to avoid malfunctions caused by small animals and the like, but it is not possible to avoid erroneous recognition due to movements of others. The safety confirmation service of the emergency call service makes an emergency call when there is no activity of the person being watched over for a certain period of time, so there is a risk that the action of the target person will be mistakenly watched and the emergency call will not be made. On the other hand, in the emergency call system 100 according to the present embodiment, since the activity of the person being watched over is grasped by performing voice authentication, the activity of the person being watched over is not misunderstood as the person being watched over, and a more reliable watch service. Can be implemented. Moreover, in this embodiment, in order to perform watching target person authentication using the voice acquired from the conversation with the watching target person who is a part of the operation of the conventional emergency call system, More reliable watching services can be implemented.

  Further, in the present embodiment, since it is configured to separate the voices of a plurality of speakers and manage the acquired voices by separating feature points, when a plurality of watching target persons respond or utter simultaneously or continuously Even so, it is possible to properly perform watching target person authentication. Therefore, according to the present embodiment, it is possible to monitor a plurality of monitoring target persons with one emergency call device 10.

  In the present embodiment, the voice feature point information is separated into the feature point information of the input voice into the permanent feature point information for watching target person authentication and the health feature point information for physical condition estimation. As a result, the accuracy of watching target person authentication can be improved and the accuracy of safety confirmation can be improved.

  Conventionally, a method for managing home / outing by using a tablet key is known, but in this method, when an elderly couple is watching, one of the couple goes out and the out-going tablet key is put out. If this happens, there is a risk that the remaining safety check will stop. On the other hand, in this embodiment, since at-home / outing states of a plurality of watching target persons are monitored and recorded for each target person, the above-described danger can be avoided.

  Conventionally, there is a known method for confirming safety based on detection of an act unrelated to the intentional act of the person such as detection of a passive sensor, or detection of an act of unknown to the person such as detection of water flow rate. . On the other hand, in this embodiment, safety is confirmed by a detection method involving voice authentication for an intentional action of the person being watched over, that is, an action of the person who responds to a call. Therefore, the state of the person can be surely grasped, and the danger that an emergency call is not made due to misunderstanding that the person is active by the action of another person can be avoided. That is, according to this embodiment, the accuracy of the safety service of the emergency call device 10 and the emergency call system 100 can be improved compared to the conventional safety confirmation method.

  Conventionally, there is a known method of continuously monitoring the amount of exercise using a passive element such as a passive sensor to watch over the health condition of the person being watched over. There is no guarantee. In addition, when there are a plurality of watching subjects such as elderly couples, the amount of exercise of the plurality of subjects cannot be separated by detection using a passive element. On the other hand, in this embodiment, since the physical condition of the watching target person is estimated after authentication as to whether or not it is the watching target person, it is possible to more accurately monitor the physical condition states of the plurality of watching target persons.

  Further, in the emergency call device 10 according to the present embodiment, the physical condition is estimated for each watching target person authentication, the judgment record is recorded in the watching target person's state information, and the watching target person's state continuation record information is the watching target. Parameter verification records for each person authentication are recorded in time series. These pieces of information can be used for more detailed judgment by an expert (for example, an operator of the emergency call center 102), so that a more detailed monitoring service is possible.

  In the watching service, human contact by conversation between the watching target person and the operator of the watching service provider is also an important element of the service. It is important to reduce the burden on the operator of the emergency call center 102 and improve the accuracy of the remote monitoring service. According to the present embodiment, it is possible to manage the states of a plurality of watching target persons with one emergency call device 10 by performing voice authentication of the watching target person. The operator of the emergency call center 102 can talk with the target person while grasping the state of the target person. Accordingly, it is possible to improve the accuracy of the remote monitoring service while reducing the load on the operator of the emergency call center 102.

  The present invention has been described based on the embodiments. This embodiment is an exemplification, and it will be understood by those skilled in the art that various modifications can be made to combinations of the respective constituent elements and processing processes, and such modifications are also within the scope of the present invention. is there.

  DESCRIPTION OF SYMBOLS 10 Emergency call device, 11 Emergency call button, 12 Consultation button, 13 Cancel button, 14 Speaker, 15 Microphone, 16 Communication part, 17 Voice part, 18 Control part, 19 Operation part, 20 Connection part, 21 Display part, 22 A / D conversion unit, 23 voice information comparison unit, 24 target person information storage unit, 40 speaker separation unit, 41 feature point separation unit, 42 acquired voice information feature point storage unit, 43 feature point comparison unit, 100 emergency call system, 101 emergency call center, 102 emergency call center, 103 communication network, 104 sensor, 106 emergency call center device, 110 watch target person, 120 center side control unit, 121 center side communication unit, 122 center side target person information storage unit, 123 Center-side audio information comparison unit.

Claims (12)

An emergency call device used to watch over the person being watched over,
An audio part for acquiring audio;
A subject information storage unit that stores in advance permanent feature point information that does not depend on the health condition of the person being watched over or has a low dependency;
By extracting the permanent feature point information from the voice acquired by the voice unit, and comparing the extracted permanent feature point information with the permanent feature point information of the person being watched stored in the subject information storage unit. A voice information comparison unit that authenticates whether or not the voice acquired by the voice unit belongs to the person being watched over,
Based on the authentication result of the voice information comparison unit, a communication unit that outputs information about the person being watched to the outside,
An emergency call device comprising:   The emergency call device according to claim 1, wherein the voice information comparison unit separates the voice acquired by the voice unit for each speaker and authenticates the person to be watched for each speaker.   The emergency notification device according to claim 1, wherein the permanent feature point information includes a configuration of a high long wave with respect to a basic voice frequency. The subject information storage unit further stores in advance health feature point information that depends on the health status of the person being watched over,
The voice information comparison unit extracts health feature point information from the voice acquired by the voice unit, and the health feature point information of the monitoring target person stored in the target person information storage unit with the extracted health feature point information; By comparing, we estimate the physical condition of the person being watched over,
The communication unit outputs information on the person to be watched to the outside based on the estimation result of the voice information comparison unit.
The emergency call device according to any one of claims 1 to 3.   5. The emergency call device according to claim 4, wherein the health feature point information includes at least one of a pitch of a voice, an interval between phonemes, and a blurring degree of a voice.   The emergency call device according to any one of claims 1 to 5, wherein the voice unit is configured to periodically call a predetermined message to a person to be watched over.   The emergency notification device according to claim 6, wherein when there is no response from the person being watched over to the voice unit, the communication unit reports an abnormal situation to a predetermined report destination.   8. The emergency call according to claim 6, wherein the voice unit is configured not to periodically call a message when the watching target person is out of the watching target person's house. 9. apparatus.   The emergency call according to any one of claims 1 to 8, wherein the target person information storage unit is configured to record, in a time series, information related to the watching target determined by the voice information comparison unit. apparatus. The emergency call device according to any one of claims 1 to 9,
An emergency call center device connected to the emergency call device via a communication network;
An emergency call system characterized by comprising: The communication unit of the emergency call device is configured to transmit voice information acquired by the voice unit to the emergency call center device,
The emergency call center device
A center-side target person information storage unit that stores permanent feature point information of the target person to be watched in advance;
The permanent feature point information is extracted from the voice acquired from the emergency call device, and the extracted permanent feature point information is compared with the permanent feature point information of the watching target person stored in the center-side target person information storage unit. Thus, the center side voice information comparison unit for authenticating whether or not the voice acquired from the emergency call device belongs to the person being watched over,
The emergency call system according to claim 10, further comprising: The center side subject person information storage unit further stores in advance the health feature point information of the person being watched over,
The center side voice information comparison unit extracts health feature point information from the voice acquired from the emergency call device, and the extracted health feature point information is stored in the center side target person information storage unit. Estimate the physical condition of the person being watched over by comparing with the feature point information.
The emergency call system according to claim 11.

Images