Monitoring system, monitoring method, and program

Abstract

To enable enhancing accuracy of supervision accompanied by continuity of the supervision.SOLUTION: A supervising system is configured to: convert sensor activity data based on output of a sensor detecting a supervision object as an object into text activity data; determine an object activity targeted for supervision on the basis of the text activity data; count the number of times in which an activity of the text activity data corresponding to the determined object activity of the text activity data an activity data storage unit stores has been performed during a unit period; transmit a report reporting on the activity in which a count value is greater than or equal to a threshold to a supervisor terminal; and cause an object activity determination model to be machine-learned so that a content of feedback information indicative of whether the activity the report reports is necessary is reflected on a determination result of the object activity.SELECTED DRAWING: Figure 8

Description

[Technical Field]

[0001] The present invention relates to a monitoring system, a monitoring method, and a program. [Background Art]

[0002] There is known a technology configured to accumulate detection data obtained by detecting the usage status of daily use equipment and facilities, register the life activity pattern of a monitored person based on the accumulated detection data, and detect an abnormality based on a result of comparing the detection data with the registered life activity pattern (see, for example, Patent Document 1). [Prior Art Documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Unexamined Patent Publication No. 2004-133777 [Summary of the Invention] [Problem to be Solved by the Invention]

[0004] As a monitoring system for monitoring a person to be monitored with a sensor, for example, it is preferable that the monitoring accuracy is improved in the process of continuing monitoring.

[0005] Accordingly, an object of the present invention is to make it possible to improve monitoring accuracy as monitoring continues. [Means for Solving the Problem]

[0006] One aspect of the present invention that solves the above-mentioned problems is a monitoring system comprising: a conversion unit that acquires text behavior data by converting sensor behavior data, which is the behavior of a person under monitoring indicated by the output of a sensor provided to detect a predetermined event for a person under monitoring, into text; a behavior data storage unit that stores the text behavior data; a reference behavior information storage unit that stores reference behavior information including a threshold for the number of times a predetermined behavior is performed in a predetermined unit period; a target behavior determination unit that determines the target behavior to be monitored based on the text behavior data acquired by the conversion unit; a counting unit that counts the number of times the behavior indicated by the text behavior data, which corresponds to the target behavior determined by the target behavior determination unit, has been performed in the unit period among the text behavior data stored in the behavior data storage unit corresponding to the person under monitoring; a report processing unit that transmits a report reporting on behaviors for which the count value from the counting unit is greater than or equal to a threshold to the monitor terminal of the monitor monitoring the person under monitoring; and a learning processing unit that causes the target behavior determination unit to learn so that the content of feedback information indicating the necessity of the behavior reported in the report is reflected in the determination result of the target behavior determination unit.

[0007] One aspect of the present invention is a monitoring method in a monitoring system, comprising: a conversion step of acquiring text behavior data by converting sensor behavior data, which is the behavior of a person under monitoring indicated by the output of a sensor provided to detect predetermined events targeting the person under monitoring, into text; a target behavior determination step of determining target behaviors to be monitored based on the text behavior data acquired in the conversion step; a counting step of counting the number of times the behavior indicated by the text behavior data, which corresponds to the target behavior determined in the target behavior determination step, has been performed in a unit period indicated by reference behavior information stored in a reference behavior information storage unit, among the text behavior data stored in the behavior data storage unit corresponding to the person under monitoring; a report processing step of transmitting a report reporting on behaviors whose count value in the counting step is greater than or equal to a threshold indicated by reference behavior information stored in the reference behavior information storage unit to the monitor terminal of the monitor monitoring the person under monitoring; and a learning processing step of training a machine learning model corresponding to the target behavior determination step so that the content of feedback information indicating the necessity of the behavior reported in the report is reflected in the determination result of the target behavior determination step.

[0008] One aspect of the present invention is a program for a computer in a monitoring system to function as a learning processing unit that causes the target behavior determination unit to learn so that the content of feedback information indicating the necessity of the behavior reported by the report is reflected in the determination result of the target behavior determination unit, among the text behavior data stored in the behavior data storage unit corresponding to the target behavior determined by the target behavior determination unit, from among the text behavior data stored in the behavior data storage unit corresponding to the target behavior, the number of times the behavior indicated by the text behavior data was performed in a unit period indicated by the reference behavior information stored in the reference behavior information storage unit, a report processing unit that reports on behaviors whose count value by the count unit is greater than or equal to a threshold indicated by the reference behavior information stored in the reference behavior information storage unit, and transmits a report to the monitor terminal of the monitor who monitors the target person, [Effects of the Invention]

[0009] According to the present invention, the effect is obtained that the accuracy of monitoring can be improved as monitoring continues. [Brief explanation of the drawing]

[0010] [Figure 1] This figure shows an example of the overall configuration of the monitoring system according to this embodiment. [Figure 2] This figure shows an example of the functional configuration of the monitoring server according to this embodiment. [Figure 3] This figure shows an example of user information according to this embodiment. [Figure 4] This figure shows an example of emergency registration information according to this embodiment. [Figure 5] This figure shows an example of behavioral data according to this embodiment. [Figure 6]This figure shows an example of reference behavior information according to this embodiment. [Figure 7] This flowchart shows an example of the processing procedure that the monitoring server according to this embodiment performs in relation to monitoring a person under monitoring. [Figure 8] This flowchart shows an example of a processing procedure for updating the target behavior decision model between the observer terminal and the observation server according to this embodiment. [Modes for carrying out the invention]

[0011] <Embodiment> Figure 1 shows an example of the overall configuration of the monitoring system of this embodiment. The monitoring system in the figure detects the daily activities of the person being monitored P who resides in the living space HS, and notifies the monitor terminal 200 used by the monitor W according to the detected activities. The monitor W may be a relative of the person being monitored P, or a caregiver, etc. In addition, notifications may be sent to the terminal of the person being monitored P according to the activities detected targeting the person being monitored P.

[0012] In the living space HS, one or more sensors 10 are provided. Each sensor 10 detects different events corresponding to predetermined daily activities of the person being monitored P in the living space HS. Specifically, the sensor 10 may include a camera. The camera-based sensor 10 may be installed in various locations within the living space HS, such as the living room, entrance hall, kitchen, and bedroom. The camera-based sensor 10 may output the image obtained through imaging as detection information.

[0013] Sensor 10 may also detect predetermined operations or states of equipment installed in the living space HS. For example, such a sensor 10 may detect states such as turning on, turning off, and illuminance of a room lighting fixture. Such a sensor 10 may also detect the open / closed state of a water faucet. Furthermore, such a sensor 10 may detect the open / closed state of doors, sliding doors, and the like. The sensor 10 may also be installed in a stove or the like to detect an ignition state, heating power, and the like. Additionally, the sensor 10 may be an electrical appliance such as a refrigerator or an air conditioner. Sensor 10 may also include portable terminals and wearable devices that move with the monitored person P, such as smartphones, smartwatches, and glasses-type devices. Sensor 10 may also include a heat-sensitive sensor or the like that measures the body temperature of the monitored person P. Sensor 10 may also be provided with a load sensor or the like that is installed on a floor, bed, or the like to detect a load applied by the monitored person P.

[0014] Detection information obtained by detection of each sensor 10 is transmitted from the gateway 20 to the monitoring server 100 via a network. The detection information transmitted from the sensor 10 to the monitoring server 100 includes a monitored person ID that uniquely identifies the corresponding monitored person P.

[0015] The monitoring server 100 executes processing corresponding to monitoring of the monitored person P based on the detection information transmitted from the sensor 10. In the figure, the sensor 10 in the living space HS corresponding to one monitored person P is shown, but the monitoring server 100 may correspond to transmission of detection information from the corresponding sensor 10 for each of a plurality of monitored persons P, and may be capable of executing processing according to monitoring for each of the plurality of monitored persons P.

[0016] The monitoring server 100 determines whether an emergency has occurred to the monitored person P based on detection information transmitted from the sensor 10. The emergency to be determined may be defined in advance. When the monitoring server 100 determines that an emergency has occurred, it is configured to transmit a notification that an emergency has occurred to the monitored person P to the monitor terminal 200 of the monitor W corresponding to the monitored person P. This allows the monitor W to recognize that an emergency has occurred to the monitored person P and take immediate action even when the monitor W is located at a place distant from the monitored person P.

[0017] In addition, in a normal state where it is determined that no emergency has occurred, the monitoring server 100 generates behavior data (sensor behavior data) based on the detection information transmitted from the sensor 10. For example, the sensor behavior data is information indicating a behavior pattern of the monitored person P, such as "turning on the faucet in the kitchen to wash dishes". The monitoring server 100 accumulates (stores) the behavior data for each monitored person P. When accumulating the behavior data, the monitoring server 100 accumulates text behavior data obtained by converting the sensor behavior data into text.

[0018] The monitoring server 100 uses the accumulated behavior data to create a monitoring result report for the monitored person P, and transmits the created report to the monitor terminal 200. The content of the monitoring result reported in the report may, for example, present behaviors of the monitored person P that do not lead to an emergency but are determined to have signs of physical or mental frailty, or are determined to correspond to frailty.

[0019] In addition, the transmission of the report from the monitoring server 100 to the monitor terminal 200 may be performed, for example, as an email, or may be transmission of a report screen displayed by a monitoring application installed in the monitor terminal 200. By periodically transmitting such a report to the monitor terminal 200, the monitor W can accurately grasp the status of the monitored person P.

[0020] Because there are individual differences among the people being monitored (P), the content that needs to be reported will also differ for each person being monitored (P). For example, the same behavior may require reporting in the case of one person being monitored (P), but not in the case of another person being monitored (P) because it does not pose a particular problem. Therefore, in this embodiment, the monitoring server 100 allows the monitor W, after reviewing the contents of the transmitted report, to specify whether the reported action is necessary or not using the monitor terminal 200, and to send feedback information indicating the necessity of the specified action to the monitoring server 100. The monitoring server 100 learns, based on the transmitted feedback information, which actions of the monitored person need to be reported and which do not, and then reports only the actions that need to be reported. Through this learning process, it becomes possible to provide reports that describe what needs to be reported for each person being monitored P. As a result, in this embodiment, it becomes possible to improve the accuracy of monitoring as monitoring of person P continues.

[0021] Figure 2 shows an example of the functional configuration of the monitoring server 100 in this embodiment. The monitoring server 100 in the figure comprises a communication unit 101, a control unit 102, and a storage unit 103. The communication unit 101 is a component that performs communication via a network. The communication unit 101 can communicate with the sensor 10 via the gateway 20 and with the observer terminal 200.

[0022] The control unit 102 performs various controls on the monitoring server 100. The control unit 102 includes an emergency determination unit 121, a conversion unit 122, a target action determination unit 123, a counting unit 124, a report processing unit 125, and a learning processing unit 126.

[0023] The emergency determination unit 121 acquires behavioral data based on detection information received by the communication unit 101 from sensors 10 in the living space HS corresponding to one person being monitored P. For example, the emergency determination unit 121 can extract image data showing the person being monitored P acting in the kitchen for cooking from image data output as detection information by a sensor 10 acting as a camera installed in the kitchen of a house in the living space HS, and acquire it as behavioral data. The emergency determination unit 121 compares the acquired behavioral data with events registered as emergencies in the emergency registration information stored in the emergency registration information storage unit 132. Based on the comparison results, the emergency determination unit 121 determines whether or not the behavior shown in the acquired behavioral data constitutes an emergency. If the emergency determination unit 121 determines that an emergency has occurred, it sends a notification of the occurrence of an emergency (emergency notification) to the corresponding monitor terminal 200. The emergency notification may also include, for example, image data acquired as behavioral data.

[0024] The conversion unit 122 performs a process (text conversion) to convert the behavioral data based on the detection information of the sensor 10 into text. Through text conversion, behavioral data (text behavioral data) is obtained that indicates the behavior through, for example, sentence structure. For example, the conversion unit 122 may perform image recognition processing on behavioral data from captured image data showing that the person being monitored P has left the faucet lever open after washing their hands in the washroom, and then perform text conversion to generate text such as "Forgot to turn off the faucet after washing hands." Furthermore, the conversion unit 122 generates tags indicating characteristics of behavior from the text behavior data obtained through text conversion, and attaches the generated tags to the text behavior data. For example, in the case of the text "Forgetting to turn off the faucet after washing hands," the conversion unit 122 may generate the tags "#handwashing," "#faucet," and "#forgettingtoturnoff." By attaching these generated tags to the text behavior data, the text behavior data can be searched efficiently. Specifically, by performing a search using "forgetting to turn off" as the search key, text behavior data tagged with "#forgettingtoturnoff," such as "Forgetting to turn off the faucet after washing hands" and "Forgetting to turn off the gas stove," can be searched together.

[0025] The target behavior determination unit 123, in creating a report, executes a process to determine which behaviors (target behaviors) can be described in the report from among the behaviors indicated by the behavior data stored in the behavior data storage unit 133 corresponding to the target person P being monitored. Such target behaviors become the behaviors that the counting unit 124 counts. Target behaviors include, for example, behaviors that do not reach the level of an emergency, but are different from normal and should be noted when observing the progress of the person P being monitored. The target action determination unit 123 can, for example, use a machine learning model constructed by machine learning the relationship between a group of behavioral data (text behavioral data) accumulated in response to one person P under observation over a certain unit period, and the behavioral data that should be selected as the target of action. Furthermore, when the target action determination unit 123 outputs the result of the target action determination, it may output it as text action data in the same format as obtained by the conversion unit 122.

[0026] The counting unit 124 counts the number of times each of the target actions determined by the target action determination unit 123 is performed within a unit period. The counting unit 124 also compares the number of times each target action has been counted with a threshold value set for the corresponding target action stored in the reference action information storage unit 134.

[0027] The report processing unit 125 creates a report as part of its report processing and sends the created report to the observer terminal 200. When creating the report, the report processing unit 125 creates the report by describing the target behaviors that were counted by the counting unit 124 more than or equal to a threshold.

[0028] As described above, the monitor terminal 200 can send feedback information to the monitoring server 100 indicating whether the actions described in the received report are necessary or not. The learning processing unit 126 trains the target action decision unit 123 so that it reflects whether or not the action indicated by the transmitted feedback information is necessary.

[0029] The storage unit 103 stores various types of information corresponding to the monitoring server 100. The storage unit 103 includes a user information storage unit 131, an emergency registration information storage unit 132, an action data storage unit 133, and a reference action information storage unit 134.

[0030] The user information storage unit 131 stores user information. Here, "user" may be a concept that includes the person being monitored P and the monitor W.

[0031] The emergency registration information storage unit 132 stores emergency registration information. Emergency registration information is information that registers actions defined as an emergency.

[0032] The behavioral data storage unit 133 stores behavioral data for each monitored person P that should be accumulated for report creation. The behavioral data stored in the behavioral data storage unit 133 may be text behavioral data obtained by converting it to text by the conversion unit 122.

[0033] The reference behavior information storage unit 134 stores reference behavior information. The reference behavior information is information that indicates a threshold value to be used as a comparison point with the number of times counted by the counting unit for each behavior defined as a possible behavior for the person being monitored P.

[0034] Figure 3 shows an example of user information stored by the user information storage unit 131 for one user (where "user" is a concept that includes the person being monitored P and the corresponding caregiver W). The user information in this figure includes areas for person being monitored information, registered sensor information, caregiver information, and caregiver terminal information. The area containing the information of the person being monitored stores information about the person being monitored P who corresponds to the user in question. The information of the person being monitored may include a person ID that uniquely identifies the person being monitored P, the name of the person being monitored, contact information, and other such information. The registered sensor information area contains information about a sensor 10 that has been registered as being installed in the living space HS corresponding to the person being monitored P. The registered sensor information may include the address of the sensor 10 on the network, unique identification information for the sensor 10 (e.g., serial number, MAC address, etc.), and the name of the sensor 10. The area containing the caregiver information stores caregiver information for caregiver W, who corresponds to the user in question. The caregiver information may include a caregiver ID that uniquely identifies the caregiver W, the caregiver W's name, contact information, and other relevant information. The area containing the caregiver terminal information stores caregiver terminal information relating to the caregiver terminal 200 used by the corresponding caregiver W. The caregiver terminal information may include, for example, an email address that can be received on the caregiver terminal 200 and the application ID of the caregiver application.

[0035] Figure 4 shows an example of emergency registration information stored in the emergency registration information storage unit 132. The emergency registration information in this figure has a structure in which information indicating the definition content is associated with a unique event ID for each event defined as an emergency.

[0036] Figure 5 shows an example of behavioral data stored by the behavioral data storage unit 133. As shown in the figure, the behavioral data storage unit 133 stores one or more behavioral dates and times and text behavioral data associated with each monitored person ID that represents the monitored person P. In other words, depending on the behavioral data stored by the behavioral data storage unit 133, it is possible to see when and what kind of actions each monitored person P took.

[0037] Figure 6 shows an example of reference behavior information stored in the reference behavior information storage unit 134. The reference behavior information in this figure has a structure in which a threshold is associated with each reference behavior data. The reference behavior data area stores the reference behavior data. The reference behavior data is behavior data that represents a single defined behavior. The reference behavior data may be in the same format as the text behavior data stored in the behavior data storage unit 133. The threshold area stores the threshold value set for the number of times the corresponding reference behavior data is counted by the counting unit. The threshold value in the reference behavior information is determined in accordance with a predetermined unit period such as a day, week, or month.

[0038] Referring to the flowchart in Figure 7, an example of the processing procedure that the monitoring server 100 performs in relation to monitoring a person P who is a person being monitored and who is a person being monitored, corresponding to one user, will be explained. Step S100: In the monitoring server 100, the emergency determination unit 121 acquires sensor behavior data, for example, at predetermined intervals, based on detection information transmitted from sensors 10 installed in the living space HS of the target person P being monitored.

[0039] Step S102: The emergency determination unit 121 determines whether or not an emergency has occurred for the person being monitored P. Specifically, the emergency determination unit 121 may determine whether the action indicated by the sensor action data acquired in step S100 corresponds to any of the events indicated by the emergency registration information in the emergency registration information storage unit 132. In this case, the emergency determination unit 121 compares the content of the action indicated by the sensor action data acquired in step S100 with the definition content for each event stored in the emergency registration information.

[0040] Step S104: If an emergency has been determined in step S102, the emergency determination unit 121 sends a notification of the occurrence of an emergency (emergency notification) to the corresponding caregiver terminal 200. At this time, the emergency determination unit 121 can identify the caregiver terminal 200 to which the emergency notification will be sent by referring to the caregiver terminal information stored in the corresponding user information stored in the user information storage unit 131. After the processing in step S104, the process returns to step S100.

[0041] Step S106: If it is determined in step S102 that no emergency has occurred, the process proceeds to the reporting transmission stage. First, the conversion unit 122 converts the sensor activity data acquired in step S100 into text.

[0042] Step S108: The conversion unit 122 generates tags from the text obtained in step S106. The conversion unit 122 obtains text behavior data with a structure in which tags are attached to the text obtained in step S106. Note that, for example, the text behavior data may not include the text from which the tags were generated, but may instead consist of a list of tags generated in step S108.

[0043] Step S110: The conversion unit 122 stores the text behavior data obtained in steps S106 and S108 in the behavior data storage unit 133, associating it with the target person P being monitored. At this time, the conversion unit 122 stores the behavior date and time in the behavior data storage unit 133 along with the text behavior data. The conversion unit 122 may obtain the date and time of the behavior corresponding to the timing when the corresponding sensor behavior data was acquired in step S100.

[0044] Step S112: The target action determination unit 123 determines the target action corresponding to the current report transmission. In other words, the target action determination unit 123 determines from the action data stored in the action data storage unit 133 the action data that the counting unit 124 in the next step S116 will count.

[0045] Step S114: The counting unit 124 counts the number of times each target action determined in step S112 was performed during the target period. Specifically, the counting unit 124 counts the number of times action data for an action corresponding to a single target action exists in each unit period. The unit period is indicated by a threshold associated with the reference action data corresponding to the same action as the target action in the reference action information stored in the reference action information storage unit 134. For example, if the unit period indicated by the threshold in the reference action information is one week, then the action data stored in the action data storage unit 133 for the period from the present to one week ago will be counted. The counting unit 124 performs the counting of such behavioral data in accordance with each target behavior.

[0046] Step S116: For each target action performed in step S114, the count unit 124 compares the count value of the number of times the action has occurred with a threshold value associated with the reference action data corresponding to the same action as the target action in the reference action information stored in the reference action information storage unit 134. Based on the comparison, the count unit 124 determines whether or not there are any target actions whose count value is equal to or greater than the threshold value. If it is determined that there are no target actions whose count value exceeds the threshold, the process returns to step S100.

[0047] Step S118: If it is determined in step S116 that there is a target action whose count value is above the threshold, the report processing unit 125 creates a report. The report created will describe that the target action whose count value was determined to be above the threshold in step S116 was performed by the person being monitored P.

[0048] Step S120: The report processing unit 125 sends the report created in step S118 to the corresponding monitor terminal 200.

[0049] In the example processing procedure shown in Figure 7, step S114 counts the number of target actions, and if the count value for any target actions exceeds a threshold in step S116, a report is created and sent. As a variation of this embodiment, the transmission of reports to the monitor terminal 200 may be performed periodically. In this case, the timing of report transmission may follow, for example, a report transmission schedule set by the monitor W.

[0050] Referring to the flowchart in Figure 8, an example of a processing procedure for updating the machine learning model corresponding to the target action decision unit 123 in this embodiment will be described.

[0051] First, we will explain an example of the processing procedure for the monitor terminal 200. Step S200: When caregiver W checked the contents of the report sent to caregiver terminal 200, they found that some of the actions reported in the report did not pose a problem for the person P they were monitoring, and therefore no further reporting was necessary. In this case, caregiver W operated caregiver terminal 200 to create feedback information describing the actions that did not need to be reported. Alternatively, caregiver W may determine, based on recent observations of the person being monitored P, that there is an action that should be reported, even though it has not been reported in a report before. In this case, caregiver W operates the caregiver terminal 200 to create feedback information describing the action that needs to be reported. The monitor terminal 200 generates feedback information indicating whether reporting via report is unnecessary or necessary, in response to the operations described above. Feedback information may be created, for example, as an email. Alternatively, feedback information may be created by the monitoring application running on the monitor terminal 200 in response to the user entering the necessary information into the feedback information input form provided by the monitoring application.

[0052] Step S202: The monitor terminal 200 transmits the feedback information created in step S200 to the monitoring server 100.

[0053] Next, an example of the processing procedure for the monitoring server 100 will be described. Step S210: In the monitoring server 100, the communication unit 101 receives the feedback information transmitted in step S202.

[0054] Step S212: In the monitoring server 100, the learning processing unit 126 inputs the actions that require reporting in a report, or actions that do not, as indicated in the feedback information received in step S202, into the machine learning model (target action decision model) used in correspondence with the target action decision unit 123, and causes the machine learning to be executed.

[0055] Step S214: The target action decision model is updated by machine learning execution. Thereafter, the target action decision unit 123 determines the target action using the updated target action decision model. Specifically, if an action requiring reporting is entered in step S212, the target action determination unit 123 will determine the entered action as a target action if it is included in the action data for the unit period. Conversely, if an action that does not require reporting is entered in step S212, the target action determination unit 123 will not determine the entered action as a target action, even if it is included in the action data for the unit period.

[0056] Furthermore, the functions of the monitoring server 100 and monitoring terminal 200 described above may be realized by recording the program for realizing these functions on a computer-readable recording medium, loading the program recorded on this recording medium into a computer system, and executing it to perform the processing of the monitoring server 100 and monitoring terminal 200. Here, "loading the program recorded on the recording medium into a computer system and executing it" includes installing the program into the computer system. Here, "computer system" includes hardware such as the OS and peripheral devices. Also, "computer system" may include multiple computer devices connected via a network including communication lines such as the Internet, WAN, LAN, and dedicated lines. Also, "computer-readable recording medium" refers to portable media such as flexible disks, magneto-optical disks, ROMs, and CD-ROMs, as well as storage devices such as HDDs and SSDs built into the computer system. Thus, the recording medium storing the program may also be a non-transient recording medium such as a CD-ROM. Also, the recording medium includes internal or external recording media that can be accessed from the distribution server for distributing the program. The program code stored on the distribution server's recording medium may be different from the program code in a format executable by the terminal device. In other words, the format in which the program is stored on the distribution server is irrelevant, as long as it can be downloaded from the distribution server and installed in a form that can be executed on the terminal device. Furthermore, the program may be divided into multiple parts, each downloaded at a different time and then combined on the terminal device, and each divided program may be distributed by a different distribution server. In addition, "computer-readable recording medium" includes volatile memory (RAM) within computer systems that act as servers or clients when a program is transmitted over a network, which retains the program for a certain period of time. Moreover, the above program may be intended to implement only a part of the functions described above.Furthermore, the above-mentioned functions may be implemented in combination with programs already recorded in the computer system, such as so-called differential files (differential programs).

[0057] <Note> (1) One aspect of this embodiment is a monitoring system comprising: a conversion unit that acquires text behavior data by converting sensor behavior data, which is the behavior of a person under monitoring indicated by the output of a sensor provided to detect predetermined events for a person under monitoring, into text; a behavior data storage unit that stores the text behavior data; a reference behavior information storage unit that stores reference behavior information including a threshold for the number of times a predetermined behavior is performed in a predetermined unit period; a target behavior determination unit that determines target behaviors to be monitored based on the text behavior data acquired by the conversion unit; a counting unit that counts the number of times the behavior indicated by the text behavior data, which corresponds to the target behavior determined by the target behavior determination unit, has been performed in the unit period among the text behavior data stored in the behavior data storage unit corresponding to the person under monitoring; a report processing unit that transmits a report reporting on behaviors for which the count value from the counting unit is greater than or equal to a threshold to the monitor terminal of the monitor monitoring the person under monitoring; and a learning processing unit that causes the target behavior determination unit to learn so that the content of feedback information indicating the necessity of the behavior reported in the report is reflected in the determination result of the target behavior determination unit.

[0058] (2) One aspect of this embodiment is the monitoring system described in (1), wherein the conversion unit generates one or more tags based on the content of the text converted from the sensor behavior data, acquires the text with the generated tags attached to the text as the text behavior data, the behavior in the reference behavior information is represented by one or more tags, and the target behavior determination unit outputs the tag corresponding to the determined target behavior.

[0059] (3) One aspect of this embodiment is the monitoring system described in (1) or (2), wherein the target action determination unit changes the target action determined based on the text action data acquired by the conversion unit, based on the notification of whether or not an action is necessary as indicated in the report transmitted by the report processing unit.

[0060] (4) One aspect of this embodiment is a monitoring system according to any one of (1) to (3), further comprising an emergency determination unit that determines whether or not a pre-registered emergency event has occurred based on the sensor behavior data, wherein if the emergency determination unit determines that the emergency event has not occurred, the conversion unit, the target behavior determination unit, the counting unit, and the report processing unit are executed.

[0061] (5) One aspect of this embodiment is a monitoring method in a monitoring system, comprising: a conversion step of acquiring text behavior data by converting sensor behavior data, which is the behavior of a person under monitoring indicated by the output of a sensor provided to detect predetermined events for a person under monitoring, into text; a target behavior determination step of determining target behaviors to be monitored based on the text behavior data acquired in the conversion step; a counting step of counting the number of times the behavior indicated by the text behavior data, which corresponds to the target behavior determined in the target behavior determination step, has been performed in a unit period indicated by reference behavior information stored in a reference behavior information storage unit, among the text behavior data stored in the behavior data storage unit corresponding to the person under monitoring; a report processing step of sending a report to the monitor terminal of a monitor who monitors the person under monitoring, reporting on behaviors whose count value in the counting step is greater than or equal to a threshold indicated by reference behavior information stored in the reference behavior information storage unit; and a learning processing step of training a machine learning model corresponding to the target behavior determination step so that the content of feedback information indicating the necessity of the behavior reported in the report is reflected in the determination result of the target behavior determination step.

[0062] (6) One aspect of this embodiment is a program that causes a computer in a monitoring system to function as a learning processing unit that causes the target behavior determination unit to learn so that the content of feedback information indicating whether or not the behavior reported in the report is necessary is reflected in the determination result of the target behavior determination unit. The program also includes a conversion unit that counts the number of times the behavior indicated by the text behavior data, which corresponds to the target behavior determined by the target behavior determination unit, has been performed in a unit period indicated by the reference behavior information stored in the reference behavior information storage unit, among the text behavior data stored in the behavior data storage unit corresponding to the target behavior. The program also includes a report processing unit that sends a report to the monitor terminal of the monitor who monitors the person under monitoring, reporting on behaviors whose count value is equal to or greater than a threshold indicated by the reference behavior information stored in the reference behavior information storage unit. [Explanation of symbols]

[0063] 10 Sensor, 20 Gateway, 100 Monitoring Server, 101 Communication Unit, 102 Control Unit, 103 Storage Unit, 121 Emergency Determination Unit, 122 Conversion Unit, 123 Target Action Determination Unit, 124 Counting Unit, 125 Report Processing Unit, 126 Learning Processing Unit, 131 User Information Storage Unit, 132 Emergency Registration Information Storage Unit, 133 Action Data Storage Unit, 134 Reference Action Information Storage Unit, 200 Monitor Terminal

Claims

1. A conversion unit that acquires text behavior data by converting sensor behavior data, which is the behavior of the person being monitored, as indicated by the output of a sensor that is installed to detect a predetermined event in the person being monitored, into text, A behavior data storage unit that stores the aforementioned text behavior data, A reference behavior information storage unit stores reference behavior information including a threshold value for the number of times a given behavior is performed within a given unit period, A target behavior determination unit determines the target behavior to be monitored based on the text behavior data acquired by the conversion unit, A counting unit counts the number of times the text behavior data corresponding to the target behavior determined by the target behavior determination unit was performed within the unit period, among the text behavior data stored in the behavior data storage unit corresponding to the person being monitored. A report processing unit transmits a report that reports on actions where the count value from the counting unit is above a threshold to the monitor terminal of the monitor who is monitoring the person being monitored. A communication unit that receives feedback information from the aforementioned monitoring terminal indicating whether the actions reported in the report are necessary, which is generated in response to the monitoring terminal's operation input, A learning processing unit is provided to train the target action decision unit so that the content of the received feedback information is reflected in the decision result of the target action decision unit. A monitoring system equipped with the following features.

2. The conversion unit generates one or more tags based on the content of the text converted from the sensor behavior data, and acquires the text with the generated tags attached as the text behavior data. The behavior in the aforementioned reference behavior information is represented by one or more tags. The aforementioned target action determination unit outputs a tag corresponding to the determined target action. The monitoring system according to claim 1.

3. A conversion unit that acquires text behavior data by converting sensor behavior data, which is the behavior of the person being monitored, as indicated by the output of a sensor that is installed to detect a predetermined event in the person being monitored, into text, A behavior data storage unit that stores the aforementioned text behavior data, A reference behavior information storage unit stores reference behavior information including a threshold value for the number of times a given behavior is performed within a given unit period, A target behavior determination unit determines the target behavior to be monitored based on the text behavior data acquired by the conversion unit, A counting unit counts the number of times the text behavior data corresponding to the target behavior determined by the target behavior determination unit was performed within the unit period, among the text behavior data stored in the behavior data storage unit corresponding to the person being monitored. A report processing unit transmits a report that reports on actions where the count value from the counting unit is above a threshold to the monitor terminal of the monitor who is monitoring the person being monitored. A learning processing unit is provided to train the target action decision unit so that the content of the feedback information indicating the necessity of the action reported in the report is reflected in the decision result of the target action decision unit. Equipped with, The conversion unit generates one or more tags based on the content of the text converted from the sensor behavior data, and acquires the text with the generated tags attached as the text behavior data. The behavior in the aforementioned reference behavior information is represented by one or more tags. The aforementioned target action determination unit outputs a tag corresponding to the determined target action. Monitoring system.

4. The target action determination unit changes the target action determined based on the text action data acquired by the conversion unit, based on the notification of whether or not the action is necessary as indicated in the report transmitted by the report processing unit. A monitoring system according to any one of claims 1 to 3.

5. The system further includes an emergency determination unit that determines whether or not a pre-registered emergency event has occurred based on the aforementioned sensor behavior data. If the emergency determination unit determines that no emergency event has occurred, the conversion unit, the target action determination unit, the counting unit, and the report processing unit are executed. A monitoring system according to any one of claims 1 to 3.

6. A monitoring method in a monitoring system, A conversion step to obtain text behavior data by converting sensor behavior data, which is the behavior of the person being monitored, as indicated by the output of a sensor installed to detect a predetermined event targeting the person being monitored, into text, A target behavior determination step in which a target behavior to be monitored is determined based on the text behavior data obtained in the conversion step, A counting step is performed to count the number of times the text behavior data corresponding to the target behavior determined in the target behavior determination step was performed within the unit period indicated by the reference behavior information stored in the reference behavior information storage unit, among the text behavior data stored in the behavior data storage unit corresponding to the target behavior of the person being monitored. A report processing step which includes sending a report to the caregiver's terminal of the caregiver who is caring for the person being cared for, reporting on actions where the count value obtained in the count step is greater than or equal to a threshold indicated by the reference action information stored in the reference action information storage unit, A receiving step of receiving feedback information from the monitoring terminal, which is generated in response to the monitoring terminal's operation input, indicating whether the action reported in the report is necessary or not, A learning process step to train a machine learning model corresponding to the target action decision step so that the content of the received feedback information is reflected in the decision result of the target action decision step, and A monitoring method including the following.

7. A monitoring method in a monitoring system, A conversion step to obtain text behavior data by converting sensor behavior data, which is the behavior of the person being monitored, as indicated by the output of a sensor installed to detect a predetermined event targeting the person being monitored, into text, A target behavior determination step in which a target behavior to be monitored is determined based on the text behavior data obtained in the conversion step, A counting step is performed to count the number of times the text behavior data corresponding to the target behavior determined in the target behavior determination step was performed within the unit period indicated by the reference behavior information stored in the reference behavior information storage unit, among the text behavior data stored in the behavior data storage unit corresponding to the target behavior of the person being monitored. A report processing step which includes sending a report to the caregiver's terminal of the caregiver who is caring for the person being cared for, reporting on actions where the count value obtained in the count step is greater than or equal to a threshold indicated by the reference action information stored in the reference action information storage unit, A learning process step to train a machine learning model corresponding to the target action decision step, such that the content of the feedback information indicating the necessity of the action reported in the report is reflected in the decision result of the target action decision step, and Includes, The conversion step includes generating one or more tags based on the content of the text converted from the sensor behavior data, and acquiring the text with the generated tags attached as the text behavior data. The behavior in the aforementioned reference behavior information is represented by one or more tags. The aforementioned target action determination step outputs a tag corresponding to the determined target action. Monitoring method.

8. Computers in surveillance systems A conversion unit acquires text behavior data by converting sensor behavior data, which is the behavior of the person being monitored, as indicated by the output of a sensor that is installed to detect predetermined events for the person being monitored, into text. A target behavior determination unit determines the target behavior to be monitored based on the text behavior data acquired by the conversion unit. A counting unit counts the number of times, within a unit period indicated by the reference behavior information stored in the reference behavior information storage unit, the text behavior data that corresponds to the target behavior determined by the target behavior determination unit, among the text behavior data stored in the behavior data storage unit corresponding to the target behavior of the person being monitored. A report processing unit transmits a report to the caregiver's terminal of the caregiver who is caring for the person being cared for, reporting on actions where the count value from the counting unit is equal to or greater than the threshold indicated by the reference action information stored in the reference action information storage unit. A communication unit that receives feedback information from the aforementioned monitoring terminal indicating whether the actions reported in the report are necessary, which is generated in response to the monitoring terminal's operation input. A learning processing unit causes the target action decision unit to learn so that the content of the received feedback information is reflected in the decision result of the target action decision unit. A program designed to function as such.

9. Computers in surveillance systems A conversion unit acquires text behavior data by converting sensor behavior data, which is the behavior of the person being monitored, as indicated by the output of a sensor that is installed to detect predetermined events for the person being monitored, into text. A target behavior determination unit determines the target behavior to be monitored based on the text behavior data acquired by the conversion unit. A counting unit counts the number of times, within a unit period indicated by the reference behavior information stored in the reference behavior information storage unit, the text behavior data that corresponds to the target behavior determined by the target behavior determination unit, among the text behavior data stored in the behavior data storage unit corresponding to the target behavior of the person being monitored. A report processing unit transmits a report to the caregiver's terminal of the caregiver who is caring for the person being cared for, reporting on actions where the count value from the counting unit is equal to or greater than the threshold indicated by the reference action information stored in the reference action information storage unit. This program functions as a learning processing unit to train the target action decision unit so that the content of the feedback information indicating the necessity of the action reported in the report is reflected in the decision result of the target action decision unit. The conversion unit generates one or more tags based on the content of the text converted from the sensor behavior data, and acquires the text with the generated tags attached as the text behavior data. The behavior in the aforementioned reference behavior information is represented by one or more tags. The aforementioned target action determination unit outputs a tag corresponding to the determined target action. A program that includes the following.

Images