server

Abstract

This technology provides accurate estimation of the condition of the person being monitored. [Solution] The service provision server 100 includes an acquisition unit 12 that acquires monitored data 22 showing the amount of electricity consumed at the target person's residence for each time period, an identification unit 13 that identifies a first period in which the target person's condition is normal and a second period in which the target person's condition is abnormal, and an estimation unit 17 that estimates the target person's condition level during the monitored period based on a first comparison result between a target waveform representing the change in electricity consumption over time during the monitored period and a first reference waveform representing the change in electricity consumption over time during the first period, and a second comparison result between the target waveform and a second reference waveform representing the change in electricity consumption over time during the second period.

Description

Technical Field

[0001] This disclosure relates to a technique for monitoring a person to be monitored.

Background Art

[0002] Japanese Patent Application Laid-Open No. 2008-112267 (Patent Document 1) discloses a method of estimating the wake-up and bedtime of a person to be monitored based on the power consumption for each fixed time period obtained from a power meter installed in a home, and determining the abnormality of the resident from the difference between the estimated time and the predicted time. The predicted time is automatically adjusted from the power usage situation in the home.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the method of Patent Document 1, the state of the person to be monitored is determined only based on the difference between the predicted time adjusted from the power usage situation and the estimated time. Therefore, further improvement in the estimation accuracy of the state of the person to be monitored is desired.

[0005] This disclosure has been made in view of the above problems, and its object is to provide a technique capable of accurately estimating the state of a person to be monitored.

Means for Solving the Problems

[0006] A server following a certain scenario comprises an acquisition unit, an identification unit, and an estimation unit. The acquisition unit acquires data showing the amount of electricity consumed at the monitored person's residence for each time period. The identification unit identifies a first period in which the monitored person's condition is normal and a second period in which the monitored person's condition is abnormal. The estimation unit estimates the monitored person's condition level during the monitored period based on a first comparison result between a target waveform representing the time-series change in electricity consumption during the monitored period and a first reference waveform representing the time-series change in electricity consumption during the first period, and a second comparison result between the target waveform and a second reference waveform representing the time-series change in electricity consumption during the second period.

[0007] In the server described above, the first comparison result shows the relationship between the first similarity of the target waveform to the first reference waveform and the first threshold. The second comparison result shows the relationship between the second similarity of the target waveform to the second reference waveform and the second threshold.

[0008] In the above server, the first similarity score represents the degree of similarity between the first statistical value calculated from the first reference waveform and the target statistical value calculated from the target waveform. The second similarity score represents the degree of similarity between the second statistical value calculated from the second reference waveform and the target statistical value.

[0009] The server described above further comprises a classification unit that classifies the person being monitored into one of several categories based on at least one of the person's functional level and lifestyle patterns, and a setting unit that sets a first threshold and a second threshold according to the category to which the person being monitored belongs.

[0010] In the server described above, the classification unit determines the lifestyle pattern of the person being monitored based on the comparison results between the waveform models of multiple lifestyle patterns and the first reference waveform.

[0011] In the server described above, the classification unit determines the functional level of the person being monitored based on the activity capacity index of the Geriatric Research Institute.

[0012] In the server described above, the configuration unit adjusts the first and second thresholds according to the estimated state level using data indicating power consumption during the verification period prior to the monitoring period.

[0013] The server described above further includes a notification unit that outputs notifications according to the status level. In the server described above, the notification unit determines the content of the notification based on the state level transitions over multiple consecutive monitoring periods.

[0014] The information processing method according to other aspects comprises the first to third steps. The first step is to acquire data showing the amount of electricity consumed at the residence of the person being monitored for each time period. The second step is to identify the first period in which the person being monitored is in a normal state and the second period in which the person being monitored is in an abnormal state. The third step is to estimate the state level of the person being monitored during the monitoring period based on the first comparison result between the target waveform representing the change in electricity consumption over time during the monitoring period and the first reference waveform representing the change in electricity consumption over time during the first period, and the second comparison result between the target waveform and the second reference waveform representing the change in electricity consumption over time during the second period.

[0015] In other situations, the program causes the computer to execute each step of the information processing method described above. [Effects of the Invention]

[0016] The technology disclosed herein allows for accurate estimation of the condition of the person being monitored. [Brief explanation of the drawing]

[0017] [Figure 1] This diagram shows a schematic of a monitoring system including a server according to this embodiment. [Figure 2] This diagram shows the hardware configuration of the service provision server. [Figure 3] This diagram shows the functional configuration of the service provider server. [Figure 4]It is a diagram showing an example of subject information. [Figure 5] It is a diagram showing an example of a method for specifying normal periods and abnormal periods. [Figure 6] It is a diagram showing an example of a method for determining the level of living function. [Figure 7] It is a diagram showing an example of a method for classifying lifestyle patterns. [Figure 8] It is a diagram for explaining a method for estimating the state level. [Figure 9] It is a diagram showing an example of estimation result data. [Figure 10] It is a flowchart showing the flow of preparatory processing. [Figure 11] It is a flowchart showing the flow of monitoring processing.

Embodiments for Carrying Out the Invention

[0018] Hereinafter, embodiments according to the present invention will be described while referring to the drawings. In the following description, the same parts and components are denoted by the same reference numerals, and redundant descriptions are not repeated. Note that the embodiments and each modification described below may be selectively combined as appropriate.

[0019] <System Configuration> FIG. 1 is a diagram showing an outline of a monitoring system including a server according to the present embodiment. As shown in FIG. 1, the monitoring system SYS includes a service providing server 100, a residence 200 of a person to be monitored (hereinafter simply referred to as “subject 400”), a management server 300, and one or more terminals 500. In the example shown in FIG. 1, the one or more terminals 500 include a terminal 500a used by the subject 400 and a terminal 500b used by a user 600 who monitors the subject 400. The user 600 is, for example, a family member of the elderly subject 400, a caregiver of the subject 400, a care manager of the subject 400, or the like.

[0020] A smart meter 210 is installed in residence 200. The smart meter 210 is linked to a supply point identification number and measures and records electricity consumption for each time period. For example, the smart meter 210 measures the electricity consumption of various electrical appliances 220 installed in residence 200 for each time period (for example, 30 minutes or 1 hour) that is divided from 0:00 to 24:00.

[0021] The management server 300 collects data (hereinafter referred to as "load survey data") showing the amount of electricity consumed for each time period measured by the smart meter 210 via the communication line 2. The load survey data includes date and time data that identifies the time period, and electricity amount data that shows the amount of electricity consumed during that time period. The date and time data indicates the year, month, day, and time period (for example, 15:00 to 15:30) when the electricity consumption was measured. The management server 300 stores the load survey data collected from the smart meter 210 in association with the supply point identification number corresponding to the smart meter 210.

[0022] The service provider server 100 obtains load survey data corresponding to the smart meter 210 installed in the residence 200 where the subject 400 lives from the management server 300, and uses the obtained load survey data to estimate the status of the subject 400. Furthermore, the service provider server 100 outputs a notification with content corresponding to the estimation result. For example, the service provider server 100 outputs a notification via the internet 3 to at least one of the terminal 500a used by the subject 400 and the terminal 500b used by the user 600 who is monitoring the subject 400.

[0023] <Hardware configuration of the service provider server> Figure 2 shows the hardware configuration of the service provision server. As shown in Figure 2, the service provision server 100 includes, as its main components, a processor 101 that executes program 110, a ROM (Read Only Memory) 102 for non-volatile data storage, a RAM (Random Access Memory) 103 for volatile data storage generated by the execution of program 110 by the processor 101 or data acquired from an external source, a hard disk drive (HDD) 104 for non-volatile data storage, a communication interface 105, and a power supply circuit 107. Each component is connected to the others by a data bus. The communication interface 105 is an interface for communication with other devices (including the management server 300 and one or more terminals 500).

[0024] Processing in the service server 100 is realized by a program 110 executed by each piece of hardware and the processor 101. Such a program 110 is pre-stored in the HDD 104. However, the program 110 may be stored on other storage media and distributed as a program product. Alternatively, the program 110 may be provided as a downloadable program product by an information provider connected to the so-called Internet. Such a program 110 is read from its storage medium by a reader device, or downloaded via a communication interface 105, etc., and then stored in the HDD 104. The processor 101 reads the program 110 from the HDD 104 and executes the program 110.

[0025] Furthermore, the number of processors 101, ROMs 102, RAMs 103, and hard disks 104 provided in the service provision server 100 is not limited to one, but may be multiple.

[0026] The service provision server 100 may be implemented by a cloud server or by one or more computers.

[0027] <Functional Configuration of the Service Provider Server> Figure 3 shows the functional configuration of the service provider server. As shown in Figure 3, the service provider server 100 comprises a storage unit 10, a registration unit 11, an acquisition unit 12, a specification unit 13, a reference data generation unit 14, a classification unit 15, a setting unit 16, an estimation unit 17, and a notification unit 18. The storage unit 10 is implemented by the ROM 102, RAM 103, and HDD 104 shown in Figure 2. The registration unit 11, acquisition unit 12, specification unit 13, reference data generation unit 14, classification unit 15, setting unit 16, estimation unit 17, and notification unit 18 are implemented by the processor 101 executing the program 110.

[0028] The registration unit 11 performs a registration process for each target person 400, registering information about that target person 400 (hereinafter referred to as "target person information 20"). Specifically, the registration unit 11 receives a registration request from the terminal 500 and displays a screen on the terminal 500 for registering new target person information 20. The registration unit 11 generates the target person information 20 in response to the input on the screen and stores the generated target person information 20 in the storage unit 10.

[0029] Figure 4 shows an example of subject information. The subject information 20 shown in Figure 4 includes ID information 20a indicating an ID that identifies the subject 400, number information 20b indicating a supply point identification number for the subject 400's residence 200, notification information 20c indicating the email address to be notified, hospitalization / outpatient information 20d, medical interview information 20e, and response information 20f for the Geriatric Research Institute's Activity Capacity Index.

[0030] The email address to which notifications are sent may, for example, be an email address owned by the subject 400 themselves, or an email address owned by user 600 who is monitoring subject 400. Multiple email addresses may also be set.

[0031] Hospitalization and outpatient information 20d shows the admission date, discharge date, outpatient visit date, severity, and urgency. Questionnaire information 20e shows age, sex, number of family members living together, underlying diseases, level of care required, wake-up time, bedtime, days of going out, time of leaving and return on days of going out. Response information 20f shows the results of the responses to each item of the Geriatric Research Institute Activity Index. The Geriatric Research Institute Activity Index is a higher-level life function assessment index composed of three aspects: instrumental independence (IADL (Instrumental Activities of Daily Living)), intellectual activity, and social role.

[0032] Furthermore, the registration unit 11 may accept requests to update the target person information 20. For example, the registration unit 11 periodically (for example, every year) sends an email to the email address included in the target person information 20, informing the recipient of the update of the target person information 20. The target person 400 or user 600 who receives the email operates the terminal 500 to send a request to update the target person information 20 to the service provider server 100. If the registration unit 11 accepts a request to update the target person information 20, it displays a screen for updating the target person information 20 on the terminal 500. The registration unit 11 updates the target person information 20 stored in the storage unit 10 according to the input on the screen.

[0033] The acquisition unit 12 shown in Figure 3 acquires load survey data corresponding to the residence 200 of each target person 400 and stores the acquired load survey data in the storage unit 10. Specifically, the acquisition unit 12 reads the supply point identification number corresponding to the target person 400 from the target person information 20. The acquisition unit 12 then acquires the load survey data corresponding to the read supply point identification number from the management server 300.

[0034] When new target information 20 is registered in the storage unit 10, or when the target information 20 is updated, the acquisition unit 12 acquires the most recent load survey data for a predetermined period (e.g., 1 year, 3 years, etc.) corresponding to the target information 20 as training data 21, and stores the training data 21 in the storage unit 10. Subsequently, every time a certain period of time has elapsed (e.g., 12 hours (half a day), 1 day), the acquisition unit 12 acquires the most recent load survey data for that period as monitored data 22, and stores the acquired monitored data 22 in the storage unit 10.

[0035] The identification unit 13 identifies, for each subject 400, a normal period (first period) in which the subject 400's condition is normal, and an abnormal period (second period) in which the subject 400's condition is abnormal. The identification unit 13 identifies the normal period and the abnormal period at the time new subject information 20 is registered in the storage unit 10, or at the time the subject information 20 is updated.

[0036] Figure 5 shows an example of a method for identifying normal and abnormal periods. The identification unit 13 reads the most recent hospitalization date t0 from the hospitalization and outpatient information 20d contained in the subject information 20. Furthermore, the identification unit 13 identifies a date a predetermined period (e.g., one month, two months) prior to the hospitalization date t0 as the boundary date t1.

[0037] As shown in Figure 5, during the period prior to the boundary day t1, subject 400 has no health problems, so electricity consumption changes over time according to subject 400's normal lifestyle. However, before hospitalization, subject 400's lifestyle may differ from normal due to their health condition. For example, even though it is a time when they would normally go for a walk, they may stay at home due to a deterioration in their health. As a result, the change in electricity consumption over time from the boundary day t1 to the hospitalization day t0 differs from the change in electricity consumption over time during the period prior to the boundary day t1.

[0038] Therefore, the identification unit 13 identifies the period from the boundary date t1 to the admission date t0 as an abnormal period, and identifies the period before the boundary date t1 as a normal period. If the admission / outpatient information 20d indicates admission and discharge dates for multiple admissions, it is preferable for the identification unit 13 to exclude from the normal period the period from a predetermined period (e.g., one month, two months) before admission dates other than the most recent one until the discharge date (or a certain period (e.g., one week) after the discharge date).

[0039] The reference data generation unit 14 shown in Figure 3 refers to the training data 21 and generates data showing a first reference waveform representing the change in power consumption over time during a normal period (hereinafter referred to as "first reference data 23"), and stores the generated first reference data 23 in the storage unit 10. Similarly, the reference data generation unit 14 refers to the training data 21 and generates data showing a second reference waveform representing the change in power consumption over time during an abnormal period (hereinafter referred to as "second reference data 24"), and stores the generated second reference data 24 in the storage unit 10.

[0040] The first reference data 23 shows representative values ​​of electricity consumption for each time period during a normal period. Representative values ​​are, for example, the mean and the median. The first reference waveform is generated by plotting the representative values ​​for each time period shown by the first reference data 23 on a graph with time on the horizontal axis and electricity consumption on the vertical axis. The first reference data 23 may further show various statistical values ​​obtained from the first reference waveform.

[0041] The second reference data 24 shows representative values ​​of power consumption for each time period during the abnormal period. Representative values ​​are, for example, the mean and the median. The second reference waveform is generated by plotting the representative values ​​for each time period shown by the second reference data 24 on a graph with time on the horizontal axis and power consumption on the vertical axis. The second reference data 24 may further show various statistical values ​​obtained from the second reference waveform.

[0042] Statistical values ​​include, for example, aggregate values ​​(total), mean, median, maximum, and standard deviation of electricity consumption over any given period (e.g., morning, afternoon, or a 3-hour period within a day).

[0043] The reference data generation unit 14 may generate the first reference data 23 and the second reference data 24 using some of the data from the training data 21.

[0044] The classification unit 15 classifies the subject 400 into one of several categories based on at least one of the subject 400's functional level and lifestyle pattern. The classification unit 15 classifies the subject 400 corresponding to the subject 400 in one of several categories when new subject information 20 is registered in the memory unit 10 or when the subject information 20 is updated.

[0045] Figure 6 shows an example of a method for determining the level of functional capacity. As shown in Figure 6, the classification unit 15 determines which of the four levels of functional capacity the subject 400 belongs to, based on the response results for each item of the Geriatric Research Institute Activity Index, which are indicated by the response information 20f included in the subject information 20. The four levels of functional capacity are: "Level of functional capacity for healthy elderly persons," "Level of functional capacity for elderly persons assumed to be pre-frail," "Level of functional capacity for elderly persons assumed to be frail," and "Level of functional capacity for persons requiring support or care." Specifically, the classification unit 15 determines the level of functional capacity of the subject 400 according to the score on the Geriatric Research Institute Activity Index.

[0046] Figure 7 shows an example of a method for classifying lifestyle patterns. The classification unit 15 classifies the lifestyle patterns of 400 subjects using three waveform models 25a to 25c. The waveform models 25a to 25c are pre-stored in the memory unit 10.

[0047] Waveform model 25a corresponds to the "standard pattern," one of the lifestyle patterns, and shows the change in power consumption over time in the residence of an elderly person who owns standard electrical appliances 220 and is active during the day. Waveform model 25b corresponds to the "energy-saving pattern," one of the lifestyle patterns, and shows the change in power consumption over time in the residence of an elderly person who is relatively energy-saving oriented or owns fewer electrical appliances 220. Waveform model 25c corresponds to the "day-night reversal pattern," one of the lifestyle patterns, and shows the change in power consumption over time in the residence of an elderly person who is active at night. Each of the waveform models 25a to 25c may be pre-generated based on load survey data obtained from the residences of elderly people with the corresponding lifestyle pattern, or may be pre-generated by simulation.

[0048] The classification unit 15 calculates the similarity between the first reference waveform, indicated by the first reference data 23, and each of the waveform models 25a to 25c. The classification unit 15 determines that the lifestyle pattern of subject 400 is a "standard pattern" if the similarity between the first reference waveform and waveform model 25a is the highest and exceeds the standard value. The classification unit 15 determines that the lifestyle pattern of subject 400 is a "power-saving pattern" if the similarity between the first reference waveform and waveform model 25b is the highest and exceeds the standard value. The classification unit 15 determines that the lifestyle pattern of subject 400 is a "day-night reversal pattern" if the similarity between the first reference waveform and waveform model 25c is the highest and exceeds the standard value. The classification unit 15 determines that the lifestyle pattern of subject 400 is "other" if none of the similarities exceed the standard value.

[0049] The multiple categories include, for example, 16 categories. Each of the 16 categories corresponds to 16 possible combinations of four functional levels and four lifestyle patterns. The classification unit 15 determines the category to which the subject 400 belongs based on the combination of the subject's functional level and lifestyle pattern.

[0050] The setting unit 16 sets thresholds (the first threshold Th1 and the second threshold Th2, described later) used to estimate the state level of each subject 400, according to the category to which the subject 400 belongs. The setting unit 16 stores threshold data 26 indicating the set thresholds in the storage unit 10.

[0051] The estimation unit 17 estimates the status level of the subject 400 during the monitoring period based on a first comparison result between the target waveform representing the change in power consumption over time during the monitoring period and the first reference waveform indicated by the first reference data 23, and a second comparison result between the target waveform and the second reference waveform indicated by the second reference data 24.

[0052] Specifically, the estimation unit 17 identifies the waveform indicated by the monitored data 22 as the target waveform. As described above, the monitored data 22 shows the change in power consumption over time over a recent period (for example, 12 hours (half a day), 1 day). Therefore, the target waveform shows the change in power consumption over time over a recent period.

[0053] The estimation unit 17 calculates a first similarity of the target waveform to a first reference waveform. Similarly, the estimation unit 17 calculates a second similarity of the target waveform to a second reference waveform.

[0054] The first similarity score represents the degree of similarity in shape and height between the target waveform and the first reference waveform for each time period. For example, the first similarity score is expressed as the sum of the squares of the differences between the target waveform and the first reference waveform for each time period. Alternatively, the first similarity score may represent the degree of similarity between the statistical values ​​calculated from the target waveform and the statistical values ​​calculated from the first reference waveform. For example, the first similarity score is expressed as the sum of the squares of the differences between the statistical values ​​calculated from the target waveform and the statistical values ​​calculated from the first reference waveform.

[0055] Similarly, the second similarity score represents the degree of similarity in shape and height between the target waveform and the second reference waveform for each time period. For example, the second similarity score is expressed as the sum of the squares of the differences between the target waveform and the second reference waveform for each time period. Alternatively, the second similarity score may represent the degree of similarity between the statistical values ​​calculated from the target waveform and the statistical values ​​calculated from the second reference waveform. For example, the second similarity score is expressed as the sum of the squares of the differences between the statistical values ​​calculated from the target waveform and the statistical values ​​calculated from the second reference waveform.

[0056] The estimation unit 17 identifies the magnitude relationship between the first similarity and the first threshold Th1 as the first comparison result. Similarly, the estimation unit 17 identifies the magnitude relationship between the second similarity and the second threshold Th2 as the second comparison result. Based on these magnitude relationships, the estimation unit 17 determines the state level of the subject 400. The first threshold Th1 and the second threshold Th2 are set by the setting unit 16.

[0057] Figure 8 illustrates the method for estimating the state level. Figure 8 shows the method for estimating the state level of 400 subjects, each classified into one of four categories whose functional capacity is equivalent to that of healthy elderly individuals.

[0058] The estimation unit 17 estimates one of the following conditions for the subject 400: "Condition Level 1" to "Condition Level 5". "Condition Level 1" is the level at which the illness has improved and the subject is assumed to be in exemplary health. "Condition Level 2" is the level at which the subject is assumed to be in normal physical condition and active. "Condition Level 3" is the level at which the subject is assumed to be able to walk and is in good health. "Condition Level 4" is the level at which the subject is assumed to be unwell and requires medical attention. "Condition Level 5" is the level at which the subject is assumed to be bedridden or collapsed. "Condition Level 1" to "Condition Level 3" indicate that the subject 400 is in a normal state. "Condition Level 4" and "Condition Level 5" indicate that the subject 400 is in an abnormal state.

[0059] The setting unit 16 sets thresholds Th1_1, Th1_2, and Th1_3 as the first threshold Th1 to distinguish between "state level 1" to "state level 3". Note that threshold Th1_1 > threshold Th1_2 > threshold Th1_3. Furthermore, the setting unit 16 sets a second threshold Th2 to distinguish between "state level 4" and "state level 5".

[0060] The estimation unit 17 estimates the state of subject 400 as "state level 1" when the first similarity exceeds the threshold Th1_1 (when the first similarity belongs to segments 1, 6, 11, or 16 shown in Figure 8). The estimation unit 17 estimates the state of subject 400 as "state level 2" when the first similarity is less than or equal to the threshold Th1_1 and exceeds the threshold Th1_2 (when the first similarity belongs to segments 2, 7, 12, or 17 shown in Figure 8). The estimation unit 17 estimates the state of subject 400 as "state level 3" when the first similarity is less than or equal to the threshold Th1_2 and exceeds the threshold Th1_3 (when the first similarity belongs to segments 3, 8, 13, or 18 shown in Figure 8). The estimation unit 17 estimates the status of subject 400 as "status level 4" if the first similarity is less than or equal to the threshold Th1_3 and the second similarity is less than or equal to the second threshold Th2 (when the first and second similarities belong to segments 4, 9, 14, and 19 shown in Figure 8). The estimation unit 17 also estimates the status of subject 400 as "status level 5" if the first similarity is less than or equal to the threshold Th1_3 and the second similarity exceeds the second threshold Th2 (when the first and second similarities belong to segments 5, 10, 15, and 20 shown in Figure 8).

[0061] Elderly people with an "energy-saving lifestyle pattern" tend to reduce their use of air conditioners, for example, and are therefore expected to be more prone to health problems compared to elderly people with a "standard lifestyle pattern." Similarly, elderly people with a "reversed day-night cycle" lifestyle are expected to be more prone to health problems compared to elderly people with a "standard lifestyle pattern" or an "energy-saving lifestyle pattern." Elderly people with an "other" lifestyle pattern have irregular lifestyles and are therefore expected to be more prone to health problems compared to elderly people with a "standard lifestyle pattern," an "energy-saving lifestyle pattern," or a "reversed day-night cycle."

[0062] Furthermore, elderly individuals whose functional capacity is at the level of "elderly individuals considered pre-frail" are expected to be more prone to health problems than elderly individuals whose functional capacity is at the level of "healthy elderly individuals." Similarly, elderly individuals whose functional capacity is at the level of "elderly individuals considered frail" are expected to be more prone to health problems than elderly individuals whose functional capacity is at the level of "elderly individuals considered pre-frail." Moreover, elderly individuals whose functional capacity is at the level of "individuals requiring support or care" are expected to be more prone to health problems than elderly individuals whose functional capacity is at the level of "elderly individuals considered frail."

[0063] Therefore, it is preferable for the setting unit 16 to set the thresholds Th1_1, Th1_2, and Th1_3 higher and the second threshold Th2 lower for categories that are expected to be more prone to illness. This improves the accuracy of estimating the state levels of the subjects 400.

[0064] In the example shown in Figure 8, if the subject 400 is classified into the "Standard Pattern" category, the setting unit 16 sets the thresholds Th1_1a, Th1_2a, Th1_3a and the second threshold Th2a, respectively, as thresholds Th1_1, Th1_2, Th1_3 and the second threshold Th2. If the subject 400 is classified into the "Power Saving Pattern" category, the setting unit 16 sets the thresholds Th1_1b, Th1_2b, Th1_3b and the second threshold Th2b, respectively, as thresholds Th1_1b, Th1_2b, Th1_3b and the second threshold Th2. If the subject 400 is classified into the "Day-Night Reversal Pattern" category, the setting unit 16 sets the thresholds Th1_1c, Th1_2c, Th1_3c and the second threshold Th2c, respectively, as thresholds Th1_1c, Th1_2c, Th1_3c and the second threshold Th2c. The setting unit 16 sets the thresholds Th1_1d, Th1_2d, Th1_3d, and second threshold Th2d respectively as thresholds Th1_1, Th1_2, Th1_3, and second threshold Th2d, respectively, when the target persons 400 are classified in the "Other" category.

[0065] The setting unit 16 should set thresholds Th1_1, Th1_2, Th1_3 and a secondary threshold Th2 according to the category to which the subject 400 belongs, so as to satisfy the following equations (1) to (4). As a result, the thresholds Th1_1, Th1_2, and Th1_3 are set higher and the secondary threshold Th2 is set lower for categories that are expected to be more prone to illness. Threshold Th1_1a < Threshold Th1_1b < Threshold Th1_1c < Threshold Th1_1d Equation (1) Threshold Th1_2a < Threshold Th1_2b < Threshold Th1_2c < Threshold Th1_2d Equation (2) Threshold Th1_3a < Threshold Th1_3b < Threshold Th1_3c < Threshold Th1_3d Equation (3) Second threshold Th2a > Second threshold Th2b > Second threshold Th2c > Second threshold Th2d Equation (4) Note that some of the inequality signs in equations (1) to (4) may be equality signs.

[0066] If the first reference data 23 and the second reference data 24 are generated using a portion of the training data 21, the setting unit 16 may use the remaining training data 21 to verify each set threshold and adjust each threshold according to the verification results.

[0067] The estimation unit 17 estimates the state level for the most recent monitoring period, and then updates the estimation result data 27 stored in the storage unit 10 according to the estimation result.

[0068] Figure 9 shows an example of estimated result data. As shown in Figure 9, the estimated result data 27 includes an ID that identifies the subject 400, the date and time of the most recent monitoring period, the status level of the most recent monitoring period, the status level of the previous monitoring period, the status level of the monitoring period two periods prior, and the status level of the monitoring period three periods prior.

[0069] The notification unit 18 outputs a notification corresponding to the state level estimated by the estimation unit 17. Specifically, the notification unit 18 sends a notification email to the email address included in the target person information 20.

[0070] For example, the notification unit 18 outputs a notification indicating the estimated state level each time the state level for the most recent monitoring period is estimated.

[0071] Alternatively, the notification unit 18 may output a notification indicating an abnormality when the status level for the most recent monitoring period is "status level 4" or "status level 5".

[0072] Alternatively, the notification unit 18 may refer to the estimated result data 27 and determine the content of the notification based on the transition of state levels over multiple consecutive monitoring periods.

[0073] For example, the notification unit 18 determines a warning text based on the number of consecutive occurrences of "Status Level 4" or "Status Level 5," and outputs a notification showing the determined warning text. Specifically, if "Status Level 4" or "Status Level 5" occurs twice in a row, the notification unit 18 determines the warning text to be "The subject may not be feeling well. Please check the subject's condition." If "Status Level 4" or "Status Level 5" occurs three or more times in a row, the notification unit 18 determines the warning text to be "The subject may have collapsed. Please check the subject's condition immediately."

[0074] The notification unit 18 may output a notification indicating that the warning will be canceled if "State Level 1" to "State Level 3" is estimated after "State Level 4" or "State Level 5" has occurred two or more times in a row.

[0075] <Server processing flow> When new person information 20 is registered in the storage unit 10, or when the person information 20 is updated, the service provider server 100 performs preparatory processing to estimate the status level of the person 400 corresponding to the person information 20, and then performs monitoring processing for the person 400. An example of the processing flow for each of the preparatory processing and monitoring processing is described below.

[0076] (Preparation process) Figure 10 is a flowchart showing the flow of the preparation process. The flow shown in Figure 10 is executed on new subject information 20 when new subject information 20 is registered in the storage unit 10, or when subject information 20 is updated.

[0077] First, the processor 101 obtains load survey data for the most recent predetermined period (e.g., 1 year, 3 years, etc.) corresponding to the subject information 20 from the management server 300 as training data 21 (step S1).

[0078] Next, the processor 101 identifies normal periods in which the subject 400's condition is normal and abnormal periods in which the subject 400's condition is abnormal, based on the hospitalization and outpatient information 20d contained in the subject information 20 (step S2).

[0079] Next, the processor 101 divides the training data 21 into normal data 21a corresponding to normal periods and abnormal data 21b corresponding to abnormal periods (step S3).

[0080] Next, the processor 101 divides the normal data 21a into reference waveform generation data 21a1 and verification data 21a2, and divides the abnormal data 21b into reference waveform generation data 21b1 and verification data 21b2 (step S4). In step S4, the processor 101 may remove error data (for example, data with abnormally high power consumption).

[0081] For example, the processor 101 randomly selects a portion of the normal data 21a and determines it as reference waveform generation data 21a1, and the remainder as verification data 21a2. Similarly, the processor 101 randomly selects a portion of the abnormal data 21b and determines it as reference waveform generation data 21b1, and the remainder as verification data 21b2.

[0082] Next, the processor 101 generates and stores first reference data 23 and second reference data 24 based on the reference waveform generation data 21a1 and 21b1 (step S5). For example, the processor 101 calculates representative values ​​of the power consumption for each time period at regular intervals throughout the day. Then, the processor 101 generates first reference data 23 and second reference data 24 by associating each time period with the representative values ​​of the power consumption for that time period. The processor 101 stores the generated first reference data 23 and second reference data 24.

[0083] In step S5, the processor 101 may calculate various statistical values ​​obtained from waveforms representing the time-dependent changes in a typical value of power consumption, and may include data showing these statistical values ​​in the first reference data 23 and the second reference data 24.

[0084] Next, the processor 101 refers to the subject information 20 to identify the activity period (step S6). For example, the processor 101 identifies the period from waking time to going to bed time as the activity period based on the medical interview information 20e (see Figure 4) included in the subject information 20. Furthermore, the processor 101 may exclude the period from leaving the house to returning home from the activity period.

[0085] Next, in step S7, the processor 101 determines the functional level of the subject 400 based on the response information 20f included in the subject information 20, and determines the lifestyle pattern of the subject 400 using the first reference data 23 and waveform models 25a to 25c. Then, the processor 101 determines the category to which the subject 400 belongs, according to the functional level and lifestyle pattern of the subject 400.

[0086] Next, in step S8, the processor 101 sets a first threshold Th1 (thresholds Th1_1, Th1_2, Th1_3) and a second threshold Th2 according to the category to which the subject 400 belongs.

[0087] Next, in step S9, the processor 101 extracts multiple first partial data 21a3 at regular intervals (e.g., 12 hours, 1 day) from the verification data 21a2, and extracts multiple second partial data 21b3 at regular intervals from the verification data 21b2. Then, for each of the multiple first partial data 21a3 and multiple second partial data 21b3, the processor 101 calculates a third similarity of the waveform shown by the partial data with respect to the first reference waveform shown by the first reference data 23. Furthermore, for each of the multiple first partial data 21a3 and multiple second partial data 21b3, the processor 101 calculates a fourth similarity of the waveform shown by the partial data with respect to the second reference waveform shown by the second reference data 24.

[0088] Next, in step S10, the processor 101 estimates the state level for each of the multiple first partial data 21a3 and multiple second partial data 21b3 using the third and fourth similarity values ​​and the first threshold Th1 and second threshold Th2. The processor 101 can estimate the state level according to the estimation method described above with reference to Figure 8.

[0089] Next, in step S11, the processor 101 determines whether or not adjustment of the first threshold Th1 and the second threshold Th2 is necessary. For example, if the ratio of the total number of first partial data 21a3 estimated to be "state level 4" or "state level 5" to the total number of second partial data 21b3 estimated to be "state level 1" to "state level 3" is less than a baseline value, the processor 101 determines that no adjustment is necessary. On the other hand, if the ratio is greater than or equal to the baseline value, the processor 101 determines that adjustment is necessary.

[0090] If adjustment is deemed necessary (YES in step S11), the processor 101 adjusts the first threshold Th1 and the second threshold Th2 (step S12). For example, if the ratio of the number of first partial data 21a3 estimated to be "state level 4" or "state level 5" to the total number of multiple first partial data 21a3 is greater than or equal to a reference value, the processor 101 adjusts the first threshold Th1 to be reduced by a certain amount. If the ratio of the number of second partial data 21b3 estimated to be "state level 1" to "state level 3" to the total number of multiple second partial data 21b3 is greater than or equal to a reference value, the processor 101 adjusts the first threshold Th1 to be increased by a certain amount and the second threshold Th2 to be increased by a certain amount. After step S12, the process returns to step S10.

[0091] If it is determined that no adjustment is necessary (NO in step S11), the processor 101 saves threshold data 26 indicating the first threshold Th1 and the second threshold Th2, and terminates processing.

[0092] (Monitoring process) Figure 11 is a flowchart showing the flow of the monitoring process. The flow shown in Figure 11 is executed repeatedly at regular intervals (for example, every 12 hours (half a day), every day).

[0093] First, in step S1, the processor 101 acquires load survey data for the most recent fixed period, which is the monitoring period, from the management server 300 as the monitoring data 22.

[0094] In the next step S22, the processor 101 calculates a first similarity of the target waveform, indicated by the monitored data 22, to the first reference waveform, indicated by the first reference data 23. Furthermore, the processor 101 calculates a second similarity of the target waveform to the second reference waveform, indicated by the second reference data 24.

[0095] In the next step S23, the processor 101 estimates the state level of the subject 400 during the monitoring period based on the first threshold Th1 and the second threshold Th2 indicated by the threshold data 26, as well as the first and second similarity scores. The processor 101 updates the estimation result data 27 according to the estimation result.

[0096] In the next step S24, the processor 101 outputs a notification corresponding to the estimated result data 27. Specifically, the processor 101 reads the email address from the subject information 20 and sends a notification email addressed to the read email address. The processor 101 may omit step S24 if the history of the state levels indicated by the estimated result data 27 satisfies predetermined conditions. For example, the predetermined conditions are that the state levels for the past four monitoring periods have all been "state level 1" to "state level 3".

[0097] <Variation> (Variation 1) In the above explanation, the processor 101, which operates as the estimation unit 17, estimates the state of the subject 400 to be "state level 5" when the first similarity is less than or equal to the threshold Th1_3 and the second similarity exceeds the second threshold Th2. However, the processor 101 may change the state level for the most recent monitoring period from "state level 4" to "state level 5" if "state level 4" occurs for a predetermined number of consecutive times (for example, 3 or 4 times).

[0098] (Modification 2) The service provider server 100 may collect weather data for the area to which the residences 200 of the target persons 400 belong, and determine the content of the notification by referring to the weather data. The weather data includes temperature, humidity, etc.

[0099] For example, the processor 101, which operates as the notification unit 18, generates a notification indicating a suspected case of heatstroke when "state level 4" or "state level 5" is estimated multiple times in a row and the maximum temperature exceeds a standard value (e.g., 30°C).

[0100] (Variation 3) The processor 101, which operates as the notification unit 18, may refer to the subject information 20 to determine the content of the notification. For example, if "state level 4" or "state level 5" is estimated multiple times in a row, and the medical interview information 20e included in the subject information 20 indicates an underlying disease "stroke", the processor 101 will generate a notification that the person may have collapsed due to a stroke.

[0101] (Modification 4) In the above explanation, the flow shown in Figure 11 is assumed to be executed repeatedly at regular intervals (for example, every 12 hours (half a day), or every day). However, the flow shown in Figure 11 may be executed excluding inactive periods, based on the questionnaire information 20e included in the subject information 20. Inactive periods include, for example, bedtime and time spent outside. Bedtime is determined by the time spent going to bed and the time spent waking up. Time spent outside is determined by the day spent outside, the time spent outside, and the time spent returning home.

[0102] (Variation 5) In the above explanation, it was assumed that the normal period and the abnormal period are identified based on the date of hospitalization. However, the processor 101, which operates as the identification unit 13, may also identify the normal period and the abnormal period according to the specifications of the subject 400 or the user 600. For example, a caregiver who is monitoring the subject 400 inputs into the terminal 500 the periods when the subject 400 was in good health and the periods when the subject 400 was in poor health. The processor 101 can then identify the periods when the subject 400 was in good health as the normal period and the periods when the subject 400 was in poor health as the abnormal period.

[0103] (Experimental variation 6) In the example above, the subjects 400 were classified into one of 16 categories corresponding to 16 possible combinations of the four functional levels and four lifestyle patterns. However, subjects 400 may also be classified into one of several categories using only one of the functional levels or lifestyle patterns. For example, only the lifestyle patterns may be determined according to the method shown in Figure 7. In this case, subjects 400 would be classified into one of the four categories corresponding to each of the four lifestyle patterns.

[0104] The embodiments disclosed herein should be considered in all respects to be illustrative and not restrictive. The scope of the present invention is indicated by the claims rather than by the foregoing description, and all modifications within the meaning and scope equivalent to the claims are intended to be included. [Explanation of Symbols]

[0105] 2 Communication line, 3 Internet, 10 Storage unit, 11 Registration unit, 12 Acquisition unit, 13 Identification unit, 14 Reference data generation unit, 15 Classification unit, 16 Setting unit, 17 Estimation unit, 18 Notification unit, 20 Target person information, 20a ID information, 20b Number information, 20c Notification destination information, 20d Hospital admission / outpatient information, 20e Medical interview information, 20f Response information, 21 Learning data, 21a Normal system data, 21a1, 21b1 Reference waveform generation data, 21a2, 21b2 Verification data, 21a3 First part data, 21b Abnormal system data, 21b3 Second part data, 22 Monitoring target data, 23 First reference data, 24 Second reference data, 25a~25c Waveform model, 26 Threshold data, 27 Estimation result data, 100 Service provision server, 101 Processor, 102 ROM, 103 RAM, 104 Hard disk, 105 Communication interface, 107 Power supply circuit, 110 Program, 200 Residence, 210 Smart meter, 220 Electrical equipment, 300 Management server, 400 Target audience, 500, 500a, 500b Terminals, 600 Users, SYS system.

Claims

1. It is a server, An acquisition unit that acquires data showing the amount of electricity consumed at the residence of the person being monitored for each time period, A specific unit that identifies a first period in which the condition of the person being monitored is normal, and a second period in which the condition of the person being monitored is abnormal. A server comprising: an estimation unit that estimates the state level of the person being monitored during the monitoring period based on a first comparison result between a target waveform representing the change in power consumption over time during the monitoring period and a first reference waveform representing the change in power consumption over time during the first period, and a second comparison result between the target waveform and a second reference waveform representing the change in power consumption over time during the second period.

2. The first comparison result shows the relationship between the first similarity of the target waveform to the first reference waveform and the first threshold. The server according to claim 1, wherein the second comparison result indicates the relationship between the second similarity of the target waveform to the second reference waveform and the second threshold.

3. The first similarity represents the degree of similarity between the first statistical value calculated from the first reference waveform and the target statistical value calculated from the target waveform. The server according to claim 2, wherein the second similarity represents the degree of similarity between the second statistical value calculated from the second reference waveform and the target statistical value.

4. A classification unit that classifies the person being monitored into one of several categories based on at least one of the person's functional level and lifestyle pattern, The server according to claim 2 or 3, further comprising a setting unit that sets the first threshold and the second threshold according to the category to which the person being monitored belongs.

5. The server according to claim 4, wherein the classification unit determines the lifestyle pattern of the person being monitored based on the comparison result between each waveform model of a plurality of lifestyle patterns and the first reference waveform.

6. The server according to claim 4, wherein the classification unit determines the functional level of the person being monitored based on the activity capacity index of the person being monitored.

7. The server according to claim 4, wherein the setting unit adjusts the first threshold and the second threshold according to the state level estimated using data indicating the amount of power consumption during a verification period prior to the monitoring period.

8. The server according to claim 1, further comprising a notification unit that outputs a notification corresponding to the aforementioned state level.

9. The notification unit determines the content of the notification based on the transition of the state level over a plurality of consecutive monitoring periods, according to claim 8.

10. The steps include obtaining data showing the amount of electricity consumed at the residence of the person being monitored, for each time period, The steps include identifying a first period in which the person being monitored is in a normal state and a second period in which the person being monitored is in an abnormal state. An information processing method comprising the steps of: estimating the state level of the person being monitored during the monitoring period based on a first comparison result between a target waveform representing the change in power consumption over time during the monitoring period and a first reference waveform representing the change in power consumption over time during the first period, and a second comparison result between the target waveform and a second reference waveform representing the change in power consumption over time during the second period.

11. A program that causes a computer to execute the information processing method described in claim 10.

Images