JP6185683B1 - Watch apparatus, watch system equipped with the same, and watch method - Google Patents

Abstract

It is possible to detect the occurrence of an abnormality in a resident regardless of various life patterns of the resident in the residence. A power data acquisition unit (31) in which a monitoring device (5) acquires power usage data related to power used in a residence (2) in time series, and current target data in the power usage data acquired in the time series. , An activity quiescence degree calculation unit 32 that calculates an index of the amount of change in power usage based on the power usage data immediately before the target data, and the amount of change in power usage calculated in the time series A divergence degree calculation unit 33 for calculating a divergence degree between a current index for a predetermined time zone and one or more past indexes corresponding to the time zone before yesterday. The configuration. [Selection] Figure 3

Description

  The present invention relates to a technique for detecting the occurrence of an abnormality of a resident in a residence, and in particular, a monitoring device suitable for detecting the occurrence of an abnormality in an elderly person living alone, a monitoring system including the same, and a monitoring system Regarding the method.

  In recent years, in Japan, the number of elderly people and elderly households living apart from the younger generation has increased with changes in the population structure and lifestyles, such as the declining birthrate and aging population. In particular, it is not easy for single elderly people to manage their own health, and even if there are any abnormalities in the elderly's body in the residence, it can be done by separated families and neighboring residents. It is difficult to detect early. Therefore, when an abnormality occurs in such an elderly person or the like, there is a need for a watching technique for detecting the occurrence of the abnormality at an early stage and notifying a family or the like living away.

  As this type of technology, for example, a monitoring device for detecting anomalies of a single person (resident) based on time series data of total power consumption measured by a watt-hour meter in a residence is installed. Based on the comparison of the past usage pattern of total power extracted from the accumulation of time series information of total power usage and the total power usage status obtained from the current time series information of total power usage For example, an anomaly detection method is known in which the health state of a single person is estimated by detecting that the current total power consumption is small or that the total power consumption is small (patent) Reference 1).

JP-A-2005-216113

  In the prior art described in Patent Document 1, a pattern that appears at a higher probability in the target residence from among several predetermined pattern extraction criteria as a pattern of the past usage situation of all power. In addition, by setting a determination condition for determining a resident's abnormality for a part of a predetermined time period based on the pattern, determination of abnormality for those predetermined time period Is done.

  On the other hand, in the above prior art, in identifying a pattern that appears at a higher probability in the target residence, for example, typical pattern extraction related to characteristic changes caused by preparation of meals in the morning, noon, and evening It is necessary to predetermine a standard, and for a resident who has a living pattern that does not meet the pattern extraction standard, a pattern that appears with high probability may not be specified. Moreover, in the said prior art, since abnormality determination is not performed about the time slot | zone which cannot identify the pattern which appears with such a high probability, a resident's abnormality cannot be detected at all.

  The present invention has been devised in view of such problems of the prior art, and is a monitoring device capable of detecting the occurrence of an abnormality in a resident regardless of various life patterns of the resident in the residence. The main purpose is to provide a watching system and a watching method provided with this.

  The monitoring device of the present invention is a monitoring device that detects an occurrence of an abnormality of a resident in a residence, and includes a power data acquisition unit that acquires power usage data related to power used in the residence in time series, and the time A power change amount calculation unit that calculates, in time series, an index of a change amount of power usage with reference to the power usage data immediately before the target data with respect to current target data in the power usage data acquired in series And an index of change in power consumption calculated in time series between a current index for a predetermined time zone and one or more past indexes corresponding to the time zone before yesterday And a divergence degree calculation unit for calculating the divergence degree.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to detect generation | occurrence | production of the abnormality in the said resident irrespective of the various life patterns of the resident in a residence.

Overall configuration diagram of the watching system 1 according to the first embodiment Hardware configuration diagram of the monitoring device shown in FIG. Functional block diagram of the monitoring device shown in FIG. Flow chart showing the flow of watching process by the watching device shown in FIG. Explanatory drawing which shows the calculation object of the activity suspension degree in step ST102 in FIG. Explanatory drawing which shows the calculation object of the deviation degree in step ST103 in FIG. Explanatory drawing which shows the calculation method of the deviation degree in step ST103 in FIG. Explanatory drawing which shows the example of notification of the degree of watching required in step ST104 in FIG. Explanatory drawing which shows the modification of the calculation object of the activity pause degree shown in FIG. Explanatory drawing which shows the modification of the calculation method of the deviation shown in FIG. Functional block diagram of the watching device according to the second embodiment The flowchart which shows the flow of the watching process by the watching apparatus shown in FIG. Explanatory drawing which shows the example of notification of abnormality occurrence in step ST104 in FIG. Functional block diagram of the watching device according to the third embodiment FIG. 14 is a functional block diagram showing details of the watching necessity level information calculation unit Explanatory drawing which shows the example of notification of the monitoring necessity degree which concerns on 3rd Embodiment The figure which shows the 1st example of the evaluation screen of the monitoring system 1 displayed on an external device The figure which shows the 2nd example of the evaluation screen of the monitoring system 1 displayed on an external device The figure which shows the 3rd example of the evaluation screen of the monitoring system 1 displayed on an external device

  1st invention made | formed in order to solve the said subject is a monitoring apparatus which detects generation | occurrence | production of the abnormality of the resident in a residence, Comprising: The electric power usage data regarding the electric power used by the said residence is acquired in time series With respect to the current target data in the power usage data acquired in the time series, a power data acquisition unit and an index of the amount of change in power usage based on the power usage data immediately before the target data 1 corresponding to the current index for a predetermined time zone and the time zone before yesterday in the power change amount calculation unit calculated in step 1 and the power usage change index calculated in time series. A divergence degree calculating unit for calculating a divergence degree with the above past index is provided.

  According to this, as an index for determining the occurrence of an abnormality in the resident, an indicator of the current amount of change in power usage for a predetermined time zone and the yesterday for the corresponding time zone Since the degree of deviation from the index of the amount of change in power consumption in the past is used, it becomes possible to detect the occurrence of abnormality in the resident regardless of the various life patterns of the resident in the residence. .

  In the second invention, the current target data includes a plurality of pieces of power usage data relating to a plurality of unit times that are continuous in time series.

  According to this, it is possible to handle an index of the amount of change in the power usage for a plurality of unit times as a vector, and it is possible to detect the occurrence of an abnormality in the resident from multiple viewpoints.

  In the third invention, the degree of divergence is calculated based on a distance between the current index and the past index in a feature space whose characteristic amount is an index of change in the amount of power usage. And

  According to this, it is possible to calculate an index for determining occurrence of abnormality in the resident by a simple method based on the distance in the feature space.

  In the fourth invention, the power usage data acquired in time series includes only power usage data within a predetermined period set in advance.

  According to this, in detecting the occurrence of an abnormality in the resident, only the power usage data within a relatively new predetermined period is used (that is, the past inappropriate power usage data after the predetermined period has been excluded). Thus, it becomes possible to appropriately respond to changes in the lifestyle patterns of residents. In this case, the predetermined period can be individually set according to the life pattern of the target resident.

  According to a fifth aspect of the present invention, the power consumption change index is an index that increases as the power consumption decreases.

  According to this, it is possible to easily grasp the situation where there is a high possibility that the occupant has an abnormality (that is, the change in power consumption is small).

  The sixth invention is characterized by further comprising an abnormality notifying unit that transmits the degree of divergence to a predetermined external device.

  According to this, the status of occurrence of an abnormality in the resident is easy for a watcher who is a user of the external device (for example, a person who should understand the occurrence of an abnormality in a family or other resident who lives separately from the resident). Can be grasped.

  The seventh invention is characterized by further comprising an abnormality determination unit that determines whether or not an abnormality has occurred in the resident based on the degree of deviation.

  According to this, it is possible to make the watcher more clearly grasp the presence or absence of an abnormality in the resident.

  In the eighth invention, when an abnormality actually occurs in the resident, the power usage data corresponding to the time zone at the time of the abnormality is excluded from the power usage data acquired in time series. It is characterized by that.

  According to this, it is possible to more accurately detect the occurrence of an abnormality in the resident thereafter by avoiding inappropriate power usage data at the time of occurrence of the abnormality being used for calculating the deviation degree. .

  In the ninth invention, the storage unit that stores the power usage data acquired by the power data acquisition unit, and the user's evaluation information related to the divergence degree are acquired from the external device, and based on the evaluation information A data operation unit for determining regular power usage data used for calculating the power usage amount change index by the power change amount calculation unit among the power usage amount data stored in the storage unit; It is characterized by having.

  According to this, it is possible to calculate the power usage amount change index and the divergence degree using appropriate power usage data based on the user evaluation information.

  In the tenth invention, the monitoring device according to any one of the first to ninth inventions is connected to the monitoring device via a network so as to be communicable and used in the residence. The watt-hour meter which transmits the electric power consumption data regarding to the said electric power data acquisition part is provided.

  The eleventh aspect of the invention is a monitoring method for detecting the occurrence of an abnormality of a resident in a residence, the power data acquisition step for acquiring power usage data related to the power used in the residence in time series, With regard to the current target data in the power usage data acquired in time series, power change calculation that calculates in time series an index of the amount of power usage change based on the power usage data immediately before the target data. And a current index for a predetermined time zone and at least one past index corresponding to the time zone before yesterday in the index of change in power usage calculated in time series And a divergence degree calculating step for calculating a divergence degree between them.

  According to the tenth and eleventh inventions, as an index for determining the occurrence of an abnormality in a resident, an index of the current amount of change in power consumption for a predetermined time zone, and a corresponding time Detecting the occurrence of anomalies in the occupants regardless of the various lifestyle patterns of the occupants in the residence, using the degree of divergence from the index of the amount of change in the power consumption in the past before yesterday. It becomes possible to do.

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

(First embodiment)
FIG. 1 is an overall configuration diagram of a watching system 1 according to the first embodiment of the present invention.

  The monitoring system 1 detects the occurrence of an abnormality of the resident 3 based on the power usage in the residence 2, and the smart meter 4 that measures the power usage related to the power used in the residence 2 and the smart meter 4 Based on the measured power consumption data (hereinafter referred to as “power consumption data”), the monitoring device 5 that detects the occurrence of abnormality in the resident 3 and the abnormality in the resident 3 detected by the monitoring device 5 And an external device 6 that acquires information related to the occurrence of the above. In addition, the abnormality of the resident 3 to be detected in the monitoring system 1 is an abnormality that can affect the power usage amount of the resident 2, such as a physical abnormality of the resident 3 due to illness or injury.

  In this embodiment, the resident 3 to be watched over is an elderly person living alone. However, the present invention is not limited to this. . Further, the resident 3 is not limited to a single person but may be a plurality of people such as elderly couples or relatives. In FIG. 1, for convenience of explanation, only one residence 2 where the resident 3 lives is shown. However, in practice, the monitoring system 1 is intended to monitor the resident who lives in each of a plurality of residences. Is possible.

  The smart meter 4 is a known watt-hour meter having a communication function capable of transmitting power usage data to a predetermined communication partner. In addition, the smart meter 4 transmits power usage data at a predetermined time interval to the monitoring device 5 that is communicably connected via the network 10. Here, the Internet is used as the network 10, but not limited to this, other known communication networks can be used, and the communication line used in the network 10 is also a mobile phone line or a dedicated line. It doesn't matter what kind.

  The watching device 5 is configured by a server capable of executing various information processing, and has a known communication function capable of communicating with a predetermined communication partner such as the smart meter 4 and the external device 6 via the network 10. In the present embodiment, the watching device 5 is installed in the monitoring center 11 managed by the administrator of the watching system 1.

  The external device 6 is an information device owned by a watcher who uses the watch system 1 (for example, a person who should understand the occurrence of an abnormality in a family or other resident who lives separately from the resident). It comprises a known mobile phone (smart phone). Also, the external device 6 can acquire various information from the watching device 5 via the network 10.

  The external device 6 is not limited to a mobile phone, and other information devices such as a tablet and a laptop computer having the same function can be used. In the monitoring system 1, the external device 6 can be omitted when the administrator of the monitoring center 11 can perform a response such as rescue when an abnormality occurs in the resident 3.

  2 and 3 are a hardware configuration diagram and a functional block diagram of the watching device 5 shown in FIG. 1, respectively.

  As shown in FIG. 2, the watching device 5 includes a server having known hardware. The main body of the watching device 5 functions as a processor 21 including a CPU (Central Processing Unit) that performs various information processing and control of peripheral devices based on a predetermined control program, and a work area of the processor 21. A random access memory (RAM) 22, a read only memory (ROM) 23 for storing control programs and data executed by the processor 21, a network interface 24 for executing communication processing via the network 10, and the like. The components are connected to each other via a bus 20.

  The monitoring device 5 is a peripheral device that includes an input device 25 that includes a keyboard, a touch panel, and the like for an administrator of the monitoring system 1 to input various commands and data, and a monitoring device that will be described in detail later. A monitor 26 for displaying various types of information related to the processing, and a storage 27 including an HDD (Hard Disk Drive) for storing various data used in the monitoring process, a calculation result of the data, and the like are provided. At least a part of the various functions of the monitoring device 5 described in detail later with reference to FIG. 3 can be realized by the processor 21 executing a predetermined control program (for example, a program for watching processing).

  The monitoring device 5 is not limited to a server, and other information devices (such as a computer) that perform the same function can also be used. Moreover, you may substitute at least one part of the function of the monitoring apparatus 5 by the process by other well-known hardware.

  As shown in FIG. 3, in the monitoring device 5, the power data acquisition unit 31 sequentially acquires power usage data in time series by communication with the smart meter 4 via the network 10. The acquired power usage data is sequentially stored in the storage unit 30 configured by the storage 27 and the like described above.

  Further, as will be described in detail later, in the monitoring device 5, the activity cessation degree calculation unit (power change amount calculation unit) 32 includes the current target data in the power usage data acquired by the power data acquisition unit 31 in time series ( With respect to the latest data that is the detection target of the occurrence of abnormality in the resident 3), an index of the amount of change in power usage based on the power usage data immediately before the target data (hereinafter referred to as “activity pause”). ) In time series. Further, the divergence degree calculation unit 33 is the same as the activity cessation degree calculated for the predetermined time zone in the activity cessation degree calculated in time series by the activity cessation degree calculation unit 32, the same as that before yesterday. The degree of divergence (degree of divergence) between one or more past activity pauses corresponding to the time zone is calculated. The calculated activity pauses and deviations are sequentially stored in the storage unit 30.

  Further, in the monitoring device 5, the display unit 34 is configured by the monitor 26 described above, and monitors the monitoring necessity level information including at least information on the divergence degree calculated by the divergence degree calculation unit 33. indicate. This monitoring necessity level information is information used to determine the possibility that an abnormality has occurred in the resident 3, and the monitoring necessity level information includes the time series power consumption corresponding to the information on the degree of deviation. Related information such as can be added. In addition, the abnormality notification unit 35 converts the monitoring necessity degree information (or part of the information) displayed on the display unit 34 into data that can be displayed on the external device 6 through communication with the external device 6 via the network 10. Convert and send. Thereby, the user of the external device 6 can use the monitoring necessity degree information received by the external device 6 as information for determining whether or not an abnormality has occurred in the resident 3. It is possible to easily grasp the current situation regarding the occurrence of abnormalities.

  In addition, it is good also as a structure which installs not only the above-mentioned structure but the information apparatus which can perform at least one part of the function of the monitoring apparatus 5 or the monitoring apparatus 5 in each residence 2, for example. In that case, for example, an information device (such as a server) having a display function equivalent to the above-described display unit 34 and a communication function equivalent to the abnormality notification unit 35 can be separately installed in the monitoring center 11. . In some cases, the smart meter 4 may be configured to execute at least a part of the functions of the monitoring device 5.

  FIG. 4 is a flowchart showing the flow of the watching process by the watching device 5, FIG. 5 is an explanatory diagram showing the calculation target of the activity pause level in step ST102 in FIG. 4, and FIGS. FIG. 8 is an explanatory diagram showing a divergence degree calculation target and a divergence degree calculation method in step ST103, and FIG. 8 is an explanatory diagram showing a notification example of the watching necessity level in step ST104 in FIG.

  As shown in FIG. 4, when the watching process is started, the watching apparatus 5 first acquires power usage data from the smart meter 4 (ST101). The acquisition of the power usage data is performed by acquiring power usage data (here, cumulative value of power used for 30 minutes) sequentially transmitted from the smart meter 4 every predetermined unit time (here, every 30 minutes). ) This is performed by the watching device 5 receiving sequentially. The received power usage data is accumulated in the storage unit 30 as time-series data to which time information (information regarding measurement date / time by the smart meter 4, information regarding reception date / time by the monitoring device 5, etc.) is added.

  In the present embodiment, the power usage amount data relates to the total power usage amount in the residence 2, but is not limited to this, and one or more predetermined electrical appliances in which the variation in the power usage amount is relatively large in the residence 2. It may be related.

  Further, the storage unit 30 can be configured to store only power usage data within a predetermined period (for example, 60 days) set in advance. As a result, in the monitoring process, only the power usage data within a relatively new predetermined period is used (that is, the past inappropriate power usage data after a long period of time is excluded), and the seasonal change Accordingly, it is possible to appropriately cope with the change in the life pattern of the resident 3 according to the above. Moreover, about a predetermined period, it can set separately according to the living pattern of the resident 3 used as object.

Next, the watching device 5 executes a process of calculating the degree of inactivity based on the power usage data accumulated as time series data in the storage unit 30 (ST102). Here, the calculation of the degree of inactivity is repeated at approximately the same time interval as the acquisition of the power usage data (that is, every time new power usage data is acquired). The watching device 5 can calculate the inactivity level N _t from the following equation (1) composed of a sigmoid function.

In the equation (1), the gain a is a positive real number that can be appropriately determined. In addition, x is the power change amount E _t −E _{t−1 at} time t (more specifically, the power usage amount E _t acquired at time t and the power usage acquired at time t−1 immediately before it). (The difference from the amount _Et-1 )) is the reciprocal of the absolute value, and the value increases as the change in the power consumption decreases. The calculation of the hiatus of N _t By using the sigmoid function, hiatus degree N _t even x diverges to infinity converges to 1.

  Normally, when an abnormality occurs in the resident 3, since the ON / OFF of the lighting fixtures and home appliances is lost, the power change is reduced. For example, since the power consumption of the luminaire is relatively small, it can be said that the situation where the power change amount is small in the watching process is a situation where the power change should be grasped more reliably.

As described above, the degree of inactivity N _t obtained by using the reciprocal of the absolute value of the power change amount as an input of the sigmoid function is amplified by a small power change amount and converges to a constant value (here, 1) (ie, Therefore, it is extremely useful for grasping the power change in a situation where the power change amount is small.

Incidentally, the convergence curve of the hiatus of N _t can be adjusted by the gain a. When the gain a is increased, the activity deactivation degree N _t rapidly approaches 1 as x increases, and conversely, when the gain a is decreased, the activity deactivation degree N _t gradually approaches 1. By appropriately setting the gain a, it is possible to more accurately detect a power change in a situation where the power change amount is small.

On the other hand, when the amount of power change is large (that is, the situation where the resident 3 is operating normally), the activity suspension degree N _t takes a small value. The discrepancy S _t to be described later is a value based on the Euclidean distance between the plurality of hiatus of N _t, but by hiatus of N _t when a large power variation is a small value, the power change amount hiatus of N _t of the greater will be difficult to affect the values of the degree of deviation S _t.

In the present embodiment, 48 pieces of power usage data (power usage amount in units of 30 minutes) are acquired from the smart meter 4 at intervals of 30 minutes per day (24 hours). For example, as shown in FIG. 5, the degree of inactivity N _t calculated at 21:00 (usually immediately after 21:00, the same applies hereinafter) is calculated based on the power usage E _t acquired at 21:00 and 20 immediately before that. It becomes a difference from the power usage amount _Et-1 acquired at 30 minutes. The calculated activity inactivity level N _t is accumulated in the storage unit 30 as time series data.

In the present embodiment uses a hiatus of N _t consisting of the above-described sigmoid function as an indication of power usage variation, not limited to this and with other known functions (e.g., logistic function) Also good. However, as described above hiatus of N _t, by setting an indication of power usage variation to be greater as the change in power consumption is small, can any abnormality residents 3 occurs It is possible to easily grasp a situation with high performance (that is, a small change in power consumption).

Subsequently, the watching device 5 executes a process of calculating a divergence degree from the degree of inactivity stored as time series data in the storage unit 30 (ST103). Here, the calculation of the degree of divergence is repeatedly performed at substantially the same time interval as the calculation of the degree of inactivity (that is, every time a new degree of inactivity is calculated). In the monitoring device 5, based on the k-nearest neighbor method, the Euclidean distance (a plurality of past activity pauses) between the newly calculated activity pause and one or more activity pauses related to the past corresponding time period accumulated in the storage unit 30. when using a hiatus degree, Euclidean distance (hereinafter, "distance" and referred.) calculating an average value of) the degree of deviation S _t. The calculated deviation degree _St is accumulated in the storage unit 30 as time series data.

For example, as shown in FIG. 6, the watch device 5, in the calculation of the deviation degree S _t, first, the latest (in this case the current date D, the power consumption of 0 30 pm to 1 o'clock calculated at 1 in based) and hiatus of n _{t_d,} date D-. 1 to date D-n (n past stored in the storage unit 30 (formerly yesterday) is corresponding to time zones of a predetermined number of days) (i.e., Based on the amount of power used in the time zone from 0:30 to 1 o'clock calculated from 1 o'clock corresponding to the target time zone of the latest activity dormancy level N _{t_d} ) activity dormancy degree N _{t_d-1} to date N _{t_d-} Calculate the distance to _n .

Therefore, the watching device 5 includes the current (latest) activity deactivation level N _{t_d} and _three of the past activity deactivation levels that are relatively close to the current activity deactivation level N _{t_d} (here, k by the k neighborhood method). = 3 to) calculates the average value of the distances between the dormant degree as deviance S _t. In the present embodiment, the activity suspension degree N _{t_d} is _{calculated with respect} to predetermined one power usage data (power usage data related to one unit time), and therefore, for example, as shown in FIG. _t is calculated on the basis of the distance between the current activity level N _{t_d} on one axis and three activity levels (shown in a circle in FIG. 7) that are relatively close.

Note that past hiatus degree used in deviance S _t is not necessarily a multiple of at least one past hiatus degree (e.g., yesterday one hiatus about time zones corresponding to the current time zone Or one activity suspension degree on the same day of the week as of the present day before yesterday regarding the time zone corresponding to the current time zone). Also, when calculating the degree of deviation S _t by using a plurality of past hiatus degree is not limited to the average value of a plurality of distance as described above, other known statistical representative value (e.g., a central It can also be with a value) and the degree of deviation S _t.

  In addition, when executing the watching process, for example, the power usage data for the past 30 days is acquired and accumulated, and the degree of inactivity is calculated and accumulated. The degree of divergence may be calculated using the degree. The past activity cessation degree is information that substantially represents a resident's life pattern and is not uniform. For example, even in the same time zone (for example, from 12:30 to 1 o'clock), a resident may be at home (using home appliances, etc., and the degree of inactivity is generally small) You may have gone out (home appliances are not used, and the degree of inactivity is generally large). In such a case, the activity suspension degree is generated based on two life patterns such as being at home or absent. Then, when the latest value of activity quiescence is calculated, if this value is close to the value of activity quiescence obtained when at home or away, the degree of divergence will be small, while at home and away The greater the difference from the value of the inactivity level acquired at the time, the greater the difference is calculated.

  Now, the process of calculating and accumulating past activity pauses is nothing but learning of the resident's life pattern. This learning is a kind of supervised learning, and it can be considered that the calculated activity pause is labeled “No Abnormal” on the assumption that there are no abnormalities in the life patterns obtained in the past. . Even if a plurality of life patterns are included in the same time zone, the divergence degree calculation unit 33 according to the present embodiment reverses a small divergence degree as a numerical value if it is similar to a life pattern that has already been learned. If it is different from the life pattern that is the subject of learning, a large degree of deviation will be output. That is, the divergence degree calculation unit 33 outputs a large degree of divergence when an activity idle degree that has not been learned so far appears (different from a known life pattern). As described above, in the present embodiment, the divergence degree is calculated on a non-rule basis.

  Next, the watching device 5 monitors (see) (transmits) the deviation degree within a predetermined period including the current (latest) deviation degree calculated in step ST103 and the power usage corresponding to the deviation degree as necessary degree information. (ST104). Here, for example, the watching device 5 can create an e-mail attached with a file including the watching necessity information, and can send the e-mail to the e-mail address for the watcher at a predetermined frequency. Alternatively, the monitoring device 5 may transmit the monitoring necessity information to the external device 6 only when there is a request from predetermined application software executed by the external device 6. As a result, the watcher who is the user of the external device 6 can display the monitoring necessity information on the monitor of the external device 6 at a desired timing via the web browser or dedicated browsing application software.

  In addition, the monitoring device 5 displays the monitoring necessity level information to the administrator of the monitoring center 11 in parallel with the notification in step ST104 (ST105). Thereby, even when the user of the external device 6 cannot cope with the occurrence of the abnormality in the resident 3, the manager of the monitoring center 11 can deal with it.

  In steps ST104 and ST105 described above, for example, as shown in FIG. 8, the monitoring necessity level information is displayed on the monitor 26 of the monitoring device 5 in the external device 6 and the monitoring center 11. In the example of FIG. 8, the degree of divergence and power consumption for the past one week including the present (July 7) are included in the monitoring necessity level information. The user of the external device 6 or the administrator of the monitoring center 11 can determine the possibility that an abnormality has occurred in the resident 3 as an index indicating the necessity of watching over the degree of deviation.

  In the watching process by the watching device 5, a configuration in which one of the steps ST104 and ST105 is omitted is also possible. In that case, the monitoring device 5 can also omit the function and hardware for executing the omitted steps in the same manner.

  In the watching process, each time new power usage data is acquired from the smart meter 4, the above steps ST101 to ST105 are repeatedly executed.

  As described above, in the monitoring system 1 according to the first embodiment, as an index for determining the occurrence of an abnormality in the resident 3, the degree of inactivity (amount of change in current power consumption) for a predetermined time zone is used. ) And the degree of cessation of activity in the past prior to yesterday for the time zone corresponding to this, the resident is resident regardless of the various life patterns of the resident 3 in the resident 2 3 can be detected.

  Moreover, in the monitoring system 1 which concerns on 1st Embodiment, the presence or absence of generation | occurrence | production of the abnormality in the resident 3 is not determined (that is, only the indicator for determining the generation | occurrence | production of the abnormality in the resident 3 is provided to a watcher). Therefore, the watching system 1 does not erroneously determine whether or not an abnormality has occurred, and there is an advantage that the degree of freedom of the watching person regarding the determination of the occurrence of the abnormality in the resident 3 is high.

  FIG. 9 is an explanatory diagram illustrating a modified example of the calculation target of the activity inactivity shown in FIG. 5, and FIG. 10 is an explanatory diagram illustrating a modified example of the method for calculating the divergence illustrated in FIG. Note that matters not specifically mentioned below are the same as those described above.

In the above-described example, the activity inactivity N _{t_d} is calculated based on the power usage data for one unit time. However, the present invention is not limited to this, and a plurality of predetermined power usage data in which the activity inactivity N _{t_d} is continuous in time series. It is also possible to calculate based on (ie, treat as a multidimensional vector). In the example shown in FIG. 9, 22 o'clock power usage _{E t} used to calculate the current hiatus of _{N t_d} is, power usage regarding three unit time _{e t, e t-1,} e t-2 ( Here, it is composed of 20:30 to 21:00, 21:00 to 21:30, and 21:30 to 22:00 power usage acquired respectively at 21:00, 21:30, and 22:00), and , The power usage amount E _t-1 at 21:30 just before that is the power usage amounts e _t-1 , e _t-2 , e _t-3 (here, _three unit times corresponding to the power usage amount E _t) Then, from 20:30 to 20:30, 20:30 to 21:00, and 21:30 to 21:30, the amount of power used acquired at 20:30, 21:00, and 21:30 respectively) Composed. In the example shown in FIG. 9, it deals with the power usage E _t as a three-dimensional vector.

In this case, as shown in FIG. 10, for example, the divergence degree calculated from the degree of inactivity is multidimensional (corresponding to the number of power usage data (number of unit times) used for calculating the degree of inactivity _{Nt_d} ). In the feature space, which is assumed to be two-dimensional here for simplicity of explanation, the current degree of inactivity N _{t — d} and three past inactivity levels (here, k = 3 by the k-nearest neighbor method). And the corresponding past activity cessation degree is shown in a circle in FIG. 10).

In such a modification, the activity inactivity N _{t_d} is calculated based on a plurality of _pieces of power usage data regarding a plurality of unit times that are continuous in time series, and therefore, an index of the amount of change in power usage is treated as a vector (feature It is possible to detect the occurrence of abnormality in the resident 3 with higher accuracy. In addition, in order to calculate the degree of divergence based on the distance between the current activity pause level and the past activity pause level in the feature space, calculate an index for determining the occurrence of abnormality in the resident by a simple method. Is possible.

  Now, by treating the change index of the power consumption as a vector, the divergence degree in the past P time (for example, 24 hours) can be calculated in a Q time period (for example, 0.5 hour). Furthermore, by applying this, for example, the divergence degree of the past 24 hours, the past 6 hours, the past 3 hours, and the past 1 hour can be calculated in a cycle of 30 minutes. This makes it possible to obtain watching necessity information from a complex viewpoint.

  In the above-described modification, steps ST101 to ST105 shown in FIG. 4 are repeatedly executed every time a predetermined number of new power usage data used for calculating one activity inactivity level is accumulated. Become.

(Second Embodiment)
FIG. 11 is a functional block diagram of the watching device 5 according to the second embodiment of the present invention, FIG. 12 is a flowchart showing a flow of watching processing by the watching device 5 shown in FIG. 11, and FIG. It is explanatory drawing which shows the example of notification of abnormality generation in middle step ST104. Regarding the second embodiment, matters that are not particularly mentioned below are the same as those in the first embodiment described above, and a detailed description thereof is omitted. In FIG. 11, the same reference numerals are shown for the same configurations as those in the first embodiment. In FIG. 12, steps ST201 to ST203 are the same as steps ST101 to ST103 shown in FIG.

  As illustrated in FIG. 11, the monitoring device 5 according to the second embodiment further includes an abnormality determination unit 41 that determines whether or not an abnormality has occurred in the resident 3 based on the deviation degree calculated by the deviation degree calculation unit 33. In the point provided, it differs from the monitoring apparatus 5 which concerns on 1st Embodiment. With the abnormality determination unit 41, the watching system 1 according to the second embodiment allows the watcher to more clearly grasp the occurrence of abnormality in the resident 3.

  More specifically, in the flow chart shown in FIG. 12, the monitoring device 5 compares the calculated current (latest) divergence with a predetermined threshold Th (see FIG. 13), so that the abnormality in the resident 3 is detected. The presence or absence of occurrence is determined (ST204). The threshold Th is set in advance by the administrator of the watching system 1, but is not limited thereto, and can be set as appropriate by the user of the external device 6 for each resident 3.

  Therefore, the monitoring device 5 determines that an abnormality has occurred in the resident 3 when the current deviation degree is equal to or greater than the threshold Th (ST205: Yes), and includes the current deviation degree calculated in step ST203. A message indicating that an abnormality has occurred in the resident 3 is added to the information on the degree of divergence within the predetermined period and the power consumption corresponding to the information, and the external device 6 is notified as abnormality occurrence information (ST206). The monitoring device 5 displays the abnormality occurrence information to the administrator of the monitoring center 11 in parallel with the notification in step ST206 (ST207). On the other hand, in step ST205, when the current divergence degree is less than the threshold value Th, it is determined that no abnormality has occurred in the resident 3 (No), and the watching process ends.

  In steps ST206 and ST207, for example, as shown in FIG. 13, the abnormality occurrence information is displayed on the monitor 26 of the monitoring device 5 in the external device 6 and the monitoring center 11.

  In the example shown in FIG. 13, there is a time zone in which the change in power consumption is relatively small at midnight on July 2 (see the region 51 indicated by the broken line). At this time, the resident 3 is in normal sleep. Therefore, the degree of divergence is less than the threshold value Th, and the watching device 5 determines that no abnormality has occurred in the resident 3. Further, on July 4, there is a time zone in which the amount of power consumption is small despite the daytime (see the area 52 indicated by the broken line), but at this time, the resident 3 is in normal activity ( Since the amount of change in the power consumption is relatively large), the degree of divergence is less than the threshold value Th, and the monitoring device 5 determines that no abnormality has occurred in the occupant 3.

  On the other hand, in FIG. 13, there is a time zone in which the change in power consumption is relatively small in the evening of July 5 (see the region 53 indicated by the alternate long and short dash line). Since the activity is stopped (the amount of change in the power consumption is relatively small and different from normal), the divergence degree is equal to or greater than the threshold value Th, and the monitoring device 5 indicates that the resident 3 has an abnormality. judge. In addition, there is a time period during which power consumption is relatively large during the day of July 7 (see the area 54 indicated by the alternate long and short dash line). (The amount of change in power consumption is relatively small and different from normal), the divergence degree is equal to or greater than the threshold value Th, and the watching device 5 determines that an abnormality has occurred in the resident 3.

  Thus, in the monitoring system 1, it is not necessary to set a specific model or condition relating to the power consumption in order to determine whether or not an abnormality has occurred in the resident 3 based on the degree of divergence as described above ( For example, it is not necessary to clearly distinguish between sleeping, at-home activity, going out, and the like, and it is possible to appropriately detect the occurrence of an abnormality in the resident 3 regardless of the apparent power usage. Also, in the monitoring system 1, it is not necessary to set a specific model or condition for the amount of change in power usage, and even if the amount of change in power usage is relatively small, the occurrence of an abnormality in the resident 3 can be accurately performed. Can be detected.

  In the monitoring process, in ST205, it is determined whether or not the current degree of inactivity is less than or equal to a predetermined threshold (or the amount of change in power usage is greater than or equal to the predetermined threshold). May be determined that no abnormality has occurred in the occupant 3 regardless of the degree of divergence, when the power consumption is less than or equal to a predetermined threshold (or the amount of change in power consumption is greater than or equal to a predetermined threshold). . This improves the accuracy of abnormality determination. Alternatively, in the above monitoring process, only when the current degree of inactivity calculated in ST202 is greater than or equal to a predetermined threshold (or the amount of change in power usage acquired in ST201 is greater than or equal to a predetermined threshold), the subsequent steps It is good also as a structure which performs. As a result, it is possible to avoid performing an unnecessary abnormality determination.

  Moreover, in the monitoring apparatus 5, as will be described in detail in the third embodiment later, the number of times that the resident 3 has erroneously determined that an abnormality has occurred (the number of times that the monitoring person or the like has pointed out as erroneous determination) is 1. When the above predetermined number of times is exceeded, all the power usage data already stored in the storage unit 30 (including the degree of inactivity and divergence associated with it) is automatically updated (including deletion). ). This improves the accuracy of abnormality determination.

  In addition, when it is erroneously determined that an abnormality has occurred, the power usage data used for the determination and the degree of inactivity may be actively accumulated in the storage unit 30 (that is, additional learning is performed). Do). In this case, dummy data may be added or weighted so that the number of data when accumulating the degree of inactivity is handled as the number (k) referred to by the k-nearest neighbor method described above. By performing such additional learning, even if a situation similar to the situation (life pattern) pointed out by the watcher as a misjudgment occurs, it will no longer be determined as abnormal (that is, in additional learning The process of accumulating the degree of inactivity in the storage unit 30 corresponds to attaching a label “no abnormality”). Of course, it is not essential to make it possible to handle the degree of inactivity as k pieces of independent data during additional learning. If at least the degree of activity suspension at the time of a misjudgment is accumulated in the storage unit 30 one by one, if the number of life patterns similar to the situation of misjudgment increases, the misjudgment will gradually be made through additional learning. Decrease.

  Further, in the monitoring device 5, when an abnormality actually occurs in the resident 3 in a situation where it is determined that no abnormality has occurred, the power consumption data corresponding to the occurrence of the abnormality or the calculation based on the power consumption data is calculated. The power usage amount data corresponding to the activity quiescence level in which the Euclidean distance is close (hereinafter, may be referred to as “neighboring power usage data”) is stored in the storage unit 30 with respect to the activity quiescence level. It is also possible to adopt a configuration in which the power consumption data is excluded (excluded from regular data). Thereby, in the monitoring apparatus 5, it becomes possible to avoid using the improper electric power consumption data at the time of abnormality occurrence for calculation of a deviation degree, and to detect the abnormality generation | occurrence | production in the resident 3 more accurately. . In the situation where the monitoring device 5 determines that an abnormality has occurred, if an abnormality actually occurs in the resident 3, the power usage data is not accumulated in the storage unit 30. In this way, in the situation where an abnormality actually occurs, the watching device 5 does not store the power usage data corresponding to the occurrence of the abnormality in the storage unit 30, regardless of whether the determination is correct or incorrect, and in the vicinity. The power usage amount data may be deleted from the storage unit 30. Hereinafter, this specific example will be described in detail in the third embodiment.

(Third embodiment)
FIG. 14 is a functional block diagram of the watching device 5 according to the third embodiment of the present invention, FIG. 15 is a functional block diagram showing details of the watching necessity level information calculation unit 40 in FIG. 14, and FIG. FIG. 17A to FIG. 17C are diagrams illustrating first to third examples of evaluation screens of the monitoring system 1 displayed on the external device 6, respectively. Regarding the third embodiment, matters not particularly mentioned below are the same as those in the first or second embodiment described above, and detailed description thereof is omitted. 14-17, the same code | symbol is shown about the structure similar to 1st or 2nd embodiment.

  As shown in FIGS. 14 and 15, in the watching device 5 according to the third embodiment, the activity quiescence degree calculation unit 32 and the divergence degree calculation unit 33 calculate a plurality of divergence degrees based on a plurality of different gains a. The monitoring according to the first and second embodiments in that a data operation unit 45 that functions as the monitoring necessity information calculation unit 40 and that can operate the data in the storage unit 30 based on the evaluation information from the external device 6 is provided. Different from the device 5. Note that various functions (at least a part thereof) by these added configurations can be realized by the processor 21 (see FIG. 2) executing a predetermined control program, as in the case described above.

  As in the case described above, the activity quiescence level calculation unit 32 calculates the activity quiescence level in time series for the current target data in the power usage data acquired by the power data acquisition unit 31 in time series. Here, as shown in FIG. 15, the activity pause level calculation unit 32 sets different gains (here, five gains a1 to a5), and calculates the activity pause level calculation based on each gain. A plurality of activity suspension degree sub-calculating units 42a to 42e. Thereby, the calculation of the activity inactivity level by the activity inactivity level calculation unit 32 is repeatedly performed on all the gains a1 to a5 based on the above formula (1) at substantially the same time interval as the acquisition of the power consumption data. .

  The value of each gain in the activity pause sub-calculating units 42a to 42e can be changed as appropriate. For example, gain a1 = 1, gain a2 = 5, gain a3 = 10, gain a4 = 50, gain a5 = 100 be able to. It is also possible to increase or decrease the number of activity suspension degree sub-calculating units 42a to 42e (that is, the number of gains set by the activity suspension degree calculating unit 32) as appropriate.

  Here, FIG. 16 shows a gain different from that in FIG. 13 (gain a = 1 in FIG. 13, gain a = 10 in FIG. 16) with respect to the transition of the power consumption shown in FIG. 13 (or FIG. 8). An example in which the degree of divergence (activity inactivity) is calculated by application is shown. However, the example of FIG. 16 is shown in contrast to the case of FIG. 13 for convenience of explanation, and the actual activity status of the resident 3 is not necessarily the same as the explanation regarding FIG. Further, the divergence shown in FIG. 16 is normalized as will be described later.

  With such a change in gain a, in FIG. 16, the divergence degree becomes equal to or greater than the threshold Th in the time zone (regions 55 and 56) in which the divergence degree was calculated to be relatively small in FIG. In the time zone (regions 52 and 54) in which the degree is calculated to be relatively large, the divergence degree is less than the threshold Th. In other words, since the abnormality detection pattern based on the degree of deviation changes due to the gain change, the abnormality detection is executed more appropriately by determining an appropriate gain according to the power usage status of each household (such as home appliances used). It becomes possible. In particular, there is an advantage that it is possible to cope with abnormality detection based on a change in the power usage status by a home appliance that consumes constant power, such as a refrigerator, which is relatively difficult to detect a change in the power usage status by changing the gain. .

  Referring to FIG. 15 again, the divergence degree calculation unit 33 is provided corresponding to the activity quiescence degree sub-calculation parts 42a-42e in the activity quiescence degree calculation part 32, and the activity quiescence degree sub-calculation parts 42a-42e There are divergence degree sub-calculating units 43a to 43e that respectively calculate the divergence degree based on each activity suspension degree calculated in series.

  Further, in the divergence degree calculation unit 33, the normalization unit 57 normalizes the divergence degrees calculated by the divergence degree sub-calculation units 43a to 43e to facilitate relative comparison of the divergence degrees (here, , Multiply each divergence by an appropriate scalar).

  Further, in the divergence degree calculation unit 33, the divergence degree adjustment unit 58 determines the maximum divergence degree among the plurality of current divergence degrees corresponding to each gain after being normalized by the normalization unit 57 as the normal divergence degree. Is extracted (calculated) as Accordingly, for example, in FIGS. 13 and 16 (examples with two gains) described above, in the time zone corresponding to the region 55, the relatively high divergence in FIG. On the other hand, in the time zone corresponding to the region 56, the relatively high divergence degree in FIG. 13 is extracted as the normal divergence degree, not the relatively low divergence degree in FIG.

  The divergence degree extracted by the divergence degree adjustment unit 58 and the degree of inactivity corresponding thereto are stored in the storage unit 30 together with the corresponding power usage data and gain value information. Further, the monitoring necessity degree information including at least information on the divergence degree extracted by the divergence degree adjustment unit 58 is transmitted to the display unit 34 and the abnormality notification unit 35, respectively.

  As described above, in the monitoring device 5 according to the third embodiment, the divergence degree based on a plurality of gains is used. Therefore, when the resident 3 has an abnormality, the divergence degree is relatively low by one gain a. Can be suppressed (that is, it is determined that there is no abnormality).

  Moreover, in the monitoring apparatus 5 which concerns on 3rd Embodiment, the abnormality notification part 35 is an evaluation screen (user) of the monitoring system 1 (namely, transmitted monitoring necessity information) with respect to the external apparatus 6 which transmitted monitoring necessity information. Information for displaying the interface screen) can be transmitted.

  For example, as shown in FIGS. 17A to 17C, the evaluation screen shown on the monitor of the external device 6 includes each time of a predetermined date and time (here, every 10 minutes on July 1, 12:00 to 12:30) ) And a result display field 61 in which a bar graph indicating the degree of need for monitoring (deviation degree) is displayed, and an evaluation input in which an evaluation result by the user (the watching degree information is “useful” or “useless”) is input. A field 62 is included. In the result display field 61, the bar graph having a longer length indicates that the degree of necessity for watching (the possibility that an abnormality has occurred in the resident 3) is high.

  The user of the external device 6 has been helpful in determining whether or not an abnormality has actually occurred in the resident 3 with respect to the degree of monitoring necessary acquired in the past from the monitoring device 5 (the actual occurrence status of the resident 3) It is possible to input the result display field 61 by determining whether or not they match.

  In the example of FIG. 17A, the watching necessity level is relatively high at 12:00 and 12:30 in the watching necessity information from the watching device 5, and at this time, an abnormality actually occurs in the resident 3 at this time. In this case, the user who has confirmed that the monitoring necessity information is appropriate has entered “useful” (touching the corresponding rectangular check box to make it black).

  In the example of FIG. 17B, the watching necessity level is relatively high at 10:00 and 12:30 in the watching necessity information from the watching device 5, and at this time, the resident 3 actually has an abnormality. This shows a case where a user who has confirmed that no occurrence has occurred (the monitoring necessity information is not appropriate) has entered “not useful” (that is, erroneous determination by the monitoring device 5).

  Further, in the example of FIG. 17C, the watching necessity level is relatively low at 11:30 and 12:00 in the watching necessity information from the watching device 5, and at this time, the resident 3 actually has an abnormality. This shows a case where a user who has confirmed that a problem has occurred (the monitoring degree information is not appropriate) has entered “not useful”.

  The data operation unit 45 of the monitoring device 5 appropriately receives the above-described input result by the user as evaluation information and uses it for selection of power usage data (that is, determination of regular power usage data).

  For example, when the degree of watching (divergence) is relatively high (indicating that the resident 3 is likely to have an abnormality), the data operation unit 45 is evaluated as “helpful” by the user. In this case (that is, when the degree of watching is high and an abnormality actually occurs, which corresponds to FIG. 17A), the corresponding power usage data is deleted from the storage unit 30. Alternatively, the data evaluated as “useful” may be excluded from the power usage data accumulated in the storage unit 30 by attaching a label “abnormal” to the data (excluded from regular data). Good. In addition, it is good also as a structure which does not accumulate | store the data evaluated as "helpful" in the memory | storage part 30 (Namely, only the electric power usage data not evaluated as "helpful" are accumulate | stored in the memory | storage part 30).

  Further, for example, when the degree of watching (divergence) is relatively high, the data operation unit 45 is evaluated as “not useful” by the user (that is, in spite of the high degree of need for watching, actually In the case where no abnormality has occurred (corresponding to FIG. 17B), in order to keep the value low in the next calculation of the degree of divergence (in the case of a similar change in power consumption), One or more pieces of data that are close (within a predetermined Euclidean distance in FIGS. 7 and 10 described above) can be added to the storage unit 30.

  In addition, for example, when the degree of watching (divergence) is relatively low (indicating that there is a low possibility that an abnormality has occurred in the resident 3), the data operation unit 45 evaluates from the user as “helpful”. (Ie, when the need for watching is low and no abnormality has actually occurred (normal state, generally the user does not input any evaluation)), the data is stored in the storage unit 30. No special processing is performed on the power consumption data. However, the data operation unit 45 may be configured to add one or more pieces of data that are the same or relatively close to the data to the storage unit 30 in order to perform greater weighting on the data evaluated as “helpful”.

  Further, for example, when the degree of watching (divergence) is relatively low, the data operation unit 45 is evaluated as “not useful” by the user (that is, although the degree of watching is low, it is actually When an abnormality occurs (corresponding to FIG. 17C), in order to make the value higher in the next calculation of the degree of divergence (in the case of a similar change in power consumption), it is relatively close to the data (described above). 7 and FIG. 10, at least a part of the data (that is, within the predetermined Euclidean distance) (that is, nearby power usage data) can be deleted from the storage unit 30.

  In this case, the activity suspension degree calculating unit 32 is based on the evaluation information acquired by the data operation unit 45 (information evaluated as “useful” from the user), for example, based on machine learning, etc. It is possible to select an appropriate gain of 1 or more (for example, stop the activity inactivity sub-calculating units 42a to 42e for which an inappropriate gain is set).

  Note that the configuration of the watching necessary degree information calculation unit 40 in the third embodiment can be similarly applied to the above-described first embodiment. Further, the data operation unit 45 in the third embodiment may be applied to the second embodiment described above.

  As mentioned above, although this invention was demonstrated based on specific embodiment, these embodiment is an illustration to the last, Comprising: This invention is not limited by these embodiment. It should be noted that the components of the monitoring device according to the present invention, the monitoring system including the monitoring device, and the monitoring method according to the present invention are not necessarily essential, and are appropriately selected as long as they do not depart from the scope of the present invention. Is possible.

1 Watching system
2 Residential 3 Resident 4 Smart meter 5 Watching device 6 External device 10 Network 11 Monitoring center 20 Bus 21 Processor 24 Network interface 25 Input device 26 Monitor 27 Storage (electricity meter)
30 Storage Unit 31 Power Data Acquisition Unit 32 Activity Inactivity Degree Calculation Unit (Power Change Calculation Unit)
33 Deviation degree calculation part 34 Display part 35 Abnormality notification part 40 Oversight necessary degree information calculation part 41 Abnormality determination part 42a-42e Activity pause degree sub calculation part 43a-43e Deviation degree sub calculation part 45 Data operation part 57 Normalization part 58 Deviation Degree adjustment unit

Claims (11)

A monitoring device for detecting the occurrence of a resident abnormality in a residence,
A power data acquisition unit that acquires, in time series, power usage data related to the power used in the residence;
Regarding the current target data in the power usage data acquired in the time series, the power change amount for calculating the power usage change index based on the power usage data immediately before the target data in time series A calculation unit;
Deviation between a current index for a predetermined time zone and one or more past indices corresponding to the time zone before yesterday in the index of change in power usage calculated in time series for example Bei and the degree of deviation calculation unit that calculates a degree,
The monitoring device according to claim 1, wherein the index of the amount of change in power usage is an index that increases as the change in power usage decreases. A monitoring device for detecting the occurrence of a resident abnormality in a residence,
A power data acquisition unit that acquires, in time series, power usage data related to the power used in the residence;
Regarding the current target data in the power usage data acquired in the time series, the power change amount for calculating the power usage change index based on the power usage data immediately before the target data in time series A calculation unit;
Deviation between a current index for a predetermined time zone and one or more past indices corresponding to the time zone before yesterday in the index of change in power usage calculated in time series A divergence degree calculation part for calculating the degree and
With
When an abnormality actually occurs in the resident, the power usage data corresponding to the time zone at the time of the abnormality is excluded from the power usage data acquired in time series. . A monitoring device for detecting the occurrence of a resident abnormality in a residence,
A power data acquisition unit that acquires, in time series, power usage data related to the power used in the residence;
Regarding the current target data in the power usage data acquired in the time series, the power change amount for calculating the power usage change index based on the power usage data immediately before the target data in time series A calculation unit;
Deviation between a current index for a predetermined time zone and one or more past indices corresponding to the time zone before yesterday in the index of change in power usage calculated in time series A divergence degree calculation unit for calculating the degree;
An abnormality notification unit for transmitting the degree of deviation to a predetermined external device;
A storage unit for storing the power usage data acquired by the power data acquisition unit;
User evaluation information related to the degree of deviation is acquired from the external device, and the amount of change in the power usage amount by the power change amount calculation unit among the power usage amount data stored in the storage unit based on the evaluation information. A data operation unit for determining regular power consumption data used for calculating the index;
A monitoring device comprising: The monitoring apparatus according to any one of claims 1 to 3, wherein the current target data includes a plurality of pieces of power usage data regarding a plurality of unit times that are continuous in time series. The discrepancy is according to claim 4, characterized in that it is calculated based on the distance between the historical metrics and the current index in the feature space, wherein the amount an indication of the variation of the power consumption Watch device. The monitoring apparatus according to any one of claims 1 to 5 , wherein the power usage data acquired in time series includes only power usage data within a predetermined period set in advance.   The monitoring apparatus according to claim 1, further comprising an abnormality determination unit that determines whether or not an abnormality has occurred in the resident based on the degree of deviation. A monitoring device according to any one of claims 1 to 7 ,
A monitoring device comprising: a power meter connected to the monitoring device via a network so as to be communicable and transmitting power usage data related to power used in the residence to the power data acquisition unit. system. A monitoring method for detecting the occurrence of a resident anomaly in a residence,
A power data acquisition step of acquiring power consumption data related to power used in the residence in time series,
Regarding the current target data in the power usage data acquired in the time series, the power change amount for calculating the power usage change index based on the power usage data immediately before the target data in time series A calculation step;
Deviation between a current index for a predetermined time zone and one or more past indices corresponding to the time zone before yesterday in the index of change in power usage calculated in time series Divergence degree calculating step for calculating degree
Including
The monitoring method according to claim 1, wherein the index of the amount of change in power usage is an index that increases as the change in power usage decreases. A monitoring method for detecting the occurrence of a resident anomaly in a residence,
A power data acquisition step of acquiring power consumption data related to power used in the residence in time series,
Regarding the current target data in the power usage data acquired in the time series, the power change amount for calculating the power usage change index based on the power usage data immediately before the target data in time series A calculation step;
Deviation between a current index for a predetermined time zone and one or more past indices corresponding to the time zone before yesterday in the index of change in power usage calculated in time series A divergence degree calculating step for calculating a degree;
When an abnormality actually occurs in the resident, the power usage data corresponding to the time zone at the time of the abnormality is excluded from the power usage data acquired in the time series; and
A method of watching characterized by comprising. A monitoring method for detecting the occurrence of a resident anomaly in a residence,
A power data acquisition step of acquiring power consumption data related to power used in the residence in time series,
Regarding the current target data in the power usage data acquired in the time series, the power change amount for calculating the power usage change index based on the power usage data immediately before the target data in time series A calculation step;
Deviation between a current index for a predetermined time zone and one or more past indices corresponding to the time zone before yesterday in the index of change in power usage calculated in time series A divergence degree calculating step for calculating a degree;
Transmitting the degree of divergence to a predetermined external device;
Storing the power usage data acquired by the power data acquisition step;
The user's evaluation information related to the divergence degree is acquired from the external device, and the power usage amount change index among the power usage data accumulated in the step of storing the power usage data based on the evaluation information. Determining regular power usage data used to calculate
A monitoring method characterized by comprising:

Images