JP5473956B2 - Monitoring device - Google Patents

Description

The present invention relates to monitoring equipment, and more particularly relates to the monitoring equipment for monitoring the residents living in the house.

  As the number of nuclear family households increases and lifestyles diversify, the number of individuals making up one household tends to decrease. In addition, with the globalization of the economy, the ratio of single employees to the working population is expected to increase. For this reason, research and development of systems for confirming the health status of family members and close relatives are being actively conducted from a distance (see, for example, Patent Documents 1 and 2).

  The device described in Patent Document 1 compares the power usage pattern of a resident who lives without abnormality in a house with the power usage pattern specified from the current value flowing through the power supply system in the house, thereby improving the health of the resident. Guess the state.

  Moreover, the apparatus of patent document 2 estimates the resident's life pattern by measuring the electric energy consumed in a house. And it is judged whether a resident has abnormality by comparing the estimated life pattern with an expected parameter.

JP 2002-109663 A JP 2008-112267 A

  In the case of using the device described in Patent Document 1, it is necessary to attach a measuring device for measuring current to all the power systems branched in the house. For this reason, there exists a problem that the manufacturing cost and running cost of an apparatus increase.

  The device described in Patent Document 2 estimates a resident's life pattern based on a change in the amount of power consumed in the house. For this reason, it is possible to determine the presence or absence of an abnormal situation from a distance by simply installing a watt-hour meter in the vicinity of the core breaker. For this reason, the structure of the apparatus can be simplified, and manufacturing costs and running costs can be suppressed.

  However, the method executed by the above apparatus has a problem that it is difficult to correctly determine whether there is an abnormality when a device such as an electric water heater or an air conditioner automatically operates.

  The present invention has been made under the circumstances described above, and an object thereof is to accurately determine the presence or absence of an abnormality in a house with a simple device.

In order to achieve the above object, the monitoring device of the present invention provides:
Power measuring means for measuring the power supplied to the house;
Storage means for storing the power measurement results in time series;
A moving average line calculating means for calculating a moving average line of a transition curve indicating the transition of the power stored in the storage means;
A difference calculating means for calculating a difference between a point on the moving average line and a point on the transition curve;
A difference curve calculating means for calculating a difference curve indicating a temporal transition of the difference;
Data string generating means for generating a data string composed of binary data from the difference curve;
First calculation means for calculating a probability that a state of a resident in the house when the power is measured when the data string is input is a predetermined state;
Based on the calculation result of the first calculation means, second calculation means for calculating a probability that the transition of the power is caused by the resident's action;
Is provided.

  According to this invention, the difference curve which shows the change of the electric power resulting from a resident's action is calculated. And based on this difference curve, the probability that transition of electric power originates in a resident's action is calculated. Therefore, it is possible to accurately determine whether there is an abnormality in the house based on this probability. Moreover, according to this invention, the presence or absence of abnormality in a house can be judged using the binarized data. For this reason, the amount of data handled when making a determination is reduced, and the apparatus can be simplified.

It is a block diagram of the monitoring device concerning this embodiment. It is a figure which shows the image displayed on a display part. It is a figure which shows electric power data. It is a flowchart which shows the process which CPU performs. It is a figure for demonstrating an example of the procedure at the time of calculating a moving average line. It is a figure which shows the vector showing a difference. It is a figure for demonstrating the calculation procedure of a difference curve. It is a figure which shows a difference curve. It is a figure for demonstrating the production | generation procedure of a data string. It is a figure which shows a data sequence. It is a figure which shows a data sequence. It is a figure which shows a Bayesian network. It is a flowchart which shows the process which CPU performs. It is a figure which shows a data table. It is a figure which shows electric power data. It is a figure which shows a difference curve. It is a block diagram which shows the modification of a monitoring apparatus.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a schematic configuration of a monitoring apparatus 10 according to the present embodiment. The monitoring device 10 is a device that monitors the state of the resident 70 from the transition of the power consumption of the electrical equipment 51 _i installed in the house 50.

  The residence 50 is a residence where the resident 70 lives. In the house 50, electric devices 51 such as a water heater, a microwave oven, a refrigerator, and an air conditioner are installed. Electric power is supplied to these electric devices 51 via a lead-in line 60 connected to a commercial power system.

  As shown in FIG. 1, the monitoring device 10 includes a CPU (Central Processing Unit) 21, a main storage unit 22, an auxiliary storage unit 23, an input unit 24, a display unit 25, an interface unit 26, and a power measurement unit 28. doing.

  The CPU 21 reads and executes the program stored in the auxiliary storage unit 23. Specific operations of the CPU 21 will be described later.

  The main storage unit 22 has a volatile memory such as a RAM (Random Access Memory). The main storage unit 22 is used as a work area for the CPU 21.

  The auxiliary storage unit 23 includes a nonvolatile memory such as a ROM (Read Only Memory), a magnetic disk, and a semiconductor memory. The auxiliary storage unit 23 stores programs executed by the CPU 21, various parameters, and the like. Further, information including processing results by the CPU 21 is sequentially stored.

  The input unit 24 has a touch panel and input keys. In the monitoring apparatus 10, a GUI (Graphical User Interface) is configured by the touch panel of the input unit 24 and the display unit 25. An instruction from a user (for example, a resident 70) is input via the input unit 24 and is notified to the CPU 21 via the system bus 27.

  The display unit 25 includes a display unit such as an LCD (Liquid Crystal Display). FIG. 2 is a diagram showing an image IM displayed on the display unit 25. When the monitoring device 10 is activated, the display unit 25 displays the image IM shown in FIG. The image IM includes, for example, selection buttons 31 and 32 for switching between the learning mode and the determination mode, and input buttons 33, 34, and 35 for inputting the current state of the resident 70.

  The user can switch the mode of the monitoring device 10 by touching a portion of the touch panel of the display unit 25 that overlaps the selection buttons 31 and 32. In addition, the current state of the occupant 70 can be input to the monitoring device 10 by touching the portion overlapping the input buttons 33, 34, and 35.

  Hereinafter, for convenience of description, a touch on a portion overlapping the selection buttons 31 and 32 is expressed as pressing of the selection buttons 31 and 32, and a touch on a portion overlapping the input buttons 33 to 35 is pressed of the input buttons 33 to 35. It expresses.

  Returning to FIG. 1, the interface unit 26 includes a serial interface, a LAN (Local Area Network) interface, and the like. The interface unit 26 connects the monitoring device 10 to the Internet. Then, the power measurement unit 28 is connected to the system bus 27.

  The power measurement unit 28 is installed on a distribution board or the like of the house 50. The power measuring unit 28 measures power (power consumption) supplied via the lead-in line 60 at a predetermined cycle. Then, the measured result is output as power data (t, P (t)) in association with the measured time. FIG. 3 is a diagram showing power data (t, P (t)). As can be seen from FIG. 3, t constituting the power data indicates a measurement time, and P (t) indicates an instantaneous value of the power supplied to the house 50.

  The power data (t, P (t)) is received by the interface unit 26 and sequentially stored in the auxiliary storage unit 23.

  Next, processing executed by the monitoring apparatus 10 configured as described above will be described. The process of the monitoring device 10 is a monitoring process for determining the state of the resident 70 from the learning process S1 (learning mode) for associating the state of the resident 70 with the transition of power consumption and the transition of the power consumption at the house 50. It is roughly divided into process S2 (monitoring mode). Each process S1, S2 is realized by the CPU 21 reading and executing the program stored in the auxiliary storage unit 23.

  The flowchart of FIG. 4 corresponds to a series of processing algorithms of a program executed by the CPU 21. Hereinafter, the learning process S1 executed by the CPU 21 will be described with reference to FIG. The series of processing shown in the flowchart of FIG. 4 is executed on condition that the selection button 31 displayed on the display unit 25 is pressed.

  In the first step S101, the CPU 21 initializes a built-in counter value n. As a result, the counter value n is reset to zero.

  In the next step S102, the display color of the selection button 31 is changed. Thereby, it is possible to confirm that the monitoring device 10 is in the learning mode.

  In the next step S103, it is confirmed whether or not any of the input buttons 33 to 35 has been pressed. If any of the input buttons 33 to 35 is pressed after the learning mode is selected (step S103: Yes), the CPU 21 proceeds to step S104.

  In the next step S104, the CPU 21 extracts power data (t, P (t)) sampled until a predetermined time (for example, 2 hours) elapses. Thereby, for example, the power data shown in FIG. 3 is extracted. When the predetermined time has elapsed and the sampling is completed, the CPU 21 proceeds to the next step S105.

  In the next step S105, the CPU 21 increments the counter value n.

In the next step S106, the CPU 21 calculates a moving average line P _n (t) for the power P (t). FIG. 5 is a diagram for explaining an example of a procedure for calculating the moving average line. As can be seen from FIG. 5, the CPU 21 calculates, for example, an average value P _n (t _n ) of two values P (t _n ) and P (t _n−1 ). Then, a moving average line P _n (t) defined by data (t _n , P _n (t _n ) consisting of the time t _n and the average value P _n (t _n ) is calculated.

Specifically, as shown in FIG. 5, the CPU 21 determines the average value P ₁ (t ₁ ) of P (t ₁ ) and P (t ₀ ), and the average value of P (t ₂ ) and P (t ₁ ). P ₁ (t ₂ ), the average value P ₁ (t ₃ ) of P (t ₃ ), P (t ₂ ),..., P (t _n ) and the average value P ₁ (P _n (t _n−1 )) t _n ) is calculated sequentially. Then, data (t ₁ , P ₁ (t ₁ )), data (t ₂ , P ₁ (t ₂ )), data (t ₃ , P ₁ (t ₃ ),..., Data (t _n , P ₁ ( The moving average line P ₁ (t) defined by t _n )) is calculated.

In the next step S107, as shown in FIG. 6, the CPU 21 sets data (t ₁ , P ₁ (t ₁ )), (t ₂ , P ₁ (t ₂ )), (t ₃ , P ₁ (t ₃ ),..., (T _n , P ₁ (t _n )) as a starting point, and corresponding power data (t ₁ , P (t ₁ )), (t ₂ , P ( t _2)), calculates the _{(t 3, P (t 3} ), ..., (t n, P (t n) vector _a 1 to a _n to ending points based on).

The magnitude of the vector _A 1 to A _n, the data and _{(t n, P 1 (t} n)), power data indicates the difference between the _{(t n, P (t n} )), the vector _A 1 to A _n Indicates the magnitude of the power data (t _n , P (t _n )) when compared with the data (t _n , P ₁ (t _n )). For example, the magnitude of the vector _{A n} is 100, when the orientation is downward, P _{(t n) indicates} that P 1 _{(t n)} less than, the difference is 100 [W].

In the next step S108, the CPU 21 calculates a difference curve f _n (t). For example CPU21, as seen with reference to FIG. 7, on the basis of the vector _A 1 to A _n, the data _{(t n, P 1 (t} n)) and the power data of the _{(t n, P (t n} )) The points F _{1 to} F _n indicating the difference are plotted. Then, a difference curve f ₁ (t) defined by the points F _{1 to} F _n is calculated. Each coordinate point F which defines the difference curve _f 1 a (t) is the magnitude of the vector _{A n} | when _{to, (t n, + | |} A n |) A n or _(t n, - | A _n |). The difference curve f ₁ (t) shown in FIG. 8 is calculated by performing the above-described processing on the power P (t) and the moving average line P ₁ (t).

In the next step S109, the CPU 21 binarizes the difference curve f _n (t) to generate a data string D (n) composed of a set of 0 and 1. For example, as shown in FIG. 9, the CPU 21 divides the difference into six levels LV1 to LV6, and sequentially determines whether or not a part of the difference curve f _n (t) exists in the range corresponding to each level. . Then, a data string D (n) is generated based on the determination results for the levels LV1 to LV6. This data string D (n) is composed of six elements d ₁ (n), d ₂ (n),... D ₆ (n). Each element d ₁ (n), d ₂ (n),..., D ₆ (n) indicates whether or not a difference curve f _n (t) exists at each level, and the difference curve f _n (t ) Becomes 1 when a part of the difference curve f _n (t) does not exist.

For example, in the example shown in FIG. 9, the range corresponding to the level LV1 and level LV6, but difference curve _f 1 (t) is not present, the level LV2～LV5, there is difference curve _f 1 (t) To do. Therefore, the values of the elements [d ₁ , d ₂ , d ₃ , d ₄ , d ₅ , d ₆ ] constituting the data string D (1) are [ ₀ , ₁ , ₁ , ₁ , ₁ , 0] and Become.

In the next step S110, the CPU 21 determines whether or not the counter value n is greater than or equal to the reference value N. When the counter value n is not equal to or greater than the reference value N (step S110: No), the CPU 21 returns to step S105, and with respect to the moving average line P _n (t) until the determination in step S110 is affirmed. Steps S106 to S109 are repeatedly executed. Accordingly, the difference curve f _{n + 1} (t) with respect to the moving average line P _n (t) is sequentially calculated, and a data string Dt (k) including N data strings D (n) is generated. As can be seen from FIG. 10, the data string Dt (k) is data in which N data strings D (n) are collected in series.

  On the other hand, when the counter value n is greater than or equal to the reference value N (step S110: Yes), the CPU 21 proceeds to step S111.

In the next step S111, the CPU 21 associates the state of the resident 70 with the data string Dt (k). The association of the state of the resident 70 is performed with respect to the elements constituting the data string Dt (k). Although the number of elements to be associated may be one or plural, it is preferable to perform the process on elements whose values change as the state of the resident 70 changes. For example, when it is known in advance that the value of the element d ₂ (1) in FIG. 10 is 1 when the state of the occupant 70 is cooking and 0 at other times, etc. The state of the resident 70 is associated with the element d ₂ (1).

For example, the table TB ₂ shown in FIG. 10 shows the state of the resident 70 when the power data that is associated with the element d ₂ (1) and is the basis of the data string Dt (1) is sampled. In table TB _2, item indicates the status of the resident 70, the value indicates whether a context of the corresponding item of status and power data. From this TB _2, the power underlying data of the data string Dt (t) it can be seen that the state of the resident 70 is one that was sampled at the time was during cooking.

The CPU 21 recognizes the situation of the resident 70 when the input buttons 33 to 35 are pressed, and generates the table TB. Then, the table TB is associated with a predetermined element d _k (n) of the data string Dt (1).

In the next step S112, the CPU 21 confirms whether or not the monitoring mode is selected. When the monitoring mode is not selected (step S112: No), the process returns to step S101, and thereafter, the processes of steps S101 to S111 are repeatedly executed until the determination in step S112 is affirmed. Thus, as shown in FIG. 11, a plurality of Dt (k) in which the table TB is associated with the predetermined element d _k (n) is generated.

  On the other hand, when the monitoring mode is selected (step S112: Yes), the CPU 21 proceeds to step S113.

In step S113, the CPU 21 constructs a Bayesian network. For example, FIG. 11 shows an example of a table associated with Dt (k) and the element d ₂ (n) of Dt (k). If the value of the element d ₂ (n) is 1 from the four elements d ₂ (n) associated with the table TB and the value of each item of the table TB, the probability of the resident is 3/4. It can be seen that the state is cooking and the user is going out with a probability of 1/4.

Therefore, based on the value of the predetermined element d _k (n) and the table TB associated with the element, the CPU 21 sets the input as the data string Dt (k) and outputs the probability as shown in FIG. A Bayesian network is constructed with Y1%, Y2%, and Y3%. The probability Y1% is a probability that the resident 70 is out when the power data that is the basis of the data string Dt (k) is sampled. The probability Y2% is the probability of cooking, and the probability Y3% is the probability of sleeping. When the construction of the Bayesian network is completed, the CPU 21 ends the learning process S1.

  Next, the monitoring process S2 executed by the CPU 21 will be described. The processing shown in the flowchart of FIG. 13 corresponds to a series of processing algorithms executed by the CPU 21. Hereinafter, the learning process S1 executed by the CPU 21 will be described with reference to FIG. A series of processing shown in the flowchart of FIG. 13 is executed on condition that the selection button 32 displayed on the display unit 25 is pressed.

  In the first step S201, the CPU 21 initializes a built-in counter value n. As a result, the counter value n is reset to zero.

  In the next step S202, the display color of the selection button 32 is changed. Thereby, it is possible to confirm that the monitoring device 10 is in the monitoring mode.

  In the next step S203, the CPU 21 extracts power data (t, P (t)) sampled until a predetermined time (for example, 2 hours) elapses. Thereby, for example, the power data shown in FIG. 3 is extracted. When the predetermined time has elapsed and the sampling is completed, the CPU 21 proceeds to the next step S204.

  In the next step S204, the CPU 21 increments the counter value n.

In the next step S205, the CPU 21 calculates a moving average line P _n (t) for the power P (t). The moving average line P _n (t) is calculated in the same procedure as in step S106 described above.

In the next step S206, as shown in FIG. 6, the CPU 21 starts from a point based on the data (t _n , P _n (t _n )) defining the moving average line, and the corresponding power data (t _n , calculates a vector _a 1 to a _n P a _(t n)) in basis point to end point.

In the next step S207, the CPU 21 calculates a difference curve f _n (t). The calculation of the difference curve f _n (t) is performed in the same procedure as in step S108 described above.

In the next step S208, the CPU 21 binarizes the difference curve f _n (t) to generate a data string D (n) composed of a set of 0 and 1. The generation of the data string D (n) is performed in the same procedure as in step S109 described above.

In the next step S209, the CPU 21 determines whether or not the counter value n is greater than or equal to the reference value N. When the counter value n is not equal to or greater than the reference value N (step S209: No), the CPU 21 returns to step S204, and with respect to the moving average line P _n (t) until the determination in step S209 is affirmed. Steps S205 to S208 are repeatedly executed. Accordingly, the difference curve f _{n + 1} (t) with respect to the moving average line P _n (t) is sequentially calculated, and a data string Dt (k) including N data strings D (n) is generated.

  On the other hand, when the counter value n is greater than or equal to the reference value N (step S209: Yes), the CPU 21 proceeds to step S210.

  In step S210, the CPU 21 is cooking with a probability Y1% that the state of the resident 70 is out when the power data (t, P (t)) that is the basis of the data string Dt (k) is sampled. The probability Y2% and the sleep probability Y3% are calculated.

  As can be seen from FIG. 12, the probabilities Y1%, Y2%, and Y3% are obtained by inputting the data string Dt (k) to the Bayesian network generated by executing the learning process S1.

  In the next step S211, the CPU 21 uses the table shown in FIG. 14 with the probability Y1%, Y2%, Y3%, and the change in the power data (t, P (t)) causes the electric device 51 to be resident. The probability Yr% considered to be caused by the operation of 70 and the probability Ym% considered to be caused by the automatic operation of the electric device 51 are calculated. Specifically, the probabilities Yr% and Ym% are calculated using the following equations (1) and (2), respectively.

Yr% = 3%, Y1% + 70%, Y2% + 5%, Y3% (1)
Ym% = 97% ・ Y1% + 30% ・ Y2% + 95% ・ Y3% ... (2)

  For example, if the probability Y1% that the resident 70 is out is 80%, the probability Y2% that the person is cooking is 10%, and the probability Y3% that the person is sleeping is 10%, the probability Yr % Is 9.9%, and the probability Ym% is 90.1%.

  In the next step S212, the CPU 21 outputs information on the probabilities Yr% and Ym% to the Internet 80 via the interface unit 26. Thereby, for example, using a communication terminal such as a personal computer, information on the probabilities Yr% and Ym% can be acquired from a distance.

  In the next step S213, the CPU 21 checks whether the learning mode is selected. When the learning mode is not selected (step S213: No), the CPU 21 returns to step S201, and thereafter repeatedly executes the processes of steps S201 to S212 until the determination in step S212 is affirmed. On the other hand, when the learning mode is selected (step S213: Yes), the CPU 21 ends the monitoring process S2.

As described above, in the present embodiment, the difference curve f _n (t) indicating the change in power caused by the resident's action is calculated. Next, a data string Dt (n) is generated based on the difference curve f _n (t). Then, based on the probabilities Y1%, Y2%, and Y3% output from the Bayesian network that receives the data string Dt (n), the change in the power data (t, P (t)) The probability Yr% that is considered to be caused by the operation of the person 70 and the probability Ym% that is considered to be caused by the automatic operation of the electric device 51 are calculated.

In the difference curve fn (t), the change in the power consumption due to the resident's action appears significantly compared to the change in the power consumption due to the automatic operation of the electrical equipment. For example, FIG. 15 shows a curve P (t) indicating the transition of the power sampled when the resident 70 is sleeping. In this curve, the peak Pe1 is attributed to the act of the resident 70, and the peaks Pe2, Pe3, Pe4 are attributed to the automatic operation of the electrical equipment. The levels of the peaks Pe1 to Pe4 are substantially equal in FIG. However, the peak Pe1 indicated by the difference curve f _n (t) shown in FIG. 16 has a significantly higher level than the peaks Pe2 to Pe4.

Therefore, by calculating the probability Yr% and the probability Ym% using the difference curve f _n (t) as in the present embodiment, the probabilities Yr% and Ym% can be calculated with high accuracy. As a result, the presence or absence of abnormality of the resident 70 can be accurately determined.

  In this embodiment, the probability Y1%, Y2%, Y3% is calculated by inputting a data string Dt (k) composed of binary data to the Bayesian network. Then, based on the probabilities Y1%, Y2%, and Y3%, the probability Yr% that is considered to be caused by the occupant 70 operating the electric device 51 and the probability Ym that is considered to be caused by the automatic operation of the electric device 51. % Is calculated. For this reason, the amount of data handled in calculating the probabilities Yr% and Ym% is reduced, the apparatus can be downsized, and the probabilities Yr% and Ym% can be calculated quickly.

  As mentioned above, although each embodiment of this invention was described, this invention is not limited by each said embodiment.

  For example, in the above-described embodiment, the monitoring device 10 includes the CPU 21, the main storage unit 22, and the auxiliary storage unit 23, and the software such as a program is stored in the auxiliary storage unit 23 like the computer. For example, as illustrated in FIG. 17, the monitoring device 10 includes the power measurement unit 28, the input unit 24, the display unit 25, the moving average line calculation unit 10a, the difference calculation unit 10b, and the difference curve calculation unit 10c. , A data string generation unit 10d, a first calculation unit 10e, a second calculation unit 10f, and a storage unit 10g.

In this case, the moving average line calculation unit 10a calculates a moving average line P _n (t) for the power P (t).

As shown in FIG. 6, the difference calculation unit 10 b starts from a point based on data (t _n , P _n (t _n )) that defines a moving average line, and corresponding power data (t _n , P (t a point-based _n)) to calculate the vector a _n to the end point.

Difference curve calculation unit 10c on the basis of the vector _{A n,} to calculate the difference curve _f n (t).

The data string generation unit 10d generates a data string Dt (n) including a set of 0 and 1 by binarizing each difference curve f _n (t).

  The first calculation unit 10e inputs the data string Dt (k) to the Bayesian network, and the resident when the power data (t, P (t)) that is the basis of the data string Dt (k) is sampled The probability Y1% that the state of 70 is going out, the probability Y2% that it is cooking, and the probability Y3% that it is sleeping are calculated.

  The second calculation unit 10f uses the probability Y1%, Y2%, Y3% and the table shown in FIG. 14 to change the electric power data (t, P (t)), so that the resident 70 A probability Yr% considered to be caused by the operation and a probability Ym% considered to be caused by the automatic operation of the electric device 51 are calculated. Then, the calculation result is output to the Internet 80.

  The storage unit 10g stores the power measurement unit 28, data input from the input unit 24, image information to be displayed on the display unit 25, and processing results of each unit.

  In the present embodiment, the case where the state of the resident 70 is classified into three states of going out, cooking, and sleeping is described as an example. Moreover, the state of the resident 70 may be divided into four or more.

  In this embodiment, the calculation result is output to the Internet 80. However, the present invention is not limited to this, and the probability Yr% that is considered to be caused by the occupant 70 operating the electric device 51 is a certain period (for example, 2, (3 days) When continuing, you may send the mail etc. which notify the abnormality of the resident 70 outside via the internet. In addition, an alarm may be issued using sound or light outside the house 50.

  Various embodiments and modifications can be made to the present invention without departing from the broad spirit and scope of the present invention. Further, the above-described embodiment is for explaining the present invention, and does not limit the scope of the present invention. That is, the scope of the present invention is shown not by the embodiments but by the claims. Various modifications within the scope of the claims and within the scope of the equivalent invention are considered to be within the scope of the present invention.

Monitoring equipment of the present invention is suitable for monitoring the resident.

DESCRIPTION OF SYMBOLS 10 Monitoring apparatus 10a Moving average line calculation part 10b Difference calculation part 10c Difference curve calculation part 10d Data sequence generation part 10e 1st calculation part 10f 2nd calculation part 10g Memory | storage part 20 Monitoring apparatus 21 CPU
DESCRIPTION OF SYMBOLS 22 Main memory part 23 Auxiliary memory part 23 Sequential auxiliary memory part 24 Input part 25 Display part 26 Interface part 27 System bus 28 Electric power measurement unit 31, 32 Selection button 33-35 Input button 50 Housing 51 Electric equipment 60 Line 70 Resident 80 Internet A Vector D, Dt Data string F point F _n point IM image LV1 to LV6 Level Pe1 to Pe4 Peak TB table

Claims (5)

Power measuring means for measuring the power supplied to the house;
Storage means for storing the power measurement results in time series;
A moving average line calculating means for calculating a moving average line of a transition curve indicating the transition of the power stored in the storage means;
A difference calculating means for calculating a difference between a point on the moving average line and a point on the transition curve;
A difference curve calculating means for calculating a difference curve indicating a temporal transition of the difference;
Data string generating means for generating a data string composed of binary data from the difference curve;
First calculation means for calculating a probability that a state of a resident in the house when the power is measured when the data string is input is a predetermined state;
Based on the calculation result of the first calculation means, second calculation means for calculating a probability that the transition of the power is caused by the resident's action;
A monitoring device comprising:   The monitoring apparatus according to claim 1, wherein the first calculation unit is a Bayesian network. An interface for inputting the status of the resident;
Bayesian network generation means for generating the Bayesian network by associating the resident state input from the interface with the elements constituting the data sequence each time the data sequence is generated;
The monitoring apparatus according to claim 2. The elements constituting the data string are:
The monitoring apparatus according to any one of claims 1 to 3, wherein a value is determined based on a relationship between a level defined for the difference and the difference curve.   The monitoring apparatus according to any one of claims 1 to 4, further comprising alarm issuing means for issuing an alarm when the probability calculated by the second calculating means is equal to or less than a threshold value.

Images