JP2022145761A - Presence/absence estimation system, power control system, and presence/absence estimation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a presence/absence estimation system capable of estimating presence/absence of a user even if there is no information indicating the presence/absence of the user whose presence/absence is estimated.

SOLUTION: A server device 1 includes: a storage unit 121 for storing a learning model 121 in which first data indicating the amount of power usage in a plurality of buildings B2 to B4 and second data indicating presence/absence of a person in the plurality of buildings B2 to B4 are generated by machine learning as teacher data; and a presence/absence estimation unit 132 for estimating the presence/absence of a user by inputting third data indicating the amount of power usage in the building B1 of the user to the learning model 121.

SELECTED DRAWING: Figure 1

COPYRIGHT: (C)2023,JPO&INPIT

Description

The present disclosure relates to a presence/absence estimation system, a power control system, and a presence/absence estimation method.

BACKGROUND Conventionally, presence/absence prediction systems are known that predict the presence/absence of a person (user) based on power usage data at a predetermined location. In the presence/absence prediction system of Patent Document 1, learning data is generated based on power data of a building past the first time and information indicating whether or not a person was actually in the building (information indicating presence/absence). (learning model) is generated. Based on the power data of the building at the first time and the learning data, it is predicted whether or not the person is in the building.

JP 2017-62765 A

As described above, in the presence/absence prediction system as disclosed in Patent Literature 1, it is necessary to acquire information indicating the presence/absence of a user whose presence/absence is to be estimated in order to generate a learning model. Therefore, if the information indicating the user's presence/absence cannot be acquired, the user's presence/absence cannot be predicted.

Accordingly, it is an object of the present invention to provide a presence/absence estimation system capable of estimating the presence/absence of a user even if there is no information indicating the presence/absence of the user whose presence/absence is to be estimated.

In order to solve the above-described problems, a presence/absence estimation system according to an embodiment of the present invention is a presence/absence estimation system for estimating the presence/absence of a user, comprising: first data indicating power usage in a plurality of buildings; a storage unit for storing a learning model generated by machine learning using second data indicating the presence/absence of a person in the plurality of buildings as teacher data; and an estimating unit for estimating the user's presence or absence by inputting third data indicating .

According to an embodiment of the present invention, a learning model is generated from first data indicating power consumption in a plurality of buildings and second data indicating presence/absence of persons in a plurality of buildings. Then, by inputting the third data indicating the power consumption of the user's building into the learning model, the presence/absence of the user is estimated. As a result, when generating a learning model, data indicating the power consumption in a plurality of buildings including buildings other than the user's building and the presence or absence of people in the plurality of buildings are used as teacher data. There is no need to acquire data indicating the presence/absence of the user for whom the is to be estimated as training data. Therefore, even if there is no information indicating the presence/absence of the user whose presence/absence is to be estimated, the user's presence/absence can be estimated.

The user's building is not included in the plurality of buildings, the third data is data indicating power consumption of the user's building in a first period including the current time, and the estimation unit Presence/absence of the user in a second period after one period may be estimated.

According to the embodiment of the present invention, data indicating the amount of electricity used in a plurality of buildings that do not include the user's building and the presence/absence of persons in the plurality of buildings is used as teacher data. The presence/absence of the user can be estimated without acquiring data indicating the presence/absence of the user as teacher data. Further, the estimating unit inputs, to the learning model, third data indicating the power consumption of the user's building during the first period including the current time, thereby Presume absence. That is, since the presence/absence of the user is estimated not only in the period immediately after the current time but also in the period after the current time (second period), the future presence/absence of the user can be estimated.

In addition, the first data is data indicating power consumption of the plurality of buildings during a third period, and the second data is data indicating the amount of power consumption of the plurality of buildings during a fourth period different from the third period. data indicating the presence or absence of

According to the embodiment of the present invention, the acquisition periods of the first and second data used as teacher data for the learning model are the third and fourth periods, respectively. no need to let This facilitates the generation of learning models.

Further, the first data and the second data are classified based on the calendar in which each data was acquired, and the learning model uses the classified first data and the second data as the teacher data. It may be used as data.

According to the embodiment of the present invention, the first data and the second data classified based on the calendar are used as teacher data for the learning model. In general, depending on the calendar, such as holidays, weekdays, days of the week, and seasons, there can be common patterns of human presence and power usage change between buildings. For example, if the building is a residence, people are generally absent during the daytime on weekdays, so the amount of electricity used in the building during the daytime is reduced. In addition, during a certain period of time from evening to night when a person returns to the building, the amount of power consumption in the building increases because the person uses devices connected to the building. On the other hand, on holidays, the hours when people are absent generally differ from those on weekdays, so the power consumption of buildings changes in a pattern different from the pattern of change in the power consumption of buildings on weekdays as described above. Therefore, by classifying the first data and the second data based on the calendar, it is possible to more accurately estimate the presence/absence of the user based on the pattern of change in the amount of electricity used in the building according to the calendar. . Note that the classification based on the calendar may be performed on the first data in the entirety of a plurality of buildings including the building of the user. and devices connected to the buildings are classified into one or more building groups according to one or more common features, and then the first data in each building group may be subjected to the first data. Examples of such classification of building groups include, in addition to the classification according to the type of building such as residences, shops, condominiums, offices, factories, warehouses, etc., the number of household members and the age group of the householder in the case of residences. , and subdivided classification according to the presence or absence of a vehicle having a charging device. According to this aspect, only the data indicating the power consumption of buildings classified into the same building group as the building of the user whose presence/absence is to be estimated is used as the first data. can be estimated more accurately based on the power usage and calendar of a group of buildings that have features in common with .

Further, a power control system according to another embodiment of the present invention is a power control system that manages the amount of power supplied to at least one of a vehicle and equipment connected to a user's building, comprising: a server device; A power control device communicably connected to a server device, wherein the server device stores first data indicating power consumption in a plurality of buildings and second data indicating presence/absence of persons in the plurality of buildings. a storage unit that stores a learning model generated by machine learning as teacher data; and the power control device controls the amount of power supplied to at least one of the vehicle and the device according to the estimation result of the estimation unit.

According to this embodiment of the present invention, the power control device controls the amount of power supplied to at least one of the vehicle and the device according to the estimation result of the estimation unit. At least one of charging the vehicle and power supply to the device is controlled according to the estimation result of the estimation unit. Therefore, for example, power generated by a solar panel or the like can be efficiently used in the building.

Further, a presence/absence estimation method according to another embodiment of the present invention is a presence/absence estimation method for estimating the presence/absence of a user, comprising: first data indicating power usage amounts in a plurality of buildings; By inputting third data indicating the power consumption of the user's building to a learning model generated by machine learning using second data indicating the presence or absence of a person in the building as teacher data, the presence or absence of the user a step of estimating

According to an embodiment of the present invention, a learning model is generated from first data indicating power consumption in a plurality of buildings and second data indicating presence/absence of persons in a plurality of buildings. Then, by inputting the power consumption of the user's building into the learning model, the presence or absence of the user is estimated. This eliminates the need to acquire the presence/absence of the user whose presence/absence is to be estimated when generating the learning model. Therefore, even if there is no information indicating the presence/absence of the user whose presence/absence is to be estimated, the user's presence/absence can be estimated.

According to the embodiment of the present invention, in a presence/absence estimation system for estimating the presence/absence of a user, it is possible to estimate the presence/absence of the user without information indicating the presence/absence of the user.

1 is a block diagram showing the configuration of a power control system according to this embodiment; FIG. FIG. 4 is a conceptual diagram of first data used as teacher data for a learning model according to the embodiment; FIG. 4 is a conceptual diagram of second data used as teacher data for a learning model according to the embodiment; 4 is a conceptual diagram of third data input to a learning model according to the embodiment; FIG. 4 is a conceptual diagram of presence probability data estimated by the presence/absence estimation system according to the present embodiment. FIG. 4 is a flowchart showing the operation of the presence/absence estimation system according to the embodiment; 4 is a flowchart showing the operation of the power control device according to the embodiment; FIG. 4 is a diagram showing an experiment record of the presence/absence estimation system according to the present embodiment;

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described in detail based on the drawings. The following description of preferred embodiments is merely exemplary in nature and is not intended to limit the invention, its applications or uses.

FIG. 1 is a block diagram showing the configuration of the power control system according to this embodiment. As shown in FIG. 1 , the power control system according to this embodiment includes a server device 1 and a power control device 21 . The server device 1 is configured to communicate with the power control device 21 via the communication network N. As shown in FIG.

The presence/absence estimation system according to this embodiment is configured as part of the power control system. Specifically, the presence/absence estimation system 100 includes a server device 1 .

The power control device 21 is for controlling the power of the user's building B1. Specifically, the power control device 21 is connected to a solar panel 22, a plurality of devices 23, and a vehicle 24 via power lines. Also, the power control device 21 is connected to the server device 1 via the communication network N. As shown in FIG. The solar panel 22 is a solar panel used for photovoltaic power generation. The equipment 23 is, for example, electrical equipment connected to the building B1, and is one or a plurality of equipment for which the amount of power supplied by the power control device 21 is controlled. For example, the devices 23 may be directly or indirectly connected to the power controller 21 by at least one wired and/or wireless communication means, respectively. The amount of power supplied to the device 23 may be controlled by switching the power supply input, increasing/decreasing the amount of power consumption, etc., by the power control device 21 via the communication means. Examples of the equipment 23 include those that are used especially when a person is in the building, such as air conditioners and lighting fixtures, those that are particularly used when a person is not in the building, security devices such as surveillance cameras, and Cleaning robots and the like, as well as items that are used regardless of the presence or absence of a person, such as refrigerators and the like.

The vehicle 24 is a vehicle equipped with a charging device, such as an electric vehicle, a hybrid vehicle, and a fuel cell vehicle.

The power control device 21 controls power supplied to each device connected to the building B1. Specifically, the DC power generated by the solar panel 22 is converted into AC power and supplied to the equipment 23 and the vehicle 24 . Further, when the amount of power generated by the solar panel 22 is greater than the amount of power used by the equipment 23 and the vehicle 24, the power control device 21 outputs surplus power to a commercial power system (not shown) and sells the power. conduct. Furthermore, when the amount of power generated by the solar panel 22 is less than the amount of power used by the equipment 23 and the vehicle 24, the power control device 21 is supplied with insufficient power from the power system.

Here, the power control device 21 may charge the vehicle 24 based on the existence probability data. Although details will be described later, the presence probability data is data indicating the probability that the user is in the building B1 for each time period (in the following description, the probability that a person is in the building is referred to as the presence probability). For example, the power control device 21 increases the amount of power supplied to the device 23 and suppresses the amount of power supplied to the vehicle 24 during a time period when the presence probability is equal to or greater than a predetermined value. On the other hand, the power control device 21 increases the amount of power supplied to the vehicle 24 and reduces the amount of power supplied to the device 23 during the time period when the presence probability is equal to or less than the predetermined value. As an example of controlling the amount of power supplied to the device 23, for example, the power control device 21 increases the amount of power supplied to the device 23, such as an air conditioner and a lighting fixture, during a time period when the presence probability is equal to or greater than a predetermined value. On the other hand, the amount of electric power supplied to the device 23 may be reduced during the time period when the presence probability is equal to or less than a predetermined value. The function of the power control device 21 allows the power obtained by the solar panel 22 to be efficiently used in the building B1.

The power control device 21 also includes a smart meter 211 . The smart meter 211 is a power meter that measures the amount of power used in the building B1. In this example, the smart meter 211 measures at least the amount of power generated by the solar panel 22 and the amount of power used by the equipment 23 and the vehicle 24 as the power usage of the building B1. The smart meter 211 measures the amount of power usage in the building B<b>1 at predetermined time intervals (generally, every 30 minutes) and transmits the power usage measurement results to the server device 1 .

(Regarding the configuration of the server device 1)
As shown in FIG. 1 , the server device 1 includes a communication section 11 , a storage section 12 and a control section 13 .

The communication unit 11 is configured by, for example, an electric circuit, and communicates with the power control system 2 via the communication network N.

The storage unit 12 is a storage medium including ROM (Read Only Memory), RAM (Random Access Memory), HDD (Hard Disk Drive), SSD (Solid State Drive), and the like. Various programs executed by the control unit 13 are stored in the storage unit 12 .

A learning model 121 is stored in the storage unit 12 . The learning model 121 is generated by machine learning using first data indicating power consumption in a plurality of buildings and second data indicating presence/absence of persons in a plurality of buildings (whether or not a person is in the building) as teacher data. It is a learning model that Machine learning models or algorithms solve classification problems, such as logistic regression (including ridge and lasso regularization), random forests, support vector machines, neural networks (including multilayer perceptrons), etc. including. The learning model generated by machine learning in the present embodiment may be one of the above-described machine learning models or algorithms that is evaluated as having high accuracy by evaluation of classification prediction accuracy. Evaluation of the accuracy of the classification prediction was performed by means of the mean squared error of the estimated conditional probabilities, the area under the curve (AUC) quantified as the area under the ROC (Receiver Operating Characteristic) curve, and the minimization of the Kullback-Leibler divergence. It may be done by one or more of logarithmic loss in the problem, and calibration plots plotting actual versus predicted values, and the like. Here, the classification prediction accuracy of a machine learning algorithm or model varies depending on the number of teacher data and feature amounts (for example, the number of days of acquisition of power consumption data used for learning and attribute information to be added). Therefore, the learning model may be selected by comparing the classification accuracy evaluation results of each machine learning algorithm or model under predetermined conditions. Also, in selecting a learning model, priority may be given to a model with a low computational cost required for learning and classification.

FIG. 2 is a conceptual diagram of first data used as teacher data for the learning model according to this embodiment, and FIG. 3 is a conceptual diagram of second data used as teacher data for the learning model according to this embodiment.

As shown in FIGS. 2 and 3, the first data includes data indicating the power consumption of buildings B2 to B4 every 30 minutes, and the second data includes data indicating the power consumption of buildings B2 to B4 every 30 minutes. Contains data indicating the presence or absence of a person. For example, the first data is generated based on data indicating power usage that is automatically transmitted every predetermined time (every 30 minutes, for example) from smart meters installed in buildings B2 to B4. Also, the second data is generated based on the results of conducting a questionnaire on the presence/absence of persons in buildings B2 to B4.

The learning model 121 is generated by machine learning using the first and second data as teacher data, and when receiving an input of the user's power consumption of the building B1, outputs the user's presence probability (details will be described later).

The control unit 13 is composed of, for example, a microcomputer including a CPU (Central Processing Unit) and a semiconductor memory. The control unit 13 controls each unit of the server device 1 by executing programs and the like stored in the storage unit 12 .

The control unit 13 also includes a power consumption acquisition unit 131 , a presence/absence estimation unit 132 , and a presence probability data delivery unit 133 .

The power usage acquisition unit 131 acquires third data indicating the power usage of the user's building B1. Specifically, the power usage amount acquisition unit 131 acquires, as the third data, data indicating the power usage amount of the building B1 transmitted every 30 minutes from the smart meter 211 installed in the building B1.

The presence/absence estimation unit 132 inputs the third data indicating the power usage of the building B1 acquired by the power usage acquisition unit 131 to the learning model 121, and estimates the presence probability of the user.

FIG. 4 is a conceptual diagram of the third data input to the learning model according to this embodiment, and FIG. 5 is a conceptual diagram of presence probability data estimated in the power control system according to this embodiment.

For example, the presence/absence estimation unit 132, as the third data, data indicating the power consumption of the building B1 in the first period (in this example, the period from 7 days ago (1/1) to the current day (1/8)). is input to the learning model 121 . Then, the learning model 121 outputs the presence probability of the user in the second period (in this example, the current day (1/8) and the next day (1/9)). The presence/absence estimation unit 132 uses the user's presence probability output from the learning model 121 as an estimation result of the user's presence/absence.

The presence probability distribution unit 133 transmits (distributes) the estimation result of the presence/absence estimation unit 132 to the power control device 21 as presence probability data. The power control device 21 performs power control of the building B1 (for example, charging of the vehicle 24, etc.) based on the received existence probability data.

For example, the power control device 21 increases the amount of power supplied to the device 23 and reduces the amount of power supplied to the vehicle 24 during the time period when the presence probability is equal to or higher than a predetermined value (here, 50% or higher). Suppress, supply processing. On the other hand, the power control device 21 increases the amount of power supplied to the vehicle 24 during the time period when the presence probability is equal to or less than a predetermined value (here, less than 50%), and increases the amount of power supplied to the device 23. to reduce the power consumption, and perform the charging process.

(About the operation of the power control system)
6 and 7 are flowcharts showing the operation of the power control system according to this embodiment. Specifically, FIG. 6 shows the operation of the presence/absence estimation system 100 (server device 1), and FIG. 7 shows the operation of the power control device 21. FIG.

First, the operation of the server device 1 will be described.

The power consumption acquisition unit 131 of the server device 1 acquires (receives) data indicating the power consumption of the user's building B1 from the smart meter 211 (step S1).

The presence/absence estimation unit 132 inputs the third data indicating the power consumption of the building B1 in the first period to the learning model 121 (step S2).

The learning model 121 outputs (calculates) the presence probability of the user in the second period (step S3). The presence/absence estimation unit 132 uses the user presence probability output from the learning model 121 as an estimation result of the user presence probability.

The presence probability distribution unit 133 transmits (distributes) the estimation result of the presence/absence estimation unit 132 to the power control device 21 as presence probability data.

Next, the operation of power control device 21 will be described.

When the power control device 21 receives presence probability data from the server device 1 (step S11), it detects whether the user is present or absent at the current time from the presence probability data. Then, the power control device 21 determines whether or not the user presence probability at the current time is equal to or greater than a predetermined value (step S12).

When the user presence probability at the current time is equal to or higher than a predetermined value (Yes in step S12), the power control device 21 increases the amount of power supplied to the device 23 and suppresses the amount of power supplied to the vehicle 24. (step S13). On the other hand, if the user presence probability at the current time is less than or equal to the predetermined value (No in step S12), the power control device 21 increases the amount of power supplied to the vehicle 24 and increases the amount of power supplied to the device 23. A charging process to suppress is performed (step S14).

(Regarding presence/absence estimation results)
FIG. 8 is a diagram showing experimental records of the presence/absence estimation system according to this embodiment. In FIG. 8, 880 households were surveyed for power usage and presence/absence every 30 minutes on a given day on weekdays and holidays. Of these, 616 households were used as training data for the learning model 121, and 264 households were used for testing the learning model 121. FIG. In addition, the "number of records" in FIG. 8 is the number of power consumption and presence/absence recorded every 30 minutes.

As shown in FIG. 8, the positive case rate for all records was 78.9% (0.789). This indicates that the percentage of "present" was 78.9% and the percentage of "absent" was 21.1% of the total number of records used in this experiment. Although not shown, in this experiment, logistic regression (ridge regularization), logistic regression (Lasso regularization), random forest, support vector machine, and neural network were used as the learning model 121 . In this experiment, the power consumption data for the current day and the past 7 days, and the power consumption data for the current day and the past 14 days were used as feature amounts to compare the five learning models 121 described above. According to this experiment, the test results of each learning model 121 were as follows.

The root-mean-square error (power consumption data for the current day and the past 7 days) is 0.145 for logistic regression (Ridge regularization), 0.143 for logistic regression (Lasso regularization), 0.130 for random forest, and support vectors 0.138 for the machine and 0.137 for the neural network. According to the results, the most suitable example of the learning model 121 is the random forest with the smallest mean squared error. In addition, the mean square error (power consumption data for the current day and the past 14 days) is 0.140 for logistic regression (Ridge regularization), 0.148 for logistic regression (Lasso regularization), 0.133 for random forest, 0.145 for support vector machines and 0.133 for neural networks. According to the results, the best examples of learning models 121 are random forests and neural networks with the lowest mean squared error.

AUC (power consumption data for current day and past 7 days) is 0.770 for logistic regression (Ridge regularization), 0.772 for logistic regression (Lasso regularization), 0.802 for random forest, and 0.802 for support vector machine 0.786, and neural network 0.790. According to the results, the best example of the learning model 121 is the random forest with the maximum AUC. In addition, AUC (power consumption data for the current day and the past 14 days) is 0.749 for logistic regression (Ridge regularization), 0.758 for logistic regression (Lasso regularization), 0.773 for random forest, and support vector 0.760 for the machine and 0.790 for the neural network. According to the results, the most suitable example of learning model 121 is the neural network with the maximum AUC.

Logarithmic loss (power consumption data for current day and past 7 days) is 0.447 for logistic regression (Ridge regularization), 0.449 for logistic regression (Lasso regularization), 0.417 for random forest, and support vector machine was 0.433, and the neural network was 0.431. According to the results, the best example of the learning model 121 is the random forest with the lowest logarithmic loss. Also, the logarithmic loss (power consumption data for the current day and the past 14 days) is 0.446 for logistic regression (Ridge regularization), 0.459 for logistic regression (Lasso regularization), 0.442 for random forest, and support The vector machine was 0.459 and the neural network was 0.423. According to the results, the best example of learning model 121 is the neural network with the lowest logarithmic loss.

According to the above experimental results, a random forest or a neural network may be selected as a suitable example of the learning model 121 in this embodiment.

As described above, the power control system according to this embodiment includes the presence/absence estimation system 100 that estimates the presence/absence of the user. This power control system is generated by machine learning using first data indicating power consumption in a plurality of buildings B2 to B4 and second data indicating presence/absence of people in a plurality of buildings B2 to B4 as teacher data. A storage unit 12 that stores a learning model 121, and a presence/absence estimation unit 132 that estimates the user's presence/absence by inputting third data indicating the power consumption of the user's building B1 to the learning model 121. Prepare.

In this configuration, the learning model 121 is generated from the first data indicating the amount of power consumption in the plurality of buildings B2-B4 and the second data indicating the presence/absence of persons in the plurality of buildings B2-B4. Then, by inputting the third data indicating the power consumption of the user's building B1 to the learning model 121, the presence/absence of the user is estimated. As a result, when generating the learning model 121, data indicating the amount of power used in a plurality of buildings including the buildings B2 to B4 other than the building B1 of the user and the presence or absence of people in the plurality of buildings B2 to B4 are used as teacher data. , there is no need to acquire data indicating the presence/absence of the user whose presence/absence is to be estimated as teacher data. Therefore, even if there is no information indicating the presence/absence of the user whose presence/absence is to be estimated, the probability of the user's presence/absence at home can be estimated.

In addition, since the learning model 121 is generated by the first data indicating the power consumption in the plurality of buildings B2 to B4 and the second data indicating the presence/absence of persons in the plurality of buildings B2 to B4, There is no need to collect other data, such as data indicating the presence or absence of a person. As a result, the load at the time of learning model generation is also reduced.

Also, the third data is data indicating the power consumption of the user's building B1 in the first period including the current day (current time). The presence/absence estimation unit 132 estimates the user's presence/absence during the second period after the first period.

In this configuration, the presence/absence of a person in the plurality of buildings B2 to B4, which does not include the user's building B1, is used as teacher data. The presence/absence of the user can be estimated without acquiring data indicating the presence/absence of the target user as teacher data. In addition, the presence/absence estimation unit 132 inputs, to the learning model 121, the third data indicating the power consumption of the user in the building B1 in the first period, thereby Estimate presence/absence. That is, since the presence/absence of the user is estimated not only in the period immediately after the current time but also in the period after the current time (second period), the future presence/absence of the user can be estimated.

The server device 1 also includes a presence/absence estimation unit 132 . The server device 1 is communicably connected to a power control system 2 that controls the charging status of a vehicle 24 connected to the user's building B1. Power control device 21 controls the amount of power supplied to vehicle 24 according to the estimation result of presence/absence estimation unit 132 .

In this configuration, power control device 21 controls the amount of power supplied to vehicle 24 according to the estimation result of presence/absence estimation unit 132 . Since the charging status of the vehicle 24 is controlled according to the estimation result of the presence/absence estimation unit 132, for example, electric power generated by a solar panel or the like can be efficiently used in the building.

(Other embodiments)
As described above, the embodiments have been described as examples of the technology disclosed in the present application. However, the technology in the present disclosure is not limited to this, and can also be applied to embodiments in which modifications, replacements, additions, omissions, etc. are made as appropriate.

In the above embodiment, as shown in FIGS. 2 and 3, the first and second data acquisition periods are the same, but they may be different. That is, the first data is data indicating the amount of power consumption in a plurality of buildings during a third period, and the second data is data indicating the presence or absence of people in a plurality of buildings during a fourth period different from the third period. It may be data indicating This makes it easy to generate a learning model because it is not necessary to match the acquisition periods of the first and second data. In this case, for example, the learning model 121 is generated by determining a reference time (same time on the same day of the week, etc.) and associating the first and second data with reference to the reference time in the first and second data. be able to.

Also, in the above embodiment, the first and second data are obtained at the same time intervals (every 30 minutes in FIG. 2), but this is not the only option. For example, the times at which the first and second data are acquired need not be every 30 minutes. Also, the first and second data may be obtained at different times.

In the above embodiment, the third data input to the learning model 211 is data indicating the amount of electricity used by the user in the building B1 from seven days ago to the current day. Data indicating the amount of electricity used in the user's building B1 may be used.

Classification is performed based on the calendar (month, date, day of the week, weekday, holiday, season, etc.) from which the first data and the second data were acquired, and the classified first data and second data are used as a teacher of the learning model 121. You may use it as data. In general, depending on holidays, weekdays, days of the week, etc., common patterns can occur between buildings in the presence or absence of people in buildings and changes in power consumption. For example, if the building is a residence, people are generally absent during the daytime on weekdays, so the amount of electricity used in the building during the daytime is reduced. In addition, during a certain period of time from evening to night when a person returns to the building, the amount of power consumption in the building increases because the person uses devices connected to the building. On the other hand, on holidays, the hours when people are absent generally differ from those on weekdays, so the power consumption of buildings changes in a pattern different from the pattern of change in the power consumption of buildings on weekdays as described above. Therefore, by classifying (distinguishing) the first data and the second data based on the calendar, the user's presence/absence can be estimated more accurately based on the pattern of changes in the power consumption of the building according to the calendar. can do. Note that the classification based on the calendar may be performed on the first data of the entire plurality of buildings including the user's building B1. , and devices connected to the buildings are classified into one or more building groups according to one or more common characteristics, and then the first data in each building group may be subjected to the first data. Examples of such classification of building groups include, in addition to the classification according to the type of building such as residences, shops, condominiums, offices, factories, warehouses, etc., the number of household members and the age group of the householder in the case of residences. , and subdivided classification according to the presence or absence of a vehicle having a charging device. According to this aspect, only the data indicating the power consumption of buildings classified into the same building group as the building of the user whose presence/absence is to be estimated is used as the first data. can be estimated more accurately based on the power usage and calendar of a group of buildings that have features in common with . Also, the learning model 121 may be generated without classifying the first data and the second data based on the calendar. Also, the first data and the second data classified in advance based on the calendar may be used as teacher data for the learning model 121 .

In the above embodiment, the power control device 21 controls the amount of electric power supplied to the vehicle 24 and the equipment 23 using the existence probability data. may be used to control the amount of power supplied to either the vehicle 24 or the device 23 .

In the above-described embodiment, the server device 1 that configures the presence/absence estimation system 100 cooperates with the power control device 21 to control charging of the vehicle 24 by using the user's presence/absence as an example. , the device (or system) that cooperates with the server device 1 is not limited to the power control device 21 . For example, the server device 1 may be linked with a delivery management system. In this case, the delivery management system may use the presence/absence of the user to determine the delivery route. Moreover, you may cooperate with the server apparatus 1 with a user's watching system. In this case, when the user's presence/absence is high, the watching system may issue a warning when the user's presence/absence cannot be confirmed. Also, the server device 1 may be linked with an advertisement delivery system. In this case, the advertisement distribution system may determine advertisements to be distributed according to the presence or absence of the user.

Further, the server device 1 may have a function of sequentially receiving the first and second data and updating the learning model 121 based on the received data.

Also, as training data for the learning model 121, data indicating the power consumption and presence/absence of the user in building B1 may be used.

Note that the power control system according to this embodiment may be linked with a credit card settlement system, an account transfer system, or the like. In this case, for example, the user's communication terminal (mobile phone, PC, etc.) accepts payment information such as the user's credit card number and account number, and registers the payment information in the credit card payment system or account transfer system. may Thereby, the user's payment information can be managed by the presence/absence estimation system.

Further, the server device 1 of the power control system according to the present embodiment may be provided with a function capable of detecting information about the credit card from an image of the credit card. For example, the server device 1 is configured to be able to communicate with a user's communication terminal (mobile phone, PC, etc.) and a credit card settlement system. The server device 1 receives an image of the user's credit card from the user's communication terminal, and reads the user's credit card number from the received image by character reading technology such as OCR (Optical character recognition). The information may be read and communicated to the credit card payment system. This saves the trouble of inputting payment information.

Further, the power control system according to the present embodiment may be linked with a customer management system that manages various information used for billing the user's power bill, such as the user's name, address, telephone number, and power consumption. good. As a result, for example, the user can be billed for electricity according to the amount of electricity used.

Also, the server device 1 may be a plurality of devices, or may be a single device. One server device may be provided for each function of the server device 1 . Further, each function of the server device 1 may be realized by cooperation of a plurality of server devices.

The presence/absence estimation system according to the present embodiment can estimate the presence/absence of the user. can be used.

REFERENCE SIGNS LIST 100 presence/absence estimation system 1 server device 12 storage unit 121 learning model 13 control unit 131 power consumption acquisition unit 132 presence/absence estimation unit 133 presence probability distribution unit 21 power control device 211 smart meter 24 vehicle

Claims (6)

A presence/absence estimation system for estimating the presence/absence of a user,
a storage unit that stores a learning model generated by machine learning using first data indicating power consumption in a plurality of buildings and second data indicating the presence/absence of people in the plurality of buildings as teacher data;
A presence/absence estimation system, comprising: an estimation unit that estimates the presence/absence of the user by inputting third data indicating the power consumption of the user's building to the learning model. In the presence/absence estimation system according to claim 1,
The user's building is not included in the plurality of buildings,
The third data is data indicating the power consumption of the user's building in a first period including the current time,
The presence/absence estimation system, wherein the estimation unit estimates the user's presence/absence during a second period after the first period. In the presence/absence estimation system according to claim 1 or 2,
The first data is data indicating power consumption of the plurality of buildings in a third period,
The presence/absence estimation system, wherein the second data is data indicating the presence/absence of the person in the plurality of buildings during a fourth period different from the third period. In the presence/absence estimation system according to any one of claims 1 to 3,
The first data and the second data are classified based on the calendar in which each data was acquired,
The presence/absence estimation system, wherein the learning model uses the classified first data and second data as the teacher data. A power control system that manages the amount of power supplied to at least one of a vehicle and equipment connected to a user's premises,
a server device;
a power control device communicably connected to the server device,
The server device
a storage unit that stores a learning model generated by machine learning using first data indicating power consumption in a plurality of buildings and second data indicating the presence/absence of people in the plurality of buildings as teacher data;
an estimating unit for estimating the presence or absence of the user by inputting third data indicating the power consumption of the user's building to the learning model;
The power control system, wherein the power control device controls the amount of power supplied to at least one of the vehicle and the device according to the estimation result of the estimation unit. A presence/absence estimation method for estimating the presence/absence of a user,
For a learning model generated by machine learning using first data indicating power consumption in a plurality of buildings and second data indicating the presence/absence of people in the plurality of buildings as training data, the user's building power A presence/absence estimation method, comprising a step of estimating the presence/absence of the user by inputting third data indicating the amount of usage.

Images