Consumer monitoring system, consumer monitoring method, and consumer monitoring program

Abstract

To provide an inhabitant monitoring system that grasps a life of an inhabitant through acquisition of a state of use of electrical apparatuses in a house, thereby enhancing support for the inhabitant.SOLUTION: An inhabitant monitoring system comprises: a power consumption state for every type acquisition unit that classifies power consumption of a plurality of electrical apparatuses installed in a house according to a type of the electrical apparatus and acquires a power consumption state of the electrical apparatuses over a predetermined period as a power consumption state for every type; a behavior mode estimation unit that estimates a behavior mode of an inhabitant on the basis of, the power consumption state for every type in the past acquired in advance over the predetermined period; a comparison creation unit that compares the power consumption state for every type in the past acquires in advance over the predetermined period with the power consumption state for every type currently being acquired, thereby creating comparison information; a behavior change estimation unit that estimates a change in the behavior mode of the inhabitant based on information on the inhabitant's health and the comparison information; and an output unit that performs output for the change in the behavior mode of the inhabitant.SELECTED DRAWING: Figure 1

Description

[[Technical Field]]

[0001] The present invention relates to a resident monitoring system, a resident monitoring method, and a resident monitoring program, and particularly relates to a resident monitoring system, method, and program for monitoring whether a resident's life is as normal based on the power consumption status of electrical appliances in the residence where the resident lives. [[Background Art]]

[0002] In the case of elderly single persons, the time spent at home becomes longer. In this case, if there is no home-visit nursing, home helper visits, etc., changes in life conditions may not be detected, and symptom worsening, death, etc., may be overlooked. However, constant visits impose a heavy financial burden on local governments such as municipalities. In addition, it is difficult to respond to abnormalities of the resident that occur outside of visit hours.

[0003] In this case, although it is possible to install a surveillance camera in the resident's home for constant monitoring, this impairs the resident's privacy and is not preferable. Instead of direct surveillance using cameras, methods have been proposed to indirectly grasp the resident's life through the usage status of electrical appliances in the home (see Patent Document 1, etc.). It is known that a specific waveform is generated during power consumption depending on the type of electrical appliance. Therefore, it is possible to identify the type of electrical appliance from the generated waveform.

[0004] The electrical appliances used for daily meals, laundry, cleaning, etc., in the home are fixed. Therefore, by clarifying the usage status of electrical appliances, it is possible to indirectly grasp the daily activities of the resident. As described above, grasping the daily activities of a resident while protecting the resident's privacy based on monitoring the usage status of electrical appliances in the home has been widely utilized. [[Prior Art Documents]] [[Patent Documents]]

[0005] [[Patent Document 1]] Japanese Patent Publication No. 2016-163445 [Overview of the project] [Problems that the invention aims to solve]

[0006] However, current monitoring equipment and systems for electrical appliance usage within homes merely acquire data on their usage without providing sufficient feedback to residents.

[0007] The present invention has been made in view of the above points, and provides a consumer monitoring system, method, and program that allows for the understanding of the individual lives and behaviors of consumers through the acquisition of usage status of electrical appliances in their homes, and facilitates the provision of support tailored to each individual consumer. [Means for solving the problem]

[0008] In other words, the embodiment is a consumer monitoring system that uses a power consumption detector connected to a distribution board installed in the consumer's home and detects the power consumption status of each type of electrical equipment in the home. The system is characterized by comprising: a type-specific power consumption status acquisition unit that classifies the power consumption status of multiple electrical equipment installed in the home according to the type of electrical equipment and acquires the power consumption status of the electrical equipment included in the classified type as type-specific power consumption status over a predetermined period; a behavioral pattern estimation unit that estimates the consumer's behavioral patterns based on past type-specific power consumption status of electrical equipment acquired in advance over a predetermined period; a comparison generation unit that generates comparison information by comparing past type-specific power consumption status of electrical equipment acquired in advance over a predetermined period with the type-specific power consumption status of electrical equipment currently being acquired; a behavioral change estimation unit that estimates changes in the consumer's behavioral patterns based on the consumer's health information and comparison information; and an output unit that outputs information regarding changes in the consumer's behavioral patterns.

[0009] Furthermore, the acquisition of power consumption status by type in the type-specific consumption status acquisition unit over a predetermined period may be at least one year, including seasonal variations.

[0010] Furthermore, the behavioral pattern estimation unit may divide the past power consumption status by type of electrical equipment into multiple time periods within a day, and estimate the behavioral patterns of consumers based on which type of electrical equipment is being used, according to the power consumption status by type of electrical equipment for each of the multiple time periods.

[0011] Furthermore, the behavioral pattern estimation unit may categorize past power consumption data by type of electrical equipment on a weekly or monthly basis, and estimate the behavioral patterns of consumers based on which types of electrical equipment are being used according to the weekly or monthly power consumption data by type of electrical equipment.

[0012] Furthermore, the comparison generation unit may generate comparison information based on the type of electrical equipment and the increase or decrease in the power consumption status of the electrical equipment, by comparing the power consumption status of each type of electrical equipment for the past with the power consumption status of each type of electrical equipment currently being acquired.

[0013] Furthermore, the output may include observations on improvements in the lifestyle habits of the individuals. It may also be assumed that the individuals are single.

[0014] Furthermore, the behavioral pattern estimation unit may include a machine learning unit, which may perform machine learning based on the type of electrical equipment and the increase or decrease in the power consumption state of the electrical equipment from past data on power consumption states by type of electrical equipment, in order to estimate the behavioral patterns of consumers.

[0015] Furthermore, the power consumption state may also be defined as the amount of power consumed and the waveform of the current when the electrical equipment is in use.

[0016] Furthermore, the behavioral pattern estimation unit may also estimate the behavioral patterns of the person based on the type of power consumption status, expressing the amount of the person's physical activity as METs-hours.

[0017] Furthermore, the output may include observations on improvements in lifestyle habits based on METs-times observed by consumers. [Effects of the Invention]

[0018] The consumer monitoring system of the present invention is a consumer monitoring system that uses a power consumption detector connected to a distribution board installed in the consumer's home and detects the power consumption status of each type of electrical equipment in the home. The system includes a type-specific power consumption status acquisition unit that classifies the power consumption status of multiple electrical equipment installed in the home according to the type of electrical equipment and acquires the power consumption status of the electrical equipment included in the classified type as type-specific power consumption status over a predetermined period, a behavioral pattern estimation unit that estimates the consumer's behavioral patterns based on past type-specific power consumption status of electrical equipment acquired in advance over a predetermined period, a comparison generation unit that generates comparison information by comparing past type-specific power consumption status of electrical equipment acquired in advance over a predetermined period with the type-specific power consumption status of electrical equipment currently being acquired, a behavioral change estimation unit that estimates changes in the consumer's behavioral patterns based on the consumer's health information and comparison information, and an output unit that outputs information regarding changes in the consumer's behavioral patterns. Therefore, by acquiring the usage status of electrical equipment in the home, it is possible to understand the individual lives and behaviors of consumers and to easily connect this to support for each individual consumer. Consumer monitoring methods and consumer monitoring programs can achieve similar effects. [Brief explanation of the drawing]

[0019] [Figure 1] This is a schematic diagram showing the configuration of the consumer monitoring system according to the embodiment. [Figure 2] (A) A schematic diagram showing electrical appliances in a house, and (B) A graph of the 24-hour power consumption of electrical appliances. [Figure 3] This is a schematic diagram showing the waveform of the electric current detected from electrical appliances in a house. [Figure 4] This is a block diagram showing the computer configuration of the consumer monitoring system. [Figure 5] This is a block diagram showing the configuration of the functional parts of a computer. [Figure 6] This is a table showing the classification of electrical appliances in a house. [Figure 7] Fig. (A) is a bar graph of power consumption of an electric device in one day, and Fig. (B) is a bar graph of power consumption obtained by classifying and showing electric devices in one day. [Figure 8] It is a schematic diagram showing generation of comparison information. [Figure 9] It is a schematic diagram showing estimation of behavior patterns. [Figure 10] Fig. (A) is a first example of an output screen, Fig. (B) is a second example of an output screen, and Fig. (C) is a third example of an output screen. [Figure 11] It is a first flow chart explaining the consumer monitoring method of the embodiment. [Figure 12] It is a second flow chart explaining the consumer monitoring method of the embodiment. DESCRIPTION OF EMBODIMENTS

[0020] The consumer monitoring system of the embodiment measures the usage status of electric devices in a house where a consumer (resident) lives, for each time, day of the week, month, and season, by using a power consumption detector connected to a distribution board installed in the consumer's house. Then, the consumer monitoring system indirectly grasps the consumer's daily behavior patterns from the usage status of electric devices (lifestyle sensing), and outputs information to relevant departments when changes or modulations occur to the consumer along with their health condition, so as to improve the consumer's Quality of Life (QOL).

[0021] Fig. 1 is a schematic diagram showing the configuration of a consumer monitoring system 1 according to an embodiment. Electric power is supplied to each house 4 from an electric wire (power line) 3 suspended on a utility pole. A distribution board 5 is installed in the consumer's house 4, and electric power is supplied to various electric devices in the house 4 (see Fig. 2) through the distribution board 5. Therefore, a power consumption detector 6 for detecting the usage status of various electric devices in the house 4 is connected to the distribution board 5. The power consumption detector 6 is a device that acquires both the power consumption amount of each electric device and the waveform generated by the change in current during operation of each electric device, as shown in the following Fig. 2 and Fig. 3. A known inspection device such as that disclosed in Patent Document 1 cited in the background art is used as the power consumption detector 6.

[0022] Each household 4's power consumption detector 6 is connected to the internet line 2 via wired or wireless connection. The resident monitoring system 1's computer 10 is also connected to the internet line 2. The resident monitoring system 1's output targets are home care facilities 7, hospitals 8 of primary care physicians (family doctors, attending physicians), and municipal offices 9 (administrative agencies), all of which are connected to the internet line 2. The content of the output from the resident monitoring system 1's computer 10 is then reported to the home care facilities 7, hospitals 8, and municipal offices 9. As a result, the efficiency of monitoring and supervision of residents is improved, and their QOL (quality of life) is more easily maintained through this monitoring.

[0023] The consumer monitoring system 1 of this embodiment assumes that the consumer is exclusively a single person. When multiple people live in a house 4, health problems of one consumer are usually noticed by the other residents. In that case, the other residents can contact home care facilities 7, hospitals 8, government offices 9, etc., at an early stage to take action. In contrast, when a consumer lives alone, there is a risk that they may not be able to promptly report any changes in their own health, which could lead to a worsening of their condition.

[0024] In such cases, the resident monitoring system 1 is effective. Furthermore, when multiple residents live in a house 4, it is difficult to identify the individual resident based on their usage of electrical appliances within the house 4. Therefore, in order to understand the usage of electrical appliances within the house 4 that is unique to each resident, it is desirable that the resident be a single person. Note that house 4 is not limited to ordinary houses, but also includes apartment buildings, dormitories, businesses, etc. In other words, any place where residents are constantly present and electrical appliances are used by those residents is acceptable. Monitoring residents through the usage of electrical appliances within house 4 enables automatic and continuous lifestyle sensing.

[0025] Figure 2(A) is a schematic diagram showing an example of electrical appliances in each house 4 shown in Figure 1 above. House 4 is equipped with a distribution board 5, and power is supplied from this distribution board to individual electrical appliances through the power lines inside the house. The illustration shows an air conditioner 51 (air conditioner, cooler), refrigerator 52, electric kettle 53, microwave oven 54 (microwave oven), and washing machine 55. Of course, other appliances not shown include televisions, personal computers (PCs), vacuum cleaners, rice cookers, induction cookers, cooking appliances (hot plates), hair dryers, fans, heaters, coffee makers, blenders, dryers, irons, dishwashers, audio equipment, lighting equipment, and even electric vehicle charging equipment.

[0026] In this embodiment, the power consumption detector 6 is connected to the distribution board 5. The power consumption detector 6 detects the power consumption status of individual electrical appliances such as air conditioners 51 and refrigerators 52 supplied via the distribution board 5 in a time series. The power consumption status refers to information that includes both the amount of power consumed when the electrical appliance is operating (e.g., watts per minute) and the waveform generated by the change in current when the electrical appliance is operating. In other words, the power consumption detector 6 detects both the amount of power consumed and the waveform of the current.

[0027] Figure 2(B) is a graph showing the 24-hour power consumption of electrical appliances in a house. The left end represents 0:00 and the right end represents 24:00, showing the fluctuation in power consumption every minute. Although the lines in the graph overlap, they represent the fluctuation in power consumption of each individual electrical appliance.

[0028] Specifically, Figure 2(B) shows the time-series changes in wattage every minute for air conditioners, washing machines, microwave ovens, refrigerators, rice cookers, televisions, vacuum cleaners, induction cooktops, high-temperature appliances (water heaters, hair dryers, etc.), standby power, and other appliances. In the graph, there is a peak rising between 7:00 and 8:00. This is due to the use of the microwave oven and rice cooker when preparing breakfast after waking up. Similarly, the peak rising between 18:00 and 19:00 is due to the use of cooking appliances such as microwave ovens when preparing dinner. The smaller peaks that appear approximately every hour are thought to be due to standby power and the operation of high-temperature appliances to maintain their temperatures.

[0029] Thus, by understanding the general types and usage times of electrical appliances within a home, it becomes possible to gain some understanding of the lifestyles and behaviors of the residents. However, because this is a general trend, it does not fully capture the specific characteristics of each individual resident's lifestyle and behavior. Furthermore, if the variety of electrical appliances becomes too large, it can actually become more difficult to understand the characteristics of each resident's lifestyle and behavior.

[0030] Therefore, the consumer monitoring system 1 of this embodiment focuses on improving the accuracy of understanding the characteristics of each consumer's lifestyle and behavior, as will be explained below. Specifically, in the consumer monitoring system 1 of this embodiment, the power consumption detector 6 is also capable of detecting the current waveform that appears when individual electrical devices are in operation. Figure 3 is a schematic diagram conceptually showing the current waveform that appears when individual electrical devices are in operation. Regarding the electrical devices (home appliances) shown in Figure 2 above, each electrical device such as the air conditioner 51, refrigerator 52, electric kettle 53, microwave oven 54, and washing machine 55 is shown on the left side of the page. The current waveforms (51W, 52W, 53W, 54W, and 55W) of each electrical device are shown on the right side of the page. When an electrical device is in operation, a unique current waveform is generated due to the influence of the internal motor, inverter, etc. Therefore, by detecting both the power consumption and the current waveform through the power consumption detector 6, it becomes easy to identify the type of individual electrical device, and the usage status of each electrical device can be objectively understood. Note that the electrical equipment and their current waveforms shown in Figure 3 are illustrative examples, and naturally, current waveforms specific to electrical equipment other than those shown will also be detected.

[0031] Figure 4 is a block diagram showing the configuration of the computer 10 that controls the consumer monitoring system 1 of the embodiment. The computer 10 is hardware-wise composed of a CPU 11, ROM 12, RAM 13, memory unit 14, and I / O 15 (input / output interface). It also includes main memory, LSI, etc. An internet line 2 (see Figure 1) is connected to the I / O 15. The computer 10 of the consumer monitoring system 1 is composed of various electronic computers (computing resources) such as personal computers (PCs), mainframes, workstations, and cloud computing systems.

[0032] When each function of the computer 10 in Figure 4 is implemented by software, the computer 10 is realized by executing instructions for a program, which is the software that implements each function. The recording medium for storing this program can be a "non-temporary tangible medium," such as a CD, DVD, semiconductor memory, or programmable logic circuit. This program may also be supplied to the computer 10 of the consumer monitoring system 1 via any transmission medium capable of transmitting the program (such as a communication network or broadcast waves).

[0033] The storage unit 14 of the computer 10 is a known storage device such as an HDD or SSD. The storage unit 14 may also be an external server (not shown). The storage unit 14 stores various data, information, a consumer monitoring program, and various data necessary for the execution of the program. In addition, each functional unit that performs calculations such as calculations and arithmetic is an arithmetic element such as a CPU 11. Furthermore, input devices such as a keyboard and mouse (not shown), a display unit (a display device such as a display), and output devices that output data may also be appropriately connected to the I / O 15 of the computer 10.

[0034] The various functional units in the CPU 11 of computer 10 are shown in the schematic block diagram in Figure 5. Each functional unit includes a type-specific consumption status acquisition unit 110, a behavior pattern estimation unit 120, a machine learning unit 121, a comparison generation unit 130, a behavior change estimation unit 140, an output unit 150, and the like. The processing and execution of computer 10 are realized in software by a consumer monitoring program loaded into main memory.

[0035] The power consumption status acquisition unit 110 classifies the power consumption status of multiple electrical appliances (see Figure 2(A)) installed in the house 4 according to the type of electrical appliance, and acquires the power consumption status of the electrical appliances included in the classified type as power consumption status by type over a predetermined period of time.

[0036] As shown in Figures 2(B) and 3 above, the time-series power consumption status (both power consumption amount and current waveform) of individual electrical devices can be obtained through the power consumption detector 6. As a result, it is possible to determine which electrical device is consuming how much power at what point in time, and the electrical device, the time of use, and the amount of power consumed are linked together. Based on this linked information of the electrical device (type, model), the time of use, and the amount of power consumed, it becomes possible to classify multiple electrical devices according to their type.

[0037] The table in Figure 6 shows an example of classification based on the types of electrical appliances. The table classifies electrical appliances into four categories: "Living," "Eating," "Activities," and "Other." The "Living" category includes electrical appliances such as air conditioners and televisions, as well as standby power. These are, in other words, electrical appliances that residents use constantly while living in their homes, from waking up to going to sleep. The "Eating" category includes electrical appliances used for cooking, such as microwave ovens, refrigerators, rice cookers, and induction cooktops. These so-called cooking appliances are used for preparing and cooking meals. The "Activities" category includes washing machines, vacuum cleaners, and high-temperature appliances (irons, hair dryers), etc. These electrical appliances are used for purposes such as washing clothes and cleaning the house, and are used in daily life that involves physical activity. The "Other" category includes types of electrical appliances that do not fall into the above categories.

[0038] Figure 7(A) is a bar graph showing the power consumption of electrical appliances over a day. This bar graph shows the power consumption of electrical appliances in a particular house over a day, simply aggregated hourly, and roughly corresponds to the period from 0:00 to 24:00 (with some omissions). From the changes in the bar graph in Figure 7(A), it can be seen that power consumption from waking up to going to sleep increases and decreases according to the activities of the occupants, and the types of electrical appliances used and their power consumption can be determined. Note that it is also possible to display the data as a bar graph on a weekly or monthly basis.

[0039] Figure 7(B) is a bar graph that recalculates the power consumption of electrical equipment over a day, according to the classification in the table in Figure 6. The bar graph, like in Figure 7(A), roughly corresponds to the period from 0:00 to 24:00 (with some omissions). Since electrical equipment is grouped into four categories, both constant power consumption and power consumption that occurs depending on the time of day can be determined. For example, in the bar graph of Figure 7(B), there are time periods when power consumption exists for electrical equipment classified as "activity" and time periods when it does not.

[0040] Therefore, the type-specific power consumption acquisition unit 110 acquires the daily power consumption by category, as shown in the bar graph in Figure 7(B), over a predetermined period. Furthermore, the type-specific power consumption acquisition unit 110 acquires data for at least one year, including seasonal variations (temperature fluctuations in spring, summer, autumn, and winter). There are differences in climate depending on the region where the house 4 is located. In addition to the different durations of summer and winter, the types of cooling and heating equipment used in summer and winter also differ. For this reason, it is preferable to acquire data for at least one year by category to understand the impact of seasonal variations.

[0041] The behavioral pattern estimation unit 120 estimates the behavioral patterns of the resident based on past power consumption data by type of electrical appliance acquired over a predetermined period. For example, in the case of the resident of house 4 shown in the bar graph of Figure 7(B), power consumption increases sharply from around 6:00 to 7:00, and then increases again from around 10:00 to 11:00, 14:00 to 15:00, and 18:00 to 19:00. The cause of the increase in power consumption is estimated to be the preparation of three meals a day. During the times when power consumption increases, the "Meals" category in the table of Figure 6 increases. In addition, during the time periods when power consumption increases, there is power consumption of electrical appliances classified as "Activities." Furthermore, the power consumption of electrical appliances classified as "Lifestyle" is almost the same from waking up to going to sleep. Of course, it is also easy to identify and classify the type of electrical appliance through the unique current waveforms included in the power consumption state of the electrical appliances.

[0042] Behavioral patterns are items that focus on the actions of the daily activities of the inhabitants of a household. In other words, they are the accumulation of various actions that occur in daily life from waking up to going to bed. Specifically, they include actions such as preparing and cleaning up meals after waking up, laundry, ironing, cleaning, bathing, and excretion, and their frequency. For example, an inhabitant of a household might have breakfast and clean up between 7am and 9am, lunch and clean up and clean between 12pm and 2pm, and dinner and clean up and bathe between 6pm and 8pm. Based on behavioral patterns that are repeated at roughly the same times every day, it is assumed that these will continue to be repeated at similar times and frequencies in the future. Estimation of behavioral patterns means that similar behavioral patterns are expected to continue in the future based on behavioral patterns that are repeated at similar times and frequencies. As shown in the table in Figure 6, by classifying according to the types of electrical appliances (four types: "daily life, meals, activities, and others"), the increase or decrease in electrical appliances in the categories of meals and activities can be comprehensively grasped, making it easier to understand the activity levels of the inhabitants of the household, such as increased or decreased activity.

[0043] Furthermore, the behavioral pattern estimation unit 120 divides past data on power consumption by type of electrical equipment into multiple time periods within a day. Based on the power consumption data by type of electrical equipment for each of these multiple time periods, it estimates the behavioral patterns of consumers based on which types of electrical equipment are being used. In addition, the behavioral pattern estimation unit 120 divides past data on power consumption by type of electrical equipment into daily or monthly categories. Based on the power consumption data by type of electrical equipment for daily or monthly categories, it estimates the behavioral patterns of consumers based on which types of electrical equipment are being used.

[0044] Figure 7 shows, for illustrative purposes, the trend in electricity consumption at different times of the day. In the case of the resident of House 4 shown in the bar graph of Figure 7(B), the waking and sleeping times can be estimated from the fluctuations in electricity consumption. Furthermore, the increase in electricity consumption under the "Eating" category suggests that the resident is cooking for themselves, taking a reasonable amount of time. In addition, the increase in electricity consumption of electrical appliances under the "Activities" category during the same time period suggests that tasks such as cleaning up after meals and cleaning the room are being performed. Therefore, it can be estimated that the resident of House 4, even if living alone, is living an independent life.

[0045] Without subdividing the day into one-hour increments as shown in Figure 7, it is possible to estimate the behavior patterns of the inhabitants of House 4 by aggregating the day into several time periods. If the detection of the power consumption status of electrical equipment is subdivided too finely, the amount of data will increase and duplication will occur. Therefore, for example, the day is divided into four time periods: 0:00 to 6:00 (midnight), 6:00 to 12:00 (morning), 12:00 to 18:00 (afternoon), and 18:00 to 24:00 (night), and the power consumption status of electrical equipment by category for each time period is aggregated.

[0046] Generally, during the late-night hours from midnight to 6am, people are asleep, and the power consumption of electrical appliances is lower compared to other times of the day. Typically, power consumption by electrical appliances is low, excluding standby power. From 6am to 12pm, many electrical appliances tend to be used in a short period for activities such as waking up, preparing breakfast, cleaning up, and doing laundry, including microwave ovens, dishwashers, and washing machines. From 12pm to 6pm, electrical appliances tend to be used for activities such as cleaning, which involves physical activity during the day, including air conditioners, vacuum cleaners, and irons. From 6pm to midnight, electrical appliances tend to be used for activities such as eating dinner, cleaning up, and bathing, including microwave ovens, dishwashers, and hair dryers.

[0047] From the types of electrical appliances and their power consumption for each time period, it is easy to determine which types of electrical appliances are being used. Therefore, for example, by looking at the changes in air conditioner operation and power consumption throughout the day, the usage and power consumption of microwave ovens, and the usage and power consumption of hair dryers, it becomes possible to estimate the behavioral patterns of the residents of House 4, such as waking up, eating, bathing, and going to sleep.

[0048] Furthermore, the classification of power consumption states by type by the behavioral pattern estimation unit 120 is not limited to the time periods within a single day as described above, but is extended to longer periods. That is, in order to grasp the fluctuations in power consumption states by type for each day of the week, the classification of power consumption states by type is set on a weekly basis. In order to mitigate the fluctuations from week to week, the classification of power consumption states by type is set on a monthly basis.

[0049] For example, even if a vacuum cleaner is used in the afternoon, it may not be used every day, but rather infrequently, such as on Wednesdays and Saturdays. In this case, detecting the type of power consumption on a weekly basis is closer to reality and is more suitable for estimating the behavior patterns of the residents. Also, when using monthly data, seasonal fluctuations (temperature changes) have a significant impact on changes in the type of power consumption. Therefore, it is possible to estimate the behavior patterns of the residents of House 4 from the changes in the monthly type of power consumption, including seasonal fluctuations. For example, focusing on the frequency of vacuum cleaner use (power consumption) in the activity classification, if it is used approximately twice a week for more than six months, it can be estimated that cleaning is continued twice a week, there are few physical ailments in daily life, and there is no onset of illness. When estimating behavior patterns, observations of multiple people are accumulated and analyzed.

[0050] Furthermore, the behavioral pattern estimation unit 120 includes a machine learning unit 121. The machine learning unit 121 performs machine learning based on the type of electrical equipment and the increase or decrease in the power consumption of the electrical equipment from past data on power consumption by type of electrical equipment, and estimates the behavioral patterns of the residents of the house 4.

[0051] As explained above, the power consumption status of each type of electrical appliance is detected sequentially hour by hour, time of day, week by week, and month by month. However, the data on power consumption status by type detected individually is subject to large fluctuations and variations each time, making it difficult to grasp trends and estimate behavioral patterns. Furthermore, if the number of residents in House 4 (the number of people equivalent to the number of households) increases, estimating the behavioral patterns of each resident in House 4 based on the power consumption status data by type becomes complicated and time-consuming and costly.

[0052] Therefore, it is possible to extract specific trends using statistical processing methods. Specifically, for each type of electricity consumption (each type and the amount of electricity consumed by each type), the daily, weekly, and monthly average values ​​(arithmetic mean), "maximum value," "minimum value," and "standard deviation" are calculated. From these calculations, the characteristics of the behavior patterns of the residents of House 4 can be extracted. In addition, the "standard deviation / mean value" is also calculated. In particular, by calculating the "standard deviation / mean value," the range of values ​​becomes larger, making it possible to extract the characteristics of the behavior patterns of the residents of House 4 more clearly.

[0053] A series of statistical calculations are performed in the behavioral pattern estimation unit 120. In addition, machine learning techniques suitable for data processing, such as support vector machines (SVMs), model trees, decision trees, neural networks, multiple linear regression, local weighted regression, established search methods, and multiple regression analysis, are used as appropriate.

[0054] The behavioral pattern estimation unit 120, equipped with a machine learning unit 121, performs calculations to detect the characteristics of how residents of the house 4 use electrical appliances from daily, weekly, and monthly power consumption data by type, thereby enabling the extraction of characteristics of the residents' behavioral patterns. More specifically, a learning model is generated by training it with training data that labels the changes in power consumption and the electrical appliances used, along with the behavioral patterns at those times. Then, data on the changes in power consumption is input to the generated learning model, and the behavioral patterns are estimated.

[0055] The comparison generation unit 130 generates comparison information by comparing the power consumption status of each type of electrical equipment acquired in advance over a predetermined period with the power consumption status of each type of electrical equipment currently being acquired. More specifically, the comparison generation unit 130 generates comparison information based on the type of electrical equipment and the increase or decrease in the power consumption of the electrical equipment, from a comparison between the power consumption status of each type of electrical equipment (mainly power consumption) acquired in the past and the power consumption status of each type of electrical equipment (mainly power consumption) currently being acquired.

[0056] Figure 8 is a schematic diagram illustrating an example of comparative information generation. The bar graph on the left side of Figure 8 shows the average total electricity consumption by type of electrical appliance (past data) for the residents of House 4 over the past week, and corresponds to the typical weekly electricity consumption by type of electrical appliance for those residents. In contrast, the bar graph on the right side of Figure 8 shows the electricity consumption by type of electrical appliance (present data) for the most recent week. In both bar graphs, there is little change in electricity consumption by type classified as "daily life" and "other." However, electricity consumption by type classified as "food," and especially "activity," has decreased significantly.

[0057] Therefore, by comparing the bar graphs of both sides, the difference in electricity consumption by category, classified as "Living," "Diet," "Activities," and "Other," can be determined. This difference becomes the comparison information. Of course, the bar graph illustration in Figure 8 is an example, and the comparison information is not limited to the illustrated bar graph; it may also be generated as a difference including a predetermined threshold from the mean and standard deviation of a line graph. Furthermore, to address seasonal fluctuations, the comparison information may be generated from the difference in electricity consumption by category between the same month of the previous year (past data) and the current month (current data).

[0058] The behavioral change estimation unit 140 estimates changes in the behavior of consumers based on the consumers' health information and comparative information. The consumers' health information refers to information about the consumers' physical condition that is communicated to home care facilities 7, hospitals 8, and operators of the consumer monitoring system 1, taking into consideration the consumers' privacy. Specifically, this includes age, sex, height, weight, waist circumference, blood pressure, medical history, presence or absence of diseases, smoking history, number of steps per day, presence or absence of dementia, and even genotype. In addition to these, regional factors (sunshine hours, precipitation, snowfall, etc.) may also be included. Various types of consumer health information are acquired at home care facilities 7, hospitals 8, etc., and then transmitted to the consumer monitoring system 1 in a timely manner.

[0059] Taking the schematic diagram in Figure 8 as an example, a comparison of the two bar graphs shows that the decrease in electricity consumption by type classified as "meals" for the current period is mainly due to a decrease in the usage time and frequency of cooking appliances such as microwave ovens and ovens. Similarly, the decrease in electricity consumption by type classified as "activities" for the current period is mainly due to a decrease in the usage time and frequency of vacuum cleaners and washing machines over the past week. From this, it can be said that the residents of House 4 have reduced the frequency of physical activity (activity level) over the past week.

[0060] As illustrated in Figure 8 above, comparative information is generated based on a comparison of power consumption by type within the power consumption status by type, the type of electrical equipment, and the increase or decrease in power consumption within the power consumption status of said electrical equipment. In addition to the weekly data disclosed in Figure 7, comparative information can also be generated daily and monthly (compared to the same month of the previous year) to ensure the estimation of behavioral patterns based on multifaceted materials and evidence.

[0061] Figure 9 is a schematic diagram illustrating the estimation of behavioral patterns. Based on multiple comparative pieces of information generated through the behavioral change estimation unit 140, changes in the behavioral patterns of the residents of house 4 are estimated. For example, taking Figure 8 as an example, the change in electricity consumption by type among the electricity consumption states classified as "daily life" and "other" is small. From this comparative information, it is estimated that the residents' behavioral patterns are such that they can carry out their daily lives and are not in a situation where their lives are immediately in danger. However, the electricity consumption by type among the electricity consumption states classified as "eating" and "activities" has decreased. From this comparative information, it is estimated that changes in the residents' behavioral patterns, such as limb disabilities, fever, diarrhea, and other infectious diseases, are occurring due to a decrease in active activities and behaviors necessary for daily life.

[0062] Furthermore, if the electricity consumption by category, categorized as "diet" and "activity," has been gradually decreasing over the past six months, it can be inferred that there has been a change in the consumer's behavior, such as a decrease in muscle strength. In addition, health information of the consumer is also added to improve the accuracy of the estimation of changes in the consumer's behavior. For example, if, along with the electricity consumption by category, categorized as "diet" and "activity," there is also a decrease in the consumer's weight and the number of steps taken per day, it can be inferred that the consumer has entered a depressive state.

[0063] In the estimation of changes in consumer behavior patterns based on consumer health information and comparative information, performed by the behavior change estimation unit 140, machine learning analysis methods such as multivariate analysis, linear regression, logistic regression, and regression analysis such as support vector machines are applied to the processing of health information from multiple consumers and multiple comparative information.

[0064] The output unit 150 outputs information regarding changes in the behavior patterns of consumers. Based on comparative information, changes in consumer behavior patterns are estimated, and if action is deemed necessary, output is sent to individuals, departments, etc., related to the consumer in household 4. The output format can be selected from a variety of methods, including sending images and documents via facsimile, voice guidance via telephone, email notifications, SNS message notifications, and chat.

[0065] As shown in the configuration of the resident monitoring system 1 in the embodiment of Figure 1, the computer 10 of the resident monitoring system 1 is connected to the internet line 2, and the output from the computer 10 of the resident monitoring system 1 is reported to the home care facility 7, hospital 8, and government office 9. Upon receiving the output from the output unit 150 of the computer 10 of the resident monitoring system 1, the home care facility 7, hospital 8, government office 9, etc., can promptly visit the resident of the house 4 and communicate with other relevant parties. In this way, it becomes easy to understand the living conditions (lifestyle sensing) of individual houses 4 and their residents through monitoring via electrical equipment. The home care facility 7, hospital 8, and government office 9 in Figure 1 are just examples of contacts for output, and other facilities, neighbors, relatives of the resident of the house 4, etc. may also be included as contacts for output.

[0066] As an example of understanding living conditions (lifestyle sensing) using the Resident Monitoring System 1, an example screen shown in Figure 10 is presented. Figure 10 is an example of the output screen, showing the output text and images. In the example screen in Figure 10(A), the text displayed is, "Mr. X in area P, decreased activity level observed in the past week. Possible heatstroke or infection." For example, if the electricity consumption by type, such as "eating" and "activity," has decreased in the past week compared to the previous week, it is thought that the person is unable to move due to poor health. In this case, the example screen in Figure 10(A) is generated as a possible scenario. In the example screen in Figure 10(B), the text displayed is, "Mr. Y in Q settlement has had the TV and lights on since last night. Please check on his condition immediately." In addition, in the example screen in Figure 10(C), the text displayed is, "Mr. Z at address R has decreased activity level. Please invite him to rehabilitation." Of course, Figure 10 is an example, and multiple types of text displays are prepared according to the estimation of changes in the resident's behavior patterns based on comparative information.

[0067] In this way, without directly monitoring the residents of house 4 at all times, it becomes possible to indirectly monitor the residents' condition based on the power consumption status of electrical appliances in the house, increases or decreases in their usage (power consumption), and changes in the types of electrical appliances. In particular, since it is possible to grasp changes over time, the output can include observations regarding improvements to the residents' lifestyle habits, specifically suggestions to increase physical activity, as shown in the example screen in Figure 9(C).

[0068] In particular, when individuals live alone, they tend to become less inclined to go outside as they age, leading to a decrease in physical activity. Furthermore, snowfall in winter restricts outdoor activity. In such cases, based on the individual's health information, preventative output aimed at improving their activity level from the perspective of maintaining muscle strength to avoid becoming bedridden is significant. Moreover, even if a change occurs in the individual's condition, information about the change in their behavior is sent to home care facilities, hospitals, government offices, etc., enabling early intervention by relevant parties. As a result, monitoring of individuals becomes more efficient, and their quality of life (QOL) is improved through improvements in lifestyle habits.

[0069] In estimating the behavior patterns of consumers in the behavior pattern estimation unit 120, in addition to the behavior patterns themselves, comparable and convertible numerical indicators are also used. Specifically, the behavior pattern estimation unit 120 estimates the behavior patterns of consumers based on the type of electricity consumption status, expressing the amount of the consumers' physical activity as METs-hours, which will be described later.

[0070] Furthermore, the comparison generation unit 130 can generate comparison information by comparing the amount of METs-hours of physical activity corresponding to the power consumption status of each type of electrical equipment, which has been acquired in advance over a predetermined period, with the amount of METs-hours of physical activity corresponding to the power consumption status of each type of electrical equipment, which is currently being acquired.

[0071] Furthermore, the behavioral change estimation unit 140 can use METs-hours, which represent the amount of physical activity of the consumer, when estimating changes in the consumer's behavioral patterns based on the consumer's health information and comparative information.

[0072] The output unit 150, when outputting information regarding changes in consumer behavior, includes in its output observations on improvements in consumer lifestyle habits based on METs-time.

[0073] For example, according to the 2006 Exercise Guidelines for Health Promotion (excerpt), the unit used to express the intensity of physical activity is "METs." METs is a unit that indicates how many times stronger physical activity is compared to resting, with sitting at rest being 1 MET and normal walking being 3 METs. The amount of physical activity is expressed as "METs-hours," which is the METs multiplied by the duration of the physical activity, or as "exercise." Here, METs is an abbreviation for "METs: metabolic equivalents," and is a unit that is widely used internationally to indicate the intensity of physical activity.

[0074] For example, if you perform physical activity at a level of 3 METs for one hour, that would be 3 METs x 1 hour = 3 METs-hours (exercise). Furthermore, if you perform physical activity at 6 METs for 30 minutes, it will be 6 METs x 0.5 hours = 3 METs-hours (exercise).

[0075] For information on the conversion between types of physical activity and METs, please refer to the revised "METs Table for Physical Activity" (https: / / www.nibiohn.go.jp / files / 2011mets.pdf), revised on April 11, 2012, published by the National Institute of Biomedical Innovation, Health and Nutrition. For example, vacuuming is 3.3 METs, ironing is 1.8 METs, sitting quietly and watching television is 1.3 METs, and sleeping is 1.0 MET. A wide range of MET values ​​are specified for each type of physical activity. The MET values ​​corresponding to physical activities are stored in the memory unit 14 of the computer 10 and can also be obtained via the internet connection 2.

[0076] Therefore, the aforementioned power consumption status by type (power consumption by type) reveals what types of electrical equipment were used and for how long. Based on the use of electrical equipment, the type of physical activity and METs corresponding to the user's actions are determined, and the METs-hour (exercise) is determined by multiplying the physical activity duration by the METs-hour. This makes it possible to numerically convert the total amount of physical activity of the user over a day, a week, a month, a six-month period, etc. As a result, when the comparison generation unit 130 generates comparison information, it is used to understand whether the user's long-term physical activity, such as on a monthly basis, is becoming more active or less active.

[0077] The behavioral change estimation unit 140 can adopt METs-hours (exercise), which is the product of the intensity and duration of physical activity, as one of the health information and comparative information items of the consumer. By using METs-hours, the total amount of the consumer's physical activity can be numerically converted, making it easier to analyze the trend of behavioral changes over a predetermined period. For example, if the frequency of ironing has increased over a predetermined period, it can be said that the person has sufficient muscle strength to iron. Also, ironing is an action performed after washing clothes. From this, it can be considered that the consumer is in a mental state where they are paying sufficient attention to their personal grooming, including washing their clothes. Therefore, it is possible to indirectly obtain inferred information such as the fact that the consumer is not in a depressed state.

[0078] As explained in the series of descriptions, METs-hours (exercise), which are the product of the intensity and duration of physical activity, can be adopted as one item of health information and comparative information for consumers. Therefore, in addition to evaluating the presence or absence of activity in each individual consumer's behavioral patterns, it is also possible to evaluate the sum of multiple types of physical activity. For example, physical activity tends to decrease in the summer due to high temperatures and in the winter due to cold temperatures. Also, the types of electrical devices used change with the seasons. In such cases, even if there is no problem with the consumer's body, some electrical devices may directly reflect variations due to annual seasonal fluctuations in usage frequency and duration. Therefore, the use of METs-hours adds objectivity to understanding consumer behavior.

[0079] Furthermore, when outputting information about changes in consumer behavior patterns using the output unit 150, it is possible to add observations to the consumer based on METs-times regarding improvements from their current lifestyle habits.

[0080] Specifically, if the frequency of vacuum cleaner use gradually decreases over a specified period, it may indicate a decline in the muscle strength necessary for vacuuming. This can help detect chronic diseases such as rheumatism in the person. In addition, since vacuum cleaners are electrical appliances used based on the person's own desire to keep their home clean, it can be considered that the person is in a mental state where they cannot pay sufficient attention to cleaning their home. Therefore, it is possible to indirectly obtain clues that the person's mental state may be depressed or otherwise.

[0081] Therefore, if a consumer is experiencing chronic illnesses such as rheumatism, a message such as "Try marching in place for 10 minutes a day" will be displayed on the consumer's device as an alternative action to improve their lifestyle, based on the corresponding METs-hours of vacuum cleaner use. Of course, this is just an example, and multiple types of messages are available depending on the estimated changes in the consumer's behavior based on comparative information. Furthermore, regarding changes in mental state, messages prompting home visits can be sent through the consumer monitoring system 1 to the consumer's primary care physician, home helper, administrative agencies, etc.

[0082] Next, using the flowcharts in Figures 11 and 12, we will explain both the consumer monitoring method and the consumer monitoring program in the consumer monitoring system 1. The consumer monitoring method is executed by the CPU 11 of the computer 10 based on the consumer monitoring program. The consumer monitoring program causes the computer 10 in Figure 4 to execute various functions, including the function of acquiring consumption status by type, the function of estimating behavior patterns, the comparison and generation function, the function of estimating behavioral change, the output function, and various machine learning functions. These functions are executed in the order shown in the diagram. Note that since each function overlaps with the explanation of the consumer monitoring system 1 above, details will be omitted.

[0083] As shown in the flowchart of Figure 11, the processing of computer 10 (CPU 11) includes various steps such as acquiring consumption status by type (S110), estimating behavioral patterns (S120), comparing and generating (S130), estimating behavioral changes (S140), and outputting (S150). Furthermore, as shown in the flowchart of Figure 12, the processing of computer 10 (CPU 11) includes a machine learning step (S121). Of course, the various steps necessary for the operation of computer 10 (CPU 11) itself are naturally included.

[0084] The Type-Specific Power Consumption Status Acquisition function classifies the power consumption status of multiple electrical appliances (see Figure 2) installed in the house 4 according to the type of electrical appliance, and acquires the power consumption status of the electrical appliances included in the classified type as Type-Specific Power Consumption Status over a predetermined period (S110; Type-Specific Power Consumption Status Acquisition Step). The Behavior Pattern Estimation function estimates the behavior patterns of the resident based on past Type-Specific Power Consumption Status of electrical appliances acquired in advance over a predetermined period (S120; Behavior Pattern Estimation Step). The Comparison Generation Function generates comparison information by comparing past Type-Specific Power Consumption Status of electrical appliances acquired in advance over a predetermined period with the Type-Specific Power Consumption Status of electrical appliances currently being acquired (S130; Comparison Generation Step). The Behavior Change Estimation Function estimates changes in the resident's behavior patterns based on the resident's health information and comparison information (S140; Behavior Change Estimation Step). The Output Function outputs information regarding changes in the resident's behavior patterns (S150; Output Step). The functions for acquiring consumption status by type, estimating behavior patterns, comparing and generating data, estimating behavioral changes, and outputting data are executed by the CPU 11 of computer 10.

[0085] The machine learning function performs machine learning based on the type of electrical equipment and the increase or decrease in the power consumption of that electrical equipment, using past data on power consumption by type of electrical equipment, to estimate the behavior patterns of consumers (S121; machine learning step). The machine learning function is executed by the CPU 11 of computer 10.

[0086] The computer program of the present invention described above may be recorded on a processor-readable recording medium, and as the recording medium, a "non-temporary tangible medium" such as a disk, card, semiconductor memory, or programmable logic circuit can be used.

[0087] The above computer program can be implemented using, for example, scripting languages ​​such as ActionScript and JavaScript®, object-oriented programming languages ​​such as Objective-C and Java®, and markup languages ​​such as HTML5. [Explanation of Symbols]

[0088] 1. Consumer Monitoring System 2. Internet connection 3. Power lines 4 House 5 Switchboard 6 Power Consumption Detector 7 Home care facilities 8 Hospitals 9. Government offices (administrative agencies) 10 Computers 11 CPU 12 ROM 13 RAM 14 Storage section 15 Input / Output Interfaces 51 Air conditioner 52 Refrigerator 53 Electric kettle 54 Microwave 55 Washing machine 110 Unit for acquiring consumption status by type 120 Behavioral Pattern Estimation Unit 121 Machine Learning Department 130 Comparison and generation unit 140 Behavioral change estimation unit 150 Output section

Claims

1. A consumer monitoring system that uses a power consumption detector connected to a distribution board installed in the home of a resident, which detects the power consumption status of each type of electrical appliance in the home, A unit for acquiring power consumption status by type classifies the power consumption status of multiple electrical appliances installed in the aforementioned house according to the type of the multiple electrical appliances, and acquires the power consumption status of the electrical appliances included in the classified type as power consumption status by type over a predetermined period of time. A behavioral pattern estimation unit estimates the behavioral patterns of the user based on the power consumption status of the aforementioned types of electrical equipment acquired in advance over a predetermined period, A comparison generation unit generates comparison information by comparing the power consumption status of electrical equipment by type for past data acquired over a predetermined period with the power consumption status of electrical equipment by type currently being acquired. A behavioral change estimation unit that estimates changes in the behavioral patterns of the consumer based on the consumer's health information and the comparative information, The system includes an output unit that outputs information regarding changes in the behavior patterns of the aforementioned consumers. A consumer monitoring system characterized by the following features.

2. The consumer monitoring system according to claim 1, wherein the acquisition of the type-specific power consumption status over the predetermined period in the type-specific power consumption status acquisition unit is for at least one year, including seasonal variations.

3. The behavioral pattern estimation unit divides the past power consumption status of electrical equipment by type into multiple time periods within a day, The consumer monitoring system according to claim 1, which estimates the consumer's behavioral patterns based on which type of electrical equipment is being used, according to the power consumption status of the electrical equipment by type for each of the multiple time periods.

4. The behavioral pattern estimation unit categorizes the past power consumption status of electrical equipment by type on a weekly or monthly basis. The consumer monitoring system according to claim 1, which estimates the consumer's behavioral patterns based on which type of electrical equipment is being used, according to the power consumption status of the electrical equipment by type on a weekly or monthly basis.

5. The consumer monitoring system according to claim 1, wherein the comparison generation unit generates comparison information based on the type of electrical equipment and the increase or decrease in the power consumption state of the electrical equipment, by comparing the power consumption state of the electrical equipment by type for past data with the power consumption state of the electrical equipment by type for currently acquired data.

6. The consumer monitoring system according to claim 1, wherein the output includes findings regarding improvements in the consumer's lifestyle habits.

7. The consumer monitoring system according to claim 1, wherein the consumer is a single person.

8. The behavioral pattern estimation unit includes a machine learning unit, The consumer monitoring system according to claim 1, wherein the machine learning unit performs machine learning based on the type of electrical equipment and the increase or decrease in the power consumption state of the electrical equipment from the power consumption state of the electrical equipment by type for the past period, and estimates the consumer's behavior patterns.

9. The consumer monitoring system according to claim 1, wherein the power consumption state is the amount of power consumed and the waveform of the current when using an electrical device.

10. The consumer monitoring system according to claim 1, wherein the behavioral pattern estimation unit estimates the behavioral pattern of the consumer as the amount of the consumer's physical activity in METs-hours based on the type of power consumption state.

11. The consumer monitoring system according to claim 10, wherein the output includes findings of improvement in the consumer's lifestyle habits based on METs-time.

12. A consumer monitoring method in a consumer monitoring system that uses a power consumption detector connected to a distribution board installed in the consumer's home and which detects the power consumption status of each type of electrical appliance in the home, Computers A step of acquiring power consumption status by type, which involves classifying the power consumption status of multiple electrical appliances installed in the aforementioned house according to the type of the multiple electrical appliances, and acquiring the power consumption status of the electrical appliances included in the classified type as power consumption status by type over a predetermined period of time. A behavioral pattern estimation step in which the behavioral patterns of the consumer are estimated based on the power consumption status of the aforementioned types of electrical equipment obtained in advance over a predetermined period, A comparison generation step that generates comparison information by comparing the power consumption status of electrical equipment by type for past data acquired in advance over a predetermined period with the power consumption status of electrical equipment by type for currently acquired data, A behavioral change estimation step that estimates changes in the behavioral patterns of the consumer based on the consumer's health information and the comparative information, The process includes an output step that outputs information regarding the changes in the behavioral patterns of the aforementioned consumers, and an output step that performs the following: A consumer monitoring method characterized by the following features.

13. A consumer monitoring program in a consumer monitoring system that uses a power consumption detector connected to a distribution board installed in a consumer's home and detects the power consumption status of each type of electrical appliance in the home, On the computer, A function for acquiring power consumption status by type, which classifies the power consumption status of multiple electrical appliances installed in the aforementioned house according to the type of the multiple electrical appliances, and acquires the power consumption status of the electrical appliances included in the classified type as power consumption status by type over a predetermined period of time, A behavioral pattern estimation function that estimates the behavioral patterns of the user based on the power consumption status of the aforementioned types of electrical equipment acquired in advance over a predetermined period, A comparison generation function that generates comparison information by comparing the power consumption status of electrical equipment by type for past data acquired over a predetermined period with the power consumption status of electrical equipment by type currently being acquired, A behavioral change estimation function that estimates changes in the behavioral patterns of the consumer based on the consumer's health information and comparative information, To realize an output function that outputs information regarding changes in the behavior patterns of the aforementioned consumers. A consumer monitoring program characterized by the following features.

Images