Monitoring system

Abstract

To provide a simple watching system that does not require large costs for installation and maintenance.SOLUTION: A watching system comprises: human sensors arranged in residences to be watched; a daily life time zone setting unit 51 that sets each daily life time zone for various lifestyles in a day; a reaction value calculation unit 52 that calculates a daily time zone reaction value from outputs of the human sensor for each divided time obtained by dividing the daily life time zone by a predetermined number; a determination condition determination unit 53 that determines determination conditions for calculating a style estimation time, which is an estimation time of a lifestyle in the daily life time zone, using a series of daily life time zone reaction values; and a lifestyle estimation time calculation unit 54 that calculates the style estimation time based on the determination conditions.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a monitoring system that estimates the lifestyle pattern of a person being monitored and monitors whether the person being monitored is living a healthy life based on the estimation results. [Background technology]

[0002] The monitoring system described in Patent Document 1 includes a number of sensors that detect the health condition and surrounding conditions of the person being monitored, a management means that uses signals output from these sensors to manage the health condition and surrounding conditions of the person being monitored according to preset management items, and a monitoring means that notifies the user when an abnormality occurs in a management item.The sensors include a sensor terminal that detects the room temperature, humidity, human presence, and illuminance in the room of the person being monitored, a bed sensor including a pressure sensor and an acoustic sensor that detects the body movements, pulse rate, respiratory rate, and ambient sound or vibration of the person being monitored, a wearable sensor that detects the body temperature of the person being monitored, and an outside air temperature sensor that detects the temperature of the air outside the room of the person being monitored. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2017-168098 Summary of the Invention [Problem to be solved by the invention]

[0004] The monitoring system shown in Patent Document 1 uses a large number of sensor signals to manage the health condition and surrounding conditions of the person being monitored, but the placement of a large number of sensors, the signal processing equipment for the large number of sensor signals, and the maintenance of these devices require large costs.Although it is important to monitor people being monitored, such as elderly people living alone or in pairs, if the installation and maintenance of such a monitoring system requires large costs, it will be difficult for the monitoring system to become widespread.

[0005] In view of the above circumstances, an object of the present invention is to provide a simple monitoring system that does not require large installation and maintenance costs. [Means for solving the problem]

[0006] The monitoring system according to the present invention comprises a human presence sensor arranged in a monitored residence, a lifestyle time zone setting unit that sets lifestyle time zones for various lifestyle behaviors in one day, a reaction value calculation unit that calculates a lifestyle time zone reaction value from the output of the human presence sensor for each divided time period obtained by dividing the lifestyle time zone by a predetermined number, and a judgment condition determination unit that determines a judgment condition for calculating a behavior estimation time, which is an estimated time of the lifestyle behavior in the lifestyle time zone, using a series of the lifestyle time zone reaction values. By determining the lifestyle time zone reaction value based on the determination condition, the specific lifestyle behavior and the estimated time when the specific lifestyle behavior was performed are obtained, and the estimated time is The estimated time of the above-mentioned as The person being watched over in this application is preferably an elderly person or a physically inactive person whose daily activities should preferably be watched over by a family member or the like.

[0007] According to this configuration, a lifestyle time zone, which is a time zone during which basic lifestyle behaviors such as waking up, breakfast, lunch, dinner, and bedtime usually occur (e.g., breakfast is eaten from 4:00 to 10:30), is divided into appropriate time intervals for evaluating the output of the human presence sensor. A lifestyle time zone response value, which is a sensor response value for each divided time period within the lifestyle time zone, is calculated from the output of the human presence sensor at each divided time period. While this sensor response value varies depending on the output form of the human presence sensor, the number of sensor response seconds, the integrated value of the time during which the sensor output exceeds the sensor threshold value within the divided time period, the integrated value of the peak sensor output, the average value of the peak sensor output, etc., are used. The judgment condition determination unit determines judgment conditions for calculating a lifestyle estimation time, which is an estimated time of a specific lifestyle behavior within the lifestyle time zone, using the lifestyle time zone response values ​​calculated in advance. These judgment conditions can be determined from the actual times of past lifestyle behaviors (ground truth data) and a set of lifestyle time zone response values ​​over time (training data). The lifestyle behavior estimated time calculation unit calculates the lifestyle behavior estimated time based on the judgment conditions determined by the judgment condition determination unit. In this monitoring system, the lifestyle time zone response value (sensor response value) obtained by processing the output of the human presence sensor in units of divided time periods of the lifestyle time zone is judged based on the judgment conditions, thereby obtaining the estimated time when a specific lifestyle behavior occurred. Since the system only processes the output of the human presence sensor, the configuration of the signal processing system is simple, which has the advantage of low costs required for installing and maintaining the monitoring system.

[0008] For example, since the lifestyle rhythm of an elderly person, who is an example of a person being watched over, is relatively constant, if there is a change in the time at which a particular lifestyle pattern, such as breakfast or lunch, is performed, it can be determined that some kind of problem has occurred. Therefore, by comparing the estimated times of various lifestyle patterns calculated daily by the lifestyle pattern estimated time calculation unit with the past lifestyle pattern times of the person being watched over and the lifestyle pattern times of the person being watched over and a healthy person, it is possible to evaluate the lifestyle condition of the person being watched over. For this reason, the present invention is provided with a lifestyle condition evaluation unit that evaluates the lifestyle condition of the person being watched over by comparing the estimated lifestyle time calculated by the lifestyle pattern estimated time calculation unit with pre-registered registered healthy times.

[0009] The output of a human sensor in a predetermined divided time period can be treated as a waveform whose output increases and decreases over time. To evaluate human behavior from such a waveform output, it is preferable to use a value obtained by statistically processing the sensor output of the human sensor, i.e., a statistically calculated value (such as a median, average, or standard deviation), as the daily time zone response value, rather than using an instantaneous peak value. Therefore, in the present invention, the daily time zone response value is a statistically calculated value of the sensor output of the human sensor in the divided time period. In particular, the inventor's experience and experimental results have revealed that if there are multiple waves during a time period (time width) during which the sensor output of the human sensor exceeds a predetermined sensor threshold, the integrated value of these waves is an appropriate statistically calculated value. Therefore, in the present invention, the statistically calculated value of the sensor output of the human sensor is an integrated value during the time during which the sensor output exceeds the sensor threshold.

[0010] Simple judgment conditions for determining the lifestyle time zone response values ​​required to obtain a lifestyle aspect estimated time are advantageous because they simplify the judgment process. In this case, when calculating a final value from multiple candidate values, a maximum value calculation that selects the largest candidate value from a group of candidate values ​​is a simple and effective calculation. To increase the flexibility of the judgment conditions, a judgment condition in which the final candidate is a lifestyle time zone response value that falls within a predetermined range from the maximum value obtained by the maximum value calculation, for example, a value obtained by calculating (e.g., simple multiplication) the maximum value with a predetermined coefficient, is also suitable. For this reason, in the present invention, a judgment threshold derived from the maximum value of the lifestyle time zone response values ​​in the lifestyle time zone and a lifestyle aspect judgment coefficient set according to the type of lifestyle aspect is used as the threshold of the judgment condition.

[0011] When there are multiple lifestyle time zone reaction values ​​that satisfy the judgment condition, the timing at which the corresponding lifestyle behaviors start is important, so it is preferable that the lifestyle time zone reaction value acquired at the earliest time is used to calculate the lifestyle behavior estimation time. For this reason, in the present invention, the earliest time at which the lifestyle time zone reaction value equal to or greater than the judgment threshold occurs is set as the lifestyle behavior estimation time.

[0012] Examples of lifestyle patterns that are important for monitoring the lifestyle of a person being monitored and that can be appropriately estimated by the system of the present invention include breakfast, lunch, and dinner. Some people being monitored may eat two meals a day instead of three, so in the present invention, the lifestyle pattern is defined as at least one of breakfast, lunch, and dinner. By observing the time each day when at least one of breakfast, lunch, and dinner is eaten, which serves as a reference for dietary habits, it is possible to monitor whether the lifestyle of the person being monitored is healthy.

[0013] It is more convenient to change the lifestyle determination coefficients and lifestyle time zones using monitoring data for each monitored residence over the past few days rather than continuing to use values ​​that have been set once. For this reason, in the present invention, the set lifestyle determination coefficients are changed using a coefficient calculated by the difference between the maximum value of the lifestyle time zone response value over the past few days and the statistically calculated value of the lifestyle time zone response value other than the divided time corresponding to the maximum value. Furthermore, the set lifestyle time zones are changed by a range that includes a time zone before and after the time when the peak of the lifestyle time zone response value over the past few days was obtained, and the lifestyle time zones are changed based on this range.

[0014] In one specific embodiment of the present invention, the lifestyle is breakfast, the lifestyle time period is from around 4:00 to around 10:30, the divided time is 30 minutes, and the lifestyle judgment condition for the lifestyle is that the estimated breakfast time candidate is the time at which the lifestyle time period response value is equal to or greater than the value obtained by multiplying the maximum value of the lifestyle time period response value by a lifestyle judgment coefficient of 0.2 to 0.4, preferably 0.3, and the earliest of the estimated breakfast time candidates is set as the estimated breakfast time. Here, "around" in around 4:00 or around 10:30 indicates the variation in the breakfast time of the person being watched over, and may be expressed as, for example, plus or minus about one hour. In this embodiment, which has proven highly reliable through experiments, the period from around 4:00 to around 10:30 is considered to be the breakfast time period, this breakfast time period is divided into 30-minute increments, and the maximum lifestyle time period response value calculated within each divided time period is multiplied by a lifestyle mode determination coefficient (breakfast determination coefficient) of 0.2 to 0.4, preferably 0.3, to determine the determination threshold, and the lifestyle time period response value that satisfies this determination threshold is determined as the estimated breakfast time candidate. Furthermore, the earliest of the estimated breakfast time candidates is determined to be the estimated breakfast time.

[0015] In another specific embodiment of the present invention, the lifestyle is lunch, the lifestyle time period is from around 11:00 to around 1:30 p.m., the divided time is 30 minutes, and the lifestyle judgment condition for the lifestyle is that the estimated lunch time candidate is the time at which the lifestyle time period response value is equal to or greater than the value obtained by multiplying the maximum value of the lifestyle time period response value by a lifestyle judgment coefficient of 0.3 to 0.5, preferably 0.4, and the earliest of the estimated lunch time candidates is set as the estimated lunch time. Here again, "around" in around 11:00 a.m. or around 1:30 p.m. indicates the variation in the lunch time of the person being watched over, and may be rephrased as, for example, plus or minus about one hour. In this embodiment, which has been shown to be highly reliable through experiments, the period from around 11:00 to around 1:30 pm is considered to be the lunch time period, this lunch time period is divided into 30-minute increments, and the maximum lifestyle time period response value calculated within each divided time period is multiplied by a lifestyle aspect judgment coefficient (lunch judgment coefficient) of 0.3 to 0.5, preferably 0.4, to determine the judgment threshold, and the lifestyle time period response value that satisfies this judgment threshold is determined to be the estimated lunch time candidate. Furthermore, the earliest of the estimated lunch time candidates is determined to be the estimated lunch time.

[0016] In another specific embodiment of the present invention, the lifestyle is dinner, the lifestyle time period is from around 4:00 PM to around 7:30 PM, the divided time period is 30 minutes, and the lifestyle judgment condition for the lifestyle is that the time at which the lifestyle time period response value is issued is equal to or greater than the value obtained by multiplying the maximum value of the lifestyle time period response value by a lifestyle judgment coefficient of 0.3 to 0.5, preferably 0.4, is set as an estimated dinner time candidate, and the earliest of the estimated dinner time candidates is set as the estimated dinner time. Here, "around" in around 4:00 PM or around 7:30 PM indicates the variation in the dinner time of the person being watched over, and may be rephrased as, for example, plus or minus about one hour. In this embodiment, which has been shown to be highly reliable through experiments, the period from around 4:00 PM to around 7:30 PM is considered to be the dinner time zone, this dinner time zone is divided into 30-minute increments, and the maximum lifestyle time zone reaction value calculated within each divided time is multiplied by a lifestyle mode reaction coefficient (dinner reaction coefficient) of 0.3 to 0.5, preferably 0.4, to determine the judgment threshold, and the lifestyle time zone reaction value that satisfies this judgment threshold is determined as the estimated dinner time candidate. Furthermore, the earliest of the estimated dinner time candidates is determined as the estimated dinner time.

[0017] In another embodiment of the present invention, which employs a practical lifestyle behavior time calculation process for realizing a simple monitoring system, the human presence sensor is disposed in a place where meals and cooking are performed, the lifestyle behavior is waking up, the wake-up time zone as the lifestyle time zone is from around 3:00 to around 11:30, and the time in the divided time immediately before the wake-up time zone is calculated. Human SensorIf the sensor response value is less than a predetermined value and the sensor response value in each of the divided times of the wake-up time band is equal to or greater than the predetermined value, the tentative wake-up time is set to before 3:00. If the wake-up determination process is not successful, the earliest time among the divided times of the wake-up time band is set to the tentative wake-up time if the maximum value of the sensor response value in the divided times of the wake-up time band is not zero. If the tentative wake-up time is later than the previously calculated estimated breakfast time, the estimated breakfast time is set to the estimated wake-up time. It is difficult to estimate the wake-up time using only a human presence sensor placed in places where meals and cooking are performed, but in a monitoring system, if the fact that someone is awake is important, the above configuration can achieve this purpose.

[0018] Similarly, it is difficult to estimate the bedtime using only a motion sensor placed in a place where meals and cooking are done, but if the fact that the person being monitored has gone to bed in the monitored residence is important in a monitoring system, it is possible to build a monitoring system to achieve this purpose. In such a monitoring system, the motion sensor is placed in a place where meals and cooking are done, the lifestyle is sleeping, the bedtime period as the lifestyle time period is from around 8:00 p.m. to around 4:30 a.m., and ... Human Sensor If the sleep determination process is successful in that the sensor response value is zero and there is a divided time at which the sensor response value is less than the sensor response value at the divided time immediately before the bedtime period, the latest divided time is set as the estimated sleep time, and if the sleep determination process is not successful, the estimated sleep time is set to 0:00 under the condition that the estimated wake-up time for the next day is calculated.

[0019] The monitoring system of the present invention can be configured using not only the output from the human presence sensor but also the power consumption from a power meter. Such a monitoring system includes a human presence sensor disposed in a residence to be monitored, a power meter acquiring the power consumption of the residence to be monitored, a lifestyle time zone setting unit setting lifestyle time zones for various lifestyles in one day, a response value calculation unit calculating a lifestyle time zone response from the output of the human presence sensor for each divided time period obtained by dividing the lifestyle time zone by a predetermined number, calculating a human presence statistically calculated value which is a statistically calculated value of the lifestyle time zone response value for the lifestyle time zone, and calculating a power statistically calculated value which is a statistically calculated value of the power consumption for the lifestyle time zone, and a judgment condition determination unit determining a human presence judgment condition for calculating a first behavior estimated time which is an estimated time of the lifestyle mode in the lifestyle time zone using the human presence statistically calculated value for the lifestyle time zone, and determining a power judgment condition for calculating a second behavior estimated time which is an estimated time of the lifestyle mode in the lifestyle time zone using the power statistically calculated value for the lifestyle time zone. The human presence statistical calculation value is determined based on the human presence determination condition to obtain the specific lifestyle and an estimated time when the specific lifestyle was performed, and the estimated time is calculated as the first lifestyle estimated time. The power statistical calculation value is determined based on the power determination condition to obtain the specific lifestyle and an estimated time when the specific lifestyle was performed, and the estimated time is calculated as the second lifestyle estimated time. From the first lifestyle estimated time and the second lifestyle estimated time, and a lifestyle behavior estimated time calculation unit that calculates a lifestyle estimated time. Going to bed and waking up, which are daily lifestyle behaviors, are determined by sleep time. During this sleep time, people barely move around, occasionally going to specific places, such as the kitchen or bathroom. Power consumption during sleep is low and constant. For this reason, a monitoring system that uses output from a human presence sensor and power consumption from a power meter to estimate wake-up and bedtimes is particularly effective.

[0020] Another embodiment of a monitoring system that uses output from a human presence sensor and power consumption from a power meter to estimate wake-up times and bedtimes includes a human presence sensor disposed in a residence to be monitored, a power meter that acquires power consumption at the residence to be monitored, a lifestyle time zone setting unit that sets lifestyle time zones for various lifestyle patterns in one day (wake-up, breakfast, lunch, dinner, bedtime), a response value calculation unit that calculates a lifestyle time zone response value (sensor response seconds) from the output of the human presence sensor for each divided time period obtained by dividing the lifestyle time zone by a predetermined number, a first judgment condition determination unit that determines a first judgment condition using a first human presence statistically calculated value (preferably an average value or a standard deviation) that is a first statistically calculated value of the lifestyle time zone response value in the lifestyle time zone, a second judgment condition determination unit that determines a second judgment condition using a second human presence statistically calculated value (preferably an average value or a standard deviation) that is a second statistically calculated value of the lifestyle time zone response value in the lifestyle time zone, and a third judgment condition determination unit that determines a third judgment condition using a power statistically calculated value (preferably an average value) that is a statistically calculated value of the power consumption in the lifestyle time zone. The first-sense statistical calculation value is determined according to the first determination condition to obtain a specific lifestyle and an estimated time when the specific lifestyle was performed, and the estimated time is calculated as a first-mode estimated time; the second-sense statistical calculation value is determined according to the second determination condition to obtain a specific lifestyle and an estimated time when the specific lifestyle was performed, and the estimated time is calculated as a second-mode estimated time; the power statistical calculation value is determined according to the third determination condition to obtain a specific lifestyle and an estimated time when the specific lifestyle was performed, and the estimated time is calculated as a third-mode estimated time; and from the first-mode estimated time, the second-mode estimated time, and the third-mode estimated time, and a lifestyle behavior estimated time calculation unit that calculates an estimated behavior time, which is an estimated time of the lifestyle behavior in the lifestyle time zone. In this configuration, a first determination condition is determined using a first human presence statistical calculation value (e.g., a value obtained by adding a standard deviation to an average value) obtained from a group of lifestyle time zone reaction values ​​based on the output of a human presence sensor, a second determination condition is determined using a second human presence statistical calculation value (e.g., a value obtained by adding three times the standard deviation to an average value) obtained from a group of lifestyle time zone reaction values ​​based on the output of the human presence sensor, and a third determination condition is determined using a power statistical calculation value, which is a statistical calculation value (e.g., an average value) of power consumption in the lifestyle time zone. From the lifestyle behavior estimated time candidates, for example, the candidate with the earliest time is calculated and set as the final lifestyle behavior estimated time (e.g., an estimated bedtime).

[0021] Other features, operations, and advantages of the present invention will become apparent from the following description of the invention using the accompanying drawings. [Brief explanation of the drawings]

[0022] [Figure 1] FIG. 1 is a functional block diagram illustrating a configuration example of a monitoring system. [Figure 2] FIG. 2 is a functional block diagram showing an example of the configuration of a data processing unit constructed in a management computer and the flow of data. [Figure 3] 10 is a flowchart illustrating an example of a monitoring process. [Figure 4] 10 is a flowchart showing an example of a daily life time zone reaction value calculation process. [Figure 5] FIG. 10 is a schematic diagram showing an example of a procedure for calculating a daily life time zone reaction value. [Figure 6] 10 is a flowchart illustrating an example of a determination condition generation process. [Figure 7] FIG. 10 is a functional block diagram showing an example of a data processing unit in a monitoring system that uses output from a human presence sensor and power consumption from a power meter. [Figure 8] 10 is a flowchart illustrating the basic logic of a practical lifestyle aspect estimated time calculation process. [Figure 9] 9 is a flowchart showing the calculation of an estimated breakfast time based on the basic logic of FIG. 8. [Figure 10] 9 is a flowchart of an estimated lunch time calculation based on the basic logic of FIG. 8. [Figure 11] 9 is a flowchart showing an estimated dinner time calculation based on the basic logic of FIG. 8. [Figure 12] 9 is a flowchart of an estimated wake-up time calculation based on the basic logic of FIG. 8. [Figure 13] 9 is a flowchart for calculating an estimated bedtime based on the basic logic of FIG. 8. DETAILED DESCRIPTION OF THE INVENTION

[0023] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A monitoring system according to an embodiment of the present invention will be described in detail below with reference to the accompanying drawings. In the following description, the monitoring system is an elderly monitoring system for monitoring elderly people.

[0024] As shown in FIG. 1, the monitoring system of this embodiment comprises a management computer 1 that functions as a cloud service that provides information to registered users, and a terminal device 20 that is located in a monitored residence (hereinafter simply referred to as residence) 2 where an elderly person lives. The management computer 1 and terminal device 20 are connected to each other via the Internet, a public line, or the like so that data can be exchanged. The management computer 1 transmits monitoring information to a user terminal 4 of a requesting user who has requested monitoring of the elderly person. The management computer 1 can also transmit monitoring information to the user terminal 4 of the elderly person who is being monitored.

[0025] A human presence sensor 3 that detects the presence of a person is disposed in a living space such as the kitchen of the residence 2. A power meter 30 is also disposed in the residence 2. The terminal device 20 is an ICT terminal and has the function of transmitting the human presence output from the human presence sensor 3 and the amount of power consumption from the power meter 30 to the management computer 1. For this purpose, the terminal device 20 includes a detection signal receiving unit 22, a signal processing unit 23, and a communication unit 24. The detection signal receiving unit 22 receives detection signals from the human presence sensor 3 and the power meter 30. The signal processing unit 23 performs necessary preprocessing on the detection signals received by the detection signal receiving unit 22, adds additional data such as time data (time stamp), performs necessary data format conversion, and generates transmission data. The communication unit 24 transmits the transmission data generated by the signal processing unit 23 to the management computer 1.

[0026] The management computer 1 includes a data transmission / reception unit 11, a data storage unit 12, and a data processing unit 5. The data transmission / reception unit 11 communicates with a terminal device 20 installed in the residence 2 of each person being monitored, and receives data related to the motion detection output from the motion detection sensor 3 and the power consumption amount from the power meter 30 from each terminal device 20. The data storage unit 12 includes a storage device capable of non-volatilely storing large amounts of data, such as a hard disk drive or solid state disk, and manages data handled by the management computer 1 in a searchable manner. The data processing unit 5 is a core element of the monitoring system, and is configured as a computer including an arithmetic unit and a storage device such as a semiconductor memory. In this embodiment (first embodiment), the motion detection output from the motion detection sensor 3 is used for the elderly monitoring process, and therefore only the motion detection output from the motion detection sensor 3 is used in the data processing unit 5.

[0027] 2, the data processing unit 5 includes a lifestyle time zone setting unit 51, a reaction value calculation unit 52, a judgment condition determination unit 53, a lifestyle estimated time calculation unit 54, a lifestyle state evaluation unit 55, and a registered user message creation unit 57. Note that the time display in this application is a 24-hour display.

[0028] The lifestyle time zone setting unit 51 sets various lifestyle patterns of the elderly person in the residence 2 to be monitored throughout the day; in this embodiment, lifestyle time zones for breakfast, lunch, and dinner (breakfast time zone, lunch time zone, and dinner time zone). As lifestyle time zones that are considered common to most elderly people, the breakfast time zone can be set from 4:00 to 10:30, the lunch time zone from 11:00 to 13:30, and the dinner time zone from 16:00 to 19:30. Of course, these lifestyle time zones can be set for each residence 2. For example, each lifestyle time zone is set or corrected based on the time when each lifestyle pattern is estimated to be occurring based on a statistical analysis of the human detection output from the human detection sensor 3 for the lifestyle time zones over the past few days for each residence 2.

[0029] The response value calculation unit 52 calculates a lifestyle time zone response value (also referred to as a sensor response value) from the output of the human presence sensor 3 for each divided time period obtained by dividing the set lifestyle time zone by a predetermined number in the lifestyle time zone setting unit 51. For example, the breakfast time zone response value (also referred to as a breakfast response value) is calculated from the output of the human presence sensor 3 for each divided time period of the breakfast time zone. The calculation of this lifestyle time zone response value will be explained later using diagrams. The divided time periods into which the lifestyle time zone is divided are determined through experiments, etc., but approximately 30 minutes is appropriate. Of course, this divided time period may also be changed depending on the lifestyle time zone and each residence 2. The divided time periods may be changed at any time by determining an appropriate divided time period through statistical calculation of a group of sensor response values ​​from the most recent few days. The response value calculation unit 52 has a statistical calculation unit 52a for performing statistical calculations on the group of lifestyle time zone response values.

[0030] The judgment condition determination unit 53 determines lifestyle mode judgment conditions (hereinafter simply referred to as judgment conditions) used when the lifestyle mode estimated time calculation unit 54 calculates lifestyle estimated times, which are estimated times of lifestyle modes, based on each lifestyle time zone reaction value calculated over time by the reaction value calculation unit 52. That is, the lifestyle mode estimated time calculation unit 54 calculates lifestyle estimated times (estimated breakfast time, estimated lunch time, estimated dinner time), which are estimated times of lifestyle modes (breakfast, lunch, dinner), using a series of lifestyle time zone reaction values ​​(breakfast reaction value group, lunch reaction value group, dinner reaction value group) calculated over time by the reaction value calculation unit 52 and the judgment conditions determined by the judgment condition determination unit 53. Specific judgment conditions for each lifestyle mode will be described in detail later.

[0031] The living condition evaluation unit 55 evaluates the living condition of the elderly person being monitored by comparing the estimated behavior times of each living condition (breakfast, lunch, dinner) calculated by the estimated lifestyle time calculation unit 54 with pre-registered registered healthy times. Based on the actual behavior times when the elderly person being monitored is in a healthy state or based on the average value of a group of actual behavior times of the elderly person, the healthy times for each living condition at which the elderly person being monitored is deemed to be living a healthy life are registered in the healthy time registration unit 56. The healthy times extracted from the healthy time registration unit 56 for evaluating the living condition are the registered healthy times.

[0032] The registered user message creation unit 57 sends a message indicating the monitoring evaluation result evaluated and output by the living condition evaluation unit 55 to the user terminal 4 of the registered user, who is the family member of the elderly person being monitored. Of course, the message indicating the monitoring evaluation result can also be sent directly to the elderly person being monitored.

[0033] The elderly monitoring process, which is one of the basic monitoring processes in this monitoring system, will be described using the flowchart in Figure 3. First, the lifestyle time zone setting unit 51 sets lifestyle time zones for each lifestyle for the elderly person living in the target residence 2 (#11). The divided time periods for calculating the sensor response value in the response value calculation unit 52 are set for each lifestyle (#12).

[0034] The response value calculation unit 52 sequentially calculates lifestyle time zone response values ​​(sensor response values) for each divided time period in the target lifestyle time zone based on the output of the human presence sensor 3 transmitted from the terminal device 20 of the target residence 2 (#13). The lifestyle aspect estimated time calculation unit 54 selects a sensor response value that satisfies the judgment condition from the series of sensor response values ​​based on the judgment condition determined by the judgment condition determination unit 53 (#14), and calculates the time at which the selected sensor response value is emitted as a lifestyle aspect estimated time candidate (#15). If there is a single lifestyle aspect estimated time candidate, the lifestyle aspect estimated time candidate becomes the final lifestyle estimated time. If there are multiple lifestyle aspect estimated time candidates, the lifestyle aspect estimated time candidate with the earliest time among them is calculated as the final lifestyle estimated time (#16). Here, the lifestyle aspect estimated time candidate becomes the breakfast estimated time candidate if the lifestyle time zone is the breakfast time zone, the lunch estimated time candidate if the lifestyle time zone is the lunch time zone, and the dinner estimated time candidate if the lifestyle time zone is the dinner time zone.

[0035] Next, the living condition evaluation unit 55 checks whether the time falls within the registered reference time range, which is the allowable time range defined from the registered healthy time extracted from the healthy time registration unit 56 (#17). If the time falls within the allowable time range (#17 Yes branch), it evaluates that there is no disturbance in the determined living behavior (#18). If the time does not fall within the allowable time range (#17 No branch), it evaluates that there is a disturbance in the determined living behavior (#19).

[0036] The evaluation results by the living condition evaluation unit 55 are notified to the corresponding registered user together with monitoring information in the form of tables or graphs for each daily living pattern (breakfast, lunch, dinner) (#20).

[0037] Next, an example of a daily time zone response value and a calculation method thereof will be described with reference to FIGS. 4 and 5. FIG. 5 schematically illustrates how the daily time zone response value is calculated. The motion sensor 3 in this embodiment is, for example, a commercially available sensor using a pyroelectric infrared sensor. The motion sensor 3 detects and outputs changes in the intensity of infrared rays emitted from a human body within the detection area of ​​the motion sensor 3, for example, at intervals of approximately 1 to 10 seconds. The top graph in the figure shows a waveform signal corresponding to the output of the motion sensor 3, and an enlarged portion of the waveform signal is shown below it. Furthermore, the bottom graph in the figure is a schematic diagram for explaining how to calculate the daily time zone response value from the output of the motion sensor 3, and the waveform signal is shown schematically. In this schematic diagram, the vertical axis represents the sensor output of the motion sensor 3, the horizontal axis represents time, and the daily time zone is divided into four time periods. The cumulative value (indicated by the subscript S) of the statistical calculation value of the time width (indicated by the subscript s) at each divided time of waves that exceed the sensor threshold set to remove noise is the daily time zone response value (sensor response value), in other words, the number of seconds for which the human presence sensor 3 responds over a 30-minute period.

[0038] When the lifestyle time zone response value calculation process is performed in real time, as shown in FIG. 4, a check is made to see if it is the start time of the lifestyle time zone to be monitored (#21). If it is the lifestyle time zone to be monitored (#21: Yes branch), the sensor output for the first divided time period is acquired (#22). As described above, the acquired sensor output is processed to calculate the duration of the sensor output wave above the sensor threshold as the sensor response value (#23), and the integrated value of the sensor response value for the divided time period is calculated as the lifestyle time zone response value (#24). The calculated lifestyle time zone response value is stored in memory. The process from step #22 to step #25 is repeated for all divided time periods. When the lifestyle time zone to be monitored ends (#26: Yes branch), the lifestyle time zone response value calculation process for this lifestyle time zone is completed, and the process waits for the start of the next lifestyle time zone to be monitored.

[0039] When the lifestyle time zone response value calculation process is batch processed, the sensor output for the lifestyle time zone to be monitored is stored in advance in memory, and the above-mentioned response value calculation process is performed on this stored sensor output. Note that this lifestyle time zone response value calculation process can also be performed by the signal processing unit 23 of the terminal device 20 installed in each residence 2. In this case, the lifestyle time zone setting unit 51 and the response value calculation unit 52 are built in the signal processing unit 23, and the lifestyle time zone response values ​​calculated therein are sent from the terminal device 20 to the data processing unit 5.

[0040] Next, the process of generating a judgment condition performed by the judgment condition determination unit 53 will be described using the flowchart of FIG. 6. First, a lifestyle time zone for a lifestyle behavior is set (#41), and then a division time for calculating a sensor response value for that time zone is set (#42). The response value calculation unit 52 calculates a series of sensor response values ​​for the lifestyle time zone (#43). From the series of sensor response values ​​for the lifestyle time zone over the past few days (e.g., three days), the sensor response value with the largest value is calculated as the maximum sensor response value (#44). Using this maximum sensor response value as a reference, a lifestyle behavior judgment coefficient is set (#45) to select a sensor response value that will be a candidate for calculating the estimated behavior time. For example, if the lifestyle behavior judgment coefficient is a numerical value less than 1 that is multiplied by the maximum sensor response value, sensor response values ​​having a value equal to or greater than the product of the maximum sensor response value and the lifestyle behavior judgment coefficient can be selected as candidates for calculating the estimated behavior time. This lifestyle behavior judgment coefficient may be a value that varies depending on the lifestyle behavior. Once the lifestyle aspect determination coefficient is set to this effect, a determination condition is generated in which the determination threshold is (maximum sensor response value) x (lifestyle aspect determination coefficient) (#45). If the calculated sensor response value is equal to or greater than this determination threshold, the sensor response value satisfies the determination condition. The time corresponding to the sensor response value that satisfies this determination condition becomes a lifestyle estimation time candidate, and the earliest lifestyle estimation time candidate among this group of lifestyle estimation time candidates becomes the lifestyle estimation time.

[0041] Regarding the setting of the lifestyle time zone, the set lifestyle time zone is variable. Once the lifestyle time zone is set, the change range is set to plus or minus 30 minutes (the range before or after) from the time when the peak sensor response value was obtained over the last three days, and the lifestyle time zone (e.g., breakfast time zone) may be changed based on this change range. For example, if the peak time of the sensor response value three days ago was 7:30, the change range would be from 7:00 to 8:00; if the peak time of the sensor response value two days ago was 6:00, the change range would be from 5:30 to 6:30; and if the peak time of the sensor response value one day ago was 8:00, the change range would be from 7:30 to 8:30. Taking both ends of these change ranges, the changed lifestyle time zone (breakfast time zone) would be from 5:30 to 8:30.

[0042] The lifestyle mode determination coefficients used in the determination conditions can be determined for each lifestyle mode by referring to experimental results. That is, a breakfast determination coefficient, a lunch determination coefficient, and a dinner determination coefficient are determined. In this embodiment, "0.3" is used as the breakfast determination coefficient, and "0.4" is used as the lunch determination coefficient and the dinner determination coefficient.

[0043] Next, we will explain how to change the set lifestyle behavior determination coefficient. First, the lifestyle behavior determination coefficient is calculated from the difference between the maximum value of the sensor response value for each lifestyle time period for several days to several weeks for each residence 2 and the statistically calculated value of the sensor response value (sensor response value group) other than the divided time period corresponding to the maximum value. Specifically, the lifestyle behavior determination coefficient can be changed using the calculated lifestyle behavior determination coefficient = {{(maximum value) - (average value of the sensor response value group other than the divided time period showing the peak - standard deviation)} / (maximum value)}.

[0044] Since waking up and going to bed are different lifestyles from the above-mentioned breakfast, lunch, and dinner, different judgment criteria are used for these lifestyles. The wake-up judgment criteria used when the lifestyle is waking up are conditions for selecting lifestyle estimation time candidates, and are called wake-up judgment thresholds, which are calculated as follows: (a) First, calculate the maximum value (here referred to as SL-max) of the sensor response value (number of seconds) of the human presence sensor 3 from 2:00 to 3:30 (a time period when the person is likely asleep) for the past three days on which the person is to be monitored. (b) The wake-up criteria are set as follows: (b1) If SL-max<5, set the wake-up determination threshold=5. (b2) If SL-max>100, the wake-up determination threshold is set to the wake-up determination threshold of the previous day. (This prevents the problem of large deviations in wake-up detection due to periods of high reaction in the middle of the night.) (b3) For other SL-max values, the wake-up determination threshold is set to Min{SL-max / 2, the previous day's wake-up determination threshold × 2}. In other words, the smaller of (SL-max / 2) and twice the previous day's wake-up determination threshold is set as the wake-up determination threshold for the current day. (c) Between 4:00 and 11:00 on the day to be monitored, if (the sensor response value of the human presence sensor 3 in the previous living time period is less than or equal to the wake-up determination threshold) and (the response value of the human presence sensor 3 in the living time period of the person to be monitored is greater than or equal to the wake-up determination threshold), the earliest estimated wake-up time candidate among the group of estimated wake-up time candidates corresponding to the selected sensor response value becomes the estimated wake-up time. (d) This sets a level that can confirm the transition from a state that is probably asleep to a state that is probably engaged in some activity. Also, in the case where a large sensor response value is generated because the person woke up in the middle of the night and went to the kitchen, the judgment threshold is designed not to fluctuate significantly between consecutive days to prevent the problem of the person being judged not to have met the wake-up conditions (not awake) for the next few days.

[0045] Furthermore, the sleep determination condition used when the lifestyle is sleeping is a condition for selecting estimated sleep time candidates, and is called a sleep determination threshold, which is calculated as follows: (a) First, calculate the average value (μs) and standard deviation (σs) of the sensor response value (number of seconds) of the human presence sensor 3 from 2:00 to 3:30 (a time when the person is likely asleep) for the past seven days on which the person is being monitored. (b) Set the first sleep threshold (judgment condition for human detection) as follows: (b1) The first person sensing sleep threshold is the minimum value of (μs+σs) calculated for each 30-minute divided time period. In the above calculation, if the value is 0, the first person sensing sleep threshold is set to 1. (c) Set the second human detection sleeping threshold (human detection judgment condition) as follows: (c1) The second person detection sleep threshold is the minimum value of (μs+3σs) calculated for each 30-minute divided time period. In the above calculation, if the value is 0, the second person detection sleep threshold is set to 1. (d) If (sensor response value of human sensor 3) < or = (first human detection sleep threshold), it is assumed that the person has fallen asleep. If (sensor response value of human sensor 3) > or = (second human detection sleep threshold) is met within one hour from the time when it was assumed that the person had fallen asleep, the judgment that the person had fallen asleep is canceled. The earliest time among the estimated sleep time candidates that satisfy this condition is assumed to be the estimated sleep time.

[0046] Next, a first embodiment of a monitoring system that estimates an estimated lifestyle time using the output of the human presence sensor 3 and the amount of power consumption by the power meter 30 as input data will be described. This monitoring system can be realized by slightly modifying the monitoring system that estimates the estimated lifestyle time using only the output of the human presence sensor 3 described above. In this monitoring system, unlike the monitoring system of the previous embodiment, the power consumption from the power meter 30 is also calculated by the reaction value calculation unit 52. That is, the reaction value calculation unit 52 also calculates a power statistical calculation value, which is a statistical calculation value of the power consumption during a lifestyle time period. Furthermore, the judgment condition determination unit 53 not only determines a human presence judgment condition for calculating a first estimated lifestyle time, which is an estimated time of a lifestyle mode during a lifestyle time period, using the human presence statistical calculation value during the lifestyle time period, but also determines a power judgment condition for calculating a second estimated lifestyle time, which is an estimated time of a lifestyle mode during a lifestyle time period, using the power statistical calculation value during the lifestyle time period. The lifestyle mode estimated time calculation unit 54 calculates the lifestyle mode estimated time based on the human detection determination conditions and the power determination conditions. For example, depending on the lifestyle mode, such as a meal mode (breakfast, lunch, dinner) or a sleep mode (bedtime, wake-up), the lifestyle mode estimated time calculation unit 54 may use either the lifestyle mode estimated time calculated based on the human detection determination conditions or the lifestyle mode estimated time based on the power determination conditions. Alternatively, the final lifestyle mode estimated time may be determined by a weighted average of both.

[0047] Next, a second embodiment of a monitoring system that estimates a lifestyle estimation time using the output of the human presence sensor 3 and the amount of power consumption measured by the power meter 30 as input data will be described. In this monitoring system, the output of the human presence sensor 3 and the amount of power consumption measured by the power meter 30 are used to estimate the lifestyle estimation time. FIG. 7 shows the data processing unit 5 of this monitoring system. This data processing unit 5 has substantially the same configuration as the data processing unit 5 in the previous embodiment shown in FIG. 2, but has additional calculation functions in the reaction value calculation unit 52 and the judgment condition determination unit 53. This monitoring system is particularly suitable for estimating bedtime as a lifestyle behavior.

[0048] In this embodiment, the response value calculation unit 52 not only processes the output (human presence output) from the human presence sensor 3 to calculate the sensor response value, which is the sensor response time width, but also has the function of acquiring data regarding power consumption from the power meter 30 and calculating statistical calculation values ​​regarding power consumption as the sensor response value, such as the average value and standard deviation over a specified period of time.

[0049] The judgment condition determination unit 53 in this embodiment includes a first judgment condition determination unit 531, a second judgment condition determination unit 532, and a third judgment condition determination unit 533. The first judgment condition determination unit 531 determines the first judgment condition using a first occupancy statistically calculated value (e.g., average value, standard deviation) that is a first statistically calculated value of the lifestyle time zone reaction value in the lifestyle time zone. The second judgment condition determination unit 532 determines the second judgment condition using a second occupancy statistically calculated value (e.g., average value, standard deviation) that is a second statistically calculated value of the lifestyle time zone reaction value in the lifestyle time zone. The third judgment condition determination unit 533 determines the third judgment condition using a power statistically calculated value (e.g., average value) that is a statistically calculated value of power consumption in the lifestyle time zone. Therefore, the lifestyle aspect estimated time calculation unit 54 calculates an estimated lifestyle aspect time that is an estimated time of the lifestyle aspect in the lifestyle time zone based on the first judgment condition, the second judgment condition, and the third judgment condition.

[0050] The first determination condition determined by the first determination condition determination unit 531 is the first person-sensing sleep threshold described above. The second determination condition determined by the second determination condition determination unit 532 is the second person-sensing sleep threshold described above. The third determination condition determined by the third determination condition determination unit 533 is the power sleep threshold (power determination condition). To calculate this power sleep threshold, first, the average and standard deviation of power consumption from 2:00 to 3:30 (a time period when the person is likely to be asleep) for the past seven days on the day to be monitored are calculated. The power sleep threshold is set to the average power consumption for the past three hours, or the average power consumption for the past week plus its standard deviation.

[0051] When selecting candidates for estimated bedtime, If (power consumption) < or = (power sleep threshold) and (sensor response value of human sensor 3) < or = (first human sleep threshold), it is assumed that the person has fallen asleep. If (sensor response value of human sensor 3) > or = (second human sleep threshold) is met within one hour from the time when it was assumed that the person had fallen asleep, the judgment that the person had fallen asleep is canceled. The earliest time among the estimated sleep time candidates that satisfy this condition is assumed to be the estimated sleep time.

[0052] A specific example of sleep determination is described below. First, under the assumption that the majority of people are asleep between 2:00 and 4:00, the sensor response value and power consumption for each divided time period between 2:00 and 4:00 over the past week are obtained for each residence 2, and the average value and standard deviation are calculated as statistically calculated values ​​for each divided time period over the past week, with the minimum value of (average value + standard deviation) being set as the threshold. The sleep determination time is set to between 9:00 PM and 3:00 AM, and if the sensor response value and power consumption fall below their respective thresholds during this time period, it is determined that the person is asleep. For example, if the average value + standard deviation of power consumption over the past week in the first divided time period from 2:00 to 2:30 (first divided time power statistical calculation value) is 0.19, the second divided time power statistical calculation value in the second divided time period from 2:00 to 2:30 is 0.22, and similarly the third divided time power statistical calculation value is 0.15, the fourth divided time power statistical calculation value is 0.17, and the fifth divided time power statistical calculation value is 0.15, then the smallest value of the fifth divided time power statistical calculation value, 0.15, is rounded up to 0.2 and set as the power threshold. However, if all values ​​are zero, the power threshold will be 0.1. Furthermore, if the average value of the sensor response value for the past week in the first divided time from 2:00 to 2:30 (first divided time human presence statistical calculation value) is 2.31, the second divided time human presence statistical calculation value in the second divided time from 2:00 to 2:30 is 1.78, and similarly, the third divided time human presence statistical calculation value is 0.49, the fourth divided time human presence statistical calculation value is 4.64, and the fifth divided time human presence statistical calculation value is 5.42, the smallest value of the third divided time power statistical calculation value, 0.49, is rounded up to set the human presence threshold to 1. However, if all values ​​are zero, the human presence threshold is set to 1. Note that the value obtained by rounding up the average value of the human presence response value + 3 × standard deviation in the time period (divided time) set as the human presence threshold, in the above example, (0.14 + 0.35 * 3 = 1.19), exceeds 2, and can be used as a sleep cancellation condition in the time period from 2:00 to 4:00.

[0053] Next, a practical calculation process for the lifestyle behavior estimated time used in the monitoring system will be described using the flowcharts in Fig. 8 to Fig. 13. In this monitoring system, the human presence sensor 3 is assumed to be placed in the kitchen or a similar location (where cooking and eating take place). Fig. 8 shows the basic logic of this practical lifestyle behavior estimated time calculation process. In Fig. 8, rectangular frames indicate the processes that are actually executed, and rounded rectangular frames indicate assumptions made when formulating the logic.

[0054] First, an assumption is made that "the lifestyle time periods of the person being monitored do not fluctuate significantly from day to day" (#100). Based on this assumption, lifestyle time periods are assumed when each lifestyle pattern (breakfast, lunch, dinner, waking up, bedtime) occurs (#101). A judgment coefficient is set for the lifestyle pattern that involves cooking (#102). Here, the breakfast judgment coefficient = 0.3, and the lunch judgment coefficient = dinner judgment coefficient = 0.4. Each judgment coefficient is calculated in increments of 0.1 through sampling statistical calculations to determine the highest judgment accuracy.

[0055] Next, an assumption is made that "the amount of time the monitored person spends in the kitchen varies from person to person" (#110), and based on this assumption, an adjustment coefficient is set as a coefficient corresponding to the lifestyle behavior determination coefficient described above, using the maximum sensor response value over the past three days, so that it differs for each monitored residence (#111). This adjustment coefficient is multiplied by the determination coefficient set in step #102 for the lifestyle time period assumed in step #101, and the value obtained by this multiplication is set as the lifestyle behavior determination threshold (#120).

[0056] Furthermore, we make the assumption that "if the activity is cooking, the sensor response value (cumulative seconds) in the kitchen should be long" (#130), and based on this assumption, we calculate the estimated time of bedtime based on the change from "response at night" to "no response at night" (#131), and the estimated time of wake-up based on the change from "no response in the morning" to "response in the morning" (#132).

[0057] Additionally, an assumption is made that "even if there is a sensor response in the kitchen, it does not necessarily mean that the activity is cooking" (#140), and based on this assumption, if the sensor response is minimal, the sensor response value is deemed to be zero (#141). Based on this assumption, the assumption of step #130, and the lifestyle behavior determination threshold set in step #120, it is determined whether the sensor response value is equal to or greater than the lifestyle behavior determination threshold, and the cooking pattern is estimated (#150).

[0058] The process of calculating each lifestyle aspect estimated time using the above basic logic will be described below.

[0059] FIG. 9 is a flowchart showing the process for calculating the estimated breakfast time. Here, the breakfast time is assumed to be from 4:00 to 11:00 (excluding 11:00). Using aggregated data acquired from each monitored residence, the maximum sensor response value (seconds) during the breakfast time for the past three days on the monitored day is calculated (this maximum value is designated as BF-max) (#201). Next, a preset breakfast determination coefficient (here, 0.3) is read, and one or more 30-minute (division time) periods during the breakfast time on the monitored day in which the sensor response value (seconds) is equal to or greater than the product of BF-max and the breakfast determination coefficient are calculated as selected time periods (selected division times) (#202). A check is made to determine whether there are any selected time periods (#203). If there are multiple selected time periods (#203, Yes branch), the earliest of those time periods is considered to be the estimated breakfast time (#204). If there is only one selected time period (#203, Yes branch), that time (departure time) is considered to be the estimated breakfast time (#204). If there is no selected time slot (No branch #203), a check is further made to see if the maximum value of all divided times is zero (#205). If the maximum value is zero (Yes branch #205), it is determined that the estimated breakfast time cannot be determined (#206). If the maximum value is not zero (No branch #205), the time of the divided time with the maximum value (departure time) is considered to be the estimated breakfast time (#207).

[0060] FIG. 10 is a flowchart showing the process for calculating the estimated lunch time. Here, the lunch time period is defined as 11:00 to 14:00 (excluding 14:00 exactly). Using the aggregated data acquired from each monitored residence, the maximum sensor response value (seconds) during the lunch time period for the past three days on the monitored day is calculated (this maximum value is designated as LU-max) (#301). Next, a preset lunch determination coefficient (here, 0.4) is read, and one or more 30-minute (divided time periods) during the lunch time period on the monitored day are calculated as selected time periods in which the sensor response value (seconds) is equal to or greater than the product of LU-max and the breakfast determination coefficient (#302). Furthermore, the earliest selected time period among the calculated selected time periods is determined as the first lunch time candidate, and the time period with the maximum value is determined as the second lunch time candidate (#303). In step #303, a check is made to see if the first and second lunch time candidates have been calculated (#304). If neither the first nor the second lunch time candidate is found in the check of #304, it is determined that the estimated lunch time cannot be determined (#305). If both the first and second lunch time candidates are found in the check of #304, a further check is made to see if the difference between the first and second lunch time candidates exceeds 60 minutes (#306). If the difference between the first and second lunch time candidates exceeds 60 minutes (#306 Yes branch), the first lunch time candidate is set as the estimated lunch time (#307). Note that the check of #304 does not result in only the second lunch time candidate, so in this case, the first lunch time candidate (= second lunch time candidate) is set. If the difference between the first and second lunch time candidates is 60 minutes or less (#306 No branch), the time in the selected time period with the largest value is given priority, and the second lunch time candidate is set as the estimated lunch time (#308).

[0061] FIG. 11 is a flowchart showing the process for calculating an estimated dinner time. Here, the dinner time period is defined as 4:00 PM to 8:00 PM (excluding 8:00 PM). This flowchart for calculating the estimated dinner time is essentially the same as the flowchart for calculating the estimated lunch time shown in FIG. 10. Using aggregated data acquired from each monitored residence, the maximum value of the sensor response value (seconds) during the dinner time period for the past three days on the monitored day is determined (this maximum value is designated as DI-max) (#401). Next, a preset dinner determination coefficient (0.4 in this case) is read, and one or more sub-periods are determined as selected time periods in which the sensor response value (seconds) for a 30-minute period during the dinner time period on the monitored day is equal to or greater than the product of DI-max and the dinner determination coefficient (#402). Furthermore, the earliest selected time period among the determined selected time periods is designated as the first dinner time candidate, and the selected time period with the maximum value is designated as the second dinner time candidate (#403). In step #403, a check is made to see if the first dinner time candidate and the second dinner time candidate have been obtained (#404). If neither the first nor the second dinner time candidate has been obtained in the check in #404, it is determined that the estimated dinner time cannot be determined (#405). If both the first and second dinner time candidates have been obtained in the check in #404, a further check is made to see if the difference between the first and second dinner time candidates exceeds 60 minutes (#406). If the difference between the first and second dinner time candidates exceeds 60 minutes (#406 Yes branch), the first dinner time candidate is determined as the estimated dinner time (#407). Note that the check in #404 does not necessarily result in only the second dinner time candidate, so in this case, the first dinner time candidate (= second dinner time candidate) is obtained. If the difference between the first and second dinner time candidates is 60 minutes or less (No branch in #406), the time in the selected time period with the maximum value is given priority, and the second dinner time candidate is set as the estimated dinner time (#408).

[0062] 12 is a flowchart showing the calculation process of the estimated wake-up time. Here, the wake-up time range is set to 3:30 to 11:30 (exact 11:30 is not included).

[0063] In the process of calculating the estimated wake-up time, first, one or more sub-periods of the wake-up time zone on the monitoring day are selected as selected time zones (#501), in which the sensor response value (seconds) in the sub-period immediately preceding the wake-up time zone is less than 60 and the sensor response value in each sub-period of the wake-up time zone is 60 or greater. Next, a first determination process is performed (#502) to determine whether the sensor response value in the sub-period immediately preceding the wake-up time zone is less than 60 and whether the sensor response value in the first sub-period of the wake-up time zone (i.e., 3:30 to 4:00) is 60 or greater. This first determination process (wake-up determination process) is intended to maximize the reliability of the wake-up time, assuming that some residences wake up around 4:00 and others go to sleep around that time. If this first determination process is "true" (#502 True branch), 3:30 is set as the tentative estimated wake-up time (#503). If this first determination is "false" (#502 False branch), a second determination is made to see if the maximum sensor response value in the wake-up time zone is zero (#504). If this second determination (wake-up determination) is "true" (#504 True branch), it is determined that the estimated wake-up time cannot be determined (#505). If this second determination is "false" (#504 False branch), the time (start time) of the earliest divided time among the divided time groups remaining in the previous determinations is set as the tentative wake-up time (#506). Next, if the estimated breakfast time was calculated prior to this process, this estimated breakfast time is compared with the tentative wake-up time calculated in step #506 (#507). If this comparison shows that the tentative wake-up time is later than the estimated breakfast time (#507 Tentative wake-up time later branch), the estimated breakfast time is set as the estimated wake-up time (#508). If the estimated breakfast time has not been calculated prior to this process, the tentative wake-up time of 3:30 set in step #503 is regarded as the estimated wake-up time, and this process ends.

[0064] FIG. 13 is a flowchart showing the process for calculating the estimated bedtime. Here, the bedtime period is assumed to be from 8:00 PM to 4:30 AM (excluding 4:30 AM). First, preprocessing is performed to consider sensor response values ​​(seconds) less than 60 during the bedtime period on the monitored day as zero (#601). Next, a determination process (sleep determination process) is performed to determine one or more subperiods of the bedtime period on the monitored day in which the sensor response value is zero and the sensor response value of the subperiod of the bedtime period is less than the sensor response value of the subperiod immediately preceding the bedtime period (#602). A check is made to determine whether or not a selected time period exists (#603). If a selected time period exists (#603: Yes), the latest selected time period is considered to be the estimated bedtime (#604). If no selected time period exists (#603: No), a further determination is made the next day to determine whether or not the estimated wake-up time for the next day has been calculated (#605). This determination is made so that even if the estimated time of going to bed cannot be calculated, if the estimated time of waking up the next day can be calculated, it can be assumed that the person went to bed at some point, so for the time being, it is determined that the person went to bed at midnight. If the result of this determination is "false" (#502 False branch), it is assumed that the estimated time of going to bed cannot be determined (#606), but if the result is "true" (#605 True branch), the estimated time of going to bed is set to midnight (#607).

[0065] [Another embodiment] (1) The functional units shown in the functional blocks of Figures 2 and 7 can be freely divided into multiple functional units, multiple functional units can be combined into one functional unit, or some functional units can be distributed to other computers or processing units. (2) The human presence sensor 3 may be a human presence sensor mounted on a gas alarm, or may be a human presence sensor that only has the function of detecting a person. (3) The message created by the message creation unit 57 for registered users includes information indicating whether the time calculated by the estimated lifestyle time calculation unit 54 is within the appropriate lifestyle time range of the person being monitored who has been registered in advance, and to what extent it deviates.

[0066] The configurations disclosed in the above embodiments (including other embodiments, the same applies below) can be applied in combination with configurations disclosed in other embodiments, as long as no contradiction arises. Furthermore, the embodiments disclosed in this specification are examples, and the embodiments of the present invention are not limited to these, and can be modified as appropriate within the scope that does not deviate from the purpose of the present invention. [Industrial Applicability]

[0067] The monitoring system of the present invention can be adapted to monitoring systems for various types of monitoring targets. [Explanation of symbols]

[0068] 1: Management computer 2: Residence 3: Human presence sensor 4: User terminal 5: Data processing unit 11: Data transmission / reception unit 12: Data storage unit 20: Terminal device 22: Detection signal receiving unit 23: Signal processing section 24: Communications Department 30: Power meter 51: Daily time zone setting section 52: Response value calculation unit 52a: Statistical calculation section 53: Judgment condition determination unit 54: Lifestyle estimated time calculation unit 55: Living Conditions Assessment Department 56: Healthy time registration section 57: Registered user correspondence creation section 531:First judgment condition determination section 532:Second judgment condition determination section 533:Third judgment condition determination section

Claims

1. A monitoring system, A human presence sensor placed in the residence to be monitored; a lifestyle time zone setting unit that sets lifestyle time zones for various lifestyles in one day; a reaction value calculation unit that calculates a daily time zone reaction value from the output of the human presence sensor for each divided time period obtained by dividing the daily time zone by a predetermined number; a determination condition determination unit that determines a determination condition for calculating a behavior estimation time, which is an estimated time of the lifestyle behavior in the lifestyle time zone, using the series of lifestyle time zone reaction values; a lifestyle behavior estimated time calculation unit that obtains a specific lifestyle behavior and an estimated time when the specific lifestyle behavior was performed by determining the lifestyle time zone reaction value using the determination conditions, and calculates the estimated time as the behavior estimated time.

2. The monitoring system of claim 1, further comprising a life status evaluation unit that evaluates the living status of the person being monitored by comparing the estimated life status time calculated by the estimated life status time calculation unit with a pre-registered registered healthy time.

3. The monitoring system according to claim 1 or 2, wherein the daily time zone response value is a statistically calculated value of the sensor output of the human presence sensor during the divided time period.

4. The monitoring system according to claim 3 , wherein the statistically calculated value of the sensor output of the human presence sensor is an integrated value of the time during which the sensor output exceeds a sensor threshold value.

5. A monitoring system as described in claim 3 or 4, wherein a judgment threshold derived from the maximum value of the lifestyle time zone reaction value in the lifestyle time zone and a lifestyle mode judgment coefficient set according to the type of lifestyle mode is used as the threshold of the judgment condition.

6. The monitoring system according to claim 5 , wherein the earliest time among times at which the daily time zone reaction value equal to or greater than the determination threshold occurs is set as the behavior estimation time.

7. The monitoring system according to claim 1 , wherein the lifestyle is at least one of breakfast, lunch, and dinner.

8. The monitoring system of any one of claims 5 to 7, wherein the lifestyle is breakfast, the lifestyle time period is from around 4:00 to around 10:30, the divided time period is 30 minutes, and the lifestyle judgment condition for the lifestyle is that the estimated breakfast time candidate is the time at which the lifestyle time period reaction value is equal to or greater than the value obtained by multiplying the maximum value of the lifestyle time period reaction value by a lifestyle judgment coefficient of 0.2 to 0.4, preferably 0.3, and further the earliest of the estimated breakfast time candidates is the estimated breakfast time.

9. The lifestyle is lunch, the lifestyle time period is from around 11:00 to around 13:30, the divided time is 30 minutes, and the lifestyle judgment condition for the lifestyle is that the time at which the lifestyle time period reaction value is issued that is equal to or greater than the value obtained by multiplying the maximum value of the lifestyle time period reaction value by a lifestyle judgment coefficient of 0.3 to 0.5, preferably 0.4, is set as the estimated lunch time candidate, and further the earliest of the estimated lunch time candidates is set as the estimated lunch time.

10. The lifestyle is dinner, the lifestyle time period is from around 4:00 p.m. to around 7:30 p.m., the divided time is 30 minutes, and the lifestyle judgment condition for the lifestyle is that the time at which the lifestyle time period reaction value is issued is equal to or greater than the value obtained by multiplying the maximum value of the lifestyle time period reaction value by a lifestyle judgment coefficient of 0.3 to 0.5, preferably 0.4, is set as an estimated dinner time candidate, and further the earliest of the estimated dinner time candidates is set as the estimated dinner time.

11. The monitoring system described in any one of claims 5, 6, 8 to 10, wherein the set lifestyle pattern determination coefficient is changed by a coefficient calculated by the difference between the maximum value of the lifestyle time zone reaction value over the last few days and the statistically calculated value of the lifestyle time zone reaction value other than the divided time corresponding to the maximum value.

12. The monitoring system of any one of claims 1 to 4, wherein the human presence sensor is positioned in a place where meals and cooking are performed, the lifestyle is waking up, the wake-up time as the lifestyle time zone is from around 3:00 to around 11:30, and if a wake-up determination process is established in which the sensor response value of the human presence sensor in the divided time immediately before the wake-up time zone is less than a predetermined value and the sensor response value in each of the divided times of the wake-up time zone is greater than or equal to the predetermined value, then a tentative wake-up time is set to before 3:00, and if the wake-up determination process is not established, then the earliest time among the divided times of the wake-up time zone is set to the tentative wake-up time if the maximum value of the sensor response value in the divided time of the wake-up time zone is not zero, and if the tentative wake-up time is earlier than a previously calculated estimated breakfast time, then the estimated breakfast time is set to the estimated wake-up time.

13. The monitoring system of any one of claims 1 to 4 and 12, wherein the human presence sensor is placed in a place where meals and cooking are performed, the lifestyle is sleeping, the bedtime period as the lifestyle time period is from around 8:00 p.m. to around 4:30 a.m., and if a sleep determination process is established in which the sensor response value of the human presence sensor at each divided time period of the bedtime period is zero and there is a divided time period in which the sensor response value is less than the sensor response value at the divided time period immediately preceding the bedtime period, the latest divided time period is set as the estimated bedtime time, and if the sleep determination process is not established, the estimated bedtime time is set to 0:00, under the condition that an estimated wake-up time for the next day is calculated.

14. A monitoring system described in any one of claims 1 to 11, wherein the set daily time zone is a time zone including a range before and after the time when the peak of the daily time zone reaction value was obtained in the last few days, and the daily time zone is changed based on this change range.

15. A monitoring system, A human presence sensor placed in the residence to be monitored; a power meter for acquiring the amount of power consumed by the monitoring target residence; a lifestyle time zone setting unit that sets lifestyle time zones for various lifestyle aspects in one day; a reaction value calculation unit that calculates a living time zone reaction value from the output of the human presence sensor for each divided time obtained by dividing the living time zone by a predetermined number, calculates a human presence statistical calculation value that is a statistical calculation value of the living time zone reaction value for the living time zone, and calculates a power statistical calculation value that is a statistical calculation value of the power consumption for the living time zone; a judgment condition determination unit that determines a human presence judgment condition for calculating a first behavior estimated time, which is an estimated time of the lifestyle mode in the lifestyle time zone, using the human presence statistical calculation value in the lifestyle time zone, and determines a power judgment condition for calculating a second behavior estimated time, which is an estimated time of the lifestyle mode in the lifestyle time zone, using the power statistical calculation value in the lifestyle time zone; a lifestyle aspect estimated time calculation unit that determines the human presence statistical calculation value according to the human presence determination condition to obtain a specific lifestyle aspect and an estimated time when the specific lifestyle aspect was performed, and calculates the estimated time as the first lifestyle aspect estimated time, and that determines the power statistical calculation value according to the power determination condition to obtain a specific lifestyle aspect and an estimated time when the specific lifestyle aspect was performed, and calculates the estimated time as the second lifestyle aspect estimated time, and calculates the lifestyle estimated time from the first lifestyle aspect estimated time and the second lifestyle aspect estimated time.

16. A monitoring system, A human presence sensor placed in the residence to be monitored; a power meter for acquiring the amount of power consumed by the monitoring target residence; a lifestyle time zone setting unit that sets lifestyle time zones for various lifestyle aspects in one day; a reaction value calculation unit that calculates a daily time zone reaction value from the output of the human presence sensor for each divided time period obtained by dividing the daily time zone by a predetermined number; a first determination condition determination unit that determines a first determination condition using a first human sensation statistically calculated value that is a first statistically calculated value of the lifestyle time zone reaction value in the lifestyle time zone; a second determination condition determination unit that determines a second determination condition using a second human sensation statistically calculated value that is a second statistically calculated value of the lifestyle time zone reaction value in the lifestyle time zone; a third determination condition determination unit that determines a third determination condition using a statistically calculated power value that is a statistically calculated value of the power consumption during the living time period; a lifestyle aspect estimated time calculation unit that determines the first-person sense statistical calculation value according to the first determination condition to obtain a specific lifestyle aspect and an estimated time when the specific lifestyle aspect was performed, and calculates the estimated time as a first lifestyle estimated time; determines the second-person sense statistical calculation value according to the second determination condition to obtain a specific lifestyle aspect and an estimated time when the specific lifestyle aspect was performed, and calculates the estimated time as a second lifestyle estimated time; determines the power statistical calculation value according to the third determination condition to obtain a specific lifestyle aspect and an estimated time when the specific lifestyle aspect was performed, and calculates the estimated time as a third lifestyle estimated time; and calculates a lifestyle estimated time that is an estimated time of the lifestyle aspect in the lifestyle time zone from the first lifestyle estimated time, the second lifestyle estimated time, and the third lifestyle estimated time.

Images