Behavior determination apparatus, method for determining behavior, and behavior determination program

Abstract

To improve determination accuracy in a technique of determining sauna bathing and resting.SOLUTION: A behavior determination apparatus 2 in which a behavior determination program 41 is incorporated acquires heart rate time-series data from a wearable terminal 1 and calculates METs time-series data from the acquired heart rate time-series data. The behavior determination apparatus 2 compares the METs time-series data with a METs threshold, determines a time during which METs exceed the METs threshold as a sauna execution time and a rest non-execution time, and determines a time during which METs are equal to or less than the METs threshold as a sauna non-execution time and a rest execution time.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The technical field disclosed in this specification relates to a behavior determination device, a behavior determination method, and a behavior determination program for determining the behavior of a sauna user. [Background technology]

[0002] By repeatedly performing a set of sauna bathing, which promotes autonomic nervous activity, followed by a cold water bath and an open-air bath (hereinafter referred to as "resting"), sauna users can achieve a sense of relaxation and mental clarity after the sauna session, in other words, a sense of "totono." It is known that this "totono" feeling can be achieved by repeatedly performing sauna bathing and resting in accordance with the user's physical condition and sauna experience. To achieve a better "totono" feeling in the sauna, sauna users have needs such as being able to see their state during the sauna, manage their sauna time, and view their sauna history later.

[0003] For example, Patent Document 1 discloses a technology that uses a wearable device worn by a sauna user to measure the sauna user's heart rate over time, and identifies and records sauna bathing periods, cold bath bathing periods, and rest periods based on the amount of change in the heart rate. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2023-160344 Summary of the Invention [Problem to be solved by the invention]

[0005] Changes in heart rate vary from person to person. For example, the baseline of a resting heart rate differs from person to person. For example, heavy sauna users who use saunas frequently tend to have better cardiopulmonary function and a lower increase in heart rate after starting sauna bathing compared to light sauna users who use saunas infrequently. For example, women tend to have more subcutaneous fat than men, and their heart rate tends to increase less after starting sauna bathing. For example, the younger the person, the faster their heart rate tends to increase. Thus, even if sauna bathing periods and rest periods are determined based on changes in heart rate, which vary from person to person, the accuracy of the determination may be low, and the technology of Patent Document 1 leaves room for improvement. [Means for solving the problem]

[0006] The behavior determination device, which has been made to solve the above-mentioned problems, is a behavior determination device equipped with a computer, and is configured to execute the following steps: a heart rate acquisition process for acquiring heart rate time series data or pulse time series data that indicates, in time series, the heart rate or pulse rate per unit time of a sauna user who uses the sauna with at least a sauna bath and a rest period as one set; a conversion process for converting the heart rate time series data or the pulse time series data acquired in the heart rate acquisition process into MET time series data; and a determination process for comparing the METs included in the MET time series data converted in the conversion process with a MET threshold value, and determining that a sauna bath is being taken if the METs are greater than the MET threshold value, and determining that the rest period is being taken if the METs are equal to or less than the MET threshold value.

[0007] A behavior determination device having such a configuration converts the sauna user's heart rate time series data or pulse rate time series data into MET time series data and determines whether the sauna user is taking a sauna bath or resting based on the MET time series data. The MET time series data reflects the sauna user's age. Therefore, a behavior determination device having the above configuration can determine whether the sauna user is taking a sauna bath or resting using standardized values, which is expected to improve the accuracy of the determination compared to determination using the heart rate time series data or pulse rate time series data itself, which still contains individual differences.

[0008] Another aspect of the behavior determination device is a behavior determination device equipped with a computer, wherein the computer executes a heart rate acquisition process to acquire heart rate time series data or pulse rate time series data that indicates, in time series, the heart rate or pulse rate per unit time of a sauna user who uses the sauna as a set of at least a sauna bath and a rest, and an acceleration acquisition process to acquire acceleration time series data that indicates the sauna user's movements; further, the computer executes a determination process to determine that the sauna user is taking the sauna bath if a determination condition is met, and to determine that the sauna user is taking the rest if the determination condition is not met, wherein the determination condition is that the heart rate time series data or the pulse rate time series data acquired in the heart rate acquisition process is on an increasing trend, and the acceleration time series data acquired in the acceleration acquisition process exceeds a threshold.

[0009] A behavior determination device having such a configuration determines that a sauna user is taking a sauna bath when the sauna user's heart rate time series data or pulse rate time series data is on an increasing trend and the acceleration time series data exceeds a threshold, satisfying a determination condition. Conversely, it determines that a sauna user is taking a rest when the determination condition is not satisfied. Heart rate or pulse rate tends to increase after starting a sauna bath. Therefore, a behavior determination device having the above configuration is expected to improve determination accuracy by filtering behavior detected based on acceleration time series data for the increasing trend of heart rate or pulse rate.

[0010] A method and a program for realizing the functions of the behavior determination device, and a computer-readable storage medium storing the program are also novel and useful. [Effects of the Invention]

[0011] According to the technology disclosed in this specification, it is possible to improve the accuracy of the behavior determination device that determines whether a person is taking a sauna bath or taking a rest. [Brief explanation of the drawings]

[0012] [Figure 1]FIG. 1 is a diagram schematically illustrating an example of a behavior determination device. [Figure 2] FIG. 10 is a diagram illustrating an example of a screen transition. [Figure 3] 10 is a flowchart illustrating an example of a control procedure for a registration process. [Figure 4] 10 is a flowchart illustrating an example of a control procedure for a determination recording process. [Figure 5] 10 is a flowchart illustrating an example of a control procedure for a MET determination process. [Figure 6] FIG. 10 is a diagram illustrating an example of data processing when using a wet sauna. [Figure 7] FIG. 10 is a diagram illustrating an example of data processing when using a wet sauna. [Figure 8] FIG. 10 is a diagram illustrating an example of data processing when using a dry sauna. [Figure 9] FIG. 10 is a diagram illustrating an example of data processing when using a dry sauna. [Figure 10] 10 is a flowchart illustrating an example of a control procedure for an acceleration determination process. [Figure 11] 10A and 10B are diagrams illustrating an example of data processing based on triaxial acceleration time series data. [Figure 12] FIG. 10 is a diagram illustrating an example of data processing based on uniaxial acceleration time series data. [Figure 13] FIG. 10 is a diagram illustrating an example of data processing based on uniaxial acceleration time series data. DETAILED DESCRIPTION OF THE INVENTION

[0013] A behavior determination device, a behavior determination method, and a behavior determination program according to this embodiment will be described below with reference to the drawings. This embodiment discloses a behavior determination device that determines sauna behavior.

[0014] In the behavior determination system 100 shown in Fig. 1, a wearable device 1 and a behavior determination device 2 are communicatively connected. The behavior determination device 2 is an example of an "information processing device." In this embodiment, the behavior determination system 100 is configured such that, for example, the heart rate and the like are measured by the wearable device 1 worn by a sauna user who uses sauna facilities or sauna equipment at home, and the behavior determination device 2 acquires the measured data and determines and records whether the user is taking a sauna bath or resting.

[0015] The wearable device 1 is a device that can be worn on the human body, such as on the arm, leg, or neck. The wearable device 1 has at least a communication function and a heart rate measurement function or a pulse rate measurement function. The wearable device 1 has a controller 10 that includes a CPU 11 and a memory 12. The wearable device 1 also has a communication interface (hereinafter referred to as "communication IF") 14, a clock unit 15, a heart rate monitor or pulse rate monitor (hereinafter referred to as "heart rate monitor") 16, and a thermo-hygrometer 17, all of which are connected to the controller 10.

[0016] The behavior determination device 2 is an information processing device having at least a communication function and a storage function. The behavior determination device 2 is, for example, a mobile terminal such as a smartphone or a tablet terminal, or a personal computer (hereinafter referred to as "PC"). The behavior determination device 2 may be owned by the sauna user, or may be loaned to the sauna user by the sauna facility. The behavior determination device 2 may also be installed in a fixed location in the sauna facility. The behavior determination device 2 may also be a server provided by a vendor of the behavior determination program 41.

[0017] The behavior determination device 2 has a controller 20 including a CPU 21 and a memory 22. The behavior determination device 2 also has a user IF 23 and a communication interface (hereinafter referred to as "communication IF") 24, which are connected to the controller 20. The CPU 21 is an example of a "computer."

[0018] The CPUs 11 and 21 execute various processes in accordance with programs read from the memories 12 and 22 and based on user operations. The memories 12 and 22 store various programs and various data. The memories 12 and 22 are also used as work areas when various processes are executed. The buffers provided in the CPUs 11 and 21 are also an example of memory. Note that examples of the memories 12 and 22 are not limited to ROM, RAM, HDD, etc. built into the wearable device 1 or the behavior determination device 2, but may also be storage media readable and writable by the CPUs 11 and 21, such as recording media, for example, CD-ROMs, DVD-ROMs, etc.

[0019] The communication IFs 14 and 24 include hardware for communicating with external devices. The wearable device 1 and the behavior determination device 2 can communicate with each other via the communication IFs 14 and 24. The wearable device 1 and the behavior determination device 2 may be connected to a network such as the Internet and be communicatively connected via the network. The communication standards of the communication IFs 14 and 24 may be wired or wireless. The communication standards of the communication IFs 14 and 24 include, for example, Ethernet (registered trademark), Wi-Fi (registered trademark), USB, Bluetooth (registered trademark), and NFC (Near Field Communication). The wearable device 1 and the behavior determination device 2 may be provided with multiple communication IFs 14 and 24 that are compatible with multiple communication standards.

[0020] The timekeeping unit 15 of the wearable device 1 has a function of measuring time. The timekeeping unit 15 may measure the elapsed time from a certain point in time in response to an instruction to start measurement. The heart rate meter 16 has a function of measuring the heart rate (HR: bpm), which is the number of times the heart beats within a certain period of time (for example, one minute). The heart rate meter 16 may measure raw heart rate signals. The thermo-hygrometer 17 has a function of detecting the temperature and humidity around the wearable device 1.

[0021] The wearable device 1 may have an acceleration sensor 18 connected to the controller 10. The acceleration sensor 18 has a function of detecting acceleration using, for example, three mutually orthogonal axes as detection axes. The acceleration sensor 18 may be a model that can identify the wearing direction, or a model that cannot. The acceleration sensor 18 may take measurements that include gravity, or may take measurements after calibrating the origin and resetting gravity before use.

[0022] The user IF 23 of the behavior determination device 2 is a touch panel having a display function for notifying the user of information and a function for the user to input operations. The user IF 23 may be a combination of a display or the like capable of displaying information and a mouse, keyboard, or the like having an input reception function.

[0023] The memory 22 of the behavior determination device 2 stores various programs including a behavior determination program 41 and various data including a determination result 42. The behavior determination program 41 in this embodiment has a determination recording function for determining whether the user is taking a sauna bath or resting, and recording the determination result. The determination result 42 includes the determination result by the behavior determination program 41. The behavior determination program 41 has a registration function for accepting registration of personal information of sauna users. The behavior determination program 41 has a display function for displaying sauna usage history. Each function will be described later.

[0024] Next, the procedure of the operation executed in the behavior determination system 100 of this embodiment will be described with reference to the drawings. Note that in this embodiment, each process other than user operation basically indicates processing by the CPUs 11 and 21 in accordance with instructions written in a program such as the behavior determination program 41. In this specification, for convenience, various processes performed by the controllers 10 and 20 or the CPUs 11 and 21 in accordance with a program such as the behavior determination program 41 may be described as each program independently performing various processes. The processing by the CPUs 11 and 21 also includes hardware control using the API (application programming interface) of the OS. In this specification, the operation of each program will be described without a detailed description of the OS. Also, the term "acquire" is used as a concept that does not necessarily require a request.

[0025] (Registration function) First, the registration function will be described. When the CPU 21 of the behavior determination device 2 receives an operation to start the behavior determination program 41, it starts the behavior determination program 41 and displays, for example, a menu screen D1 as shown in FIG. 2(A) on the user IF 23. The menu screen D1 displays a registration button SW11, a record button SW12, and a display button SW13. The registration button SW11 is an operator for receiving a registration instruction to register personal information. The record button SW12 is an operator for receiving a record instruction to record a sauna usage history. The display button SW13 is an operator for receiving a display instruction to display a sauna usage history.

[0026] For example, when a sauna user operates the registration button SW11, the CPU 21 accepts the registration instruction and executes the registration process shown in FIG. 3. The CPU 21 first displays a registration screen on the user IF 23 (S11). For example, the registration screen D2 shown in FIG. 2(B) displays a setting area SA21 and an OK button SW21. The setting area SA21 can accept parameter settings for each personal information item. Examples of personal information include gender, age, height, and weight. The setting area SA of the registration screen D2 may include fields for accepting information such as physical condition and mood before bathing, the name of the hot spring facility, and posture during rest.

[0027] 3, the CPU 21 that has displayed the registration screen D2 determines whether or not a registration execution instruction has been received (S12). The CPU 21 waits until the OK button SW21 is operated (S12: NO). When the CPU 21 receives a registration execution instruction in response to the operation of the OK button SW21 (S12: YES), the CPU 21 stores the personal information set in the setting area SA21 in a non-volatile area of ​​the memory 12 (S13), and ends the registration process.

[0028] In this embodiment, the CPU 21 displays the registration screen at any timing in response to the operation of the registration button SW11, and can accept the setting or change of personal information any number of times. The registration process may be automatically executed when the behavior determination program 41 is installed, and the registration of personal information may be accepted.

[0029] The behavior determination program 41 can accept the registration of multiple people and may accept the selection of the person to be registered via the registration screen D2. The behavior determination program 41 may display the registration screen D2 on the condition that the login authentication is successful, and may store the personal information entered on the registration screen D2 in association with the account information (e.g., user identification information, user name, etc.) for which the login authentication was successful. In this case, the behavior determination program 41 can manage the personal information for each sauna user.

[0030] (Sauna behavior) For example, a sauna user puts on wearable device 1 on their arm in a changing room and repeats a set of sauna bathing, cold water bathing, and resting multiple times. After completing the final set, the sauna user returns to the changing room and removes wearable device 1 from their arm.

[0031] The wearable device 1 is initialized before being worn. The wearable device 1 measures the sauna user's heart rate using a heart rate monitor 16. The wearable device 1 also measures the ambient temperature and humidity using a thermo-hygrometer 17. The wearable device 1 associates the measured heart rate, temperature, and humidity with time information (e.g., time of day, elapsed time) measured by a timer 15, and stores them in the memory 12 as heart rate time series data, temperature time series data, and humidity time series data. In other words, the wearable device 1 acquires heart rate time series data that indicates the sauna user's heart rate per unit time in time, and environmental information that includes at least one of the room temperature and humidity when using the sauna.

[0032] The triggers for the wearable terminal 1 to start or end measurement include, for example, when a switch provided on the wearable terminal 1 is operated, when it detects that the wearable terminal 1 has been worn or removed by a sauna user, or when it receives a measurement start signal or measurement end signal from an external device.

[0033] When the wearable terminal 1 includes the acceleration sensor 18, the acceleration measured by the acceleration sensor 18 may be associated with time information measured by the clock unit 15 and stored in the memory 12 as acceleration time-series data.

[0034] (Judgment recording function) Next, the judgment recording function will be described. For example, after using the sauna, a sauna user operates the record button SW12 on the menu screen D1 shown in Fig. 2(A) when recording their sauna behavior in the behavior determination device 2. Then, the CPU 21 of the behavior determination device 2 receives the recording instruction and executes the judgment recording process shown in Fig. 4.

[0035] The CPU 21 acquires time-series data (S21). Specifically, the CPU 21 establishes communication with the wearable device 1 worn by the sauna user using the communication IF 24 and requests the transmission of the time-series data. Upon receiving the request using the communication IF 14, the wearable device 1 reads the time-series data from the memory 12 and transmits it to the requesting behavior determination device 2. Upon receiving the time-series data using the communication IF 24, the CPU 21 of the behavior determination device 2 temporarily stores it in the memory 22. The time-series data includes, for example, heart rate time-series data, temperature time-series data, and humidity time-series data.

[0036] If the wearable device 1 has an acceleration sensor 18, the time series data may include acceleration time-series data. S21 is an example of a "heart rate acquisition process," a "heart rate acquisition step," an "environment information acquisition process," an "acceleration acquisition process," and an "acceleration acquisition step."

[0037] The CPU 21 determines whether acceleration time series data has been acquired (S22). If the time series data acquired in S21 does not include acceleration time series data, the CPU 21 determines that acceleration time series data has not been acquired (S22: NO). In this case, the CPU 21 executes MET determination processing (S23) and proceeds to S24. The MET determination processing is processing for determining whether to take a sauna bath or rest based on the MET time series data converted from the heart rate time series data. In contrast, if the time series data acquired in S21 includes acceleration time series data, the CPU 21 determines that acceleration time series data has been acquired (S22: YES). In this case, the CPU 21 executes acceleration determination processing (S26) and proceeds to S24. The acceleration determination processing is processing for determining whether to take a sauna bath or rest based on the acceleration time series data. The MET determination processing and the acceleration determination processing will be described later.

[0038] The CPU 21 stores the determination result of S23 or S26 in a non-volatile area of ​​the memory 22 in association with the sauna use date and time and the sauna user's user identification information (S24). The CPU 21 may store the determination result 42 in an external storage device such as a server. In this case, the CPU 21 may attach the sauna user's user identification information, user name, and device identification information that identifies the behavior determination device 2 to the determination result 42, and the external storage device may store these in association with each other. This allows the external storage device to determine who made the determination in the determination result 42 and which behavior determination device 2 was used, and makes it possible to provide the determination result 42 in response to a request that specifies the sauna user or the behavior determination device 2.

[0039] Thereafter, the CPU 21 notifies the user that the recording of the sauna usage history has been completed (S25). For example, the CPU 21 causes the user IF 23 to display a notification screen D3 shown in FIG. 2(C). The notification screen D3 includes a completion message and an OK button SW31. When the CPU 21 accepts the operation of the OK button SW31, it ends the determination recording process shown in FIG. 4 and returns to the menu screen D1.

[0040] (METS determination process) An example of the control procedure for the MET determination process will be described with reference to Fig. 5. For example, if the wearable device 1 does not include the acceleration sensor 18 and the behavior determination device 2 cannot acquire acceleration time-series data from the wearable device 1, the MET determination process is executed.

[0041] First, the CPU 21 determines whether the sauna user's personal information has been registered (S31). If the sauna user's personal information has been registered in the memory 22, the CPU 21 determines that the sauna user's personal information has been registered (S31: YES) and proceeds to S32. On the other hand, if the sauna user's personal information has not been registered in the memory 22, the CPU 21 determines that the sauna user's personal information has not been registered (S31: NO) and executes registration processing (S44). The registration processing is the same as in FIG. 3, so its description will be omitted. The CPU 21 proceeds to S32 on the condition that the personal information has been registered.

[0042] The CPU 21 converts the heart rate time series data acquired in S21 into MET time series data (S32). METs is a well-known index that indicates the intensity of physical activity, and indicates how many times the energy consumed when sitting still, with 1 MET being the energy consumed when sitting still. According to the "METs Table for Daily Activities" published by the Ministry of Health, Labor and Welfare, slow walking (less than 3.2 km / h) is 2.0 METs, and normal walking (4.0 km / h) is 3.0 METs.

[0043] In this embodiment, the MET time series data is generated by converting each heart rate included in the heart rate time series data acquired in S21 into METs using the age information received in the registration process in S44 or the registration process shown in FIG. 3 executed in response to the operation of the registration button SW11.

[0044] METs can be calculated, for example, by the relative heart rate reserve %HRR × maximum oxygen uptake VO2max (ml / kg / min) ÷ oxygen uptake at rest.

[0045] Of these, the relative heart rate reserve value %HRR can be calculated by (heart rate - resting heart rate) ÷ heart rate reserve HRR (bpm). The heart rate is the heart rate included in the heart rate time series data acquired in S21, the resting heart rate is the minimum heart rate included in the heart rate time series data acquired in S21, and the heart rate reserve HRR (bpm) can be calculated by the maximum heart rate - the resting heart rate. The maximum heart rate HRmax (bpm) can be calculated by 220 - age. The age is the age accepted in the registration process in S44 or the registration process shown in FIG. 3 executed in response to the operation of the registration button SW11.

[0046] The maximum oxygen intake (VO2max) (ml / kg / min) can be calculated by dividing the maximum heart rate by the resting heart rate x 15.3. The resting heart rate for the maximum heart rate can be calculated as described above. Furthermore, the resting oxygen intake (VO2rest) is a constant, 3.5 ml / kg / min.

[0047] Therefore, the CPU 21 can automatically calculate METs using the minimum heart rate included in the heart rate time-series data and the age accepted in the registration process, and convert the heart rate time-series data into METs time-series data. Note that there are various formulas for calculating METs and maximum oxygen uptake, but the calculation method shown here is just one example.

[0048] The behavior determination program 41 may also reflect the resting posture registered via the registration screen D2 in the MET time-series data. For example, resting may be performed while sitting in a chair, i.e., in a seated position, or while lying in a reclining chair, i.e., in a supine position. It is known that the supine position is less affected by gravity than the seated position, resulting in a lower resting heart rate. Therefore, a weighting coefficient may be applied to the information registered via the registration screen D2 so that the MET time-series data value is lower when the user is in a supine position. In this way, the MET time-series data will be more in line with the actual behavior of sauna users.

[0049] The CPU 21 calculates a moving average of the converted MET time series data to calculate the MET moving average time series data (S33). The MET moving average time series data is an example of "MET time series data." S32 and S33 are examples of "conversion processing" or "conversion step."

[0050] For example, Figures 6 and 7 are diagrams illustrating an example of data processing based on METs when using a wet sauna. HW in Figure 6(A) shows an example of heart rate time series data of a sauna user who performed sauna activities, with one set consisting of a wet sauna bath, a cold water bath, and a break (open air bath). MW in Figure 6(A) shows an example of METs moving average time series data based on the heart rate time series data HW. The horizontal axis of Figure 6(A) represents time, the left vertical axis represents heart rate (bpm), and the right vertical axis represents METs.

[0051] As shown in Figure 6(A), the MET moving average time series data MW and the original heart rate time series data HW show similar trends. However, the heart rate time series data HW is affected by the age and resting heart rate, which vary from sauna user to sauna user, so a threshold must be set for each individual. On the other hand, the MET moving average time series data MW is standardized taking into account age and resting heart rate, so only one threshold needs to be set.

[0052] Therefore, as shown in FIG. 5, the CPU 21 determines whether the sauna user has used a dry sauna or a wet sauna (S34), and sets the MET threshold value according to the determination result.

[0053] In a dry sauna, the temperature is generally maintained between 70°C and 100°C, and the humidity is maintained at 10% or less. In a wet sauna, the temperature is generally maintained between 40°C and 70°C, and the humidity is maintained at 50% or more, by heat radiation from steam or mist. In other words, a wet sauna is used in a hotter and more humid environment than a dry sauna. Therefore, it is possible to automatically distinguish between a dry sauna and a wet sauna based on at least one of the temperature and humidity. In this embodiment, based on at least one of the temperature time series data and humidity time series data acquired in S21, it is automatically determined whether the sauna user has used a wet sauna or a dry sauna.

[0054] For example, CPU 21 determines that the sauna is a wet sauna if the maximum temperature in the temperature time series data is less than a temperature threshold (e.g., 60°C) and the maximum humidity in the humidity time series data is equal to or greater than a humidity threshold (e.g., 50%). On the other hand, CPU 21 determines that the sauna is a dry sauna if the maximum temperature in the temperature time series data is less than the temperature threshold and the maximum humidity in the humidity time series data is not equal to or greater than the humidity threshold. S34 is an example of a "determination process."

[0055] When the CPU 21 determines that the sauna user has used a wet sauna (S34: wet sauna), it acquires the wet threshold value N1 (S35) and proceeds to a determination process (S36). In the determination process, the CPU 21 compares the METs moving average time series data MW with the wet threshold value N1 to determine whether the sauna user is taking a sauna bath or resting. S36 is an example of a "determination process" or a "determination step."

[0056] The sauna flag time series data FS2 in Figure 6(B) shows the results of a determination process performed on the MET moving average time series data MW. The sauna flag time series data FS2, shown by a solid line in Figure 6(B), is data that indicates whether sauna bathing was performed or not in a time series. The rest flag time series data FR2, shown by a dashed line in Figure 6(B), is data that indicates whether rest was performed or not in a time series. The horizontal axis in Figure 6(B) shows time. The left vertical axis shows the sauna determination value of the sauna flag time series data FS2, where a sauna determination value of "0" indicates no sauna bathing and a sauna determination value of "1" indicates sauna bathing. The right vertical axis shows the rest determination value of the rest flag time series data FR2, where a rest determination value of "0" indicates no resting and a rest determination value of "1" indicates resting. Note that the sauna determination value "1" and the rest determination value "1" are examples of "first values," and the sauna determination value "0" and the rest determination value "0" are examples of "second values."

[0057] In addition, because cold water bathing involves an extremely short exposure time, drastic changes in body movement and posture, and some people do not bathe in cold water, cold water bathing was included in sauna bathing, and the behavior of sauna users was separated into sauna bathing and rest.

[0058] The CPU 21 compares the METs moving average time series data MW in Fig. 6(A) with the wet threshold N1. In this embodiment, the wet threshold N1 is set to 2.0 METs, which corresponds to a slow walk, but the value of the wet threshold N1 is not limited to this.

[0059] As shown in section N11 of FIG. 6(A), when the MET moving average time series data MW exceeds the wet threshold N1, the CPU 21 determines that a sauna bath was taken and that a rest period was not missed, and sets the sauna determination value of the sauna flag time series data FS2 to "1" and the rest determination value of the rest flag time series data FR2 to "0" as shown in FIG. 6(B). On the other hand, when the MET moving average time series data MW is equal to or less than the wet threshold N1 as shown in N12 of FIG. 6(A), the CPU 21 determines that a rest period was taken and that a sauna bath was not taken, and sets the sauna determination value of the sauna flag time series data FS2 to "0" and the rest determination value of the rest flag time series data FR2 to "1" as shown in FIG. 6(B). This classifies the sauna user's behavior into sauna bathing and rest.

[0060] The CPU 21 corrects the determination result of S36 (S37), as shown in Fig. 5. For example, the CPU 21 calculates the moving average of each of the sauna determination value and rest determination value shown in Fig. 6(B), compares each moving average with a threshold value, and corrects the sauna implementation time N41 for performing the sauna and the sauna implementation time N52 for performing the rest. In this embodiment, the calculation range for the moving average is 120 seconds before and after (240 seconds in total), but the calculation range is not limited to this embodiment.

[0061] The sauna flag moving average time series data FSs2 shown by the solid line in Figure 7(A) is data that shows, in time series, the sauna flag moving average value obtained by taking the moving average of the sauna judgment values ​​of the sauna flag time series data FS2 shown by the solid line in Figure 6(B). The rest flag moving average time series data FRs2 shown by the dashed line in Figure 7(A) is data that shows, in time series, the rest flag moving average value obtained by taking the moving average of the rest judgment values ​​of the rest flag time series data FR2 shown by the dashed line in Figure 6(B). The horizontal axis of Figure 7(A) represents time, the left vertical axis represents the sauna flag moving average value, and the right vertical axis represents the rest flag moving average value.

[0062] FSr2, shown by a solid line in Figure 7(B), is sauna flag corrected time series data that shows, in time series, the correction results based on the sauna flag moving average time series data FSs2 shown in Figure 7(A). FRr2, shown by a dashed line in Figure 7(B), is rest flag corrected time series data that shows, in time series, the correction results based on the rest flag moving average time series data FRs2 shown in Figure 7(A). The horizontal axis in Figure 7(B) represents time. The left vertical axis represents the sauna determination value of the sauna flag corrected time series data FSr2, with a sauna determination value of "0" indicating no sauna use and a sauna determination value of "1" indicating use of a sauna. The right vertical axis represents the rest determination value of the rest flag corrected time series data FRr2, with a rest determination value of "0" indicating no rest use and a rest determination value of "-1" indicating use of a rest.

[0063] The CPU 21 compares the sauna flag moving average time series data FSs2 with a sauna threshold N91, which is a criterion for determining whether a person is having a sauna bath. The heart rate gradually increases after the sauna bath begins, and then suddenly decreases when the sauna bath ends and the person takes a cold water bath. Therefore, the sauna determination value is likely to fluctuate between the start and end of the sauna bath. Therefore, in this embodiment, the sauna threshold N91 is set to 0.5. Note that the sauna threshold N91 may be different from that in this embodiment.

[0064] As shown in Fig. 7(A), if the sauna flag moving average value of the sauna flag moving average time series data FSs2 is equal to or greater than the sauna threshold N91, the CPU 21 determines that sauna bathing has been performed, and sets the sauna determination value to "1" as shown in Fig. 7(B). Also, as shown in Fig. 7(A), if the sauna flag moving average value of the sauna flag moving average time series data FSs2 is less than the sauna threshold N91, the CPU 21 determines that sauna bathing has not been performed, and sets the sauna determination value to "0" as shown in Fig. 7(B). This divides sauna behavior into sauna bathing time N41, when sauna bathing is performed, and non-sauna bathing time N42, when sauna bathing is not performed.

[0065] The CPU 21 compares the rest flag moving average time series data FRs2 with the rest threshold N92, which is the criterion for determining whether or not a rest has occurred. The heart rate stabilizes around the resting heart rate during rest. Therefore, the rest determination value does not fluctuate much during rest. Therefore, the rest threshold N92 is set to a value smaller than the sauna threshold N91. In this embodiment, the rest threshold N92 is set to 0.2. The rest threshold N92 may be different from that in this embodiment.

[0066] As shown in Fig. 7(A), when the rest flag moving average value of the rest flag moving average time series data FRs2 is equal to or less than the rest threshold value N92, the CPU 21 determines that a rest is not being taken, and sets the rest determination value to "0" as shown in Fig. 7(B). Also, as shown in Fig. 7(A), when the rest flag moving average value of the rest flag moving average time series data FRs2 exceeds the rest threshold value N92, the CPU 21 determines that a rest is being taken, and sets the rest determination value to "-1" as shown in Fig. 7(B). This divides the sauna activity into a rest implementation time N52 when a rest is taken and a non-rest implementation time N51 when a rest is not taken.

[0067] As shown by N44 and N45 in Figure 7(B), in the sauna flag-corrected time series data FSr2 and the rest flag-corrected time series data FRr2, the small fluctuations in sauna bathing and resting shown by N31 in Figure 6(B) are treated as a single continuous sauna bathing and resting, respectively. Furthermore, as shown by N46 and N47 in Figure 7(B), the sauna flag-corrected time series data FSr2 and the rest flag-corrected time series data FRr2 have removed the noise of the sauna bathing and resting shown by N33 in Figure 6(B). S37 is an example of a "moving average process," a "first correction process," and a "second correction process."

[0068] As shown in FIG. 5, the CPU 21 determines whether or not there is an overlapping time between the waveform of the sauna flag corrected time series data FSr2 and the waveform of the rest flag corrected time series data FRr2 (S38).

[0069] The heart rate gradually increases after the sauna bath begins. Therefore, for example, as shown in FIG. 7(B), an overlap time N61 may occur in which the sauna bathing time N41 in the sauna flag corrected time series data FSr2 and the rest time N52 in the rest flag corrected time series data FRr2 overlap. In this case (S38: YES), the CPU 21 regards the overlap time N61 as the sauna bathing time (S43) and proceeds to S39. In this case, the CPU 21 changes the rest determination value for the overlap time N61 from "-1" to "0." S43 is an example of "overlap processing."

[0070] For example, as shown in FIG. 7(B), if there is a break implementation time N71 of a predetermined time or less (e.g., 3 minutes or less) between adjacent sauna implementation times N41, the CPU 21 adjusts the sauna implementation time by regarding the adjacent sauna implementation times N41 as one continuous sauna implementation time (S39). Specifically, the CPU 21 changes the sauna determination value corresponding to the break implementation time N71 from "0" to "1" and changes the break determination value from "-1" to "0." After completing the adjustment of the sauna implementation time, the CPU 21 proceeds to S40 in FIG. 5. S39 is an example of a "third correction process."

[0071] The CPU 21 calculates the determination time series data based on the data processed in S36 to S39 and S43 (S40).

[0072] FIG. 7(C) shows an example of the judgment time series data C2. The horizontal axis indicates time, and the vertical axis indicates the judgment value. A judgment value >0 indicates that sauna bathing was performed but rest was not performed, and a judgment value <0 indicates that sauna bathing was not performed and rest was performed. Furthermore, judgment values ​​"1" and "-1" indicate the first set of sauna bathing and resting. Judgment values ​​"2" and "-2" indicate the second set of sauna bathing and resting.

[0073] For example, the CPU 21 adds the sauna judgment value and the rest judgment value shown in Fig. 7(B) for each time series, thereby setting a judgment value of "1" for the sauna implementation time TS21 during which the sauna is implemented, and setting a judgment value of "-1" for the rest implementation time TS22 during which the rest is implemented.

[0074] The CPU 21 counts the number of sets, where one set is a sauna bath and a rest. Then, it multiplies the calculated judgment value by the corresponding number of sets. As a result, for the first set, the sauna implementation time TS21 is set to a judgment value of "1" and the rest implementation time TS22 is set to a judgment value of "-1." For the second set, the sauna implementation time TS21 is set to a judgment value of "2" and the rest implementation time TS22 is set to a judgment value of "-2."

[0075] There is a time lag between when the heart rate measurement starts and when the sauna user actually enters the sauna room and sits down. Therefore, the CPU 21 sets the judgment value for the first sauna session time TS21 and before to "0."

[0076] As a result, the determination time series data C2 shown in Fig. 7(C) is calculated. After calculating the determination time series data C2, the CPU 21 ends the Mets determination process of Fig. 5 and returns to S24 of Fig. 4.

[0077] In contrast, as shown in FIG. 5, if the CPU 21 determines that the sauna user has used a dry sauna (S34: dry sauna), it acquires a dry threshold M1 (S42) and proceeds to S36. A dry sauna has a higher temperature than a wet sauna and places a greater strain on the circulatory system. Therefore, it is desirable to set the dry threshold M1 to a value greater than the wet threshold N1. In this embodiment, the dry threshold M1 is set to 3.0 METs, which corresponds to normal walking. The dry threshold M1 may be different from this embodiment.

[0078] The CPU 21 performs the processes from S36 onward to determine whether the sauna user is taking a sauna bath or resting, in the same manner as when it is determined that the sauna user has used a wet sauna.

[0079] 8 and 9 are diagrams illustrating an example of data processing based on METs during dry sauna use. HD in FIG. 8(A) shows an example of heart rate time series data of a sauna user who performed sauna activities consisting of a dry sauna bath, a cold water bath, and a rest (open air bath) as one set. MD in FIG. 8(A) shows an example of METs moving average time series data based on the heart rate time series data HD. The horizontal axis of FIG. 8(A) represents time, the left vertical axis represents heart rate (bpm), and the right vertical axis represents METs. FIGS. 8(B), 9(A), 9(B), and 9(C) correspond to FIGS. 6(B), 7(A), 7(B), and 7(C), respectively.

[0080] From S36 onwards, data processing is carried out in the same way as when using a wet sauna. In Figures 8 and 9, the same symbols as in Figures 6 and 7 are used for data processed in the same way as when using a wet sauna, and their explanations will be omitted.

[0081] Because the room temperature in a dry sauna is kept higher than in a wet sauna, the heart rate is less likely to drop suddenly during sauna bathing, as shown in the heart rate time series data HD in Figure 8(A). Therefore, as shown in Figures 8(A) and 8(B), when using a dry sauna, the METs moving average time series data MD, sauna flag time series data FS1, and rest flag time series data FR1 are less likely to fluctuate rapidly compared to when using a wet sauna.

[0082] However, during the first sauna bath, the body is not yet accustomed to the temperature of the dry sauna, so as shown in Figure 8(B) N31, the sauna judgment value and rest judgment value may fluctuate slightly immediately after the first sauna bath begins.

[0083] Even in this case, the CPU 21 compares the sauna flag moving average time series data FSs1 and rest flag moving average time series data FRs1 shown in Fig. 9(A), which are calculated by taking the moving averages of the sauna determination value and rest determination value shown in Fig. 8(B), with the sauna threshold N91 and rest threshold N92, respectively, to calculate the sauna bathing time N41, non-sauna bathing time N42, rest bathing time N52, and non-rest bathing time N51, and corrects the sauna determination value and rest determination value as shown in the sauna flag corrected time series data FSr1 and the rest flag corrected time series data FRr1 in Fig. 9(B). As a result, the small fluctuations shown at N31 in Fig. 8(B) are regarded as one continuous sauna bath, as shown at N44 and N45 in Fig. 9(B).

[0084] Furthermore, the CPU 21 regards the overlap time N61 between the sauna flag corrected time series data FSr1 and the rest flag corrected time series data FRr1 shown in Fig. 9(B) as the sauna implementation time. If there is no rest implementation time of a predetermined time (e.g., 3 minutes) or less between the sauna implementation times N41, as shown in the sauna flag corrected time series data FSr1 and the rest flag corrected time series data FRr1 in Fig. 9(B), the CPU 21 skips S39 in Fig. 5 and executes S40 to calculate the determination time series data C1.

[0085] Therefore, by performing the METs determination process, the behavior determination device 2 can determine whether a sauna bath or a rest period is being performed based on the METs time-series data, whether a dry sauna or a wet sauna is being used, and can calculate the sauna time TS21, the rest time TS22, and the number of sets.

[0086] (Acceleration determination process) Next, the acceleration determination process will be described with reference to Fig. 10. For example, if the wearable device 1 has an acceleration sensor 18 and the behavior determination device 2 can acquire acceleration time-series data from the wearable device 1, the acceleration determination process is executed.

[0087] First, CPU 21 determines whether the wearing state of wearable device 1, i.e., the direction setting of acceleration sensor 18, can be identified (S202). This is because the direction of acceleration depends on the direction setting of acceleration sensor 18. The determination of whether the wearing state can be identified differs depending on, for example, whether wearable device 1 is worn on the trunk of the body, where the orientation of wearable device 1 is the same as the orientation of the body, and whether wearable device 1 is worn on any of the limbs such as the arms or legs, or the head, where the orientation of wearable device 1 may differ from the orientation of the body.

[0088] For example, if the CPU 21 does not acquire acceleration direction setting information from the wearable device 1 when acquiring time-series data in S21 of FIG. 4, the CPU 21 determines that the wearing state of the wearable device 1, i.e., the direction setting of the acceleration sensor 18, cannot be identified (S202: NO). In this case, the CPU 21 acquires triaxial acceleration time-series data (acceleration time-series data on mutually orthogonal x-, y-, and z-axes) from the time-series data acquired in S21 of FIG. 4, and calculates a composite vector in time series based on the acquired triaxial acceleration time-series data (S203). The CPU 21 acquires a composite threshold P1 (S204) and performs a determination process (S205). The acceleration time-series data on the x-, y-, and z-axes are examples of "first acceleration time-series data," "second acceleration time-series data," and "third acceleration time-series data."

[0089] FIG. 11 is a diagram illustrating an example of data processing based on triaxial acceleration when using a wet sauna. The dashed line in FIG. 11(A) indicates heart rate time series data HW, and the solid line indicates composite vector time series data BM. The horizontal axis of FIG. 11(A) indicates time, the left vertical axis indicates heart rate (bpm), and the right vertical axis indicates composite vector acceleration (m / s 2 ) is shown.

[0090] When a sauna user is resting in the same posture, the composite vector BodyMotion is 0 m / s as shown in the composite vector time series data BM in Figure 11(A). 2On the other hand, when the sauna user makes body movements, such as when washing the body, entering and sitting in the sauna room, standing up and leaving the sauna room, lying down to rest, or standing up from a lying position, the composite vector BodyMotion of the composite vector time series data BM fluctuates wildly.

[0091] Therefore, the CPU 21 extracts a peak time P21 at which the composite vector BodyMotion of the composite vector time series data BM exceeds a composite threshold value P1, as shown in Fig. 11(A). In this embodiment, the composite threshold value P1 is 0.75 m / s 2 The composite threshold value P1 can be set in accordance with the sensor characteristics of the acceleration sensor 18, and may be different from that in this embodiment.

[0092] PBM in Figure 11(B) shows the body movement presence / absence time series data PBM indicating whether the sauna user is moving. C3 in Figure 11(B) shows the judgment time series data. The horizontal axis in Figure 11(B) shows time. The left vertical axis in Figure 11(B) shows the body movement presence / absence judgment value of the body movement presence / absence time series data PBM, where a body movement presence / absence judgment value of "1" indicates "body movement present" and a body movement presence / absence judgment value of "0" indicates "body movement absent." The right vertical axis in Figure 11(B) shows the judgment value of the judgment time series data C3. A judgment value > 0 indicates sauna use and no rest, and a judgment value < 0 indicates no sauna use and rest. Furthermore, judgment values ​​"1" and "-1" indicate the first sauna bath and rest. Judgment values ​​"2" and "-2" indicate the second sauna bath and rest.

[0093] 11B, the CPU 21 sets a body motion presence / absence determination value of "1" at the extracted peak time P21. The CPU 21 also sets a body motion presence / absence determination value of "0" at a time P22 when the composite vector BodyMotion of the composite vector time series data BM is equal to or smaller than the composite threshold.

[0094] Body movement during time P11 when the heart rate is increasing is likely to be an action associated with sauna bathing. Therefore, during time P11 when the heart rate is increasing, CPU 21 considers the period from the first peak time P21f to the last peak time P21e as one continuous sauna bathing period P41 and sets the body movement presence / absence determination value to "1." As a result, the body movement presence / absence determination value "1" is set for time P41 when the heart rate is increasing and the composite vector time series data BM exceeds the composite threshold value P1, which satisfies the determination conditions, making it possible to determine that the person is sauna bathing. Furthermore, the body movement presence / absence determination value "0" is set for time periods that do not satisfy these determination conditions, making it possible to determine that the person is resting.

[0095] As shown in Fig. 10, the CPU 21 adjusts the sauna session time P41 (S206). For example, as shown in Fig. 11(B), the CPU 21 changes the body movement presence / absence determination value from "0" to "1" for time P62, during which the composite vector BodyMotion fluctuates, among time P31, for which the body movement presence / absence determination value is set to "0." As a result, the times P41 adjacent to and before time P62, for which the body movement presence / absence determination value is set to "1," are joined together and considered to be one continuous sauna session time.

[0096] For example, when taking a wet sauna bath while bathing the lower half of the body, a slight change in posture can cause a temporary drop in heart rate. Even in such a case, the process in S206 determines that the action is a sauna bath, avoiding the misidentification of the action as a rest.

[0097] As shown in Fig. 10, CPU 21 calculates the determination time-series data (S207). As shown in Fig. 11(B), CPU 21 sets a determination value of "1" at the time when the body movement presence / absence determination value is set to "1", and sets a determination value of "-1" at the time when the body movement presence / absence determination value is set to "0".

[0098] Then, the CPU 21 calculates the number of sets, where one set is the sauna time and the rest time. The CPU 21 multiplies the judgment value by a value corresponding to the number of sets. As a result, the sauna time TS21 of the first set is set to a judgment value of "1," the rest time TS22 of the first set is set to a judgment value of "-1," the sauna time TS21 of the second set is set to a judgment value of "2," and the rest time TS22 of the second set is set to a judgment value of "-2." Therefore, the CPU 21 can also determine the sauna time TS21, the rest time TS22, and the number of sets based on the judgment values ​​of the judgment time-series data C3. The CPU 21 sets the judgment value before the sauna time TS21 of the first set to "0."

[0099] As a result, the determination time series data C3 is calculated. After calculating the determination time series data C3, the CPU 21 ends the acceleration determination process shown in Fig. 10 and returns to S24 in Fig. 4.

[0100] 10, for example, when acquiring time-series data in S21 of FIG. 4, if CPU 21 acquires acceleration direction setting information from wearable device 1, CPU 21 determines that the wearing state of wearable device 1, i.e., the direction setting of acceleration sensor 18, can be identified (S202: YES). In this case, CPU 21 acquires vertical acceleration time-series data with the vertical direction as the detection axis from the triaxial acceleration time-series data acquired in S21 of FIG. 4 (triaxial acceleration time-series data acquired from the acceleration sensor in this embodiment) (S211). This is because sauna bathing and resting movements involve vertical movements, such as entering the sauna room and sitting down, leaving the sauna room and entering the cold bath, and lying down in a rest area.

[0101] The CPU 21 acquires the uniaxial threshold value P2 (S212). After that, the CPU 21 proceeds to the determination process (S205). The vertical acceleration is a value larger than the composite vector. Therefore, the uniaxial threshold value P2 is set to a value higher than the composite threshold value P1. In this embodiment, the uniaxial threshold value P2 is set to 1.0 (m / s 2) The single-axis threshold P2 may be different from that in this embodiment depending on the sensor characteristics, etc. The processing from S205 onwards has been described above, but will be briefly explained below.

[0102] FIG. 12 is a diagram illustrating an example of data processing based on uniaxial acceleration during use of a wet sauna. FIG. 12(A) shows heart rate time series data HW, with the horizontal axis representing time and the vertical axis representing heart rate (bpm). FIG. 12(B) shows vertical acceleration time series data AC, with the horizontal axis representing time and the vertical axis representing vertical acceleration (m / s 2 ) PA in Figure 13(A) shows an example of body movement presence / absence time series data, with the horizontal axis showing time and the vertical axis showing body movement presence / absence judgment value. A body movement presence / absence judgment value of "1" indicates the presence of movement, and a body movement presence / absence judgment value of "0" indicates the absence of movement. C4 in Figure 13(B) shows judgment time series data, with the horizontal axis showing time and the vertical axis showing judgment value. A judgment value > 0 indicates sauna bathing and no resting, and a judgment value < 0 indicates sauna bathing and no resting. Furthermore, judgment values ​​"1" and "-1" indicate the first set of sauna bathing and resting. Judgment values ​​"2" and "-2" indicate the second set of sauna bathing and resting.

[0103] The CPU 21 compares the vertical acceleration time series data AC in Fig. 12(B) with the uniaxial threshold P2, and sets a body movement presence / absence determination value of "1" at peak time P21 when the vertical acceleration time series data AC exceeds the uniaxial threshold P2, as shown in Fig. 13(A), and sets a body movement presence / absence determination value of "0" at time P22 when the vertical acceleration time series data AC is equal to or less than the uniaxial threshold P2, as shown in Fig. 12(B). The CPU 21 connects peak times P21 at time P11 when the heart rate is increasing as shown in Fig. 12(A) to form a single continuous time P41 (S205).

[0104] Furthermore, as shown in Figure 13(A), for time P62 during which the composite vector BodyMotion fluctuates, among time P31 during which the body movement determination value "0" is set, CPU21 changes the body movement determination value "0" to "1" and adjusts the sauna duration (S206).

[0105] The CPU 21 calculates a determination value based on the body movement presence / absence time-series data PA in Fig. 13(A), and calculates determination time-series data C4 that can determine the sauna time TS21, the rest time TS22, and the number of sets, as shown in Fig. 13(B) (S207). Thereafter, the CPU 21 ends the acceleration determination process shown in Fig. 10, and returns to S24 in Fig. 4.

[0106] (Display processing) The display process will be described with reference to Fig. 2. For example, when the display button SW13 on the menu screen D1 shown in Fig. 2(A) is operated, the behavior determination program 41 receives a display instruction and causes the user IF 23 to display the sauna usage history display screen D4 shown in Fig. 2(D) based on the determination result 42.

[0107] The sauna usage history display screen D4 includes, for example, a usage date specification field SA that accepts the specification of sauna usage dates. The sauna usage history display screen D4 may allow the user to specify multiple sauna usage dates by inputting the display period and the number of sauna usage dates that can be displayed starting from the most recent, in response to the operation of button SB. The behavior determination program 41 reads from memory 22 the determination result 42 corresponding to the specified sauna usage date and displays it in the history display field SC of the sauna usage history display screen D4.

[0108] The sauna user can check whether they were able to use the sauna in a way that "refreshed" them by looking at the sauna bath and rest times displayed in the history display field SC. When the behavior determination program 41 receives an operation of the end button SW41 on the sauna usage history display screen D4, it ends the display process.

[0109] (Effectiveness confirmation test) The inventors conducted a test to confirm the accuracy of the judgment based on METs, triaxial acceleration, and vertical acceleration. In the test, a wet sauna was used, which is expected to increase the judgment error due to METs. The subjects repeated two sets of sauna bathing, cold water bathing, and rest (open air bathing). The subjects wore a wearable device 1 with a heart rate measurement function and an acceleration sensor with a triaxial acceleration measurement function. The subjects also provided information on the correct behavior of the start and end times of the sauna bathing, and the start and end times of the resting period.

[0110] The inventors input the heart rate time series data measured by the wearable device 1 and the subject's age into a PC equipped with the behavior determination program 41, and performed a determination based on the MET time series data. Then, the determination result was compared with the correct behavior provided by the subject, and the determination accuracy was calculated.

[0111] As a result, the accuracy of the determination was 93% for the sauna time TS21 of the first set, 92% for the rest time TS22 of the first set, 88% for the sauna time TS21 of the second set, and 83% for the rest time TS22 of the second set. Overall, the accuracy of the determination was 90% for the sauna time TS21 and 87% for the rest time TS22, with regard to sauna bathing and resting.

[0112] The inventors also input the heart rate time series data measured by the wearable device 1 and the triaxial acceleration time series data measured by the acceleration sensor into a PC equipped with the behavior determination program 41, and performed a determination based on the triaxial acceleration time series data. Then, the determination result was compared with the correct behavior provided by the subject, and the determination accuracy was calculated.

[0113] As a result, the accuracy of the determination was 98% for the sauna time TS21 of the first set, 96% for the rest time TS22 of the first set, 99% for the sauna time TS21 of the second set, and 98% for the rest time TS22 of the second set. Overall, the determination of sauna bathing and resting was possible with an accuracy of 97% for the sauna time TS21 and 99% for the rest time TS22.

[0114] The inventors also input the heart rate time series data measured by the wearable device 1 and the vertical acceleration time series data measured by the acceleration sensor into a PC equipped with the behavior determination program 41, and performed a determination based on the vertical acceleration time series data. Then, the determination result was compared with the correct behavior provided by the subject, and the determination accuracy was calculated.

[0115] As a result, the accuracy of the first sauna session (TS21) was 99%, the first rest session (TS22) was 93%, the second sauna session (TS21) was 97%, and the second rest session (TS22) was 92%. Overall, the accuracy of the sauna session and rest session was 92%.

[0116] The results of the above effectiveness confirmation test showed that judgments based on MET time series data, triaxial acceleration time series data, and vertical acceleration time series data can all determine whether a sauna bath or rest is taking place with an accuracy of over 87%.

[0117] Judgments based on acceleration time series data are more accurate than judgments based on MET time series data, and by giving it a higher priority of use than judgments based on MET time series data, it is expected that highly accurate judgment results will be obtained.

[0118] There are cases where a wearable device with a heart rate measurement function does not have a built-in acceleration sensor 18, or where an acceleration sensor 18 is built in but the acceleration time series data cannot be acquired by an external device such as the behavior determination device 2 or a server. In such cases, a determination based on MET time series data is used. The determination based on MET time series data showed an average error of -1.2 minutes for the first sauna session (TS21), an average of 1.6 minutes for the first rest session (TS22), an average of 0.9 minutes for the second sauna session (TS21), and an average of -4.0 minutes for the second rest session (TS22) compared to the correct behavior. However, in reality, the MET time series data includes the time lag between entering the sauna room and sitting down, and the time lag between moving to the changing room after the rest. Therefore, an error of about one minute can be considered within the acceptable range. Therefore, the accuracy of the determination based on MET time series data is not a problem in practical use.

[0119] As described above, the behavior determination device 2 of this embodiment converts the sauna user's heart rate time series data HW, HD into MET moving average time series data MW, MD, and determines whether the sauna user is taking a sauna bath or resting based on the MET moving average time series data MW, MD. The MET moving average time series data MW, MD reflects the sauna user's age and resting heart rate. Therefore, the behavior determination device 2 of this embodiment can determine whether the sauna user is taking a sauna bath or resting by taking into account individual differences in the sauna user's heart rate variability, and is expected to achieve improved determination accuracy compared to determination that uses the heart rate time series data HW, HD directly.

[0120] Furthermore, the behavior determination device 2 of this embodiment determines that the sauna user is taking a sauna bath if the sauna user's heart rate time series data HW is on an increasing trend and the composite vector time series data BM exceeds the composite threshold P1 or the vertical acceleration time series data AC exceeds the uniaxial threshold P2, and determines that the sauna user is taking a rest if the determination conditions are not met. Heart rate tends to increase after starting a sauna bath. Therefore, with the behavior determination device 2 of this embodiment, improved determination accuracy can be expected by filtering the behavior detected based on the composite vector time series data BM or the vertical acceleration time series data AC by the increasing heart rate trend.

[0121] It should be noted that this embodiment is merely an example and does not limit the present invention in any way. Therefore, the present invention can naturally be improved and modified in various ways without departing from the spirit and scope of the present invention. For example, the behavior determination program 41 may use pulse time-series data instead of heart rate time-series data. The wearable device may be equipped with a pulse meter instead of the heart rate meter 16.

[0122] For example, the behavior determination program 41 may upload the determination result 42 to a server in S24 of Fig. 4. Then, the behavior determination device 2 may access the server at any timing and display the determination result 42 on the user IF 23. This prevents the memory 22 of the behavior determination device 2 from being overwhelmed by the determination result 42.

[0123] For example, if the wearable device 1 includes a user interface, the user interface may display the determination result 42. The displayed determination result 42 may be determined by the wearable device 1 itself, or may be determined by an external device such as a server or the behavior determination device 2.

[0124] For example, the behavior determination program 41 may use a value input by the sauna user or a value measured in advance as the resting heart rate to be applied to the calculation of the MET time-series data in S32 of Fig. 5. However, by using the minimum value of the heart rate time-series data measured by the wearable device 1 for that day as the resting heart rate, not only is the physical condition of that day reflected with high accuracy, but the effort of manually inputting and measuring the resting heart rate is also reduced.

[0125] For example, the behavior determination program 41 may omit the processes of S31 and S44 in Fig. 5, and personal information may not be registered. However, by registering personal information in the behavior determination program 41, the sauna user can avoid the trouble of having to input personal information each time they record.

[0126] For example, the behavior determination program 41 may omit the processes of S34, S35, and S42 in Fig. 5 and use a common MET threshold for both dry sauna and wet sauna to determine whether sauna bathing or resting is being performed. However, wet sauna and dry sauna have different loads on the circulatory system. Therefore, by distinguishing between wet sauna and dry sauna and using appropriate wet threshold and dry threshold for each, improved determination accuracy can be expected.

[0127] 5 may distinguish between a dry sauna and a wet sauna in response to a user operation. For example, the behavior determination program 41 may, prior to S34, cause the behavior determination device 2 to display a type selection screen that allows the sauna user to select the type of sauna used, and in S34, distinguish between a dry sauna and a wet sauna based on the type selected on the type selection screen. However, by distinguishing between a dry sauna and a wet sauna based on at least one of the temperature time-series data and the humidity time-series data acquired from the wearable device 1, it is more likely that an appropriate MET threshold will be used, and this can be expected to improve the accuracy of the determination.

[0128] 5 may be omitted. However, by calculating a moving average of the sauna determination value and the moving average of the rest determination value, comparing them with a sauna threshold N91 and a rest threshold N92, and correcting the sauna and rest times determined in the determination process of S36, noise N33 contained in the determination result can be removed. Furthermore, if there is an overlap time N61 where the corrected sauna and rest times N41 and N52 overlap, the overlap time N61 can be regarded as the sauna time. This increases the likelihood of properly detecting sauna bathing, and improves the accuracy of the determination, even when the tendency for heart rate or pulse rate to increase after sauna bathing differs depending on the characteristics of sauna users, such as heavy and light users.

[0129] For example, step S39 in Fig. 5 may be omitted. However, if there is a rest period N62 of a predetermined length or less between adjacent sauna periods N41 after correction, correcting the adjacent sauna periods N41 to one continuous sauna period increases the likelihood of determining that a sauna bath has occurred even if, for example, the heart rate or pulse rate decreases for some reason during sauna bathing, which is expected to improve the accuracy of the determination.

[0130] 10, the composite vector of the triaxial acceleration time series data and the vertical acceleration time series data may be compared with the same threshold value. However, the acceleration of the vertical acceleration time series data will be greater than the acceleration of the composite vector. Therefore, if the wearing state of the wearable device 1 can be identified, the triaxial acceleration time series data and the vertical acceleration time series data may be compared with the composite threshold value P1 and a uniaxial threshold value P2 greater than the composite threshold value P1, respectively, which is expected to improve the accuracy of the determination.

[0131] For example, the judgment time series data C1 and C2 calculated in S40 of Fig. 5 and the judgment time series data C3 and C4 calculated in S207 of Fig. 10 do not have to be values ​​obtained by multiplying the judgment value by the number of sets. However, by using a judgment value obtained by multiplying the judgment value by the number of sets, it is easier to understand the relationship between the number of sets, sauna bathing, and rest.

[0132] For example, S33 in Fig. 5 may be omitted, and the MET time series data before the moving average may be used directly in the determination process of S36. However, by determining whether to sauna bath or rest based on the MET moving average time series data MW, which is the moving average of the MET time series data, erroneous determination due to noise N43 can be avoided, and improved determination accuracy can be expected.

[0133] Furthermore, in any flowchart disclosed in the embodiments, the execution order of multiple processes in any multiple steps can be changed or they can be executed in parallel as long as no contradiction occurs in the processing content.

[0134] The processes disclosed in the embodiments may be executed by hardware such as a single CPU, multiple CPUs, or ASIC, or a combination thereof. The processes disclosed in the embodiments may be realized in various ways, such as a recording medium on which a program for executing the processes is recorded, or a method. [Explanation of symbols]

[0135] 2 Behavior determination device 21 CPU 41 Behavioral Judgment Program

Claims

1. A behavior determination device including a computer, The computer a heart rate acquisition process for acquiring heart rate time series data or pulse rate time series data that indicates the heart rate or pulse rate per unit time of a sauna user who uses the sauna as a set of at least a sauna bath and a rest; a conversion process for converting the heart rate time series data or the pulse rate time series data acquired in the heart rate acquisition process into MET time series data; a determination process of comparing the METs included in the MET time-series data converted by the conversion process with a MET threshold, and determining that the sauna bath is being performed if the METs are greater than the MET threshold, and determining that the rest is being performed if the METs are equal to or less than the MET threshold; To execute The behavior determination device is configured as follows.

2. 2. The behavior determination device according to claim 1, The computer Execute a registration process to register personal information of the sauna user, the personal information including at least age; In the conversion process, converting the heartbeat time series data or the pulse time series data into the MET time series data based on age and resting heart rate or resting pulse rate, wherein the age is the age registered in the registration process, and the resting heart rate or resting pulse rate is the minimum value of the heartbeat time series data or the pulse time series data; The behavior determination device is configured as follows.

3. 2. The behavior determination device according to claim 1, The computer further comprises: Execute a determination process to determine whether the sauna user has used a dry sauna or a wet sauna; In the determination process, If the discrimination process determines that the sauna user has used the wet sauna, the wet threshold value is used as the MET threshold value, and if the discrimination process determines that the sauna user has used the dry sauna, the dry threshold value, which is greater than the wet threshold value, is used as the MET threshold value. The behavior determination device is configured as follows.

4. 4. The behavior determination device according to claim 3, The computer executes an environmental information acquisition process to acquire environmental information including at least one of room temperature and humidity; In the determination process, Distinguishing between the dry sauna and the wet sauna based on the environmental information acquired in the environmental information acquisition process. The behavior determination device is configured as follows.

5. 2. The behavior determination device according to claim 1, In the determination process, calculating sauna flag time series data in which the state of taking a sauna bath is represented by a first value and the state of not taking a sauna bath is represented by a second value; and further calculating rest flag time series data in which the state of taking a break is represented by the first value and the state of not taking a break is represented by the second value; The computer a moving average process for calculating sauna flag moving average time series data by taking a moving average of the sauna flag time series data and rest flag moving average time series data by taking a moving average of the rest flag time series data; a first correction process for comparing the sauna flag moving average value included in the sauna flag moving average time series data calculated by the moving average process with a sauna threshold value, and correcting the sauna implementation time when the sauna is implemented and the sauna non-implementation time when the sauna is not implemented; a second correction process for comparing a break flag moving average value included in the break flag moving average time series data calculated by the moving average process with a break threshold value, and correcting a break implementation time during which the break is implemented and a break non-implementation time during which the break is not implemented; Run Furthermore, the computer If there is an overlapping time between the sauna session time corrected by the first correction process and the break session time corrected by the second correction process, an overlapping process is performed to regard the overlapping time as the sauna session time. The behavior determination device is configured as follows.

6. 2. The behavior determination device according to claim 1, In the determination process, calculating sauna flag time series data in which the state of taking a sauna bath is represented by a first value and the state of not taking a sauna bath is represented by a second value; and further calculating rest flag time series data in which the state of taking a break is represented by the first value and the state of not taking a break is represented by the second value; The computer a moving average process for calculating sauna flag moving average time series data by taking a moving average of the sauna flag time series data and rest flag moving average time series data by taking a moving average of the rest flag time series data; a first correction process for comparing the sauna flag moving average value included in the sauna flag moving average time series data calculated by the moving average process with a sauna threshold value, and correcting the sauna implementation time when the sauna is implemented and the sauna non-implementation time when the sauna is not implemented; a second correction process for comparing a break flag moving average value included in the break flag moving average time series data calculated by the moving average process with a break threshold value, and correcting a break implementation time during which the break is implemented and a break non-implementation time during which the break is not implemented; Run Furthermore, the computer If there is a break time of a predetermined length or less between adjacent sauna times among the sauna times corrected by the first correction process, a third correction process is executed to correct the adjacent sauna times into one continuous sauna time. The behavior determination device is configured as follows.

7. A behavior determination device including a computer, The computer a heart rate acquisition process for acquiring heart rate time series data or pulse rate time series data that indicates the heart rate or pulse rate per unit time of a sauna user who uses the sauna as a set of at least a sauna bath and a rest; an acceleration acquisition process for acquiring acceleration time series data indicating the movements of the sauna user; Run Furthermore, the computer If a determination condition is satisfied, it is determined that the sauna bathing is being performed, and if the determination condition is not satisfied, it is determined that the resting is being performed, and the determination condition is that the heart rate time series data or pulse rate time series data acquired in the heart rate acquisition process is on an increasing trend and the acceleration time series data acquired in the acceleration acquisition process exceeds a threshold. The behavior determination device is configured as follows.

8. 8. The behavior determination device according to claim 7, the acceleration time series data includes first acceleration time series data, second acceleration time series data, and third acceleration time series data based on three axes orthogonal to each other; In the determination process, If the direction setting of the acceleration sensor measuring the acceleration time series data can be identified, data corresponding to the up and down direction from the first acceleration time series data, the second acceleration time series data, and the third acceleration time series data is acquired, and the acquired data is compared with a uniaxial threshold value to determine whether the sauna bath is taking place or whether the rest is taking place; on the other hand, if the direction setting of the acceleration sensor cannot be identified, a composite vector of the first acceleration time series data, the second acceleration time series data, and the third acceleration time series data is calculated, and the composite vector is compared with a composite threshold value that is smaller than the uniaxial threshold value to determine whether the sauna bath is taking place or whether the rest is taking place. The behavior determination device is configured as follows.

9. a heart rate acquisition step of acquiring heart rate time series data or pulse rate time series data showing the heart rate or pulse rate per unit time of a sauna user who uses the sauna as a set of at least a sauna bath and a rest; a conversion step of converting the heart rate time series data or the pulse rate time series data acquired in the heart rate acquisition step into MET time series data; a determination step of comparing the METs included in the MET time-series data converted in the conversion step with a MET threshold, and determining that the sauna bathing is being performed if the METs are greater than the MET threshold, and determining that the rest is being performed if the METs are equal to or less than the MET threshold; To execute The behavior determination method is configured as follows.

10. a heart rate acquisition step of acquiring heart rate time series data or pulse rate time series data showing the heart rate or pulse rate per unit time of a sauna user who uses the sauna as a set of at least a sauna bath and a rest; an acceleration acquisition step of acquiring acceleration time series data indicating the movement of the sauna user; Run moreover, a determination step is executed in which it is determined that the sauna bathing is being performed if a determination condition is satisfied, and it is determined that the resting is being performed if the determination condition is not satisfied, and the determination condition is that the heart rate time series data or the pulse rate time series data acquired in the heart rate acquisition step is on an increasing trend and the acceleration time series data acquired in the acceleration acquisition step exceeds a threshold value; The behavior determination method is configured as follows.

11. A behavior determination program executable by a computer of an information processing device, The computer, a heart rate acquisition process for acquiring heart rate time series data or pulse rate time series data that indicates the heart rate or pulse rate per unit time of a sauna user who uses the sauna as a set of at least a sauna bath and a rest; a conversion process for converting the heart rate time series data or the pulse rate time series data acquired in the heart rate acquisition process into MET time series data; a determination process of comparing the METs included in the MET time-series data converted by the conversion process with a MET threshold, and determining that the sauna bath is being performed if the METs are greater than the MET threshold, and determining that the rest is being performed if the METs are equal to or less than the MET threshold; Execute A behavioral judgment program configured as follows.

12. A behavior determination program executable by a computer of an information processing device, The computer, a heart rate acquisition process for acquiring heart rate time series data or pulse rate time series data that indicates the heart rate or pulse rate per unit time of a sauna user who uses the sauna as a set of at least a sauna bath and a rest; an acceleration acquisition process for acquiring acceleration time series data indicating the movements of the sauna user; Execute Furthermore, the computer A determination process is executed to determine that the sauna bathing is being performed if a determination condition is satisfied, and to determine that the resting is being performed if the determination condition is not satisfied, the determination condition being that the heart rate time series data or the pulse rate time series data acquired in the heart rate acquisition process is on an increasing trend and the acceleration time series data acquired in the acceleration acquisition process exceeds a threshold. A behavioral judgment program configured as follows.

Images