Core body temperature estimation device, core body temperature estimation method, core body temperature estimation program

Abstract

The accuracy of estimation of deep body temperature is improved compared to when only one deep body temperature estimation model is used. [Solution] The core body temperature estimation device comprises: a core body temperature acquisition unit that acquires the core body temperature of a target subject to be estimated at an initial time; a pulse acquisition unit that acquires the pulse of the target subject to be estimated; and an estimation unit that estimates the core body temperature of the target subject at a predetermined time based on the core body temperature acquired by the core body temperature acquisition unit at the initial time and the pulse acquired by the pulse acquisition unit from the initial time to a predetermined time, using a core body temperature estimation model for estimating the core body temperature of the target subject. The estimation unit of this core body temperature estimation device uses two or more core body temperature estimation models for the target subject to be estimated, depending on predetermined conditions.

Description

[Technical Field]

[0001] The present invention relates to a deep body temperature estimation device, a deep body temperature estimation method, and a deep body temperature estimation program. [Background technology]

[0002] Various techniques for preventing heat stroke have been disclosed (see, for example, Patent Documents 1, 2, and 3).

[0003] The technology described in Patent Document 1 relates to an air-conditioned suit that drives an air blower based on the results of body temperature measurement.

[0004] The technology described in Patent Document 2 is a technology for detecting the risk of heat stroke in a worker before an abnormality occurs in the worker's core body temperature.

[0005] The technology described in Patent Document 3 is a technology for estimating core body temperature based on physiological indices. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Publication No. 2017-166075 [Patent Document 2] Japanese Patent Publication No. 2017-27123 [Patent Document 3] WO2021 / 049573 publication Summary of the Invention [Problem to be solved by the invention]

[0007] However, the technology described in Patent Document 1 does not take into account core body temperature. On the other hand, it is desirable to measure a worker's core body temperature in order to determine the risk of heatstroke or the like. However, it is difficult to measure a worker's core body temperature while they are performing some kind of work. Therefore, there is a need to estimate core body temperature using physiological indicators that are easier to measure than core body temperature.

[0008] The technology described in Patent Document 2 measures eardrum temperature as the deep body temperature, which is easier to measure than rectal temperature and less affected by sweating, to avoid the hassle of measurement. The technology described in Patent Document 3 constructs a deep body temperature estimation model using an index reflecting heart rate variability as an explanatory variable and deep body temperature as a target variable.

[0009] The technology described in Patent Document 3 uses this deep body temperature estimation model to estimate the current deep body temperature from the above current indices. These technologies estimate deep body temperature based on other measurements, physiological indices, etc., without directly measuring rectal temperature, and therefore can reduce the burden on the worker compared to technologies that measure rectal temperature.

[0010] However, when using a deep body temperature estimation model, the correlation between physiological indices and deep body temperature may differ, for example, between when the deep body temperature is rising and when it is falling. Therefore, in all cases, if the deep body temperature is estimated from the physiological indices of heart rate variability using a deep body temperature estimation model constructed using actual measurements of when the deep body temperature is rising according to the conventional method described in Patent Document 3 (hereinafter also referred to as the "conventional method"), the deep body temperature may not be accurately estimated during the period when it is actually falling.

[0011] The present invention aims to improve the accuracy of estimation of deep body temperature based on physiological indices compared to when only one deep body temperature estimation model is used. [Means for solving the problem]

[0012] In one embodiment, the present invention provides a core body temperature estimation device comprising: a core body temperature acquisition unit that acquires the core body temperature of a target subject to be estimated at an initial time; a pulse acquisition unit that acquires the pulse of the target subject to be estimated; and an estimation unit that estimates the core body temperature of the target subject at a predetermined time based on the core body temperature acquired by the core body temperature acquisition unit at the initial time and the pulse acquired by the pulse acquisition unit from the initial time to a predetermined time, using a core body temperature estimation model for estimating the core body temperature of the target subject, wherein the estimation unit uses two or more core body temperature estimation models for the target subject to be estimated according to predetermined conditions.

[0013] In a preferred embodiment, each of the core body temperature estimation models is constructed from different datasets and estimates the change in core body temperature for each reference period divided into predetermined time intervals, wherein the datasets are created according to the change in core body temperature for each reference period, and the predetermined conditions are determined based on the pulsations of the reference period that include at least the predetermined time obtained by the pulsation acquisition unit.

[0014] In a preferred embodiment, each of the core body temperature estimation models is a model constructed by regression analysis in which an index obtained from the analysis of Poincaré plots of pulse intervals, which are the intervals between pulses of multiple model-building subjects, for each reference period, and the estimated core body temperature of the model-building subjects at the start of each reference period are used as explanatory variables, and the change in the core body temperature of the model-building subjects for each reference period is used as the dependent variable.

[0015] In a preferred embodiment, the estimation unit obtains an index from the analysis of the Poincaré plot of the pulse intervals of the pulses acquired by the pulse acquisition unit for each of the reference periods, and uses a core body temperature estimation model selected from two or more core body temperature estimation models according to the index for at least the reference period including the predetermined time.

[0016] In a preferred embodiment, the core body temperature estimation model includes an ascending model constructed from a data set when the change in core body temperature for each reference period is a positive value, and a descending model constructed from a data set when the change is a negative value, and the estimation unit uses the ascending model when the ratio of the index for the reference period including the specified time to the index for the reference period immediately preceding the reference period exceeds a threshold, and uses the descending model when it does not exceed a threshold.

[0017] In a preferred embodiment, the estimation unit estimates the deep body temperature of the subject to be estimated after the initial time by adding the change in deep body temperature estimated for each reference period after the initial time to the deep body temperature at the initial time.

[0018] In a preferred aspect, a pulse meter including electrodes for measuring the pulse of the estimated subject is provided, and the pulse acquisition unit acquires the pulse of the estimated subject measured by the pulse meter.

[0019] In one aspect, the present invention provides a deep body temperature estimation method in which a computer acquires the deep body temperature at an initial time of a subject whose deep body temperature is to be estimated, acquires the pulse of the subject, and uses a deep body temperature estimation model to estimate the deep body temperature of the subject, based on the deep body temperature at the initial time acquired by the deep body temperature acquisition unit and the pulse from the initial time to a predetermined time acquired by the pulse acquisition unit, estimates the deep body temperature of the subject at the predetermined time, and when estimating the deep body temperature, uses two or more deep body temperature estimation models for the subject depending on predetermined conditions.

[0020] In one embodiment, the present invention provides a core body temperature estimation program in which a computer acquires the core body temperature of a subject to be estimated at an initial time, acquires the pulse of the subject to be estimated, and estimates the core body temperature of the subject to be estimated at a predetermined time based on the core body temperature acquired by the core body temperature acquisition unit and the pulse acquired by the pulse acquisition unit from the initial time to a predetermined time, and when estimating the core body temperature, uses two or more core body temperature estimation models for the subject to be estimated, depending on predetermined conditions. [Effects of the Invention]

[0021] According to the present invention, a technique for estimating core body temperature based on physiological indices can be provided. Furthermore, according to the present invention, the accuracy of estimation can be improved compared to when only one core body temperature estimation model is used to estimate core body temperature. [Brief explanation of the drawings]

[0022] [Figure 1] FIG. 2 is a diagram showing an example of functional blocks of the deep body temperature estimation model construction device according to the present embodiment. [Figure 2] A diagram showing an example of the functional block of the core body temperature estimation device of this embodiment. [Figure 3] FIG. 1 is a diagram showing an example of the hardware configuration of an information processing apparatus. [Figure 4] This diagram shows an example of the operation flow for building a deep body temperature estimation model in a deep body temperature estimation model building device. [Figure 5] A figure showing an example of a Poincaré plot for RRI. [Figure 6] A diagram illustrating an example of the relationship between estimated core body temperature and the change in core body temperature. [Figure 7] A diagram illustrating an example of the operation flow for core body temperature estimation in a core body temperature estimation device. [Figure 8] A graph showing an example of actual core body temperature measurements during the test and estimated values ​​using conventional methods. [Figure 9]10 is a graph showing an example of actual core body temperature measurements during a test and estimated values ​​using an improved estimation method. [Figure 10] FIG. 10 is a diagram showing an example of a second coefficient determination method. [Figure 11] 10 is a graph showing actual measured values ​​and estimated values ​​when the second coefficient determination method is adopted, and a comparative example. DETAILED DESCRIPTION OF THE INVENTION

[0023] Embodiments will be described below with reference to the drawings. The configurations of the embodiments are illustrative, and the configuration of the invention is not limited to the specific configurations of the disclosed embodiments. In carrying out the invention, specific configurations may be adopted as appropriate depending on the embodiment.

[0024] <Embodiment> Figure 1 shows an example of the functional blocks of the core body temperature estimation model construction device of this embodiment. The core body temperature estimation model construction device 100 in Figure 1 includes a core body temperature acquisition unit 102, a heart rate information acquisition unit 104, an estimation model construction unit 106, and a storage unit 108.

[0025] The core body temperature estimation model building device 100 acquires core body temperature and heart rate information from multiple model building subjects during work under predetermined workloads, and builds a core body temperature estimation model that estimates core body temperature.

[0026] The core body temperature acquisition unit 102 acquires the core body temperature of a model construction subject measured by a thermometer or the like. Core body temperature is, for example, rectal temperature.

[0027] The heart rate information acquisition unit 104 acquires heart rate information of the model-building subject during work, measured by an electrocardiograph or the like, which measures heart rate information. Heart rate information includes, for example, electrocardiogram data and heart rate data. An electrocardiograph is an example of a pulsatric meter that measures heartbeats.

[0028] The estimation model construction unit 106 constructs a core body temperature estimation model that estimates core body temperature by regression analysis based on the core body temperature and heart rate information of the model construction subjects acquired from the initial time onward.

[0029] The storage unit 108 stores programs, models, data, etc. used in the deep body temperature estimation model construction device 100. The storage unit 108 stores the constructed deep body temperature estimation model, deep body temperature, heart rate information, etc.

[0030] 2 is a diagram showing an example of functional blocks of a core body temperature estimation device of this embodiment. The core body temperature estimation device 200 in FIG. 2 includes an initial core body temperature acquisition unit 202, a heartbeat information acquisition unit 204, an estimation unit 206, and a storage unit 208.

[0031] The core body temperature estimation device 200 acquires the core body temperature of the subject to be estimated at an initial time and the heart rate information of the subject during the work, and estimates the core body temperature of the subject using the core body temperature estimation model constructed by the core body temperature estimation model construction device 100. The subject used for model construction and the subject to estimation are collectively referred to simply as subjects.

[0032] The initial core body temperature acquisition unit 202 acquires the core body temperature of the target subject to be estimated before the work is performed, as measured by a thermometer or the like. This "before the work" refers to any time (first time), which is the initial time when the estimation process performed by the core body temperature estimation device 200 begins. In other words, the initial core body temperature acquisition unit 202 is an example of a core body temperature acquisition unit that acquires the core body temperature of the target subject to be estimated at an initial time.

[0033] The heartbeat information acquiring unit 204 acquires heartbeat information of the estimation target subject during work measured by an electrocardiograph or the like that measures heartbeat information. The heartbeat information is, for example, electrocardiogram data and heartbeat data. The heartbeat information acquiring unit 204 is an example of a heartbeat acquiring unit that acquires the heartbeat of the estimation target subject.

[0034] The estimation unit 206 estimates the change in core body temperature during each reference period using the core body temperature estimation model constructed by the core body temperature estimation model construction device 100, based on the core body temperature of the subject to be estimated at a first time (initial time = arbitrary time) and heart rate information for the period including the first time. Here, the "period including the first time" is divided into "reference periods," and the estimation unit 206 accumulates the change in core body temperature during each reference period to estimate the core body temperature at the end time (predetermined time) of the period including the first time. In other words, the estimation unit 206 is an example of an estimation unit that estimates the core body temperature of the subject to be estimated at a predetermined time using a core body temperature estimation model that estimates the core body temperature of the subject to be estimated, based on the core body temperature at the initial time acquired by the core body temperature acquisition unit and the heartbeat from the initial time to the predetermined time acquired by the heartbeat acquisition unit. Furthermore, the core body temperature estimation model described above is an example of a model that estimates the change in core body temperature for each reference period divided into predetermined time intervals.

[0035] The storage unit 208 stores programs, models, data, etc. used in the deep body temperature estimation device 200. The storage unit 208 stores the deep body temperature estimation model constructed by the deep body temperature estimation model construction device 100, the amount of change in estimated deep body temperature, deep body temperature, heart rate information, etc.

[0036] Here, heart rate information is used, but other types of pulse information may be used instead. Heart rate is the beating of the heart. Heart rate is a type of pulse. A pulse is a movement that occurs when the internal organs repeatedly contract and relax periodically. Examples of other types of pulses include fingertip pulse waves and earlobe pulse waves. Heart rate is the number of heartbeats in a given period. Pulse is the beating of an artery. Pulse interval is the length of one cycle of a pulse.

[0037] The deep body temperature estimation model construction device 100 and the deep body temperature estimation device 200 can be implemented using a dedicated or general-purpose computer such as a PC (Personal Computer), workstation (WS), smartphone, mobile phone, tablet terminal, car navigation system, or PDA (Personal Digital Assistant), or an electronic device equipped with a computer.

[0038] Figure 3 shows an example of the hardware configuration of an information processing device. The information processing device 90 shown in Figure 3 has the configuration of a typical computer. The deep body temperature estimation model construction device 100 and the deep body temperature estimation device 200 are realized by the information processing device 90 as shown in Figure 3. The information processing device 90 in Figure 3 has a processor 91, memory 92, storage unit 93, input unit 94, output unit 95, and communication control unit 96. These are connected to each other by a bus. The memory 92 and storage unit 93 are computer-readable recording media. The hardware configuration of the computer is not limited to the example shown in Figure 3, and components may be omitted, replaced, or added as appropriate.

[0039] The information processing device 90 can achieve a function that matches a predetermined purpose by having the processor 91 load a program stored on the recording medium into the working area of ​​the memory 92 and execute it, thereby controlling each component through the execution of the program.

[0040] Processor 91 is, for example, a CPU (Central Processing Unit) or a DSP (Digital Signal Processor).

[0041] Memory 92 includes, for example, RAM (Random Access Memory) and ROM (Read Only Memory). Memory 92 is also called main memory.

[0042] The storage unit 93 is, for example, an erasable programmable read only memory (EPROM) or a hard disk drive (HDD). The storage unit 93 may also include removable media, i.e., portable recording media. The removable media is, for example, a universal serial bus (USB) memory or a disc recording medium such as a compact disc (CD) or a digital versatile disc (DVD). The storage unit 93 is also called a secondary storage device.

[0043] The storage unit 93 stores various programs, various data, and various tables on a recording medium in a readable and writable manner. The storage unit 93 stores an operating system (OS), various programs, various tables, etc. The information stored in the storage unit 93 may be stored in the memory 92. Furthermore, the information stored in the memory 92 may be stored in the storage unit 93.

[0044] The operating system is software that mediates between software and hardware, manages memory space, manages files, and manages processes and tasks. The operating system also includes a communication interface. The communication interface is a program that exchanges data with other external devices connected via the communication control unit 96. Examples of external devices include other computers and external storage devices.

[0045] The input unit 94 includes a keyboard, pointing device, wireless remote control, touch panel, etc. The input unit 94 may also include video or image input devices such as a camera, and audio input devices such as a microphone.

[0046] The output unit 95 includes a display device such as an LCD (Liquid Crystal Display), an EL (Electroluminescence) panel, a CRT (Cathode Ray Tube) display, a PDP (Plasma Display Panel), a printer, etc. The output unit 95 may also include an audio output device such as a speaker.

[0047] The communication control unit 96 connects to other devices and controls communication between the information processing device 90 and the other devices. The communication control unit 96 is, for example, a LAN (Local Area Network) interface board, a wireless communication circuit for wireless communication, or a communication circuit for wired communication. The LAN interface board and the wireless communication circuit are connected to a network such as the Internet.

[0048] The computer that realizes the core body temperature estimation model construction device 100 realizes the functions of the core body temperature acquisition unit 102, heart rate information acquisition unit 104, and estimation model construction unit 106 by having the processor load a program stored in the auxiliary storage device into the main storage device and execute it. Meanwhile, the storage unit 108 is provided in the storage area of ​​the main storage device or the auxiliary storage device.

[0049] The computer that realizes the core body temperature estimation device 200 realizes the functions of the initial core body temperature acquisition unit 202, heart rate information acquisition unit 204, and estimation unit 206 by having the processor load a program stored in the auxiliary storage device into the main storage device and execute it. Meanwhile, the storage unit 208 is provided in the storage area of ​​the main storage device or the auxiliary storage device.

[0050] <Construction of a model for estimating core body temperature> 4 is a diagram showing an example of an operational flow for constructing a deep body temperature estimation model in a deep body temperature estimation model construction device. Here, the deep body temperature estimation model construction device 100 acquires deep body temperature from the start time to the end time of a predetermined period and heart rate information from the start time to the end time of each model construction subject when the subject performs a predetermined exercise stress test over the predetermined period, and constructs a deep body temperature estimation model that estimates the subject's deep body temperature. The deep body temperature estimation model is a model that estimates deep body temperature for each reference period divided into predetermined time intervals (time P) from an arbitrary first time, for example, based on deep body temperature at the first time and heart rate information for a period including the first time.

[0051] More specifically, the deep body temperature estimation model is a model that estimates the amount of change in deep body temperature from time t=nP to time t=(n+1)P (n is an integer equal to or greater than 0) based on the deep body temperature at time t=nP (which may be an estimated deep body temperature) and heart rate information from time t=(n-1)P to time t=(n+1)P. During the exercise stress test, each subject for model construction performs work with a predetermined load in a predetermined environment (predetermined temperature, predetermined humidity, etc.) from the start time to the end time of a predetermined period.

[0052] The exercise stress test is performed, for example, in an artificial climate chamber controlled at a temperature of 35°C and humidity of 50%, with the following sequence: 6 minutes of rest, 18 minutes of exercise load, 18 minutes of rest, 24 minutes of exercise load, 18 minutes of rest, and 42 minutes of rest. Of these, the final 42 minutes of rest is performed in a separate room controlled at a temperature of 25°C and humidity of 50%. The exercise load is administered using an ergometer at 80 kW. During the exercise stress test, each model-building subject's core body temperature is continuously measured in parallel using a thermometer or the like, and heart rate information is continuously measured using an electrocardiograph or the like. The exercise stress tests for multiple model-building subjects may be performed simultaneously or separately.

[0053] In S101, the core body temperature acquisition unit 102 of the core body temperature estimation model building device 100 acquires the core body temperature of the model building subjects measured by a thermometer or the like. The core body temperature acquisition unit 102 acquires the core body temperature of multiple model building subjects from the start time to the end time of the exercise stress test, either directly or via a network or the like, from a thermometer or the like connected to the core body temperature estimation model building device 100. Alternatively, the core body temperature acquisition unit 102 may acquire the core body temperature from the start time to the end time of the exercise stress test by having the model building subjects or other administrators input their core body temperature using an input means such as a keyboard. The core body temperature acquisition unit 102 stores the acquired core body temperature in the storage unit 108.

[0054] In S102, the heart rate information acquisition unit 104 acquires heart rate information from the start time to the end time of the exercise stress test of multiple model construction subjects, measured by an electrocardiograph or the like that measures heart rate information. The electrocardiograph is, for example, a device that includes electrodes attached to the skin of the model construction subject and a transmitting unit for storing and transmitting the heart rate information measured by the electrodes. Alternatively, the electrocardiograph may be a wearable terminal that includes electrodes attached to clothing (such as a T-shirt) and a transmitting unit for transmitting electrocardiogram data measured by the electrodes. In addition, the electrocardiograph may be a wearable terminal that is worn on the wrist, such as a wristwatch or wristband that acquires heart rate information, or a clip-type wearable terminal that is attached to the fingertip or ear. The heart rate information acquisition unit 104 acquires heart rate information from the start time to the end time of the exercise stress test of the model construction subjects from an electrocardiograph or the like that is connected directly to the core body temperature estimation model construction device 100 or via a network or the like. Furthermore, the heart rate information acquisition unit 104 may acquire heart rate information via a network or the like from other information processing devices that have acquired heart rate information from an electrocardiograph or the like. The heart rate information acquisition unit 104 acquires heart rate information from the initial time onward. The heart rate information acquisition unit 104 stores the acquired heart rate information of each model construction subject in the storage unit 108.

[0055] The order of processing S101 and S102 may be reversed. Also, processing S101 and S102 may be performed in parallel. The core body temperature acquisition unit 102 and the heart rate information acquisition unit 104 may acquire core body temperature and heart rate information, respectively, in parallel with the measurement of core body temperature and heart rate information during the exercise stress test of the model construction subject. The core body temperature and heart rate information are each stored in the storage unit 108 along with time information indicating the time of measurement.

[0056] In S103, the estimation model building unit 106 determines whether the core body temperature has risen based on at least the most recent change in the core body temperature of the model building subject acquired. Here, "most recent change" refers to the change determined by comparing the reference period to which the current time belongs with the reference period immediately preceding it. If the estimation model building unit 106 determines that the core body temperature has risen, it proceeds to S104. On the other hand, if the estimation model building unit 106 determines that the core body temperature has not risen, it proceeds to S105.

[0057] In S104 and S105, the estimation model construction unit 106 constructs core body temperature estimation models by regression analysis based on the acquired core body temperature and heart rate information of the subject used for model construction from the initial time onward. The core body temperature estimation model constructed in S104 is called the rising model. The core body temperature estimation model constructed in S105 is called the falling model. These two core body temperature estimation models are, for example, models that estimate the amount of change in core body temperature for each predetermined reference period from the initial time. The core body temperature estimation model is constructed, for example, with heart rate information and core body temperature at the start of one reference period as explanatory variables, and the amount of change in core body temperature during that reference period as the dependent variable. The dependent variable may be an estimated value of core body temperature. These two core body temperature estimation models may be, for example, models that estimate core body temperature for each reference period from the initial time.

[0058] In S103, the estimation model construction unit 106 distinguishes between data from when the model construction subject's core body temperature is rising and data from when it is falling. Then, in S104 and S105, the estimation model construction unit 106 uses these distinguished data to construct the rising model and the falling model, respectively. In other words, the rising model is constructed from a dataset consisting of data from when the model construction subject's core body temperature is rising. The falling model is constructed from a dataset consisting of data from when the model construction subject's core body temperature is falling.

[0059] Therefore, these two core body temperature estimation models are both examples of core body temperature estimation models constructed from different datasets. Furthermore, the different datasets used to construct the rising and falling models are examples of datasets created based on the change in core body temperature over a base period.

[0060] Furthermore, when the core body temperature has risen recently, the change in core body temperature during the most recent reference period is a positive value. Conversely, when the core body temperature has fallen recently, the change in core body temperature during the most recent reference period is a negative value. Therefore, the rising model described above is an example of a rising model constructed from a dataset where the change in core body temperature for each reference period is a positive value. Similarly, the falling model described above is an example of a falling model constructed from a dataset where the change in core body temperature for each reference period is a negative value.

[0061] Furthermore, while this explanation focuses on the case where there are only two core body temperature estimation models—an rising model and a falling model—the core body temperature estimation models may be divided into three or more models. For example, the core body temperature estimation models may be divided into four models, combining the rise and fall of core body temperature with the rise and fall of the derivative (or mean rate of change) of core body temperature. That is, the rising and falling phases are each further divided into two cases: one where the change is convex upwards, and another where the change is convex downwards.

[0062] 4, in step S103, the data is divided into two types of data sets by determining whether the deep body temperature is rising, and models corresponding to those data sets are constructed. However, the process of dividing the data sets and constructing the models does not have to be performed consecutively for each data set, as in this operational flow. For example, the deep body temperature estimation model construction device 100 may store the deep body temperature and heart rate information measured during an exercise stress test of a model construction subject as separate data sets, distinguishing between when the deep body temperature is rising and when it is falling. The deep body temperature estimation model construction device 100 may then read out each of the stored data sets and construct a model for when the deep body temperature is rising and a model for when the deep body temperature is falling, respectively.

[0063] The estimation model construction unit 106 calculates the RRI (RR interval) from the acquired heartbeat information. The RRI is the time difference between the occurrence time of an R wave in an electrocardiogram and the occurrence time of the R wave immediately before it. The RRI is the length of one heartbeat (one cycle). The RRI is an example of a beat interval. Fluctuations in the RRI are called heartbeat variability. The estimation model construction unit 106 calculates the RRI from the acquired heartbeat information. The estimation model construction unit 106 divides the calculated RRI of each heartbeat into reference periods (e.g., 3 minutes) from the initial time and creates a Poincaré plot for each reference period. A length of one heartbeat calculated by another known method may be used instead of the RRI. The length of the reference period is not limited to 3 minutes and may be longer or shorter than 3 minutes. The length of the reference period is preferably 1 minute or longer from a statistical perspective for calculating an index reflecting heartbeat variability (described below). Furthermore, the length of the reference period is preferably 15 minutes or shorter from the perspective of accuracy in estimating core body temperature and operational aspects.

[0064] 《Example of a Poincaré plot》 Figure 5 shows an example of a Poincaré plot for RRI. The Poincaré plot shown in Figure 5 is an example created at baseline intervals (e.g., 3 minutes) while subjects used for model construction are fitted with electrocardiographs or other devices to measure heart rate information and perform a predetermined task (exercise).

[0065] In the Poincaré plot in Figure 5, the horizontal axis represents the RRI (ms) of the nth beat, and the vertical axis represents the RRI (ms) of the (n+1)th beat, plotting the relationship between the nth and (n+1)th beats within a single reference period. In Figure 5, lines where the horizontal and vertical axis coordinates are equal are shown as identities. Furthermore, Figure 5 shows the centroids of all plotted points. The coordinates (X,Y) of the centroid are calculated as (average value of the coordinates of each point on the horizontal axis, average value of the coordinates of each point on the vertical axis). The line passing through the centroid and parallel to the identity is called the first centroid line, and the line passing through the centroid and perpendicular to the identity is called the second centroid line. Here, let vi be the distance from the i-th point (horizontal axis coordinate: RRI of the i-th beat, vertical axis coordinate: RRI of the (i+1)th beat) to the identity line, and let hi be the distance from the i-th point to the second centroid line. Using these, the indices SD1 and SD2 based on the Poincaré plot of RRI are obtained as follows:

[0066]

number

[0067] Here, N is the total number of points in a single Poincaré plot. SD1 and SD2 are indices that reflect heart rate variability. Note that SD1 and SD2 are either 0 or positive values, respectively. SD1 and SD2 are examples of indices obtained from the analysis of the Poincaré plot of RRI.

[0068] 《Explanation of the estimation model》 The estimation model construction unit 106 calculates SD2 for each reference period based on RRI. Let SD2(x) be the SD2 for the x-th reference period (the x-th reference period). The estimation model construction unit 106 constructs a core body temperature estimation model to estimate core body temperature at intervals of reference periods (e.g., 3 minutes) from the initial time. Here, if the length of the reference period is P, then the time after the first reference period has elapsed from the initial time (let's say t=0) is t=1×P, and the time after the j-th reference period has elapsed is t=jP. That is, the first reference period is from the initial time (t=0) to the time after the first reference period has elapsed (t=1×P). The j-th reference period is from the time after the j-1-th reference period has elapsed (t=(j-1)×P) to the time after the j-th reference period has elapsed (t=j×P). The estimation model construction unit 106 constructs a core body temperature estimation model by performing a regression analysis with SD2(j) and SD2(j-1) obtained from the analysis of the Poincaré plots of RRI for the j-th base period and the j-1 base period, and the estimated core body temperature at time t=(j-1)P at the end of the j-1 base period (start of the j-th base period) (let's call it TMP(j-1)) as explanatory variables, and the change in core body temperature from time t=(j-1)P to time t=jP (change in core body temperature during the j-th base period) as the dependent variable. Statistical analysis methods such as multiple regression analysis and logistic regression analysis can be used for the regression analysis. If the change in core body temperature from time t=(j-1)P to time t=jP (change in core body temperature during the j-th base period) is denoted as ΔTMP(j), it can be expressed as follows.

[0069]

number

[0070] Here, f(SD2(j), SD2(j-1), TMP(j-1)) is a function of SD2(j), SD2(j-1), and TMP(j-1). Also, TMP(0) is the core body temperature at the initial time. For example, f(SD2(j), SD2(j-1), TMP(j-1)) can be expressed as follows:

[0071]

number

[0072] Here, A1, A2, A3, and A4 are coefficients obtained by regression analysis. log represents the common logarithm. The equation obtained by regression analysis is not limited to the one shown here. Furthermore, the estimated core body temperature TMP(j) at time t=jP is obtained as follows.

[0073]

number

[0074] Here, TMP(0) is the deep body temperature at the initial time (t=0) obtained by measurement. The estimated value TMP(j) of the deep body temperature at time t=jP is obtained by adding the change in the estimated deep body temperature for each reference period from time t=0 to time t=jP to the deep body temperature TMP(0) at the initial time. The estimation model construction unit 106 stores the constructed deep body temperature estimation model (ΔTMP, TMP) in the storage unit 108. This allows the deep body temperature estimation model to be constructed in the deep body temperature estimation model construction device 100. In the deep body temperature estimation model, ΔTMP and TMP are expressed as functions of TMP(0) and SD2. By including SD2 in the deep body temperature estimation model, the deep body temperature can be calculated based on heart rate variability.

[0075] FIG. 6 is a diagram showing an example of the relationship between the estimated value TMP of core body temperature and the change ΔTMP in core body temperature. The horizontal axis of the graph in FIG. 6 represents time t, and the vertical axis represents the estimated value TMP of core body temperature. Here, TMP(0) is the measured value of core body temperature at the initial time. Each ΔTMP(j) can be estimated for each reference period. For example, the change ΔTMP(2) in core body temperature during the second reference period is the change in core body temperature from the time the first reference period has elapsed (t=P) to the time the second reference period has elapsed (t=2P). Furthermore, for example, the estimated value TMP(2) of core body temperature at the time the second reference period has elapsed can be expressed as TMP(1) + ΔTMP(2) = TMP(0) + ΔTMP(1) + ΔTMP(2). The time at which the nth reference period has elapsed (t=nP) is the same as the start of the n+1th reference period (t=(n+1)PP=nP).

[0076] The estimation model construction unit 106 may use other physiological indicators as explanatory variables when constructing a core body temperature estimation model. Other physiological indicators include, for example, sex, age, skin temperature, blood pressure, exhaled gas, pulse wave, blood oxygen concentration, blood flow rate, exercise level, and respiratory rate. Furthermore, clothing temperature, WBGT (Wet Bulb Globe Temperature), meteorological data, etc., may also be added as explanatory variables. By constructing a core body temperature estimation model that takes these explanatory variables into account, more accurate estimations can be made. The above is the mathematical formula used by the estimation model construction unit 106 in the model construction process.

[0077] Explanation of the improved estimation method The improved core body temperature estimation model incorporates two improvements. The first improvement is that it distinguishes between two or more datasets and constructs a separate core body temperature estimation model for each dataset.

[0078] The second improvement involves multiplying the change in core body temperature (ΔTMP) for each reference period, calculated by the core body temperature estimation model, by a coefficient C that depends on the initial state. The improvements are explained below.

[0079] Explanation of the first improvement The improved estimation model construction unit 106 calculates multiple sets of coefficients A1, A2, A3, and A4, which constitute the aforementioned f(SD2(j),SD2(j-1),TMP(j-1)), for each dataset. In other words, A1, A2, A3, and A4 are determined for each of the multiple distinguished datasets, and the function f is determined accordingly. The datasets are distinguished into two types, for example, as shown in Figure 6: data where the core body temperature TMP monotonically increases as time t progresses, and data where it decreases (not shown).

[0080] Here, we define A1, A2, A3, and A4 as the sets of coefficients obtained from the dataset where the time-dependent change in measured core body temperature is positive, i.e., when core body temperature is rising (during the rise). Then, we define the function determined by the above set of coefficients obtained from this rising data set as f(SD2(j),SD2(j-1),TMP(j-1)).

[0081] Furthermore, here, we define the coefficients B1, B2, B3, and B4 as the set of coefficients obtained from the dataset where the time-dependent change in measured core body temperature is negative, that is, when core body temperature is decreasing (during the decrease). The function determined by the above set of coefficients obtained from this decreasing dataset is defined as g(SD2(j),SD2(j-1),TMP(j-1)).

[0082] In other words, the function g of the descending model, which shows the core body temperature estimation model during a decrease, is distinguished from the function f of the ascending model, which shows the core body temperature estimation model during an increase, and is shown as follows.

number

[0083] This function g is expressed as follows using coefficients B1, B2, B3, and B4 obtained from the descending data set:

number

[0084] <<Explanation of the second improvement>> Furthermore, the improved core body temperature estimation model calculates the estimated value TMP(j) of the core body temperature at time t=jP as follows:

[0085]

number

[0086] The coefficient C in this equation is a coefficient that depends on the initial state. That is, the estimated value TMP(j) of the core body temperature at time t=jP is obtained by multiplying the sum of the changes in the core body temperature ΔTMP(i: i=1 to j) estimated by the core body temperature estimation model by this coefficient and adding the result to the core body temperature TMP(0) in the initial state.

[0087] In other words, the deep body temperature estimation device 200 that estimates deep body temperature using this formula is an example of a deep body temperature estimation device that has an estimation unit that corrects each change in deep body temperature estimated for each reference period using a deep body temperature estimation model by multiplying it by a coefficient.

[0088] This coefficient C can be determined in two ways, for example. The first method of determining the coefficient is to adopt a predetermined constant as the coefficient C, depending on the subject's environment in the initial state. In this method, for example, the following table is referred to, and the coefficient C is determined according to the conditions at the time of estimation. Specifically, if the initial environment is "below the discomfort index threshold" and the subject is "lightly dressed," then C11 is adopted as the coefficient C.

[0089] [Table 1]

[0090] In this case, coefficient C is an example of a coefficient that is a constant associated with the environment of the estimated subject at the initial time.

[0091] The second coefficient determination method is a method of determining the coefficient C using a predetermined function of the subject's index SD2 in the initial state (referred to as "initSD2"). In this coefficient determination method, the coefficient C is calculated, for example, by the following formula:

[0092]

number

[0093] In this case, the predetermined function for determining the coefficient C is a linear function of the index. In this equation, C0 is the constant term (intercept) of the linear function, and C1 is the coefficient (slope) of the linear term. These parameters (C0 and C1) are determined by, for example, regression analysis (multiple regression analysis, simple regression analysis). Note that the predetermined function does not have to be a linear function.

[0094] In this case, the coefficient C is an example of a coefficient represented by a function whose independent variable is an index of the estimated subject obtained from analysis of a Poincaré plot of the beat interval, which is the beat interval of the estimated subject during a reference period including the initial time.

[0095] <Estimation of core body temperature> FIG. 7 is a diagram showing an example of the operation flow of deep body temperature estimation in a deep body temperature estimation device. Here, the deep body temperature estimation device 200 acquires the deep body temperature at an initial time (first time) of the estimation target subject and heart rate information for a period including the first time, and estimates the deep body temperature of the estimation target subject using the deep body temperature estimation models for rising and falling constructed by the deep body temperature estimation model construction device 100 according to, for example, the operation flow of FIG. 4. The estimation target subject has had heart rate information measured using an electrocardiograph or the like. The estimation target subject also has their deep body temperature measured at the initial time (first time) using a thermometer or the like. The estimation target subject's deep body temperature is estimated based on the deep body temperature at the initial time and heart rate information for a period including the initial time. The estimation target subject's deep body temperature is not measured after the initial time. It is assumed, for example, that the estimation target subject is performing some kind of work.

[0096] In S201, the initial core body temperature acquisition unit 202 of the core body temperature estimation device 200 acquires the core body temperature of the subject to be estimated at an initial time, measured by a thermometer or the like that measures core body temperature. The initial time is the start time of the period for which the core body temperature is estimated. The initial core body temperature acquisition unit 202 acquires the core body temperature of the subject to be estimated at an initial time from a thermometer or the like that is connected to the core body temperature estimation device 200 directly or via a network or the like. Alternatively, the initial core body temperature acquisition unit 202 may acquire the core body temperature by having the subject to be estimated or other administrators input their core body temperature using an input means such as a keyboard. The initial core body temperature acquisition unit 202 stores the acquired core body temperature in the storage unit 208. The initial core body temperature acquisition unit 202 acquires the core body temperature of the subject to be estimated at an initial time and does not acquire core body temperature after the initial time. If it is difficult to directly measure the deep body temperature of the subject at the initial time, the deep body temperature may be estimated from the eardrum temperature or skin surface temperature of the subject at the initial time. Alternatively, the average deep body temperature of people of the same age and sex as the subject at the initial time may be obtained as the deep body temperature of the subject at the initial time.

[0097] In S202, the heartbeat information acquisition unit 204 acquires heartbeat information of the subject to be estimated measured by an electrocardiograph or the like that measures heartbeat information. The electrocardiograph used in S102 of FIG. 4 may be the same as the electrocardiograph used in S102. The heartbeat information acquisition unit 204 acquires the heartbeat information of the subject to be estimated from an electrocardiograph or the like connected to the core body temperature estimation device 200 directly or via a network or the like. The heartbeat information acquisition unit 204 may also acquire heartbeat information via a network or the like from another information processing device or the like that has acquired heartbeat information from the electrocardiograph or the like. The heartbeat information acquisition unit 204 acquires heartbeat information for a period including the initial time. For example, the heartbeat information acquisition unit 204 acquires heartbeat information from a time one reference period before the initial time to an arbitrary time after the initial time. The length of the reference period is, for example, three minutes. The length of the reference period here is the same as the length of the reference period when the core body temperature estimation model is constructed. The heart rate information acquisition unit 204 stores the acquired heart rate information of the subject to be estimated in the storage unit 208.

[0098] Core body temperature and heart rate information are stored in the storage unit 208 along with time information indicating the time of measurement.

[0099] In S203, the estimation unit 206 determines whether the index of the reference period including the predetermined time satisfies a predetermined condition. This predetermined condition will be described later. If it is determined that the index of this reference period satisfies this "predetermined condition," the estimation unit 206 proceeds to S204 and adopts the falling model of the core body temperature estimation model. On the other hand, if it is determined that the index of this reference period does not satisfy the above condition, the estimation unit 206 proceeds to S205 and adopts the rising model of the core body temperature estimation model.

[0100] Here, the reference period to which the current time (predetermined time) belongs is the jth reference period, and the reference period immediately preceding it is the j-1th reference period. The index SD2 of the jth reference period is SD2(j), and the index SD2 of the j-1th reference period is SD2(j-1). In this case, for example, the estimation unit 206 determines the core body temperature estimation model to be adopted depending on whether dSD2(j), which is the ratio of SD2(j) to SD2(j-1), satisfies the above condition. dSD2(j), for example, is expressed as follows:

[0101]

number

[0102] The "specified conditions" mentioned above refer to, for example, the condition that both (Condition 1) and (Condition 2) below are met. (Condition 1) The dSD2 of the current base period (the j-th base period) is 1 or greater and less than 3 (1 ≤ dSD2(j) < 3). (Condition 2) The dSD2 of the previous reference period (the j-1 reference period) is 1 or greater and less than 3 (1 ≤ dSD2(j-1) < 3). Note that "1" and "3" in (Condition 1) and (Condition 2) are merely examples of threshold values, and other numerical values ​​may be used.

[0103] In S206, the estimation unit 206 estimates the core body temperature using the core body temperature estimation model (rising model or falling model) adopted in S204 or S205, which is constructed according to the operation flow in Figure 4, based on the acquired core body temperature at the initial time and heart rate information from the initial time onward. The estimation unit 206 can estimate the core body temperature by adding the change in core body temperature for each reference period to the core body temperature at the initial time (first time step), using the change in core body temperature estimated by the core body temperature estimation model. In other words, this estimation unit 206 is an example of an estimation unit that uses two or more core body temperature estimation models for a subject to be estimated, depending on predetermined conditions.

[0104] Furthermore, the above-mentioned "predetermined condition" is an example of a condition that is determined based on at least the pulsations during a reference period that includes the predetermined time acquired by the pulsation acquisition unit.

[0105] The estimation unit 206 extracts the core body temperature estimation model stored in the storage unit 208. The estimation unit 206 also calculates the RRI for each heartbeat from the acquired heart rate information. The estimation unit 206 divides the calculated RRI for each heartbeat into reference periods (e.g., 3 minutes) starting from the time one reference period prior to the initial time, and creates a Poincaré plot for each reference period. Based on the Poincaré plot of RRI, the estimation unit 206 calculates the SD2 index for each reference period.

[0106] For example, the estimation unit 206 calculates the ratio (i.e., dSD2(j)) of the SD2 index for each of the j-th reference period and the j-1-th reference period (i.e., SD2(j) and SD2(j-1)) from the SD2 index calculated for each reference period. Then, the estimation unit 206 selects the rising model when this ratio exceeds a threshold and adopts it as the core body temperature estimation model. On the other hand, the estimation unit 206 selects the falling model when the calculated ratio does not exceed the threshold and adopts it as the core body temperature estimation model.

[0107] The estimation unit 206 then estimates the change in core body temperature (ΔTMP) for each reference period using the adopted core body temperature estimation model, based on the core body temperature of the subject to be estimated at an initial time (an arbitrary time) and the index SD2 for each reference period. The estimation unit 206 also estimates the core body temperature (TMP) of the subject to be estimated at each reference period (the end time of each reference period) by adding the change in core body temperature for each reference period to the core body temperature at the initial time. The estimated core body temperature and other data are stored in the storage unit 208.

[0108] Therefore, this estimation unit 206 is an example of an estimation unit that, for each reference period, obtains an index obtained from the analysis of the Poincaré plot of the pulse interval of the pulse acquired by the pulse acquisition unit, and uses a core body temperature estimation model selected from two or more core body temperature estimation models according to the index of the reference period that includes at least a predetermined time.

[0109] Furthermore, this estimation unit 206 is an example of an estimation unit that uses an up-time model when the ratio of an index for a reference period including a specified time to an index for the reference period immediately preceding the reference period exceeds a threshold, and uses a down-time model when the ratio does not exceed a threshold.

[0110] In this embodiment, there are only two core body temperature estimation models: an rising model and a falling model. However, if the core body temperature estimation model construction device 100 has constructed three or more core body temperature estimation models using different datasets, the core body temperature estimation device 200 may appropriately select and use these three or more core body temperature estimation models. In this case as well, the estimation unit 206 of the core body temperature estimation device 200 divides the calculated RRI of each heartbeat into reference periods (e.g., 3 minutes) starting from the time one reference period prior to the initial time, and creates a Poincaré plot for each reference period. Furthermore, the estimation unit 206 calculates an index SD2 for each reference period based on the Poincaré plot of RRI, and selects one core body temperature estimation model from the three or more core body temperature estimation models based on this index SD2. The conditions for determining the index SD2 include, for example, comparison with a threshold.

[0111] Furthermore, the estimation unit 206 may issue a warning of the risk of heat stroke or the like when the estimated deep body temperature is equal to or higher than a predetermined threshold. The estimation unit 206 may output a warning of the risk of heat stroke or the like using output means such as a display or a speaker. This allows the estimation subject or the like to understand the risk of heat stroke or the like. The estimation unit 206 may simply output the estimated deep body temperature using output means. This allows the estimation subject or the like to understand the estimated value of the deep body temperature based on the heart rate information.

[0112] The estimation unit 206 may calculate SD2 from the RRI for each reference period and estimate the deep body temperature. Furthermore, the estimation unit 206 may output the estimated deep body temperature for each reference period using an output means. This allows the estimation target subject or the like to grasp the deep body temperature in real time using the output means of the deep body temperature estimation device 200.

[0113] The core body temperature estimation device 200 may be integrated in advance with the core body temperature estimation model construction device 100, which constructs the core body temperature estimation model according to the operational flow of Fig. 4. The core body temperature estimation device 200 may also include an electrocardiograph that measures heartbeat information of the subject to be estimated. In this case, the heartbeat information acquisition unit 204 of the core body temperature estimation device 200 acquires heartbeat information from the electrocardiograph.

[0114] In the core body temperature estimation device 200, SD2(0) is required when calculating ΔTMP(1), etc., but SD2(0) is calculated from heart rate information from a time P before the initial time (t=-P) to the initial time (time t=0). Here, for example, the core body temperature TMP(1) at time t=P may be measured using a thermometer or the like (or the core body temperature TMP(1) at time t=P may be estimated from the eardrum temperature or skin surface temperature at time t=P), and ΔTMP(j) and TMP(j) from time t=2×P(j=2) onward may be calculated. Alternatively, the core body temperature TMP(1) at time t=P may be considered to be the same as the core body temperature TMP(0) at time t=0, and TMP(1) = TMP(0). In other words, heart rate information from time t=-P to time t=0 is not required.

[0115] (others) Figure 8 is a graph showing an example of measured core body temperatures during the test and estimated values ​​using a conventional method. In the graph of Figure 8, the horizontal axis represents time from the start of the exercise stress test, and the vertical axis represents core body temperature. Also in the graph of Figure 8, the dotted line represents the average measured core body temperatures for all subjects, and the solid line represents the average estimated core body temperatures for all subjects. Note that error bars indicating standard error are attached to each of the dotted and solid lines in Figure 8. During the exercise stress test, each subject performs work under a specified load in a specified environment (specified temperature, specified humidity, etc.) from the start time to the end time of a specified period. The subjects in this test were 22 healthy adult males aged 21 to 52. Of these, 20 were under 25 years old, and the remaining two were 30 and 52 years old, respectively.

[0116] This exercise stress test was conducted in a climate chamber maintained at a temperature of 35°C and humidity of 50% during exercise and rest periods. The test consisted of 6 minutes of rest (R1), 18 minutes of exercise (E1), 18 minutes of rest (R2), 24 minutes of exercise (E2), 18 minutes of rest (R3), and 42 minutes of rest (R4). However, R4 was performed after the subject was moved to a separate room maintained at a temperature of 25°C and humidity of 50%. The exercise stress was administered using an ergometer at 80 kW or 60% of each subject's pre-measured maximum oxygen uptake. During the exercise stress test, each subject's core body temperature was continuously measured using a thermometer or other device, and heart rate information was measured using an electrocardiograph or other device. The subject's actual core body temperature at the start of the exercise stress test was used to estimate core body temperature.

[0117] In the graph of Figure 8, it can be seen that the estimated core body temperature of each subject generally matches the actual measured core body temperature of each subject only when the actual measured core body temperature value increases. However, when the actual measured core body temperature value decreases, the slopes of the actual measured value and the estimated value diverge and cross. In other words, it is clear that it is difficult to estimate the core body temperature of a subject using conventional methods.

[0118] Figure 9 is a graph showing an example of the actual measured core body temperature during the test and the estimated value by the improved estimation method. The estimation method used to obtain the estimated value in Figure 9 reflects the first improvement described above. In other words, this estimated value is obtained using two or more core body temperature estimation models for the subject to be estimated, depending on predetermined conditions.

[0119] The estimated values ​​shown in Figure 9 use the rising model and falling model as core body temperature estimation models. The rising model is a core body temperature estimation model constructed using a dataset in which the change in the actual measured value of core body temperature over time is a positive value. The rising model uses the above-mentioned function f when calculating the change in core body temperature for each reference period (ΔTMP). In this test, the coefficients used to make up function f were A1 = 0.9528, A2 = -0.0748, A3 = 0.0052, and A4 = -0.0219.

[0120] The descending model is a core body temperature estimation model constructed using a data set in which the change in core body temperature measured over time is a negative value. The descending model uses the above-mentioned function g when calculating the change in core body temperature (ΔTMP) for each reference period. In this test, the coefficients used to make up function g were B1 = 0.4565, B2 = -0.1381, B3 = 0.0144, and B4 = -0.0070.

[0121] This improved estimation method switches between these two core body temperature estimation models, the rising model and the falling model, depending on whether the dSD2 described above meets the predetermined conditions. dSD2 is the ratio of the index SD2 in the current base period to the index SD2 in the immediately preceding base period.

[0122] In this study, if the condition that dSD2 was between 1 and 3 for two consecutive periods was met, the descending model was adopted as the core body temperature estimation model; otherwise, the ascending model was adopted.

[0123] Furthermore, the estimated values ​​in Figure 9 reflect the second improvement described above, and the first coefficient determination method is used to determine the coefficient C. Here, the coefficient C is given as a constant (e.g., "1.1") depending on the type of clothing worn by the subject.

[0124] As shown in Figure 9, during the R4 period when the measured core body temperature decreases, the estimated core body temperature also decreases in the same way as the measured value, and the difference between them is smaller compared to Figure 8. In other words, this result shows that the improved estimation method provides a higher accuracy in estimating core body temperature than the conventional method.

[0125] This allows for flexible adjustment of differences when the core body temperature estimation model is constructed, such as when a subject is wearing a T-shirt, but when the subject is dressed in different clothing, such as long-sleeved work clothes, during the core body temperature estimation process.

[0126] Figure 10 shows an example of the second coefficient determination method described above. In the graph shown in Figure 10, the index SD2 (i.e., initSD2) of the initial reference period is shown on the horizontal axis, and the optimal coefficient is shown on the vertical axis. This optimal coefficient is obtained by finding the coefficient C for which the variance, standard deviation, etc., between the measured and estimated core body temperature values ​​are optimal for each subject. Then, the parameters C1 and C2 used to calculate the coefficient C are obtained by fitting the plots of the optimal coefficients obtained for each subject to a linear function with initSD2 as the independent variable using the least squares method. For example, in the example shown in Figure 10, C1 is 0.0083 and C0 is 0.5628. The coefficient of determination R2 of the linear function with these parameters is 0.507.

[0127] Figure 11 is a graph showing the measured values, estimated values, and a comparison example when the second coefficient determination method is adopted. In this case as shown in Figure 11, during the R4 period when the measured core body temperature decreases, the estimated core body temperature also decreases in the same way as the measured value, and the difference between them is smaller compared to Figure 8. Furthermore, in this case, the difference in the slopes of the measured and estimated values ​​during the decrease is smaller compared to Figure 9. In other words, this result shows that the accuracy of core body temperature estimation is improved with the improved estimation method compared to the conventional method.

[0128] (Effects and mechanisms of the embodiment) In conventional methods that use only one core body temperature estimation model, as shown in Figure 8, even when the measured core body temperature was on a downward trend, the estimated value sometimes failed to accurately capture this trend. On the other hand, as shown in Figures 9 and 11, the core body temperature estimation device 200 of the present invention can calculate an estimated value that follows the downward trend in the measured core body temperature of the subject during the R4 period of this graph. The core body temperature estimation device 200 uses two or more core body temperature estimation models for the target subject, depending on predetermined conditions, such as an rising model and a falling model. Therefore, this core body temperature estimation device 200 can estimate the core body temperature of the target subject with higher accuracy compared to other core body temperature estimation devices that use only one core body temperature estimation model.

[0129] <Computer-readable recording medium> A program that enables a computer or other machine or device (hereinafter referred to as "computer, etc.") to perform any of the above functions can be recorded on a recording medium that the computer, etc. can read. By having the computer, etc. read and execute the program on this recording medium, it can be made to provide that function.

[0130] Here, a computer-readable recording medium refers to a recording medium that stores information such as data and programs electrically, magnetically, optically, mechanically, or chemically and can be read by a computer, etc. Such a recording medium may be provided with elements that constitute a computer, such as a CPU and memory, and the CPU may be made to execute the program. Among such recording media, those that can be removed from a computer or the like include, for example, flexible disks, magneto-optical disks, CD-ROMs, CD-R / Ws, DVDs, DATs, 8mm tapes, memory cards, and the like.

[0131] In addition, there are hard disks, ROMs, SSDs, and other storage media that are fixed to computers and other devices.

[0132] Although the embodiments of the present invention have been described above, these are merely examples, and the present invention is not limited to these. Various modifications based on the knowledge of those skilled in the art, such as combinations of the various components, are possible, provided that they do not deviate from the spirit of the claims. [Explanation of symbols]

[0133] 100...Deep body temperature estimation model construction device, 102...Deep body temperature acquisition unit, 104...Heartbeat information acquisition unit, 106...Estimation model construction unit, 108...Storage unit, 200...Deep body temperature estimation device, 202...Initial deep body temperature acquisition unit, 204...Heartbeat information acquisition unit, 206...Estimation unit, 208...Storage unit, 90...Information processing device, 91...Processor, 92...Memory, 93...Storage unit, 94...Input unit, 95...Output unit, 96...Communication control unit

Claims

1. a deep body temperature acquisition unit that acquires the deep body temperature of a subject at an initial time, the subject being the subject whose deep body temperature is to be estimated; a pulse acquisition unit that acquires the pulse of the subject; An estimation unit estimates the core body temperature of the subject to be estimated at a predetermined time, using a core body temperature estimation model for estimating the core body temperature of the subject to be estimated, based on the core body temperature at the initial time acquired by the core body temperature acquisition unit and the pulse acquired by the pulse acquisition unit from the initial time to a predetermined time. Equipped with A deep body temperature estimation device, wherein the estimation unit uses two or more deep body temperature estimation models for the subject to be estimated according to predetermined conditions.

2. Each of the aforementioned core body temperature estimation models is constructed from different datasets and estimates the change in core body temperature for each reference period divided into predetermined time intervals. The data set is created according to the amount of change in core body temperature for each reference period, The predetermined conditions are determined based on the pulsations of the reference period, including at least the predetermined time, which are acquired by the pulsation acquisition unit. The deep body temperature estimation device according to claim 1.

3. All of the aforementioned core body temperature estimation models are constructed by regression analysis using an index obtained from the analysis of Poincaré plots of pulse intervals (the intervals between pulses of multiple model-building subjects) for each reference period, and the estimated core body temperature of the model-building subjects at the start of each reference period as explanatory variables, with the change in the core body temperature of the model-building subjects for each reference period as the dependent variable. The deep body temperature estimation device according to claim 2 .

4. The estimation unit obtains an index from the analysis of the Poincaré plot of the pulse intervals of the pulses acquired by the pulse acquisition unit for each of the reference periods, and uses a core body temperature estimation model selected from two or more core body temperature estimation models according to the index for at least the reference period including the predetermined time. The deep body temperature estimation device according to claim 3 .

5. The core body temperature estimation model includes an rising model constructed from a dataset where the change in core body temperature for each reference period is a positive value, and a falling model constructed from a dataset where the change is a negative value. The estimation unit uses the rising model when the ratio of the index for the reference period including the predetermined time to the index for the reference period immediately preceding the reference period exceeds a threshold, and uses the falling model when it does not exceed a threshold. The deep body temperature estimation device according to claim 4.

6. The estimation unit corrects the amount of change estimated for each reference period using the deep body temperature estimation model by multiplying the amount of change by a coefficient. The deep body temperature estimation device according to claim 2 .

7. The coefficient is a constant associated with the environment of the subject to be estimated at the initial time. The deep body temperature estimation device according to claim 6.

8. The coefficient is expressed as a function with an independent variable that is an index of the estimated subject obtained from the analysis of a Poincaré plot of the pulse interval, which is the interval between the pulses of the estimated subject during a reference period including the initial time. The deep body temperature estimation device according to claim 6.

9. The aforementioned function is a linear function. The deep body temperature estimation device according to claim 8.

10. The estimation unit estimates the core body temperature of the subject to be estimated after the initial time by adding the amount of change in core body temperature estimated for each reference period after the initial time to the core body temperature at the initial time. The deep body temperature estimation device according to claim 2 .

11. a pulsatility meter including electrodes for measuring the pulsation of the estimated subject; The pulse acquisition unit acquires the pulse of the estimation subject measured by the pulse meter. The deep body temperature estimation device according to any one of claims 1 to 10.

12. The computer The core body temperature of the subject to be estimated is obtained at the initial time, Acquire the heartbeat of the subject to be estimated; Using a core body temperature estimation model that estimates the core body temperature of the subject to be estimated, the core body temperature of the subject to be estimated at the predetermined time is estimated based on the core body temperature obtained at the initial time and the pulse from the initial time to the predetermined time. When estimating the deep body temperature, two or more deep body temperature estimation models are used for the subject to be estimated according to predetermined conditions. Method for estimating core body temperature.

13. The computer Acquire the deep body temperature of the subject at an initial time whose deep body temperature is to be estimated; The pulse of the estimated target subject is obtained, Using a core body temperature estimation model that estimates the core body temperature of the subject to be estimated, the core body temperature of the subject to be estimated at the predetermined time is estimated based on the core body temperature obtained at the initial time and the pulse from the initial time to the predetermined time. When estimating the deep body temperature, two or more deep body temperature estimation models are used for the subject to be estimated according to predetermined conditions. Core body temperature estimation program.

Images