Human body core body temperature detection device and method

Abstract

A human body core body temperature detection device comprises a fixing band, and a single-channel heat flow method sensor, a respiration sensor, an electrocardio sensor and a data processing unit which are fixed on one side of the fixing band, the single-channel heat flow method sensor, the respiration sensor and the electrocardio sensor are electrically connected with the data processing unit. The invention further provides a human body core body temperature detection method. The method comprises the steps that original body temperature data are collected through the single-channel heat flow method sensor; collecting respiration rate data through a respiration sensor; heart rate data are collected through an electrocardio sensor; the data processing unit dynamically adjusts the weights of the original body temperature, the heart rate and the respiration rate, and a pre-core body temperature and the body temperature compensation amount are calculated through a two-stage fusion algorithm; and correcting the pre-core body temperature data through the core body temperature compensation amount to obtain the core body temperature of the human body. By integrating the dynamic changes of the heart rate (HR) and the respiratory rate (RR), the accuracy and robustness of body temperature detection are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the field of core body temperature detection, and particularly relates to a human core body temperature detection device and method. Background Art

[0002] Measuring the human core body temperature is an important means for evaluating physiological status, disease diagnosis, and health management, and its core significance is reflected in multi-dimensional human body function monitoring and adaptive regulation. From the perspective of clinical medicine, the core body temperature can be used as a sensitive indicator for infections, inflammations, endocrine disorders, and metabolic abnormalities: for example, persistent fever may indicate bacterial infection or autoimmune diseases, while low body temperature may be related to hypothyroidism, severe trauma, or sepsis. Timely monitoring helps early intervention and reduces the risk of complications. In the field of sports science, core body temperature monitoring is crucial for ensuring the safety of athletes. If the body temperature exceeds 40°C during high-intensity training, heat stroke may occur. Through real-time data, the training plan can be optimized and the heat dissipation strategy can be adjusted to avoid muscle damage or cognitive function decline. Marathon runners often rely on oral or rectal thermometers to prevent heat stroke. In addition, workers in extreme environments such as steelworkers or polar explorers need to rely on core body temperature monitoring to prevent heat stress or hypothermia. The development of wearable devices has enabled dynamic early warning and emergency response. In basic metabolic research, the core body temperature is closely related to energy consumption. Approximately 60-70% of the basic metabolism is used to maintain body temperature. Abnormal resting body temperature may reflect thyroid function abnormalities or metabolic syndrome, providing a basis for obesity management and chronic disease prevention and treatment. The progress of modern medical technology has further promoted the precision of core body temperature monitoring. For example, during surgery, the body temperature is regulated through minimally invasive sensors to reduce postoperative complications. Abnormal body temperature in chemotherapy patients can indicate drug side effects and be treated in a timely manner.

[0003] Generally speaking, the core body temperature is not only the "biological clock" of the health status, but also a bridge connecting physiological mechanisms and external interventions. Its applications in disease prevention, sports optimization, environmental adaptation, and precision medicine have significantly improved humans' understanding and management ability of their own health, and have become an indispensable observation dimension in modern life science.

[0004] The prior art mainly measures the human core body temperature by using the single-channel heat flux method: According to the laws of thermodynamics, the temperature difference between the core part of the human body and the human body surface will cause heat to continuously transfer from the inside of the human body to the human body surface, thereby forming a heat flux. The single-channel heat flux method is to attach the measurement model to the human body surface, and the heat of the human body will be transferred to the measurement model, forming a heat flux channel from the human body surface to the measurement model. Without considering the lateral dissipation of heat, the heat flow inside the human body and the heat flow transferred to the measurement model are equal. According to this principle, the human core body temperature is calculated.

[0005] However, there are problems with the single-channel heat flux method for measuring human core body temperature, such as limited accuracy, lag in dynamic response, and narrow applicable scenarios. It relies on local skin or cavity temperature, is easily affected by environmental temperature, exercise state, and measurement location, resulting in large data fluctuations. Moreover, it can only measure temperature at a single point and cannot comprehensively reflect the true heat distribution in the body, especially with significant errors during strenuous exercise or complex pathological conditions. In addition, some invasive measurement methods (such as rectal probes) have poor comfort, limiting the application of long-term continuous monitoring, while surface patch devices may reduce accuracy due to poor contact or sweat interference, making it difficult to meet the high requirements of precision medicine and real-time health management. Summary of the Invention

[0006] The object of the present invention is to solve the problems of limited accuracy and lag in dynamic response existing in the existing single-channel heat flux method for measuring human core body temperature.

[0007] The object of the present invention is achieved by adopting the following technical solutions:

[0008] A human core body temperature detection device, the device includes: a fixing band, and a single-channel heat flux method sensor, a respiration sensor, an electrocardiogram sensor, and a data processing unit fixed on one side of the fixing band; the single-channel heat flux method sensor, the respiration sensor, and the electrocardiogram sensor are respectively electrically connected to the data processing unit.

[0009] Preferably, the single-channel heat flux method sensor includes: a heat conductor, a skin temperature sensor fixed at the bottom of the heat conductor, a microenvironment temperature sensor fixed at the top of the heat conductor, and a heat insulator covering the periphery and top surface of the heat conductor.

[0010] Preferably, the respiration sensor includes an inductive respiration sensor.

[0011] Preferably, the inductive respiration sensor includes an elastic band and two continuously bent wires fixed along the length direction of the elastic band.

[0012] Preferably, the electrocardiogram sensor includes a plurality of electrocardiogram electrodes.

[0013] Based on the same inventive concept, the present invention also provides a human core body temperature detection method, which uses the above-mentioned human core body temperature detection device, and the method includes the following steps:

[0014] Collect original body temperature data through the single-channel heat flux method sensor;

[0015] Collect respiration rate data through the respiration sensor;

[0016] Collect heart rate data through the electrocardiogram sensor;

[0017] The data processing unit dynamically adjusts the weights of the original body temperature, heart rate, and respiratory rate, and calculates the pre-core body temperature and the body temperature compensation amount using a two-stage fusion algorithm;

[0018] The human core body temperature is obtained by correcting the pre-core body temperature data with the core body temperature compensation amount.

[0019] Preferably, the two-stage fusion algorithm includes: a multi-physiological parameter fusion algorithm based on deep learning; a dynamic compensation algorithm based on the perception of the change rates of heart rate, respiratory rate, and pre-core body temperature.

[0020] Preferably, the multi-physiological parameter fusion algorithm based on deep learning includes: a ridge regression model.

[0021] Preferably, the formula of the multi-physiological parameter fusion algorithm is as follows:

[0022] T pre =w o +w1T ori +w2RR+w3HR+w4T ori 2 +w5RR 2 +w6HR 2 +w7T ori RR+w8T ori HR+w9RRHR;

[0023] Where:

[0024] T pre is the pre-core body temperature;

[0025] T ori is the original body temperature calculated by the single-channel heat flux method;

[0026] RR is the respiratory rate;

[0027] HR is the heart rate;

[0028] w0 is the reference offset, and its value range is 36.5 - 37.2;

[0029] w1 is the body surface temperature weight, and its value range is 0.6 - 0.9;

[0030] w2 is the respiratory rate weight, and its value range is 0.01 - 0.05;

[0031] w3 is the heart rate weight, and its value range is 0.001 - 0.005;

[0032] w4 is the square term coefficient, and its value range is on the order of 1e-2;

[0033] w5 is the square term coefficient, and its value range is on the order of 1e-3;

[0034] w6 is the coefficient of the square term, and its value range is on the order of 1e-4;

[0035] w7 to w9 are the weights of the interaction terms, usually on the order of 1e-4.

[0036] Preferably, the dynamic compensation algorithm includes the following formula:

[0037]

[0038] Where:

[0039] is the core body temperature compensation amount;

[0040] α is the global learning rate, which controls the scaling amplitude of the compensation amount;

[0041] x is the set of all physiological parameters that need to be compensated, specifically including the original body temperature, respiratory rate (RR), and heart rate (HR);

[0042] is the weight function;

[0043] sgn(x): The sign function, indicating the change direction (positive for increase, negative for decrease);

[0044] Δx (t) : The change amount of the parameter x at the current moment;

[0045] Rx: The physiological safety range of the parameter (to avoid exceeding the physiological limit).

[0046] Compared with the prior art, the beneficial effects of the present invention are:

[0047] The core body temperature detection device and method of the present invention can non-invasively and continuously monitor the core body temperature of the human body. By using the method of multi-source data fusion, it can better avoid the disadvantages of single-parameter estimation; at the same time, through dynamic compensation, the weights of the original body temperature, heart rate, and respiratory rate measured based on a single channel are adjusted, enhancing the problem of poor adaptability of estimating the core body temperature by a single method due to different manifestations of core body temperature changes in different individuals.

[0048] The "HR-RR combined compensation heat flux method body temperature detection model" proposed by the present invention significantly improves the accuracy and robustness of body temperature detection by integrating the dynamic changes of heart rate (HR) and respiratory rate (RR). Traditional heat flux methods rely on the linear relationship between body surface heat conduction and environmental parameters and are susceptible to external interferences (such as room temperature fluctuations and patient body movements) or individual differences (such as metabolic rate and skin condition), resulting in measurement deviations. After introducing HR and RR as compensation parameters, the model can capture the associated effects of physiological activities on heat distribution in real time: for example, an increase in heart rate may be accompanied by an increase in local blood flow, indirectly affecting the body surface heat dissipation efficiency; an increase in respiratory rate may change the heat loss rate through evaporation. By establishing a non-linear coupling relationship between HR-RR and heat flux density, the model can dynamically correct the systematic errors in body temperature calculation, especially in dynamic scenarios such as intensive care and perioperative management, effectively alleviating the lag and misjudgment risks caused by relying on a single parameter.

[0049] The present invention uses a directional weighting algorithm. Combining the physiological allowable ranges of each parameter (such as the 100 bpm threshold for HR and the 30 bpm threshold for RR), the contribution weights of parameter changes are quantified through relative amplitude (Δx / Rx) and the sign function (sgn(x)), which not only avoids the interference of outliers but also retains the dynamic characteristics of physiological signal changes. Clinical verification shows that this technology can reduce the false alarm rate of body temperature when the environmental temperature difference exceeds ±1°C, greatly reducing the adverse impact of environmental temperature fluctuations on the accuracy of core body temperature measurement; moreover, the "HR-RR combined compensation heat flux method body temperature detection model" proposed by the present invention has higher sensitivity for predicting the body temperature trend in complex pathological states such as shock and fever, providing a solution with both real-time and reliability for precision medicine or health status monitoring. Brief Description of the Drawings

[0050] Figure 1 It is a schematic structural diagram of a human core body temperature detection device according to the present invention;

[0051] Figure 2 It is a schematic diagram of the principle of a human core body temperature detection device according to the present invention;

[0052] Figure 3 It is a schematic structural diagram of a single-channel heat flux method sensor according to the present invention;

[0053] Figure 4 It is a detection structure and respiratory signal acquisition flow chart of a respiratory sensor according to the present invention;

[0054] Figure 5 It is a block diagram of a respiratory signal acquisition circuit according to the present invention;

[0055] Figure 6 It is a schematic diagram of the principle of a central electrocardiogram detection circuit according to the present invention;

[0056] Wherein: 1. fixing band, 2. single-channel heat flux method sensor, 3. respiration sensor, 4. electrocardiogram sensor, 5. data processing unit. Detailed implementation mode

[0057] The following further describes the technical solution in conjunction with the drawings and specific embodiments to help understand the content of the present invention.

[0058] Embodiment 1

[0059] As Figure 1-2 shown, a human core body temperature detection device, the device includes: a fixing band 1, and a single-channel heat flux method sensor 2, a respiration sensor 3, an electrocardiogram sensor 4 and a data processing unit 5 fixed on one side of the fixing band 1; the single-channel heat flux method sensor 2, the respiration sensor 3, and the electrocardiogram sensor 4 are respectively electrically connected to the data processing unit 5.

[0060] The human core body temperature detection device further includes an acquisition unit, and the acquisition unit includes signal acquisition circuits of the above temperature sensor, electrocardiogram sensor, and respiration sensor.

[0061] The fixing band 1 serves as a carrier for the sensor, the acquisition unit, the data processing unit 5, and the communication unit.

[0062] The data processing unit 5 processes the physiological signals from the data acquisition unit and processes the data.

[0063] The human core body temperature detection device further includes a communication unit, and the communication unit transmits physiological signal data, device parameters, etc.

[0064] As Figure 3 shown, the single-channel heat flux method sensor 2 is composed of two temperature sensors, including a skin temperature sensor and a microenvironment temperature sensor. Between these two temperature sensors is a heat-conducting material with a thermal conductivity similar to that of human tissues. The outside of the sensor is an insulating material to isolate the influence of external heat flux on the sensor as much as possible. The two temperature sensors are integrated temperature chips with a sampling rate of 1 Hz, directly generating digital temperature values and transmitting them to the data processing unit.

[0065] The respiration sensor of this device adopts the respiratory inductive plethysmography method (RIP). By using a coil, the change in the chest / abdominal perimeter caused by respiration is converted into the inductance change of the coil, thereby realizing the detection of the respiration waveform. Compared with other methods, the coil in the RIP method can be better embedded in the device body, improving the wearing comfort.

[0066] The electrocardiogram sensor is composed of an electrocardiogram electrode and an electrocardiogram lead wire.

[0067] The body temperature acquisition will adopt an improved method based on the single-channel heat flux method, with the addition of double-layer heat insulation materials. The single-channel heat flux method consists of two temperature sensors and a prediction algorithm.

[0068] The body temperature acquisition circuit is encapsulated in the body temperature module. The body temperature acquisition circuit adopts the design method of rigid-flexible boards. The body temperature sensor is placed on the rigid board, and the flexible board is equivalent to a connector, and then the acquisition circuit is connected to the host computer.

[0069] Figure 4 It is a schematic diagram of the breathing rate acquisition process and the main structure classification.

[0070] Figure 5 It is a circuit schematic diagram of the respiratory inductive plethysmography (RIP).

[0071] The core of the breathing measurement circuit lies in the oscillation generator. The stability and linearity of the oscillation generator directly affect the quality of the measured breathing waveform. In the past, traditional RIP sensors measured breathing waveforms by using RCL oscillators or Wien bridge oscillators to generate oscillation waveforms for breathing measurement, but these traditional oscillators all have the problem of unstable startup conditions. The present invention uses an LC ring oscillator to well solve the problems of instability and poor frequency linearity of traditional RCL and Wien bridge oscillators.

[0072] The LC ring oscillator of the present invention is a circuit that can make its own signal change according to a fixed period in a self-excited manner, and is composed of an odd number of inverters or differential inverters. Its oscillation frequency is controlled by the delay of the inverter, and of course is also affected by inductors, capacitors, resistors, etc. in the circuit. The performance characteristics of the ring oscillator make it conducive to designing oscillators with high frequency, wide adjustment range, and high linearity. It is precisely the characteristic of the wide adjustment range of the ring oscillator that makes the circuit less affected by temperature, humidity, vibration, sharp changes in inductance coils, etc. during actual use, the circuit starts up stably, and the working state is good.

[0073] The circuit of the present invention also has great advantages in anti-motion interference. The sine wave generated by the LC ring oscillator (A1) is input into a counter. After the counter (A2) counts over, it will cause the level of the monostable resonant circuit to flip, and at the same time, the counter will be automatically reset (CLS pin). And so on, the level of the monostable resonant circuit (A3) keeps flipping, forming a square wave output. When the cross-sectional area of the chest changes due to breathing movement, the inductor coil will be stretched and contracted due to chest movement, resulting in a change in inductance, thus changing the oscillation frequency of the sine wave. While the resonant frequency of the sine wave caused by breathing changes, it also changes the duty cycle of the square wave output by the monostable resonant circuit. Therefore, the breathing signal of the human body can be measured by using the average voltage per unit time of the square wave. This working method can well reduce the influence of motion on the measurement accuracy of breathing, and has a large dynamic range of AC signal input, because the amplitude change of the input signal is converted into a frequency change, thus avoiding saturation caused by too large an AC input signal.

[0074] In addition to the above advantages, the components of this breathing measurement circuit are simple and the quantity is small (5 chips). The circuit design area is only 15*15mm, and the working current is only 1mA, which can well meet the requirements of miniaturization and extremely low power consumption of wearable physiological monitoring devices.

[0075] There are 4 ECG electrodes in this embodiment, and the ECG data of 2 channels are collected.

[0076] The present invention will select ADS1292 of Texas Instruments (Ti) company. It belongs to the ADS129x series, but is different from ADS1298 and ADS1299. It supports 2-lead ECG signal detection, just meeting the requirements of this project, integrating functions such as 24bits ADC, and connecting to the MCU through the SPI interface.

[0077] The data processing unit is used to obtain ECG data through the SPI interface, obtain body temperature data through the IIC interface, and obtain breathing data through the ADC. At the same time, the processor also performs necessary data processing on the obtained data.

[0078] The specific body temperature calculation model is as follows:

[0079] Table 1 Body Temperature Model Coefficients

[0080]

[0081]

[0082] For the first stage (before 1000 seconds):

[0083] T ori= 0.302 * Te - 2.020 * Ts - 0.821 * Ts * Te - 0.011 * Te 2 + 0.030 * Ts 2 + 0.007 * Ts * Te + 0.816 * log

[0084] (Te) + 0.387 * log(Ts) + 0.007 * dif(Ts) - 0.086 * (init_Ts) + 63.467

[0085] where log is the natural logarithm, dif is the forward first-order difference, and init_Ts is the starting value of the Ts temperature sensor.

[0086] For the second stage (after 1000 seconds):

[0087] T ori = 11.400 * Te - 21.798 * Ts + 0.782 * Ts * Te - 0.074 * Te 2 + 0.418 * Ts 2 - 0.197 * Ts * Te - 0.597 * log(Te) + 0.095 * log(Ts) + 0.087 * dif(Ts) + 221.254

[0088] where log is the natural logarithm, dif is the forward first-order difference.

[0089] The specific steps for calculating the respiratory rate are as follows:

[0090] The main steps for calculating the respiratory rate are: downsampling and filtering the respiratory signal, differentiating the signal, determining whether the differentiated signal is a peak point, calculating the interval for the qualified peak points, then finding the average value of the peak point intervals within the window period, and further calculating the respiratory rate.

[0091] The specific steps for calculating the heart rate are as follows:

[0092] Filtering, differentiating, averaging, and downsampling the electrocardiogram signal, finding the maximum value for the downsampled signal, determining whether the maximum value meets the peak point condition, then finding the average value of the peak point intervals within the window period, and further calculating the heart rate.

[0093] Embodiment 2

[0094] Based on the same inventive concept, the present invention also provides a method for detecting the core body temperature of a human body, which uses the described device for detecting the core body temperature of a human body, and the method includes the following steps:

[0095] Collecting the original body temperature through a single-channel heat flux sensor;

[0096] Collecting respiratory rate data through a respiratory sensor;

[0097] Collect heart rate data through an electrocardiogram sensor;

[0098] The data processing unit dynamically adjusts the weights of the original body temperature, heart rate, and respiratory rate, and calculates the pre-core body temperature and the body temperature compensation amount using a two-stage fusion algorithm;

[0099] Correct the pre-core body temperature data with the core body temperature compensation amount to obtain the human core body temperature.

[0100] This algorithm adopts a two-stage fusion architecture, including:

[0101] 1) A multi-physiological parameter fusion algorithm based on deep learning or a deep neural network based on static physiological parameters (static model)

[0102] 2) A dynamic compensation algorithm for heart rate, respiratory rate, and pre-calculated core body temperature or a weight adaptive mechanism based on change rate perception (dynamic compensation)

[0103] Specific implementation process of the algorithm:

[0104] (1) Multi-source data synchronization

[0105] Heat flux body temperature (1Hz) → Original sampling

[0106] Respiratory rate (0.1Hz) → Linear interpolation to 1Hz

[0107] Heart rate (0.166Hz) → Cubic spline interpolation to 1Hz

[0108] (2) Static model algorithm

[0109] 1) Standardization processing

[0110]

[0111] where x i is the set of all physiological parameters participating in the calculation of the pre-core body temperature, specifically including the original body temperature (T ori ), respiratory rate (RR), and heart rate (HR), μ i is the mean value of each feature, σ i is the standard deviation, and x′ i is the original body temperature (T ori ), respiratory rate (RR), and heart rate (HR) after standardization processing.

[0112] 2) Feature selection

[0113] Construct a second-order polynomial feature space:

[0114] Φ(X) = [1, T ori , RR, HR, Tori 2 , RR 2 , HR 2 , T ori *RR, T ori *HR, RR * HR

[0115] Φ(X) is composed of T ori , HR, RR and their quadratic terms and cross - product terms, and captures the non - linear interaction between physiological parameters through feature crossing.

[0116] 3) Model selection

[0117] Adopt the Ridge Regression model:

[0118] min w ||Y - Φ(X)w|| 2 + α||w|| 2

[0119] Where:

[0120] Y is the true value of core body temperature; w is the parameter vector to be solved; α is the regularization coefficient (determined by cross - validation). Regularization coefficient α: Optimize in the logarithmic space of [0.01, 100] through 5 - fold cross - validation.

[0121] 4) Final formula

[0122] The core body temperature prediction formula obtained through training is as follows:

[0123] T pre = w o + w1T ori + w2RR + w3HR + w4T ori 2 + w5RR 2 + w6HR 2 + w7T ori RR + w8T ori HR + w9RRHR;

[0124] Where:

[0125] T pre is the predicted value of core body temperature;

[0126] T ori is the original body temperature calculated by the single - channel heat flux method;

[0127] RR is the respiratory rate;

[0128] HR is the heart rate;

[0129] w0 is the reference offset, and its value range is 36.5 - 37.2;

[0130] w1 is the body surface temperature weight, and its value range is 0.6 - 0.9;

[0131] w2 is the respiratory rate weight, and its value range is 0.01 - 0.05;

[0132] w3 is the heart rate weight, and its value range is 0.001 - 0.005;

[0133] w4 is the square term coefficient, and its value range is on the order of 1e - 2;

[0134] w5 is the square term coefficient, and its value range is on the order of 1e - 3;

[0135] w6 is the square term coefficient, and its value range is on the order of 1e - 4;

[0136] w7 - w9 are the interaction term weights, usually on the order of 1e - 4.

[0137] (3) Dynamic compensation algorithm

[0138] Step 1: Parameter change rate calculation

[0139] The sliding window difference method is adopted (window length w = 5 seconds):

[0140]

[0141] where x ∈ {T ori , RR, HR}, x (t-i) is T ori , RR, HR at time (t - i), x (t-i-1) is T ori , RR, HR at time (t - i - 1), Δx (t) is the difference value calculated by the sliding window method at time t, retaining the sign information:

[0142]

[0143] Step 2: Standardization processing

[0144] Dynamic z - score standardization is adopted:

[0145]

[0146] is the mean value of the previous t moments, is the standard deviation of the previous t moments.

[0147] Step 3: Weight calculation

[0148] Introduce the normalized exponential function (Softmax function) to calculate T ori , the weights of three physiological parameters RR, HR, and HR participating in compensation:

[0149]

[0150] Among them, the numerator is the natural logarithm of the compensation parameter to be calculated, and the denominator is the sum of the natural logarithms of the three compensation parameters.

[0151] Step 4: Compensation amount calculation

[0152] Compensation formula considering the directionality of change:

[0153]

[0154] Among them:

[0155] α is the global learning rate (default 0.1);

[0156] Rx is the physiological range of the parameter (T pre : 2°C, RR: 30 bpm, HR: 100 bmp).

[0157] (4) Final result of core body temperature

[0158]

[0159] is the body temperature compensation amount at time t.

[0160] The core body temperature detection device and method of the present invention can non-invasively and continuously monitor the core body temperature of the human body. By using the method of multi-source data fusion, the disadvantages of single-parameter estimation can be better avoided; at the same time, through the method of dynamic compensation, the weights of the pre-core body temperature, heart rate, and respiratory rate measured based on a single channel are adjusted, enhancing the problem of poor adaptability of estimating the core body temperature by a single method due to different manifestation methods of the core body temperature change in different individuals.

[0161] The necessary technical content not mentioned in the above embodiments is the prior art and shall be subject to the well-known technical content, so it will not be elaborated here.

[0162] The above are only embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention are included within the scope of the claims of the present invention.

Claims

1. A human body core temperature detection device, characterized in that, The device includes: a fixing belt, and a single-channel heat flux method sensor, a respiration sensor, an electrocardiogram sensor and a data processing unit fixed on one side of the fixing belt; the single-channel heat flux method sensor, the respiration sensor and the electrocardiogram sensor are respectively electrically connected to the data processing unit.

2. The human core body temperature detection device according to claim 1, characterized in that The single-channel heat flux method sensor includes: a heat conductor, a skin temperature sensor fixed at the bottom of the heat conductor, a microenvironment temperature sensor fixed at the top of the heat conductor, and a heat insulator covering the periphery and top surface of the heat conductor.

3. The human core body temperature detection device according to claim 1, characterized in that, The respiration sensor includes an inductive respiration sensor.

4. The human core body temperature detection device according to claim 3, characterized in that The inductive respiration sensor includes an elastic band and two continuously bent wires fixed along the length direction of the elastic band.

5. The human core body temperature detection device according to claim 1, characterized in that, The electrocardiogram sensor includes a plurality of electrocardiogram electrodes.

6. A method for detecting the core body temperature of a human body, characterized in that, Using a human core body temperature detection device as described in any one of claims 1-5, the method includes the following steps: Collect the original body temperature data T through the single-channel heat flux method sensor ori ; Collecting respiration rate data RR through the respiration sensor; Collecting heart rate data HR through the electrocardiogram sensor; The data processing unit dynamically adjusts the weights of the original body temperature T ori , heart rate HR, and respiratory rate RR, and calculates the pre-core body temperature and the body temperature compensation amount using a two-stage fusion algorithm; Correcting the pre-core body temperature data with the core body temperature compensation amount to obtain the human core body temperature.

7. The method for detecting the core body temperature of a human body according to claim 6, wherein The two-stage fusion algorithm includes: a multi-physiological parameter fusion algorithm based on deep learning; a dynamic compensation algorithm based on the perception of the change rates of heart rate, respiration rate, and pre-core body temperature.

8. The method for detecting the core body temperature of a human body according to claim 7, wherein, The multi-physiological parameter fusion algorithm based on deep learning includes: a ridge regression model.

9. The human core body temperature detection method according to claim 7, wherein The formula of the multi-physiological parameter fusion algorithm is as follows: T pre = w o + w1T ori + w2RR + w3HR + w4T ori 2 + w5RR 2 + w6HR 2 + w7T ori RR + w8T ori HR + w9RRHR; Where: T pre is the pre-core body temperature; T ori is the original body temperature calculated by the single-channel heat flux method; RR is the respiration rate; HR is the heart rate; w0 is the reference offset, and its value range is 36.5-37.2; w1 is the original body temperature weight, and its value range is 0.6-0.9; w2 is the respiration rate weight, and its value range is 0.01-0.05; w3 is the heart rate weight, and its value range is 0.001-0.005; w4 is the square term coefficient, and its value range is of the order of 1e-2; w5 is the square term coefficient, and its value range is of the order of 1e-3; w6 is the square term coefficient, and its value range is of the order of 1e-4; w7-w9 are the interaction term weights, usually of the order of 1e-4.

10. A method for detecting the core body temperature of a human body according to claim 7, characterized in that, The dynamic compensation algorithm includes the following formula: Where: is the core body temperature compensation amount; α is the global learning rate, which controls the scaling amplitude of the compensation amount; x is the set of all physiological parameters participating in the compensation, specifically including the original body temperature, respiration frequency (RR), and heart rate (HR); is a weight function; sgn(x): the sign function, indicating the change direction (positive for increase, negative for decrease); Δx (t) : The change in the parameter x at the current time t; Rx: the physiological safety range of the parameter (to avoid exceeding the physiological limit).

Images