Detailed Description
The present application will be described in detail with reference to the accompanying drawings.
The core body temperature monitoring experimental equipment comprises two parts: a body temperature simulation generator and a core body temperature measuring device; the core body temperature measuring device is arranged on the peripheral wall of the barrel container of the body temperature simulation generator to perform experiments.
Body temperature simulation generator
A simulated body temperature generator for use in the present application, comprising: a drum container 11, a first water pump 12a, a second water pump 12b, and a constant temperature water tank 13.
The drum container 11 is made of ABS plastic; the drum container 11 is filled with an aqueous medium; the end of the drum container 11 is closed by a heat insulating material.
The constant temperature water tank 13 is arranged outside the barrel container 11, and a first pipeline and a second pipeline are formed between the constant temperature water tank 13 and the barrel container 11; the first and second lines are each formed of an insulated hose 14; the first water pump 12a is installed in the first pipeline, and the second water pump 12b is installed in the second pipeline; the first line is for supplying an aqueous medium from the thermostatic water bath 13 to the drum container 11, and the second line is for supplying an aqueous medium from the drum container to the thermostatic water bath, thereby constituting water circulation between the thermostatic water bath and the drum container.
A temperature measuring unit 15 is arranged in the barrel container 11. The temperature measuring unit 15 is a thermistor to measure the temperature of the aqueous medium in the drum 11, which is the simulated core body temperature.
The inner wall or the outer wall of the barrel container 11 is provided with a radiation-proof layer to prevent heat loss caused by heat radiation, so that the simulated core body temperature is stable.
The first water pump 12a and the second water pump 12b are installed in the heat insulation box body, so that heat loss caused by the water pumps is avoided as much as possible.
The drum container 11 comprises a plurality of thermometry locations 11a, 11b, each of which corresponds to a different wall thickness. The wall thickness corresponding to one of the temperature measuring positions is 11mm.
The barrel container 11 had a height of 250mm, an inner diameter of 80mm and a bottom thickness of 25mm.
The constant temperature water tank is controlled to generate constant heat, the water in the constant temperature water tank is sent to the barrel container through the first water pump, the water in the barrel container is sent to the constant temperature water tank through the second water pump, and heat balance is formed between the barrel container and the constant temperature water tank, so that a uniform and stable body core heat source is generated in the body temperature generator, the body core heat source is a body core heat source for simulating a human body, the heat of the body core heat source is conducted to the surface of the barrel container through the barrel wall, and the surface temperature changes to reflect the change of the internal body core temperature, and the surface temperature is used for simulating the body surface temperature of the human body. The barrel wall of the barrel container is used for simulating the thermal resistance of skin (fat, muscle, epidermis and the like) with a certain thickness of human body. The temperature measuring positions corresponding to different wall thicknesses represent the thermal resistances of the skins with different thicknesses.
The calculation process of the barrel wall thickness is as follows:
The human skin heat conductivity coefficient lambdas =0.47W/(m.k), the fat heat conductivity coefficient lambdaf =0.21W/(m.k), the thickness of normal abdomen male skin fold (skin+subcutaneous fat) is 5-15 mm, and the thickness of female skin fold is 12-20mm. We take the average skin thickness δ=10mm, where skin thickness δs =2mm, fat thickness δf =8mm, according to the steady state phase of conduction the total thermal resistance is equal to the sum of the series thermal resistances, equation 1:
Substituting the data to calculate the average heat conductivity coefficient of the skin wrinkles is lambdaa =0.236W/(m.k). The barrel wall of the body temperature simulation generator is made of ABS material, and the heat conductivity coefficient lambdaabs = 0.2512W/(m.k) of the ABS material is equal to the thickness of the ABS material with equivalent 10mm skin wrinkling thickness according to the principle of equal heat resistanceThe body thickness of the barrel wall of the body temperature simulation generator is 11mm. The thickness of other temperature measuring positions can be 9mm, 10mm, 12mm, 13mm and the like.
In order to meet the experiment requirement, save materials and facilitate transportation, the overall size is set to be 250mm in barrel height, 80mm in inner diameter and 25mm in thickness of an ABS base, and heat insulation materials are added for bedding in the experiment to reduce uneven barrel wall temperature caused by bottom heat loss. The top of the barrel container is provided with a sealing ball valve with the thickness of 10mm, and a unidirectional heat insulation soft pipe hole and an exhaust hole are reserved. The heat transfer medium in the barrel container is water, and a 12V 4W water pump is arranged in the heat insulation box body, so that heat loss caused by the water pump is avoided. Stirring the water flow to make the temperature in the generator uniform. The temperature control precision of the constant temperature water tank is within +/-0.05 ℃, and a constant temperature curve is drawn to test the temperature control capability.
Under the condition that the core body temperature is stable, different temperature measuring parts are easy to cause measurement errors due to the fact that the thickness of skin folds is changed. Therefore, it is necessary to provide different wall thicknesses on the body temperature generator to simulate the change of the temperature measuring part of the human body or the difference of the thickness of different skin folds of the human body. In addition, different wall thicknesses can also facilitate a plurality of temperature measuring devices to carry out multipoint measurement, and parallel experiments are carried out to accelerate the calibration of the thermometer.
Core body temperature measuring device
The core body temperature measuring device includes: the device comprises a containing element 7, a first thermistor 4, a second thermistor 3, a micro-electric heating sheet 6, a control unit 81, a temperature detection unit 82 and a shell 1.
The receiving element 7 is made of an anisotropic material with a thermal conductivity between 0.05 and 0.3W/mK, preferably a Polyetheretherketone (PEEK) material, which is also easy to injection mold while ensuring a low thermal conductivity. Polyethylene (PE), polymethyl methacrylate (PMMA), polycarbonate (PC), polysulfone (PSU), or the like may also be selected. From the first end to the second end of the containing element 7, a plurality of ducts are formed, which are closed at the first end and the second end, respectively, and contain air or foam (between 0.01 and 0.1W/mK) inside them. The pipeline can reduce radial heat conduction and reduce radial heat conduction errors. The tubing allows the containment element to have significant anisotropy, which is beneficial to provide a radial conduction level lower (2-20 times) than the axial conduction, ensuring accuracy of the measurement.
The first thermistor 4 is arranged at a first end of the accommodating element 7 and is used for measuring the body surface temperature; the second thermistor 3 is arranged at a second end of the housing element 7 for measuring the ambient temperature; a micro-hotplate 6 is arranged in the vicinity of the second thermistor for generating a disturbance temperature, which actively simulates the change in ambient temperature.
The control unit 81 controls the micro electric heating plate 6 to work; the first thermistor 4 and the second thermistor 3 are connected to a temperature detection unit 82 through a switch circuit 84, and the temperature detection unit 82 is connected to a control unit 81; the control unit 81 may be connected with a bluetooth unit 83 to communicate with the outside. Battery 85 provides power to the entire circuit.
The control unit adopts 36-pin STM32F103TBU6, and the chip has the advantages of less pins, small volume, lower power consumption, lower price, mature technology and stable performance. The temperature detection unit adopts MAX31865, the chip has high integration level and high precision, the theoretical precision is 0.003, and the actual measurement precision is about 0.008. The switching circuit is an analog switch for switching the inputs of the two sensors. The Bluetooth module adopts a mature Bluetooth 4.0 module, and the power consumption is lower.
The shell 1 is made of a heat insulating material, is formed into a cylinder shape, is opened at a first end and closed at a second end, and is used for thermally isolating the interior of the core body temperature measuring device from the external environment, so that the disturbance of the passive environment temperature is as small as possible; the housing element 7, the first thermistor 4, the second thermistor 3, the micro-heat sink 6, the control unit 81, and the temperature detection unit 82 are enclosed in the casing 1, and thermally isolated from the outside as much as possible.
The plurality of pipes of the accommodation element 7 are of a straight line shape or a curved line shape. The curved channel can be formed such that the height of the receiving element 7 is as small as possible while ensuring the channel length.
The first ends of the plurality of tubes are concentrated around the first thermistor 4; the second ends of the plurality of tubes are concentrated around the second thermistor 3 so that the influence of the ambient temperature on the body surface temperature is as small as possible. The plurality of tubes are equally spaced and parallel to one another.
A biocompatible layer is provided at a first end of the containment element 7; the biocompatible layer closes the first end of the housing 1. The biocompatible layer in the present application may be the above biocompatible colloid. The biocompatible layer is arranged under the circuit structure and is used for adhesively connecting the human skin with the core body temperature measuring device. The biocompatible layer is tightly attached to the heat-insulating shell, so that the requirement of long-term monitoring can be met while the severe temperature disturbance of the isolated environment is ensured.
A good heat conducting layer 5 is connected between the second thermistor 3 and the micro-hotplate 6 so that the active environmental disturbance applied by the micro-hotplate 6 can be quickly reflected on the second thermistor 3.
The receiving member 7 is located at a central position of the housing 1, and a predetermined interval is formed between an inner surface of the housing 1 and an outer surface of the receiving member 7. The inside of the housing 1 may be evacuated of air, further avoiding the influence of external temperature on the temperature inside the housing.
A control unit 81, a temperature detection unit 82, a bluetooth unit 83, and a switching circuit 84 are provided on the circuit board 8.
The circuit board 8 is located at a first end of the receiving element 7 and covers the first thermistor 4.
The first thermistor 4 is arranged in parallel with the second thermistor 3.
The measurement of the thermistor was calibrated using a Fluke thermometer as a reference device. Calibration experiment test records are shown in table 1. The error value is between about 0.003-0.01 c, and is substantially stabilized to an accuracy of 0.01 c using kalman filtering. The temperature measurement value is transmitted to the upper computer through Bluetooth in a wireless mode.
Table 1 temperature sensing reading calibration experiment table
Annotation: the values were read 10 times at each bath temperature and were found to be mean+ -sd.
Establishing a dual-channel heat flow model
Assuming that the heat transfer coefficient of human tissue is kg approximately 45W/m2 K, the axial heat transfer coefficient kv approximately 50W/m2 K of the structured containing element 7, the heat transfer coefficient ks approximately 401W/m2 K of the good heat conducting layer 5, the measuring device detects the near-body skin temperature Ts, the far-body environment temperature Te, wherein the main heat flow enters the heat flow insulator from the body tissue and flows from the near-body temperature sensing element to the far-body temperature sensing element, the human core temperature Tcore can be calculated by equation 2.
Establishing a variable parameter thermal model
In brief, the physiological temperature field consists of an internal temperature Tc, an epidermal temperature Ts, an ambient temperature Te, and a correlation between the three.
The internal temperature field mainly comprises substances (skin, fat and the like) with similar heat source substances and thermal properties generated in body surface tissues, and on one hand, the internal temperature field comprises energy conversion efficiency (physiological state changes such as feeding, exercise and the like) in the process of converting chemical energy of the internal substances into heat energy, and according to different physiological state levels, the energy provided by chemical reactions of different substances is uneven; on the other hand, when physiological activities are in a normal state, the deep tissues and the body surface tissue structures are highly consistent with the thermal characteristics of the heat insulating materials in thermal characteristics, and the two parts jointly reflect the structure with the most sensitive internal temperature change, namely the composition of an internal temperature field. The skin temperature field is determined by the skin surface and the skin structure shape, the thermal properties of different skin colors, hairs and parts show larger difference, the heat absorption and heat conduction efficiency shows the thermal property of the skin, and the external shape measuring and calculating method of the skin can be specifically designed according to the placed part, the hair length and the placed surface area combined with the internal temperature field, so that the aim is to reduce the contact surface area as much as possible and the heat dissipation efficiency so as to obtain more accurate temperature measurement values. The external temperature field is determined by the external environment, and the external environment factors including the air temperature, the humidity and the external air flow rate affect the heat transfer rate to different degrees, so that the external temperature field distribution is affected.
The physiological temperature field analysis is integrated, and factors influencing the change of the temperature field mainly comprise the change of an internal central thermal characteristic area, uneven internal heat conduction rate, uneven heat dissipation at different positions on the skin surface, heat dissipation rate influenced by external environment and the like; in addition, the physiological temperature field changes to some extent can be numerically represented as changes in internal temperature Tc, skin temperature Ts, and ambient temperature Te. In conclusion, different physiological thermal models can be obtained by processing the temperature field variation influencing factors in different modes.
Thus, after analysis of the physiological temperature field, the physiological Lumped Parameter (Lumped Parameter) thermal model is currently the main model used to describe the thermal behavior in physiological states and to enable estimation of internal temperature. The model mainly comprises a body core heat generating source, internal heat conduction, surface heat absorption, external heat transfer and environmental temperature compensation, wherein a physiological temperature field is described by using a thermoelectric similarity principle, the whole temperature field is equivalent to an internal heat capacity, a heat generating rate and internal and external heat conduction rates which are expressed by thermal resistance, and the lumped parameter thermal model is shown in figure 4. In fig. 4, Q is an internal heat source, cc and Cs are the internal heat capacity and the surface heat capacity, respectively, and Ri and Ro are the internal thermal resistance and the external thermal resistance, respectively; in the physiological lumped parameter thermal model, the thermal model parameters Cc, cs, ri and Ro are processed in a constant manner, and various factors affecting the physiological temperature field change cannot be sufficiently considered and properly processed.
In order to solve two major problems of the lumped parameter model: 1) Ri is different according to individual differences and physiological states of factors such as body surface fat, muscles and the like of a human body, and the Ri needs to be measured and calculated in a personalized way; 2) The Ts variation is simultaneously affected by both Tc and Te variations, requiring consideration of the dynamic switching of the temperature field. If Te is maintained constant, the variation of Ts directly reflects the dynamic variation of Tc, and therefore, it is necessary to remove the influence of the ambient temperature variation from the variation of Ts.
To achieve accurate estimation of the physiological internal temperature, it is most important to build a reasonable physiological thermal model. According to Semenov combustion theory, the thermal model regards the internal temperature of the object as a uniform and periodic and stable structure, which is a scientific method suitable for analyzing and describing the internal heat distribution of the object, and the theoretical model is one of the main models currently used for simulating and estimating the internal temperature of the object, so the invention regards the internal temperature of the physiology as uniform steady-state variation.
Meanwhile, on the basis of the basic structure of the dual thermistor shown in fig. 1, a micro electric heating plate 6 is additionally arranged outside the accommodating element 7 and the second thermistor 3, the change of the ambient temperature is simulated through the micro electric heating plate, the ratio of Rc to Re is obtained, personalized correction is realized, and the quantized coefficient of the influence of the ambient temperature on the skin temperature is obtained.
Firstly, analyzing the path of an environmental heat source (micro electric heating plate), wherein the main effect is that the Joule heat generated by the equivalent ohmic internal resistance in the resistance wire is generated, and other heat generation and heat absorption in the environmental heat field can be ignored relative to the Joule heat, so that a heat generation rate calculation formula of a formula 3 is adopted. The heat transfer between physiology and external environment is mainly carried out by a heat exchange mode, and the heat transfer between physiology and internal environment is mainly carried out by a heat conduction mode.
Qh=i2 r (formula 3)
Wherein QH is the heat generation rate (unit: W) of an environmental heat source (micro electric heating sheet), I is the working current (unit: A) of the micro electric heating sheet, and r is the equivalent ohmic internal resistance (unit: omega) of the resistance wire.
Secondly, factors such as internal central area change, internal heat conduction rate non-uniformity, heat dissipation non-uniformity at different positions of the surface skin, heat dissipation rate influenced by external environment and the like are comprehensively considered, and the current thermal model is optimized. The study treated physiological internal temperature as equivalent homogeneous and there was some degree of variation in this region, and in experiments to show the effect of this variation on internal temperature, internal heat capacity was treated as variable. Meanwhile, the internal heat conduction rate and the surface heat dissipation rate have larger variation along with the position change, and although certain variation exists, the two have the main effect of heat transfer from the whole perspective, so that the two are uniformly and equivalently treated respectively in the modeling of the two in the research. In addition, to reflect the special change of the external environment, the external heat transfer rate is considered to be uniform and variable, and meanwhile, the influence on the core temperature is measured according to the heat flow of the electric heating plate. The problem of inconsistent heat capacity of the epidermis caused by different skin types is solved, and the equivalent unified treatment of the heat capacity of the epidermis is also realized, so that the modeling is convenient.
Based on analysis of physiological internal heat generation mechanism, heat transfer mechanism and classical lumped parameter thermal model, the improved content of the research is combined, so that the physiological thermal model can be constructed to be close to a real thermal model as much as possible in order to reflect the influence of environmental temperature on the thermal model parameters. Meanwhile, the invention builds a simplified variable parameter thermal model integrating variable internal heat capacity and external thermal resistance models for reducing the complexity of the model and increasing the applicability, as shown in figure 5.
Cc and Cs in FIG. 5 are the internal heat capacity of the battery and the surface heat capacity of the battery (unit: J/K), respectively, and represent the amount of absorption and heat dissipation capacity of the substance; tc, ts and Te are respectively physiological internal, surface and environment temperature (unit is DEG C); ri represents the physiological internal heat conduction rate for the internal thermal resistance; ro is the external thermal resistance that characterizes the rate of heat exchange of the skin surface to the environment (in K/W); q is an internal heat source (physiological heat generation rate); QH is an external heat source (micro-hotplate heat generation rate). Based on a simplified variable parameter thermal model, a first-order RC filter network is analogized by a battery internal and external energy conservation equation, a Fourier heat law and a Newton heat dissipation law, and a system characteristic equation is shown as a formula 4 and a formula 5:
to reflect the effect of Ta on Cc and Ro, the internal heat capacity and external thermal resistance were modeled with equation 6:
To sum up, a core temperature estimation formula is obtained:
And the thermal model parameter measurement and calculation comprises an equivalent ohmic internal resistance r, an individuation correction coefficient P, an environment temperature correction coefficient alpha and an environment temperature Te to internal heat capacity Cc and external thermal resistance Ro relation equation. After a reasonable physiological thermal model is established, accurate identification of thermal model parameters and micro-electric heating plate characteristic parameters is a key for realizing core temperature estimation, and parameters are modeled on the basis, so that a parameter model mechanism is analyzed.
Equivalent ohmic internal resistance modeling r
Errors in thermal resistor design are mainly due to lead resistance, self-heating effects, non-linear errors, and minor errors on some circuits. In order to improve the accuracy of estimating the heat production rate of an environmental heat source, the equivalent ohmic internal resistance of the micro electric heating plate is further measured and modeled by considering the nonlinear influence of the environmental temperature on the internal ohmic internal resistance, meanwhile, in order to avoid the nonlinear temperature measurement influence of a thermosensitive temperature measurement element, experiments are limited to narrower heating temperature, and in order to simplify calculation and enable the fitting result to be more approximate to the real internal resistance characteristic, the fitting relation is as follows:
r=l×te+m×Te2+n×R+o×R2+z×Te ×r+y+ε (formula 7)
L, m, n, o, z in the formula 7 is a coefficient to be fitted, y is a constant term, and ε is an error.
Designing an experiment, measuring and calculating at the temperature range of 25-30 ℃, and building a multiple regression model for fitting.
The regression fit equation is:
r= -9.7441 ×te-0.0377×Te2+0.2399×R-0.0002×R2+0.0107×Te ×r (formula 8)
Confidence intervals are respectively [-23804.2019862140,4316.03520634591],[-92.1913388073052,16.7386726429827],[-96.0680748222936,575.938694895014],[-0.575625139000544,0.0960727813991399],[-4.68413278354100,26.0187150992556],[0,0].
Table 2 thermistor reading calibration experiment table
Individual correction coefficient P and ambient temperature correction coefficient alpha measurement and calculation
The skin temperature sensor steady-state measurement Ts is affected by four factors, namely, the body core temperature Tc, the ambient temperature Te, the thermal insulation material physical properties (thermal resistance Ro, density ρ and specific heat capacity c) and the body surface tissue equivalent thermal resistance Ri.
The transient thermally conductive circuit model is equation 9:
the steady state thermal circuit model is equation 10:
keeping the core temperature Tc unchanged, changing the ambient temperature Te to Te+DeltaTe by heating the micro-electric heating plate, and recording the variation DeltaTs of Ts, wherein the individuation correction coefficient is shown as formula 11:
the ambient temperature correction factor is equation 12:
The experiment measurement can obtain the individuation correction coefficient P, and meanwhile, the numerical table or the relation function of Cs (heat capacity of the heat insulating material) at different temperatures can be obtained by referring to the data.
The core temperature Tc was designed to be 35 ℃, the ambient temperature Te was heated by a micro-hotplate, and the variation DeltaTs of Ts was recorded, and the experimental record is shown in Table 3.
TABLE 3 variation of skin temperature Ts with ambient temperature Te with core temperature Tc maintained at 35℃
Modeling the relationship of internal heat capacity Cc and external thermal resistance Ro by ambient temperature
As can be seen from analysis of the internal temperature equilibrium, when the physiological internal temperature is in a thermal steady state, the internal temperature change rate is approximately zero, and first, the internal and external thermal resistances Ri and Ro at different environmental temperatures Te are identified, and the flow is shown in FIG. 6.
From Cs (heat capacity of the heat insulating material) and the individualization correction coefficient P (i.e., the ratio of the internal resistance and the external resistance, R0) at different ambient temperatures Te have been obtained. Based on the method, the thermoelectric parameters are identified by adopting a least square method through a transient process of internal temperature change before thermal steady state and combining with a formula 6. Values of internal heat capacity and external thermal resistance at different ambient temperatures Te are recorded as shown in table 1 and non-linear fit was performed as equation 13:
g. s, u, j, q, h are coefficients to be fitted.
The fitting equation is:
TABLE 4 calculation of tissue heat capacity and external thermal resistance at different Te' s
Mechanism analysis of internal heat capacity Cc and external thermal resistance Ro
According to the built variable internal heat capacity model, cc changes along with Te, and the characteristics of inconsistent internal temperature rising rate and obvious change of the ratio of Tse to Tcs are shown under different Te, because when Te is low, the composition range of the similar region of the internal central thermal characteristic is smaller, so that the internal structure heat capacity of the physiological epidermal tissue is smaller, and along with the rising of Te, the composition range of the similar region of the internal central thermal characteristic is larger along with the rising of Te, and the characteristic along with Te changes reflects that Te has some influence on Cc and cannot be simply reduced by a fixed value.
The heat exchange rate of the heat insulating material of the dual temperature sensor is represented by external thermal resistance, wherein the heat conducting medium (air molecules) around the surface has larger difference in density due to different Te, so that the external heat exchange rate shows non-uniform characteristics due to temperature layering at the periphery of the surface, and when Te is lower, the heat dissipation rate of the heat conducting medium around the surface is lower, and the heat dissipation rate tends to change in the opposite direction along with the rising of Te. For the convenience of calculation, and as much as possible, the dynamic nonlinear relation of Cc and Ro changing along with Te is shown, cc and Ro are equivalent to be uniform and variable, and equation 7 is used for fitting the micro-electric heat transfer rate (equivalent ohmic internal resistance).
Establishing an extended Kalman filter estimation model and determining a dynamic measurement algorithm of a core body temperature
An internal temperature estimation algorithm is established, a variable parameter thermal model established based on the steps is combined with a dynamic characteristic equation of a double temperature sensor and the complexity of the whole physiological temperature measurement system, and a dynamic measurement algorithm of a core body temperature based on an Extended KALMAN FILTER (EKF) is adopted.
Modeling of state equations and observation equations
The internal temperature estimation algorithm based on EKF needs to accurately describe the variable parameter thermal model system characteristics of the physiological temperature measurement system by using a state space model. To embody the dynamic characteristics of the variable parametric thermal model system, a state equation (equation 15) is established according to equation 4:
the core temperature (Tc) is used as a system state variable, and the micro-electric heating plate current (I) is used as a system input
The observation vector of the dynamic system is the basis of the adaptive filtering of the system, the accuracy and the reliability of the system are directly influenced, the skin temperature (Ts) and the environment temperature (Te) are taken as the observation vector in combination with the formula (6), and meanwhile, for representing the influence of Tc and Te on the variation of Ts, the observation equation is expressed by the formula 16:
In formulas 15 and 16, tc (t) and Ts (t) are Tc and Ts at time t, tc (0) and Ts (0) are Tc and initial values (reference amounts) of Ts, I is the working current of the micro-electric heating sheet, and r is the equivalent ohmic internal resistance of the micro-electric heating sheet.
EKF discretization
In practical application, the built continuous dynamic system model is often subjected to discretization in advance, and because the discrete model has the special advantages, the discrete model is convenient for realizing operation on a computer, and the discrete model can describe the dynamic characteristics of the system more intuitively. Based on the above analysis, discretizing the state equation, taking model errors into account, adding process noise, and substituting the variable internal heat capacity equivalent model built in the previous step into formula 15 to obtain a discretized state equation (formula 17):
Secondly discretizing a system observation equation, simultaneously adding observation noise caused by observation errors, and bringing the variable external thermal resistance equivalent model built above into a formula 16 to further obtain a discretized observation equation (formula 18):
In formulas 17 and 18, g, s and u are variable internal heat capacity fitting coefficients; l, m, n, o, z, y and epsilon are equivalent ohmic internal resistance fitting coefficients; j. q and h are variable external thermal resistance fitting coefficients; Δt is the sampling time(s); wk、vk is independent zero-mean Gaussian white noise; k-1 and k are the k-1 and k calculation sequence numbers; tc (k-1) and Ts (k-1) are the internal temperature and the surface temperature calculated in the kth time, and Tc (k-1) and Ts (k) are the internal temperature and the surface temperature calculated in the kth time.
EKF algorithm flow
Specifically, xk is the kth calculation Tc (k), uk is the k time input ik,yk is the kth calculation Ts (k)
Definition:
Setting a filtering initial value:
setting a variance initial value:
State prediction:
Error covariance prediction:
Gain matrix:
And (5) updating the state:
error covariance update:
unit time delay: k=k+1
Temperature sensor temperature measurement accuracy test
The core body temperature is designed to fluctuate from 26-46 ℃ by using the body temperature simulation generator, the estimated value of the core body temperature by the core body temperature detection device is recorded, and Bland-Altman is drawn as shown in figure 7. Observing a 95% consistency limit (95%limits of agreement,95% LoA), it is believed that these two methods have better consistency when the vast majority of the differences lie within this interval. The average value of the temperature measurement difference values of the core body temperature detection device in the temperature range of 26-46 ℃ is-0.14 ℃, the consistency limit is 0.41 ℃, and the core body temperature detection device has higher temperature measurement precision.
Unless defined otherwise, all technical and/or scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application relates. The materials, methods, and examples mentioned herein are illustrative only and not intended to be limiting.
Although the present application has been described in connection with specific embodiments thereof, those skilled in the art will appreciate that various substitutions, modifications and changes may be made without departing from the spirit of the application.