Method for comprehensive evaluation of thermal tolerance by fusing multi-dimensional variable data

Abstract

The present application relates to the technical field of endurance evaluation, and discloses a thermal endurance comprehensive evaluation method fusing multi-dimensional variable data, comprising: collecting multi-dimensional variable data of an evaluation object, and calculating integral skin conductance; performing extreme value normalization processing on the multi-dimensional variable data, and constructing a composite environmental stress index, a standardized load index, a physiological stress index and a psychological subjective score based on the normalized data and the integral skin conductance; constructing an individual thermal endurance score model by using the four indexes, calculating an individual thermal endurance comprehensive score, and comparing the comprehensive score with a threshold to output a thermal endurance grade. The present application incorporates high-altitude parameters and hypoxia parameters into model calculation, adapts to a highland composite thermal environment, combines with military special operation load characteristics, fuses four-dimensional variables of environment, load, physiology and psychology, makes the model adapt to a combat scene, provides a clear grading conclusion, and supports quantitative evaluation and troop deployment.

Description

Technical Field

[0001] This invention relates to the field of endurance assessment technology, specifically to a comprehensive thermal endurance assessment method that integrates multidimensional variable data. Background Technology

[0002] Heat tolerance refers to the overall level of human endurance in maintaining normal physiological functions and work capacity under thermal stress. In the field of military operations, the physical environment in which missions are carried out is highly complex, ranging from plains to high-altitude areas. The heat tolerance level of personnel directly affects the efficiency of military missions and personnel safety. With the development of military medicine and environmental physiology, heat tolerance assessment has evolved from early single-dimensional early warning relying on conventional meteorological parameters to a stage of comprehensive analysis combining multidimensional environmental characteristics and dynamic changes in human physiological signs. Currently, assessment variables are gradually expanding to include multiple dimensions such as environmental characteristics, workload, physiological state, and psychological perception, aiming to map an individual's comprehensive physiological and psychological tolerance under complex stress environments through quantitative mathematical models and parameter indicators.

[0003] Current technologies and methods in this field primarily revolve around conventional meteorological parameters and basic physiological data, forming two main technical approaches. The first is an assessment scheme based on the environmental stress index, which relies on basic environmental variables such as ambient temperature, relative humidity, wind speed, and mean radiant temperature to construct a thermal index model for risk prediction. The second is a physiological assessment approach based on single-indicator monitoring, which utilizes sensors to continuously monitor physiological parameters such as heart rate, body surface temperature, or estimated core temperature, and performs state assessment by comparing these parameters to fixed physiological safety thresholds. Both of these mainstream approaches typically involve data calibration and model validation in conventional environments and ordinary work scenarios, meeting the needs of routine heat stress assessment in basic scenarios.

[0004] However, existing technologies are still imperfect in extreme military combat scenarios. The main problem lies in the fact that the underlying architecture of existing assessment models is mostly applicable to conventional plains scenarios, failing to incorporate the hypoxia sensitivity effect caused by high altitude into the quantification process of environmental stress. Furthermore, they fail to comprehensively reflect the combined interaction between the total weight of combat gear, mission metabolic equivalent, and extreme operational time. This leads to serious distortion of assessment results and an inability to accurately guide military operations in high-altitude or high-intensity combat scenarios. Therefore, a comprehensive thermal endurance assessment method integrating multi-dimensional variable data is urgently needed to solve these problems. Summary of the Invention

[0005] To address the problems in related technologies, this invention provides a comprehensive thermal endurance evaluation method that integrates multidimensional variable data, thereby overcoming the aforementioned technical problems in existing related technologies.

[0006] To solve the aforementioned technical problem, the present invention is achieved through the following technical solution:

[0007] This invention provides a comprehensive thermal endurance assessment method that integrates multidimensional variable data, specifically including:

[0008] Collect multidimensional variable data of the assessment subjects and calculate the integral skin conductance; the multidimensional variable data includes at least environmental variables, load variables, physiological variables, and psychological variables;

[0009] The multidimensional variable data is subjected to extreme value normalization, and a pruning rule is introduced to handle the extrapolation bias of extreme conditions, so as to obtain the normalized multidimensional variable data.

[0010] Based on the normalized multidimensional variable data and the integral skin conductance, a composite environmental stress index, a standardized load index, a physiological stress index, and a psychological subjective score were constructed.

[0011] An individual heat tolerance scoring model is constructed based on the composite environmental stress index, the standardized load index, the physiological stress index, and the psychological subjective score. The individual heat tolerance comprehensive score is calculated, and the individual heat tolerance comprehensive score is compared with a threshold to determine the level classification and output the heat tolerance level.

[0012] As a preferred embodiment of the thermal endurance comprehensive evaluation method integrating multidimensional variable data described in this invention, the environmental variables include ambient temperature, relative humidity, wind speed, mean radiant temperature, altitude, and oxygen partial pressure.

[0013] The load variables include total combat gear weight, mission metabolic equivalent, duration, and rest-to-work ratio; the rest-to-work ratio is the proportion of rest time to total time.

[0014] The physiological variables include core temperature, heart rate, blood oxygen saturation, and transient skin conductance; the transient skin conductance is integrated by performing a time integration operation on the transient skin conductance;

[0015] The psychological variables include thermal perception score, military fatigue score, individual thermal perception baseline, and individual fatigue baseline.

[0016] As a preferred embodiment of the thermal endurance comprehensive assessment method integrating multidimensional variable data described in this invention, the core temperature is determined based on the ambient temperature, specifically including: collecting the resting core temperature baseline, average body surface temperature, real-time heart rate and resting heart rate baseline, and body surface ambient temperature gradient in a seated state; constructing a nonlinear physiological process reconstruction model based on ridge regression; and calculating the core temperature.

[0017] As a preferred embodiment of the thermal endurance comprehensive evaluation method integrating multidimensional variable data described in this invention, the acquisition of the normalized multidimensional variable data includes:

[0018] For positive indicators, the difference between the original collected value and the actual minimum value in the training dataset is divided by the difference between the actual maximum value and the actual minimum value in the training dataset to obtain a dimensionless value.

[0019] For inverse indicators, divide the difference between the actual maximum value and the original collected value in the training dataset by the difference between the actual maximum value and the actual minimum value in the training dataset to obtain a dimensionless value.

[0020] The clipping rules are applied to handle extrapolation bias under extreme conditions. When the dimensionless value is less than zero, it is taken as zero; when the dimensionless value is greater than one, it is taken as one. The normalized multidimensional variable data is then output.

[0021] As a preferred embodiment of the thermal resistance comprehensive evaluation method integrating multidimensional variable data described in this invention, the construction of the composite environmental stress index includes:

[0022] By multiplying the normalized values ​​of ambient temperature and mean radiation temperature by the corresponding meteorological weighting coefficients, a nonlinear physical approximation model that nonlinearly couples relative humidity and wind speed is constructed to obtain the basic thermal environment function.

[0023] An equivalent composite stress variable is constructed based on the flat characteristic region of the oxygenated hemoglobin dissociation curve at low altitude. The hypoxia sensitivity index of the equivalent composite stress variable is calculated to a power. The composite environmental stress index is obtained by multiplying the basic thermal environment function with the power result.

[0024] As a preferred embodiment of the comprehensive thermal endurance assessment method integrating multidimensional variable data described in this invention, the calculation formula for constructing the standardized load index is as follows:

[0025]

[0026] In the formula, For standardized load index, To evaluate the normalized value of effective work time within a time window, The normalized value for the total weight of combat attire. This is the normalized value of the task metabolic equivalent. Compared to rest, This is the coefficient of the negative weight effect. This is the metabolic effect coefficient. This represents the interaction coefficient between load and metabolism.

[0027] As a preferred embodiment of the comprehensive thermal endurance assessment method integrating multidimensional variable data described in this invention, the formula for calculating the physiological stress index is as follows:

[0028] ;

[0029] In the formula, Physiological stress index For the core temperature, The resting core temperature, Heart rate, Resting heart rate This is the upper limit threshold temperature. This is the upper limit threshold of heart rate. Blood oxygen saturation This refers to the resting blood oxygen level at low altitude. This is the amplification factor for the hypoxia hazard index. The normalized integral skin conductance, For temperature coefficient, Heart rate coefficient Blood oxygenation coefficient This represents the skin conductivity coefficient.

[0030] As a preferred embodiment of the comprehensive thermal endurance assessment method integrating multidimensional variable data described in this invention, the calculation formula for the subjective psychological score is as follows:

[0031] ;

[0032] In the formula, For psychological subjective rating, For thermal perception scoring, As an individual thermal perception baseline, As the baseline for individual fatigue, and These represent the maximum and minimum values ​​of the thermal perception scale, respectively. Scoring military fatigue and These represent the maximum and minimum values ​​of the fatigue scale, respectively. For heat sensing weights, This represents the fatigue weight.

[0033] As an optimal scheme for extracting corresponding samples in the comprehensive thermal endurance assessment method that integrates multidimensional variable data as described in this invention, the construction of the individual thermal endurance scoring model includes: introducing the hierarchical analysis method based on prior knowledge to determine the theoretical benchmark weights of the four dimensions of environment, load, physiology, and psychology; constructing a risk penalty multiplier based on real-time dimensionless index to dynamically adjust the benchmark weights; and multiplying and summing the normalized dimensionless indexes of each dimension with the corresponding dynamic weight coefficients to construct the individual thermal endurance scoring model.

[0034] As a preferred embodiment of the comprehensive thermal resistance assessment method integrating multidimensional variable data described in this invention, the method involves comparing the individual comprehensive thermal resistance score with a threshold to determine the grade classification and outputting the thermal resistance grade, including:

[0035] Set a first threshold, a second threshold, and a third threshold;

[0036] If the individual's comprehensive heat resistance score is less than or equal to the first threshold, the heat resistance level is determined to be excellent.

[0037] If the individual's comprehensive heat resistance score is greater than the first threshold and less than or equal to the second threshold, the heat resistance level is determined to be qualified.

[0038] If the individual's comprehensive heat resistance score is greater than the second threshold and less than or equal to the third threshold, the heat resistance level is determined to be of concern.

[0039] If the individual's overall heat tolerance score is greater than the third threshold, the heat tolerance level is determined to be high risk.

[0040] The present invention has the following beneficial effects:

[0041] 1. This invention constructs a composite environmental stress index by collecting environmental variables including altitude and oxygen partial pressure, and constructs an equivalent composite stress variable based on the flat characteristic region of the oxyhemoglobin dissociation curve, eliminating the collinearity of air pressure reduction and oxygen partial pressure reduction, and incorporating high-altitude parameters and low-oxygen parameters into the model calculation, thus solving the problem of distortion in assessment results caused by the failure to integrate altitude and air oxygen content in existing technologies, making the assessment method adaptable to the complex thermal environment of plateaus; it constructs a standardized load index by collecting load variables including total weight of combat clothing, task metabolic equivalent, effective working time, and rest-to-work ratio, and introduces a three-dimensional joint interaction term for nonlinear reconstruction based on net effective muscle work time, thus solving the problems of existing technologies not considering military-specific loads and the lack of biomechanical logic in the linear superposition of single load variables.

[0042] 2. This invention integrates the composite environmental stress index and the standardized load index, combined with the physiological stress index obtained from core temperature, heart rate, blood oxygen saturation, and transient skin conductance response, and the psychological subjective score obtained from thermal perception score and military fatigue score, to construct an individual thermal endurance scoring model and calculate the individual comprehensive thermal endurance score. By fusing four-dimensional variables—environmental variables, load variables, physiological variables, and psychological variables—and employing military operational critical thresholds and extreme value pruning operators in the calculation of the physiological stress dimension to prevent extreme data from causing model divergence, the invention introduces the analytic hierarchy process based on prior knowledge to determine theoretical baseline weights and combines them with real-time risk penalty multipliers for dynamic adjustment of weights, making the calculation model closely resemble real combat scenarios.

[0043] 3. This invention sets a dynamic threshold and determines and outputs the heat endurance level based on the comparison between the individual's comprehensive heat endurance score and the dynamic threshold, thereby providing a clear heat endurance grading conclusion. This can support the quantitative assessment of individual soldiers to adapt to military needs and guide training and troop deployment.

[0044] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0045] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, the drawings can be obtained from these drawings without creative effort.

[0046] Figure 1 The flowchart of the method for comprehensive evaluation of thermal endurance by integrating multidimensional variable data is provided by the present invention.

[0047] Figure 2 A schematic diagram of the curve of altitude and oxygen content correction coefficient provided for this invention. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Example 1

[0050] Existing heat tolerance assessments are mostly based on a single environmental factor or a single physiological indicator, without incorporating altitude and oxygen content, or considering military-specific loads. This leads to serious distortion of assessment results and an inability to accurately guide military operations in high-altitude or high-intensity combat scenarios.

[0051] To solve the above technical problems, such as Figure 1 As shown, this embodiment provides a comprehensive assessment method for thermal tolerance that integrates multidimensional variable data. This method aligns the various multidimensional input variables over time and normalizes their extreme values. It then constructs a composite environmental stress index, a standardized load index, a physiological stress index, and a subjective psychological score. Using these four indices, an individual thermal tolerance scoring model is built to calculate the individual's comprehensive thermal tolerance score. Finally, a thermal tolerance level is output through threshold comparison. Specifically, the method includes the following steps:

[0052] S1. Align all input variables to the same evaluation time window, collect multidimensional variable data of the aligned evaluation object, and calculate the core temperature. With integral skin conductance Specifically, it includes the following sub-steps:

[0053] S11. Set the evaluation time window For environmental variables, load variables, and physiological variables, the average or cumulative value within a time window is used; for psychological variables, the average of multiple samples within a time window is used. Specifically:

[0054] Environmental variables, load variables, physiological variables, and psychological variables of the aligned assessment subjects were collected. The environmental variables included ambient temperature. relative humidity Wind speed Mean radiant temperature Altitude oxygen partial pressure The load variable includes the total weight of combat attire. Task metabolic equivalent Duration Rest to homework ratio Define duration With the evaluation time window All within the same time scale. The physiological variables include core temperature. Heart rate Blood oxygen saturation Transient skin conductance response The psychological variables include heat perception ratings. Military fatigue score Individual thermal perception baseline Individual fatigue baseline .

[0055] As one possible implementation, a temperature sensor is used to obtain the ambient temperature. The unit is ℃; relative humidity is obtained using a humidity sensor. The unit is %; wind speed is obtained using an anemometer. The unit is m / s; the average radiation temperature is calculated after measurement using a black sphere thermometer. Altitude is measured in degrees Celsius (°C) using a barometric altimeter. The unit is m; oxygen partial pressure is measured using an oxygen sensor. The unit is kPa.

[0056] The total weight of combat attire is obtained by weighing the user before the mission. The unit is kg; the metabolic equivalent of the task is estimated by using a wearable triaxial accelerometer to collect triaxial acceleration signals during the operation. The unit is MET, and the specific calculation method is as follows: High-pass filtering is performed on the raw signal containing the gravity component output from the triaxial accelerometer to extract the dynamic acceleration component, and the evaluation time window is calculated. Signal amplitude area within , the area of ​​signal amplitude Substitute into the metabolic mapping model to calculate the task metabolic equivalent The calculation formula is as follows: ; In the formula, , , They are respectively in The dynamic acceleration components along the x, y, and z axes are extracted after high-pass filtering, with units of m / s². The evaluation time window is measured in seconds. This is the resting metabolic constant, with a value of 1.0. This is the exercise metabolic conversion coefficient, with units of s² / m.

[0057] Use task logs to record the rest-to-work ratio Dimensionless, with a value range of [0,1], the specific calculation method is as follows: extract the evaluation time window from the task log. The time-series data of work and rest states within the evaluation window is used to filter out all time segments in the rest state. The duration of each rest time segment is summed and then divided by the evaluation time window. The calculation formula is as follows: In the formula, Evaluation time window for task log recording Inner The duration of each rest period, measured in seconds (s). For the assessment time window The total number of rest time segments.

[0058] Heart rate is continuously monitored using an electrocardiogram sensor. The unit is bpm; blood oxygen saturation is continuously monitored using a wearable pulse oximeter. The unit is %; transient skin conductivity responses are continuously monitored using wearable skin conductance sensors. The unit is μS.

[0059] The ASHRAE seven-point thermal perception scale was used during the assessment time window. The thermal sensing score is obtained by taking the arithmetic mean of multiple equally spaced samples. The scale of the ASHRAE seven-point thermal perception scale is defined as follows: to consecutive integers, to These correspond to cold, cool, slightly cool, moderate, slightly warm, warm, and hot, respectively; the Military Training Fatigue Assessment Scale is used within the assessment time window. The military fatigue score is obtained by taking the arithmetic mean of multiple equally spaced samples. This military training fatigue assessment scale is a subjective evaluation scale that has undergone reliability and validity testing, and its internal consistency coefficient is greater than [value missing]. The scale is defined as to consecutive integers, It means no fatigue. to This indicates mild fatigue. to This represents moderate fatigue. to This indicates severe fatigue. This represents complete exhaustion; the tools used to collect individual baseline data must be absolutely consistent with the assessment tools used during the work process, and the assessment subject must have been in an environment with a temperature of [temperature value missing] before the day's work. And the relative humidity is Sit quietly in a comfortable environment The single thermal perception score measured immediately after the minute ends serves as the individual's thermal perception baseline. The subjects being evaluated sat quietly in the same comfortable environment for a full period of time. The single fatigue score measured immediately after the minute is used as the individual fatigue baseline. .

[0060] Simultaneous data collection and evaluation of the subject in a comfortable environment (temperature) relative humidity Resting heart rate after 10 minutes of sitting meditation With resting core temperature .

[0061] Simultaneous collection and assessment of resting blood oxygen saturation of subjects in a plain environment and while seated. (Unit: %), as the blood oxygen level value.

[0062] It should be noted that the evaluation time window The determination is based on the dynamic response delay period of physiological variables such as human core temperature and transient skin conductance, as well as the short-term signal stability boundary of environmental and load variables, to ensure that complete continuous physiological feedback features can be extracted within the window and transient high-frequency noise interference can be effectively filtered. In this embodiment, its value range is set to 300 to 900 seconds. To achieve accurate alignment of multi-source heterogeneous data on the time axis, a synchronization reference timestamp is assigned to each type of sensor using the system global clock, and a linear interpolation and anti-aliasing filtering resampling algorithm is used to uniformly resample the data streams of all underlying devices with heterogeneous initial sampling frequencies to a standard reference frequency of 1 Hz. Then, the timing node of this standard reference frequency is used as the alignment anchor point within the set evaluation time window. Internally, it strictly performs time-scale feature extraction operations on multidimensional variable data, thereby eliminating time phase misalignment and computational dimension deviation caused by differences in the sampling rate of the underlying hardware.

[0063] S12. A core temperature estimation model is reconstructed based on the nonlinear physiological process of dynamic balance between heat production and heat dissipation. The core temperature is calculated using the nonlinear term of the ratio of body surface temperature to heart rate and the temperature gradient of the body surface environment. The calculation formula is as follows:

[0064] ;

[0065] In the formula, Core temperature, in °C; The resting core temperature baseline is expressed in °C. Average body surface temperature, in °C; The constant representing the baseline of physiologically neutral body surface temperature is set to a value of ; Real-time heart rate, in units of ; This is the baseline resting heart rate, in units of... ; This refers to the ambient temperature, expressed in °C. A dimensionless coefficient for regulating body surface temperature; is the nonlinear conversion coefficient for metabolic heat production, in °C; The dimensionless coefficient for convective-radiative heat dissipation; The system is calibrated using a compensation constant, expressed in °C.

[0066] As one possible implementation method, general estimation parameters , , With constant The calibration method is as follows:

[0067] A large-sample calibration dataset was collected during the early stages of research and development, and the total number of samples extracted was [number missing]. The synchronous measurement data defines the first Body surface temperature deviation characteristics of individual samples Nonlinear metabolic heat production characteristics Heat dissipation temperature gradient characteristics Construct the feature matrix The true core temperature deviation was simultaneously measured using a swallowable temperature-sensitive capsule. Construct the target vector The ridge regression algorithm is used to handle the multicollinearity between heat dissipation temperature gradient characteristics and body surface temperature deviation characteristics, and a parameter vector of general estimation parameters is defined. Its analytical calculation formula is: In the formula Characteristic matrix The transpose of the matrix, for An identity matrix of dimension 1 Let be the regularization penalty constant. The generalized cross-validation method is used to determine the regularization penalty constant. To determine the specific values, construct a generalized cross-validation function. The calculation formula is as follows:

[0068] ;

[0069] In the formula, for An identity matrix of dimension 1 This is the trace operation on a matrix. The logarithmic search interval for the regularization parameter is set to... and with Generate a discrete optimization sequence for the step size, and traverse the discrete optimization sequence to calculate the generalized cross-validation function. The value, extract The optimal penalty constant value with the global minimum The optimal penalty constant. Substitute the parameter vector In the analytical calculation formula, the parameter vector with a unique analytical solution is obtained. Deconstructing the parameter vector Extract the corresponding mapped values ​​and solidify them into general estimation parameters for the deployment phase. , , The system calibration compensation constant is obtained by calculating the mean of the global prediction residuals of the calibration dataset. .

[0070] It should be noted that the total number of samples in the calibration dataset The number of synchronous measurement data sets is set to be no less than 10,000. The data set is based on statistical power analysis and cross-combination distribution of typical military operation scenarios. The aim is to fully cover the boundaries of individual physiological differences in the target group and ensure the asymptotic unbiasedness and generalization ability of the ridge regression model parameter estimation.

[0071] S13, Based on transient skin conductance response Calculate the evaluation time window Integral skin conductance It is used to reflect the amount of sweating, and the calculation formula is:

[0072] ;

[0073] In the formula, It is the integral variable.

[0074] S2. Perform extreme value normalization on the environmental variables, load variables, and physiological dynamic variables collected in step S1, and introduce pruning rules to handle extrapolation bias under extreme conditions. This specifically includes the following sub-steps:

[0075] S21. The extreme value normalization formula used for positive indicators is:

[0076] ;

[0077] The extreme value normalization formula used for the contrarian indicator is:

[0078] ;

[0079] In the formula, These are the original collected values ​​of environmental variables, load variables, and physiological dynamic variables; and These are the actual minimum and maximum values ​​of the variable in the training dataset, respectively; The dimensionless values ​​are normalized values ​​of environmental variables, load variables, and physiological dynamic variables.

[0080] As one possible implementation method, the normalized dimensionless value Includes: normalized values ​​of ambient temperature The relative thermal stress level of the current ambient temperature in the training samples; the normalized value of relative humidity. The current ambient humidity level is represented in the training samples; the normalized wind speed value... , representing the relative level of the current wind speed in the training sample; mean radiant temperature normalized value Characterizing the relative level of the current environmental mean radiation temperature in the training sample; normalized value of total combat clothing weight. Characterizes the relative load level of the current external load in the training samples; normalized value of task metabolic equivalent. This represents the relative level of the current exercise intensity within the training samples; the normalized value of the effective working time. This represents the relative level of task duration within the current evaluation window compared to the training samples; the rest-to-task ratio normalized value. This represents the relative level of the current rest time as a percentage of the total time in the training sample; the integrated skin conductance normalized value. This represents the relative level of average skin conductance activity within the current evaluation window in the training samples.

[0081] S22. Introduce clipping rules to handle extrapolation bias under extreme conditions. If the normalized result... Then take ,like Then take .

[0082] S3. Based on the normalized data of each variable, construct the dimensionless composite environmental stress index, standardized load index, physiological stress index, and psychological subjective score, respectively. This includes the following sub-steps:

[0083] S31. Construct the composite environmental stress index, the calculation formula is as follows: ;

[0084] in, This is the basic thermal environment function. Because the standard wet-bulb black-bulb temperature index is calculated using a linear superposition of environmental parameters, it cannot accurately characterize the nonlinear abrupt change in the obstruction of evaporative heat dissipation by the human body under high humidity conditions, and it does not reflect the nonlinear attenuation effect of heavy combat clothing on wind speed convective heat dissipation gain. Therefore, By employing a nonlinear physical approximation model that introduces exponential decay, the linear baseline structure of ambient temperature and mean radiation temperature in the standard wet-bulb black-bulb temperature index can be preserved. The linear wet-bulb temperature component in the standard wet-bulb black-bulb temperature index is improved into a nonlinear coupling component of relative humidity and wind speed, thereby accurately quantifying the complex thermodynamic boundary effects in complex scenarios. The calculation formula is as follows:

[0085] ;

[0086] In the formula, It is the basic thermal environment function; This is the normalized value of the ambient temperature. This is the normalized value of the mean radiation temperature; This is the normalized value for relative humidity; This is the normalized value for wind speed; , , These are dimensionless meteorological weighting coefficients; The humidity index constant is used to characterize the exponential hindering of heat dissipation under high humidity conditions, and its mesh optimization interval is set to... The optimization step size is set to ; The wind speed attenuation constant is used to characterize the marginal decrease in wind speed heat dissipation gain, and its mesh optimization interval is set to... The optimization step size is set to .

[0087] Humidity index constant With wind speed attenuation constant A nonlinear least squares method is used for grid optimization. The specific optimization logic is as follows: The objective is to minimize the mean square error between the predicted values ​​of the basic thermal environment function in the calibration dataset and the measured baseline values ​​of the heat stress level. Within the set grid optimization interval and step size, a discrete ergonomic search is performed to determine the optimal solution. Meteorological weighting coefficients are also used. , , Locking in the optimal humidity index constant With wind speed attenuation constant Based on this, the ridge regression algorithm is used to analytically calculate and obtain constant values ​​after eliminating collinearity among environmental variables. The effective range of the normalized wind speed values ​​applicable to the above nonlinear physical approximation model is as follows: .

[0088] in, The correction factor for altitude and oxygen content is calculated using the following formula:

[0089] ;

[0090] In the formula, This is a correction factor for altitude and oxygen content; Altitude, in units of ; Oxygen partial pressure, unit: ; This is a baseline altitude constant, its value is based on the fact that the human oxyhemoglobin dissociation curve has a flat characteristic region at low altitudes, and usually increases at higher altitudes. Only after this will the body substantially activate physiological stress compensation mechanisms such as increased ventilation and changes in heart rate, thus requiring a fixed setting. As a critical threshold height for the manifestation of physiological altitude stress; The standard oxygen partial pressure constant at sea level is given by a fixed value. ; The stress source weighting coefficient is used to balance the physical and mechanical effects of the decrease in air pressure at altitude and the physiological hypoxia effect of the decrease in oxygen partial pressure. It is determined after eliminating collinearity interference in the calibration dataset using partial least squares regression. The hypoxia sensitivity index is used to characterize the nonlinear decay tolerance of the human body to equivalent compound stress variables. It considers continuously changing altitude... With multiple gradient oxygen partial pressures Substituting the values ​​into the calculation formula for the altitude and oxygen content correction factor, a numerical simulation was performed to obtain a schematic diagram of the altitude and oxygen content correction factor curve, as shown below. Figure 2 As shown.

[0091] For composite environmental stress index Perform extreme value normalization and pruning again to obtain dimensionless values. .

[0092] As one possible implementation method, a regularized regression method is used to determine the meteorological weighting coefficients. , , Eliminate collinearity among environmental variables and simultaneously determine the humidity index constant. With wind speed attenuation constant Specifically: Establish an independent meteorological calibration dataset and extract multiple sets of dimensionless ambient temperatures. relative humidity Wind speed Mean radiant temperature and standard wet-bulb black bulb temperature reference value Construct a nonlinear regression loss function The calculation formula is:

[0093] ;

[0094] In the formula, Let be the ridge regression regularization hyperparameter. The loss function is optimized using a nonlinear iterative optimization algorithm. The solution is minimized, and the constant values ​​at the convergence state are extracted and solidified as meteorological weighting coefficients. , , .

[0095] As one possible implementation method, a logarithmic multiple linear regression model is used to determine the altitude stress index. With hypoxia sensitivity index Specifically: Establish an independent, large-sample hypobaric chamber combined stress calibration dataset, and extract the simulated altitude of exogenous regulation. Input oxygen partial pressure Dimensionless objective scale for population physiological surge We construct a log-transform regression equation that eliminates exponential coupling, expressed as:

[0096] ;

[0097] In the formula, The function is the natural logarithm. The least squares method is used to analytically calculate the logarithmic transformation regression equation to obtain stable fitting weights, which are set as a fixed altitude stress index. With hypoxia sensitivity index .

[0098] S32. Constructing a standardized load index Since the evaluation time window includes rest periods with physiological recovery characteristics, normalized values ​​of effective working time are extracted. The proportion of dynamic tasks The product of net effective muscle work time is extracted to isolate metabolic distortion caused by rest. Simultaneously, based on the principles of human biomechanics, the additional equipment weight carried by the work object not only generates static posture maintenance energy consumption but also acts as a multiplier during dynamic displacement, drastically increasing the physiological energy consumption per unit task metabolic equivalent. Furthermore, the deep neuromuscular fatigue caused by this additional mechanical work is strictly controlled by the cumulative time of dynamic work. A three-dimensional joint interaction term reconstruction model is introduced, incorporating the total weight of combat gear, task metabolic equivalent, and net effective muscle work time, to calculate the standardized load index. :

[0099] ;

[0100] In the formula, Standardized load index; Normalized value for the total weight of combat attire; Normalized value for task metabolic equivalent; To evaluate the normalized value of effective work time within the time window; The ratio of rest to work; The static support bearing coefficient characterizes the isometric contraction fatigue effect of muscles caused by the equipment mass maintaining a resting posture throughout the entire evaluation time window. The dynamic baseline work coefficient characterizes the baseline energy loss of the work object in overcoming its own weight to perform a predetermined task intensity. The biomechanical coupling multiplier characterizes the nonlinear physical amplification effect of additional load on overall metabolic burden under specific exercise intensity and duration of continuous work. This is relative to the standardized load index. Extreme value normalization and pruning are performed to obtain dimensionless values. .

[0101] As one possible implementation method, the load effect coefficient in the standardized load index model Metabolic effect coefficient With the interaction effect coefficient The method for determining it is as follows:

[0102] S321. Select a large sample group to conduct standardized live-fire tests covering different total weights of combat gear, mission metabolic equivalents, and rest-to-work ratios. The total sample size is [missing data]. Independent offline test data were collected. To overcome the nonlinear decoupling bias between absolute total energy consumption and actual human heat load, a portable metabolic analysis device was used to simultaneously collect the absolute energy consumption of subjects within the corresponding assessment time window. This data was then combined with the core temperature drift rate obtained in real time from the temperature-sensitive capsule to calculate the net metabolic heat storage rate, characterizing the difference in actual cumulative heat load between continuous and intermittent work under different work modes. Extreme value normalization was performed on the net metabolic heat storage rate to obtain a dimensionless target heat load benchmark value. Define the target vector. Its expression is: In the formula, The transpose operation of a matrix.

[0103] S322. Extract the dimensionless loading input variables for the corresponding samples, and define the first... Static support bearing characteristics of individual samples , Dynamic basic work characteristics , Biomechanical coupling characteristics . Constructing the feature matrix Its expression is:

[0104] ;

[0105] S323, Given the characteristic matrix The biomechanical coupling characteristics and individual work characteristics exhibit endogenous mathematical collinearity; the ridge regression penalty term is used to limit weight divergence. A load coefficient vector is defined. Load factor vector The analytical calculation formula is: In the formula, Characteristic matrix The transpose of the matrix, for An identity matrix of dimension 1 This is the regularization penalty constant. Establish... The discrete logarithmic grid search sequence is used to traverse the grid search sequence and perform the operation on the load calibration dataset. Cross-validation. Extract the optimal penalty constant that minimizes the global mean squared error of cross-validation. The optimal penalty constant. Substitute the load factor vector In the analytical calculation formula, the load factor vector with a unique analytical solution is obtained by solving. Deconstructing the load factor vector Extract the corresponding mapping values ​​and solidify them into the static support bearing capacity coefficient in the standardized load index calculation formula for the actual deployment phase. Dynamic basic work coefficient Biomechanical coupling multiplier .

[0106] S33. Constructing a physiological stress index using extreme value clipping constraints and logistic nonlinear functions. The calculation formula is:

[0107] ;

[0108] In the formula, Physiological stress index; Core temperature, in °C; Resting core temperature, in °C; The core temperature is the critical threshold for military operations, representing the temperature safety boundary for maintaining continuous combat capability, and is fixed at a value of [value missing]. ; Heart rate, unit: ; This is the resting heart rate, in units of... ; The critical threshold for heart rate in military operations, in units of The target population dynamic heart rate reserve model is used for calibration, and its calculation formula is as follows: ; Blood oxygen saturation, measured in units of ; This represents the resting blood oxygen level at low altitude, in units of... ; The hypoxia physiological deterioration rate coefficient is dimensionless. Its physiological significance is to characterize the steepness of the step decline in comprehensive physiological functions when the human body crosses from the hypoxia compensation period to the decompensation period. The deviation threshold of the midpoint of blood oxygen decompensation is used to characterize the nonlinear inflection point of the logistic function curve; The normalized integral skin conductance is dimensionless. , , , These are the temperature coefficient, heart rate coefficient, blood oxygen coefficient, and skin conductivity coefficient, respectively, all dimensionless. In the formula, when a soldier experiences an instantaneous physiological extreme value under extreme combat scenarios, resulting in a real-time core temperature... breakthrough or real-time heart rate breakthrough At that time, the pruning operator forcibly truncates the corresponding dimensionless normalized characteristic term to the limit value. This physical truncation mechanism ensures that abnormal data exceeding the safety threshold will not trigger mathematical divergence in the global assessment model. Simultaneously, by continuously outputting full individual stress weights, it ensures the system can stably trigger the highest level of risk warning intervention. Regarding the aforementioned physiological stress indices... After performing extreme value normalization and pruning again, the final dimensionless value for subsequent calculations is obtained. .

[0109] As one possible implementation method, the hypoxia hazard index amplification factor With temperature coefficient Heart rate coefficient blood oxygen coefficient Skin conductivity The determination method is as follows: An independent physiological stress calibration dataset is established, and a large sample group is selected to perform a standard treadmill exercise test covering multiple gradients of ambient temperature, relative humidity, and simulated altitude within an artificial climate oxygen chamber. Resting baseline values ​​and real-time physiological parameters, including core temperature, are collected simultaneously from the subjects at different test stages. Heart rate Blood oxygen saturation Normalized integral skin conductance By simultaneously measuring the blood lactate concentration of subjects using independent medical testing equipment, a dimensionless clinical physiological stress benchmark value representing the degree of approximation to the objective physiological limit is generated. The definition includes The calibration target vector of each sample The expression is: In the formula, Characterize the matrix transpose operation. Construct the outer nonlinear parameters. Discrete logarithmic grid search space, setting constants The traversal interval and optimization step size. For each candidate constant in the grid search space. Substitute the sample data to calculate the nonlinear blood oxygen saturation stress term. Construct the inner linear ridge regression feature matrix. , characteristic matrix Each row contains four dimensionless physiological input features of the corresponding sample, and the core temperature normalization term of the corresponding sample is extracted and calculated. Heart rate normalization term Nonlinear blood oxygen saturation stress item Normalized integral skin conductance This constitutes a feature column vector. A physiological weighted coefficient vector is defined. The expression is: ; The current candidate constant is calculated using the ridge regression analytical formula. The optimal physiological weighted coefficient vector , The analytical calculation formula is: In the formula, Characteristic matrix The transpose of the matrix, for An identity matrix of dimension 1 The inner regularization penalty constant is determined using internal cross-validation. The model predictions and clinical physiological stress baselines are calculated for the current candidate parameter combinations. The root mean square error between them. Extract the optimal candidate constant that minimizes the global root mean square error. Its corresponding optimal physiological weighting coefficient vector Deconstructing the optimal physiological weighted coefficient vector Extract the corresponding mapping values ​​and define them as the hypoxia hazard index amplification factor in the standardized physiological stress index calculation formula for the actual deployment phase. With temperature coefficient Heart rate coefficient blood oxygen coefficient Skin conductivity .

[0110] S34, Introducing thermal sensing weights With fatigue weight Based on weight constraints The subjective psychological rating is defined as a weighted average of two standardized ratings, and a subjective psychological rating is constructed. The calculation formula is:

[0111] ;

[0112] In the formula, As the baseline for individual thermal perception; This serves as the baseline for individual fatigue. and These are the maximum and minimum values ​​of the thermal sensing scale, respectively. and These are the maximum and minimum values ​​of the fatigue scale, respectively; the calculation results are dimensionless values. .

[0113] As one possible implementation method, thermal perception weights in psychological subjective rating models With fatigue weight The method for determining the sample population is as follows: A large sample population is selected and subjected to standardized tasks in a specific test environment (covering a full matrix cross-test environment consisting of an altitude of 0m to 4500m, an ambient temperature of 15℃ to 45℃, and a relative humidity of 30% to 90%). Within the set test time window, a total of [number missing] samples are collected. Subjective scale data from the sample. Extract the first... Thermal perception score of each sample Military fatigue score Individual thermal perception baseline Individual fatigue baseline For each psychological indicator, intra-individual standardized calculations are performed, and a translation transformation is applied to eliminate non-positive values. The first value is defined as... Thermal sensing features of each sample With fatigue characteristic terms The calculation formula is: ; In the formula, To prevent the logarithmic operation from becoming meaningless due to the value of zero, the value of the extremely small constant is set to zero. Definition of the first The sample at the th The proportion under the feature The calculation formula is: In the formula, Values Represents thermal sensing characteristics. Values Represents fatigue characteristics. Definition 1 Information entropy of a feature The calculation formula is: In the formula The natural logarithm function is used. Feature difference coefficients are calculated based on the information entropy of each feature, and then normalized to obtain weights that satisfy the normalization constraint. Objective weighting. Thermal sensing weighting. The calculation formula is: Fatigue weight The calculation formula is: The numerical values ​​corresponding to the mappings obtained from the weight calculation results are extracted and fixed as constants, which are then solidified into the thermal perception weights in the psychological subjective scoring formula during the actual deployment phase. With fatigue weight .

[0114] S4. Construct an individual heat tolerance scoring model based on the composite environmental stress index, standardized load index, physiological stress index, and psychological subjective score, and calculate the individual's comprehensive heat tolerance score. An individual's overall heat tolerance score is determined based on a set threshold. The process involves classifying and outputting heat resistance ratings. This includes the following sub-steps:

[0115] S41. Construct an individual thermal tolerance scoring model, the expression of which is:

[0116] ;

[0117] In the formula, , , , These are the normalized weighting coefficients for the four dimensions of environment, load, physiology, and psychology.

[0118] As one possible implementation method, environmental weights Load weight Physiological weight Psychological weight The method for determining it is as follows:

[0119] S411. Establish a judgment matrix comprising four dimensions: environment, load, physiology, and psychology. Based on the basic principles of military operational physiology, perform pairwise comparisons and scoring of the importance of each dimension. Calculate the largest eigenvalue of the judgment matrix and its corresponding eigenvector. Perform a consistency check, extract the eigenvectors that pass the consistency check, and normalize them to obtain the theoretical baseline weight vector. Its expression is: In the formula, , , , These are the theoretical baseline weights for the environmental, load, physiological, and psychological dimensions, respectively. The transpose operation of a matrix.

[0120] S412. To adapt to the dynamic risk evolution patterns under different application scenarios such as seasons and regions, a real-time risk penalty mechanism is introduced to adjust weights. The real-time risk penalty multiplier for each dimensionless index is defined. The calculation formula is as follows: In the formula, These correspond to four dimensions: environment, load, physiology, and psychology. For the real-time dimensionless exponent of the corresponding dimension, i.e. , , , ; It is an exponential function with the natural constant as its base; It is a risk sensitivity coefficient used to regulate the degree of nonlinear amplification of weights under high stress conditions.

[0121] S413: Applying the theoretical baseline weight vector each dimensional element in it to the corresponding real-time risk penalty multiplier to perform element-wise multiplication operation, so as to obtain unnormalized dynamic weight coefficients , and the calculation formula is: ; wherein, is the i-th dimensional element in the theoretical baseline weight vector .

[0122] S414: Performing global normalization processing on the unnormalized dynamic weight coefficients to obtain environmental weights, load weights, physiological weights, and psychological weights finally used for model calculation , and the calculation formulas are: , , , .

[0123] S42: Setting a first threshold , a second threshold and a third threshold to grade the individual heat tolerance comprehensive score , and output the heat tolerance grade. When , the heat tolerance grade is determined as excellent. When , the heat tolerance grade is determined as qualified. When , the heat tolerance grade is determined as requiring attention. When , the heat tolerance grade is determined as high risk.

[0124] As a possible implementation, collect complete evaluation data of no less than 300 military personnel in representative training scenarios (covering an altitude of 0-4500 m and a temperature of 15-45°C), and record the individual heat tolerance comprehensive score HES of each evaluation and the corresponding gold standard event labels (heatstroke, heat exhaustion, severe altitude reaction, etc.);

[0125] The determination method of the third threshold is as follows: taking the gold standard event as a positive label, drawing the receiver operating characteristic curve of HES, and determining by maximizing the Youden's index . If there are multi-level gold standard events, the weighted Youden's index is adopted, and higher weights are assigned to severe events such as heat stroke.

[0126] The first threshold and the second threshold ​​​​The determination method is as follows: For samples in the training set without gold standard event labels, calculate their HES values ​​and sort them in ascending order; use the 25th percentile of the reliable samples after dissociation index screening as the HES value. The 50th percentile as Calculate the dissociation index for each sample. Remove For samples larger than the training set mean plus twice the standard deviation, the percentile is recalculated as the final threshold.

[0127] Furthermore, when the cumulative number of newly added valid assessment samples reaches 50, or the altitude change exceeds 500m, or the season changes, all thresholds are recalculated according to the above method.

[0128] The preferred embodiments of the invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A comprehensive evaluation method for thermal endurance integrating multidimensional variable data, characterized in that, include: Collect multidimensional variable data of the evaluation subjects and calculate the integral skin conductance. The multidimensional variable data includes at least environmental variables, load variables, physiological variables, and psychological variables; The multidimensional variable data is subjected to extreme value normalization, and a pruning rule is introduced to handle the extrapolation bias of extreme conditions, so as to obtain the normalized multidimensional variable data. Based on the normalized multidimensional variable data and the integral skin conductance, a composite environmental stress index, a standardized load index, a physiological stress index, and a psychological subjective score were constructed. An individual heat tolerance scoring model is constructed based on the composite environmental stress index, the standardized load index, the physiological stress index, and the psychological subjective score. The individual heat tolerance comprehensive score is calculated, and the individual heat tolerance comprehensive score is compared with a threshold to determine the level classification and output the heat tolerance level.

2. The comprehensive thermal endurance assessment method integrating multidimensional variable data according to claim 1, characterized in that: The environmental variables include ambient temperature, relative humidity, wind speed, mean radiant temperature, altitude, and oxygen partial pressure. The load variables include total combat gear weight, mission metabolic equivalent, duration, and rest-to-work ratio; the rest-to-work ratio is the proportion of rest time to total time. The physiological variables include core temperature, heart rate, blood oxygen saturation, and transient skin conductance; the transient skin conductance is integrated by performing a time integration operation on the transient skin conductance; The psychological variables include thermal perception score, military fatigue score, individual thermal perception baseline, and individual fatigue baseline.

3. The comprehensive thermal endurance evaluation method integrating multidimensional variable data according to claim 2, characterized in that, The core temperature is determined based on the ambient temperature, specifically including: collecting the resting core temperature baseline, average body surface temperature, real-time heart rate and resting heart rate baseline, and body surface ambient temperature gradient in a seated state; constructing a nonlinear physiological process reconstruction model based on ridge regression; and calculating the core temperature.

4. The comprehensive thermal endurance assessment method integrating multidimensional variable data according to claim 1, characterized in that, The acquisition of the normalized multidimensional variable data includes: For positive indicators, the difference between the original collected value and the actual minimum value in the training dataset is divided by the difference between the actual maximum value and the actual minimum value in the training dataset to obtain a dimensionless value. For inverse indicators, divide the difference between the actual maximum value and the original collected value in the training dataset by the difference between the actual maximum value and the actual minimum value in the training dataset to obtain a dimensionless value. The clipping rules are applied to handle extrapolation deviations under extreme conditions. When the dimensionless value is less than zero, it is taken as zero; when the dimensionless value is greater than one, it is taken as one. The normalized multidimensional variable data is then output.

5. The method for comprehensive thermal endurance evaluation by integrating multidimensional variable data according to claim 1, characterized in that, The construction of the composite environmental stress index includes: By multiplying the normalized values ​​of ambient temperature and mean radiation temperature by the corresponding meteorological weighting coefficients, a nonlinear physical approximation model that nonlinearly couples relative humidity and wind speed is constructed to obtain the basic thermal environment function. An equivalent composite stress variable is constructed based on the flat characteristic region of the oxygenated hemoglobin dissociation curve at low altitude. The hypoxia sensitivity index of the equivalent composite stress variable is calculated to a power. The composite environmental stress index is obtained by multiplying the basic thermal environment function with the power result.

6. The comprehensive thermal endurance assessment method integrating multidimensional variable data according to claim 1, characterized in that, The formula for constructing the standardized load index is as follows: ； In the formula, For standardized load index, To evaluate the normalized value of effective work time within a time window, The normalized value for the total weight of combat attire. This is the normalized value of the task metabolic equivalent. Compared to rest, This is the coefficient of the negative weight effect. This is the metabolic effect coefficient. This represents the interaction coefficient between load and metabolism.

7. The method for comprehensive thermal endurance evaluation by integrating multidimensional variable data according to claim 1, characterized in that, The formula for calculating the physiological stress index is as follows: ； In the formula, Physiological stress index For the core temperature, The resting core temperature, Heart rate, Resting heart rate This is the upper limit threshold temperature. This is the upper limit threshold of heart rate. Blood oxygen saturation This refers to the resting blood oxygen level at low altitude. This is the amplification factor for the hypoxia hazard index. The normalized integral skin conductance, For temperature coefficient, Heart rate coefficient Blood oxygenation coefficient This represents the skin conductivity coefficient.

8. The method for comprehensive thermal endurance evaluation by integrating multidimensional variable data according to claim 1, characterized in that, The formula for calculating the subjective psychological rating is as follows: ； In the formula, For psychological subjective rating, For thermal perception scoring, As an individual thermal perception baseline, As the baseline for individual fatigue, and These represent the maximum and minimum values ​​of the thermal perception scale, respectively. Scoring military fatigue and These represent the maximum and minimum values ​​of the fatigue scale, respectively. For heat sensing weights, This represents the fatigue weight.

9. The method for comprehensive thermal endurance evaluation by integrating multidimensional variable data according to claim 1, characterized in that, The construction of the individual heat tolerance scoring model includes: introducing the hierarchical analysis method based on prior knowledge to determine the theoretical benchmark weights of the four dimensions of environment, load, physiology, and psychology; constructing a risk penalty multiplier based on real-time dimensionless index to dynamically adjust the benchmark weights; and multiplying and summing the normalized dimensionless indexes of each dimension with the corresponding dynamic weight coefficients to construct the individual heat tolerance scoring model.

10. The method for comprehensive thermal endurance evaluation by integrating multidimensional variable data according to claim 1, characterized in that, The individual's comprehensive heat resistance score is compared with a threshold to determine the grade classification, and the heat resistance grade is output, including: Set a first threshold, a second threshold, and a third threshold; If the individual's comprehensive heat resistance score is less than or equal to the first threshold, the heat resistance level is determined to be excellent. If the individual's comprehensive heat resistance score is greater than the first threshold and less than or equal to the second threshold, the heat resistance level is determined to be qualified. If the individual's comprehensive heat resistance score is greater than the second threshold and less than or equal to the third threshold, the heat resistance level is determined to be of concern. If the individual's overall heat tolerance score is greater than the third threshold, the heat tolerance level is determined to be high risk.

Images