Wearable intelligent clothing health monitoring system

Abstract

The invention relates to the technical field of health monitoring, and discloses a wearable intelligent clothing health monitoring system, which is characterized in that sweat biochemistry, sweat rate and microclimate signals are continuously acquired through a multi-mode sensing array integrated on electronic textile clothing, a processing module calculates fluid transmission flux of a clothing sweat passage according to microclimate and sweat rate data, and the fluid transmission flux of the clothing sweat passage is calculated according to the microclimate and sweat rate data. In-situ identification is carried out on a fabric pore transport retention coefficient representing the blocking characteristic of the porous structure of the fabric; a sensing interface biological passivation time constant is combined to construct system comprehensive response delay time reflecting double delay, and then a transmission distortion correction factor is calculated; and finally, inverting and restoring the sweat signal subjected to time domain aliasing into a real concentration trend of the skin end by using a fractional order inverse model based on the retention coefficient, and introducing a transmission distortion correction factor as a dynamic confidence penalty term into a health risk assessment model, thereby realizing accurate and real-time health early warning.

Description

Technical Field

[0001] This invention relates to the field of health monitoring, and more specifically, to a wearable smart clothing health monitoring system. Background Technology

[0002] With the deep integration of flexible electronics technology and smart healthcare, wearable health monitoring devices are transitioning from rigid wristbands and patches that cause discomfort to fully flexible, integrated electronic textiles. Sweat, a bodily fluid rich in electrolytes (sodium, chloride, potassium), metabolic products (lactic acid, glucose), and trace elements, can non-invasively reflect the body's water-salt balance, heat stress levels, and metabolic state. Continuous sweat biochemical analysis technology based on smart clothing, leveraging the natural full-body coverage of clothing to achieve all-weather, imperceptible physiological monitoring, demonstrates irreplaceable application value in competitive sports load management, special operations (such as firefighting and border defense) safety protection, and chronic disease monitoring.

[0003] Existing wearable sweat monitoring technologies primarily follow the design principles of traditional microfluidic chips. This involves using photolithography or molding processes to create closed microchannels on the surface of polymers (such as PDMS and PET), or simply utilizing the capillary effect of fabric yarns to guide sweat to the surface of electrochemical electrodes. In signal processing, these systems typically assume that the flow of sweat within the channels is a uniform plug flow, and the values ​​read by the sensor represent the current biochemical state of the skin surface. Current warning algorithms often employ simple threshold-based methods, such as triggering an alarm when the detected ion concentration or loss exceeds a certain fixed value, or simply introducing a fixed linear delay constant to roughly compensate for transmission time.

[0004] However, after fully textileizing microfluidic systems, existing monitoring methods face failure problems caused by the special structure of fabrics, mainly in the following two aspects:

[0005] First, the non-uniform retention in porous media leads to signal temporal aliasing (the so-called long-tailed memory phenomenon). Unlike standard microfluidic channels with smooth inner walls, electronic textiles consist of a complex hierarchical porous network composed of macroscopic yarn gaps (macropores) and microscopic internal voids within fiber bundles (micropores). When sweat flows through this structure, it travels rapidly in macropores, but is extremely slow in micropores due to strong capillary forces. This speed difference means that newly secreted sweat cannot completely push away old sweat like a piston; instead, it continuously mixes and dilutes with the old sweat retained in the micropores during transmission. This mechanism means that the signal measured by the sensor is not a single component at the current moment, but a weighted mixture of sweat components over a long period of time, resulting in severe tailing and smoothing of the signal waveform on the time axis, making it impossible to accurately capture rapid changes in physiological indicators.

[0006] Second, the bio-passivation of open fabric interfaces leads to sluggish response. The sweat pathways in smart clothing are located in an open microclimate environment with a very large specific surface area. This makes it easy for sebum (lipids), protein fragments in sweat, and microorganisms on the skin surface to adsorb onto the rough textile electrode surface, forming a high-resistivity biofouling film. This film not only hinders the diffusion of target ions to the electrode surface but also consumes some biochemical substances due to the metabolic activities of microorganisms, effectively causing a dynamic decay of sensor sensitivity and a non-linear extension of response time.

[0007] In summary, existing technologies neglect the dual distortion effects of transport mixing and interface passivation unique to textile structures, resulting in severely delayed and distorted biochemical signals. Especially in critical situations such as acute dehydration or heatstroke, when physiological parameters change drastically, the system often fails to promptly remove these environmental disturbances and reconstruct the true skin-level trend, leading to delayed or missed warnings. Summary of the Invention

[0008] This invention provides a wearable smart clothing health monitoring system, which solves the technical problems mentioned in the background art.

[0009] This invention provides a wearable smart clothing health monitoring system, comprising:

[0010] A multimodal sensing unit integrated into the electronic textile garment body is used to continuously collect sweat biochemical signals, sweat rate-related signals, and garment microclimate signals; and

[0011] The data processing module connected to the multimodal sensing unit is configured to perform the following operations: calculate the fluid transport flux of the clothing sweat pathway based on the microclimate signal and the sweat rate correlation signal, and identify the fabric pore transport retention coefficient, which characterizes the fluid resistance properties of the fabric pore structure.

[0012] By combining the biopassivation time constant of the sensing interface with the fluid transport flux, a comprehensive system response lag time is constructed, and a transmission distortion correction factor is calculated based on this lag time and the dynamic change cycle of physiological parameters.

[0013] A porous media transport inversion model is established using the fabric pore transport retention coefficient to restore the sweat biochemical signal to real-time concentration data at the skin end. At the same time, the transmission distortion correction factor is introduced as a confidence weighting term into the health risk assessment function, and an early warning signal is output when the assessment result exceeds a preset safety threshold.

[0014] The beneficial effects of this invention are as follows: By establishing a power-law-biological response correlation distortion correction mechanism, this invention can not only restore the real rapid physiological changes on the human skin surface from sensor signals that are slowed down and smoothed by the fabric structure, but also adaptively improve the early warning sensitivity under high distortion conditions according to the current fluid transport state and biological passivation degree. This effectively avoids delayed or missed reports caused by slow hardware response in critical moments such as acute dehydration and heatstroke, significantly improving the monitoring reliability and safety of smart clothing in complex dynamic scenarios. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the implementation process of the present invention;

[0016] Figure 2 This is a schematic diagram of an implementation scenario of the present invention. Detailed Implementation

[0017] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0018] like Figure 1 As shown, a wearable smart clothing health monitoring system includes:

[0019] A multimodal sensing unit integrated into the electronic textile garment body is used to continuously collect sweat biochemical signals, sweat rate-related signals, and garment microclimate signals; and

[0020] The data processing module connected to the multimodal sensing unit is configured to perform the following operations: calculate the fluid transport flux of the clothing sweat pathway based on the microclimate signal and the sweat rate correlation signal, and identify the fabric pore transport retention coefficient, which characterizes the fluid resistance properties of the fabric pore structure.

[0021] By combining the biopassivation time constant of the sensing interface with the fluid transport flux, a comprehensive system response lag time is constructed, and a transmission distortion correction factor is calculated based on this lag time and the dynamic change cycle of physiological parameters.

[0022] A porous media transport inversion model is established using the fabric pore transport retention coefficient to restore the sweat biochemical signal to real-time concentration data at the skin end. At the same time, the transmission distortion correction factor is introduced as a confidence weighting term into the health risk assessment function, and an early warning signal is output when the assessment result exceeds a preset safety threshold.

[0023] In a preferred embodiment, a multimodal sensing unit integrated into the electronic textile garment body is used to continuously collect sweat biochemical signals, sweat rate-related signals, and garment microclimate signals, including:

[0024] Used to acquire sodium ion potential signals and chloride ion potential signal Ion-selective electrode assembly;

[0025] Used to acquire impedance modulus or capacitance readings. The main perspiration rate sensing channel, and the background ion concentration data acquisition channel. Ion interference compensation channel;

[0026] And skin temperature sensors deployed inside the clothing in contact with the skin surface to collect skin temperature. An ambient temperature sensor deployed on the outer surface of the garment is used to collect ambient temperature data. A relative humidity sensor is used to monitor the humidity inside clothing and to collect relative humidity data. ;

[0027] in This represents the discrete sampling time.

[0028] Preferably, the multimodal sensing unit integrated into the electronic textile garment body is constructed using a flexible electronic fabric manufacturing process to solve the signal distortion problem caused by environmental coupling and biochemical interference in dynamic physiological monitoring using traditional single-modal sensing. Structurally, this multimodal sensing unit includes components for acquiring sodium ion potential signals. and chloride ion potential signal The system includes an ion-selective electrode assembly, a sensing channel for collecting perspiration rate data, and a temperature and humidity sensing array for microclimate monitoring. For the ion-selective electrode assembly, the electrochemical response mechanism is based on the Nernst equation, where the electrode potential is linearly related to the logarithm of the target ion activity.

[0029] In practical implementation, the component consists of a solid-state contact working electrode coated with a specific ion carrier (such as Na-X zeolite for sodium ions or quaternary ammonium salt derivatives for chloride ions) and an Ag / AgCl reference electrode. Preferably, conductive carbon paste and a functionalized sensitive film are printed onto a fabric substrate using screen printing technology. The signal output by this component... and All are open-circuit potential values ​​relative to the reference electrode, typically expressed in millivolts (mV). Indicates the discrete sampling time and sampling period. The preferred setting is between 0.2 seconds and 5.0 seconds, with the most preferred setting being 1.0 seconds, so as to satisfy the coverage of the physiological signal bandwidth by the Shannon sampling theorem while also taking into account the low power consumption requirement.

[0030] Due to the slope of the Nernst response Directly affected by temperature Impact (among others) The gas constant is It is Faraday's constant. (for valence), therefore the skin temperature described later must be introduced. Real-time slope correction is performed to eliminate measurement errors caused by thermodynamic temperature fluctuations.

[0031] Furthermore, regarding the sweat rate monitoring section, this embodiment employs a dual-channel differential or compensation mechanism, specifically including methods for acquiring impedance modulus readings or capacitance readings. The main perspiration rate sensing channel, and the background ion concentration data acquisition channel. The ion interference compensation channel. Specifically, the fluid volume (i.e., sweat volume) within the fabric microfluidic channel is typically characterized by changes in capacitance or impedance of the interdigitated electrodes, and the calculation formula can be approximated as follows: ,in The effective surface area wetted by sweat directly corresponds to the amount of sweat accumulated. However, the dielectric constant of sweat... and conductivity It is not a constant, but a function that fluctuates dramatically with the concentration of electrolytes (mainly Na+ and Cl-) in sweat. If only the main channel reading is relied upon... (Preferably operating in the high-frequency range of 100kHz to 1MHz, exhibiting predominantly dielectric properties), concentrated sweat can lead to artificially high readings, resulting in a falsely high sweat rate. Therefore, by introducing an ion interference compensation channel operating in the low-frequency range (preferably 1kHz to 10kHz, exhibiting predominantly resistive / conductive properties), a signal reflecting ion intensity can be obtained. Specifically, low-frequency impedance is mainly controlled by ion mobility and is strongly correlated with electrolyte concentration. In subsequent processing, this is utilized... right Decoupling correction is fundamental to achieving high-precision sweat rate measurement. The electrode structure of the main sweat rate sensing channel preferably uses a micrometer-level spacing (e.g., 50 μm). Up to 200 Interdigitated gold electrodes or graphene composite fibers are used to maximize sensitivity to trace amounts of sweat.

[0032] Regarding the acquisition of microclimate signals, this embodiment imposes the following limitations on the spatial layout. Specifically, a skin temperature sensor deployed inside the clothing, in contact with the skin surface, is used to collect skin temperature data. The sensor preferably employs an NTC thermistor or platinum resistance thermometer (RTD) encapsulated in a highly thermally conductive flexible polymer, and its measurement accuracy needs to be superior to that of other sensors. ,because It is used not only to calculate the saturated vapor pressure on the skin surface, but also directly as a temperature compensation source for electrochemical sensors. An ambient temperature sensor deployed on the outer surface of clothing is used to collect ambient temperature data. This sensor requires heat insulation to block body heat conduction and only responds to external environmental thermal fields. A relative humidity sensor used to monitor the humidity inside clothing collects relative humidity data. It is preferably installed between fabric layers or near the evaporation interface of microfluidic channels, using a humidity-sensitive capacitive element, with a measurement range covering 0% to 100% RH.

[0033] The joint measurement of sweat biochemical signals, sweat rate-related signals, and clothing microclimate signals was conducted to construct the driving force gradient required for Fick's Law, based on... ,in This represents the saturated vapor pressure. Discrete sampling time. Triggered uniformly by the system master clock, ensuring all the above signals , ,Signal , and signals , , Align on the timeline.

[0034] In a preferred embodiment, calculating the fluid transport flux of the clothing sweat pathway based on the microclimate signal and the sweat rate correlation signal includes:

[0035] Based on the readings of the ion interference compensation channel Compared with the reference baseline Calculation of ion strength compensation factor based on the difference The ion intensity compensation factor was used to measure the readings of the main perspiration rate sensing channel. The time-varying rate is corrected to obtain the compensated instantaneous perspiration rate. :

[0036]

[0037]

[0038] in, The ion interference sensitivity coefficient, The proportionality constant of the sweat rate sensing channel. The sampling period is determined by the empirical exponential relationship between temperature and saturated vapor pressure, based on the readings of the skin temperature sensor. and ambient temperature sensor readings Calculate the skin-side saturated water vapor pressure With environmental saturated water vapor pressure :

[0039]

[0040] The relative humidity sensor readings are used to calculate the skin-side saturated water vapor pressure and the ambient-side saturated water vapor pressure. The corrected difference is then multiplied by a preset equivalent mass transfer coefficient. Effective gas-liquid interface area The equivalent volumetric flow rate of evaporation is obtained. :

[0041]

[0042]

[0043] The compensated instantaneous sweat rate Equivalent volumetric flow rate of evaporation Together they constitute the fluid transport flux.

[0044] Preferably, the processing module calculates the fluid transport flux of the sweat pathway in the clothing based on the collected microclimate signal and sweat rate related signal, so as to quantify the total hydraulic driving force that drives the migration of sweat in the porous medium of the fabric. The total hydraulic driving force is composed of active secretion pressure and passive evaporation tension.

[0045] First, for the calculation of the active secretion part, the measurement error caused by the ion-capacitance / impedance coupling effect must be addressed. That is, the increase in electrolyte concentration in sweat will parasitically increase the capacitance reading or decrease the impedance modulus, causing the system to misjudge the amount of sweat as an increase.

[0046] Therefore, by employing a decoupling algorithm, the ion strength compensation factor is first calculated. ,in As a reference baseline, this represents the background conductivity / impedance reading of the sensor in deionized water or simulated sweat at very low concentrations (e.g., 10 mM NaCl). Its preferred value depends on the sensor electrode spacing and is typically within [specific range]. to between; The ion interference sensitivity coefficient characterizes the gain shift in sweat rate readings caused by a unit change in background signal. It is obtained through a standard salt solution gradient calibration experiment, with an optimal range of [value missing]. to This factor was used to analyze the readings of the main perspiration rate sensor channel. The time-determined component (such as the interdigital capacitance value) is corrected to obtain the compensated instantaneous sweat rate. , here The sampling period (e.g., 1 second). This is the proportionality constant of the sweat rate sensing channel, used to map the rate of change of the electrical signal to volumetric flow rate (units). This parameter depends on the geometric volume of the fabric microfluidic channel and the electrode sensitivity, with an optimal value in [value missing]. to between.

[0047] Secondly, the calculations for the passive evaporation portion are based on the gas-liquid interface evaporation driving mechanism in mass transfer, namely, the evaporation flux is proportional to the water vapor partial pressure difference across the interface. The examples utilize Tetens' empirical formula. skin temperature and ambient temperature Converting to the corresponding saturated vapor pressure (in kPa), the constants 0.61078, 17.27, and 237.3 are empirical vapor pressure constants applicable to the 0-100℃ range. Then, the evaporation flux per unit area is calculated. ,in Relative humidity (normalized value 0-1). The preset equivalent mass transfer coefficient characterizes the moisture permeability and boundary layer convection intensity of the fabric material. Its optimal value is obtained by testing the fabric evaporation rate in a constant temperature and humidity chamber, with a typical range of [value missing]. to Ultimately, the evaporation equivalent volumetric flow rate... ,in The effective gas-liquid interface area (i.e., the area of ​​the humid fabric exposed to air) determined for clothing design is preferably taken as a value of to .

[0048] In conclusion, Characterizes the source term injected into the skin. The fluid transport flux, which is characterized by the terminal suction term, the skin injection source term, and the terminal suction term, can describe the true hydraulic state of sweat under complex conditions of unsteady injection and continuous evaporation.

[0049] In a preferred embodiment, identifying the fabric pore transport retention coefficient, which characterizes the fluid retardation properties of the fabric pore structure, includes:

[0050] A very small volume tracer pulse is released within a clothing microfluidic module, and the tracer response concentration at the sensing site is recorded. Divide the tracer response concentration by the total integral of the tracer response concentration and the sampling period. The product of these factors yields the normalized residence time distribution curve. :

[0051]

[0052] Select the tail time interval of the dwell time distribution curve Calculate the logarithmic decay slope within this interval. :

[0053]

[0054] in and These are the mean values ​​within the interval; the logarithmic decay slope is... The negative value after adding one is used as the pore transport retention coefficient of the fabric. :

[0055] .

[0056] Preferably, the processing module identifies the fabric pore transport retention coefficient, which characterizes the fluid impediment properties of the fabric pore structure, to quantify the time memory effect caused by irregular porous media in electronic textiles and apparel through in-situ system identification. This is based on a combination of Residence Time Distribution (RTD) theory and anomalous diffusion dynamics.

[0057] First, a tiny volume of tracer pulse is released within the garment's microfluidic module, typically during system initialization or periodic self-calibration. The microfluidic structure integrates a miniature, electrically controlled reservoir containing a highly conductive tracer solution (such as a high-concentration NaCl solution or a bio-inert electrolyte). Upon receiving a command, this tracer pulse is instantaneously injected into the main sweat channel via a thermally actuated or electrowetting valve, with the volume precisely controlled. to The system then transmits pulsed fluid through downstream sensing points at a high frequency ( The tracer response concentration is recorded (preferably from 0.1 seconds to 1.0 seconds). The concentration curve exhibits a rapid rise from zero to a peak in the time domain, followed by a slow, trailing decline. To convert the concentration into a probability density function, normalization is necessary; the calculation formula is as follows. ,in To the total number of sampling points sufficient to cover the complete effluent process of the tracer (typically) (more than 30 minutes), denominator term The normalization operation ensures that the total amount of tracer passing through the cross section (mass or moles) is represented by the amount of tracer. , making The residence time of the corresponding fluid element in the system is The probability density.

[0058] Subsequently, The extraction of the tail features of the curve is crucial because capillary retention and fractal network effects in fabric pores are primarily manifested in the long-tail region. The processing module selects the tail time interval of the residence time distribution curve. The selection of this interval avoids the initial peak dominated by convection and the end dominated by noise; the preferred selection criterion is based on... Decline to peak Start of time (corresponding) (Until the signal-to-noise ratio decreases to) The time (corresponding) Within this interval, the moisture-wicking properties of the fabric follow a power-law decay model. This model represents a typical characteristic of fractal porous media transport.

[0059] To extract the power-law exponent, the system calculates the linear regression slope in a double logarithmic coordinate system. Its calculation formula adopts the analytical solution form of the least squares method: ,in Represents the natural logarithm of the sampling time index. The natural logarithm of the normalized concentration. and These are the arithmetic mean of the logarithm of time and the logarithm of concentration within the interval, respectively. The numerator of this formula is the covariance of the logarithm of time and the logarithm of concentration, and the denominator is the variance of the logarithm of time. Their ratio gives the slope of the fitted line in the double logarithmic coordinate system.

[0060] Finally, based on the theoretical derivation of the anomalous diffusion equation, the decay power and the order of the fractional derivative of the probability density function are... There is an algebraic relationship, namely the slope. Therefore, the processing module will use the logarithmic decay slope. The negative value after adding one is used as the pore transport retention coefficient of the fabric. ,Right now This coefficient Usually located Within the range, the smaller the value, the stronger the effect of the fabric structure on dragging and retaining sweat, that is, the longer the memory.

[0061] In a preferred embodiment, the overall system response lag time is constructed by combining the biopassivation time constant of the sensing interface with the fluid transport flux, including:

[0062] Utilizing sampling time Multiply by the aforementioned residence time distribution curve The hydraulic residence time, which characterizes the fluid transport flux, is then calculated by integrating and summing the results. :

[0063]

[0064] Calculate the sensor sensitivity corresponding to sweat biochemical signals The natural logarithm of the attenuation ratio within the observation window, using the sampling period. The negative of the ratio to the natural logarithm is used to calculate the biopassivation time constant of the sensing interface. :

[0065]

[0066] The hydraulic residence time Biopassivation time constant of the sensing interface Add one to the ratio, then multiply by the hydraulic residence time. The overall system response lag time is obtained. :

[0067] .

[0068] Preferably, the processing module combines the biopassivation time constant of the sensing interface with the fluid transport flux to construct the system's comprehensive response lag time, thereby coupling and modeling the simple fluid dynamics stagnation with the sensor response lag caused by biochemical contamination, in order to obtain a comprehensive time scale that reflects the system's true signal delay.

[0069] First, the processing module calculates the hydraulic residence time, which characterizes the fluid transport flux. Hydraulic residence time This is equivalent to the average residence time (MeanResidenceTime) of sweat droplets in a porous textile network, from the secretion point to the sensing site. The calculation is based on the principle of statistical moments, utilizing sampling time. Multiply by the normalized residence time distribution curve And then perform integration and summation, the formula is: ,in For discrete-time indexing, The sampling period is preferably 1.0 second. This is the cutoff point for integration; its value must ensure coverage. The curve contains over 99% of the total energy, typically corresponding to a time length of 20 to 40 minutes.

[0070] Secondly, to address the sensitivity attenuation caused by biofilm formation on the electrode surface due to the adsorption of lipids, proteins, and microorganisms in sweat, the system calculates the biopassivation time constant of the sensing interface in real time. Based on a first-order kinetic deactivation model, sensor sensitivity... (defined as potential signal) For concentration partial derivatives Or through charge transfer resistance by electrochemical impedance spectroscopy The inversion yields an exponential decay over time. At a length of... Within the sliding observation window (preferably 5 to 10 minutes), the sensitivity is calculated at the end of the window. and the start time The natural logarithm of the decay ratio This value is negative, representing the cumulative degree of decay. Using the formula... Mapping this decay rate to a time constant, here As a normalization factor, The smaller the value, the faster the passivation rate caused by contamination.

[0071] Finally, based on the reaction-transport coupling theory, the transport time and chemical reaction time are combined in a dimensionless manner, and the calculation formula is the system's overall response lag time. Specifically, when biological passivation is extremely slow ( )hour, Approaching the hydraulic residence time ; while when biological passivation is faster ( When the concentration is relatively small, the sensor's response to concentration changes will be significantly slowed down by the surface retardation process, resulting in a lag time in the overall system response. Non-linear growth.

[0072] In a preferred embodiment, the transmission distortion correction factor is calculated based on the lag time and the dynamic change period of physiological parameters, including:

[0073] Select a fixed-length time window Calculate the skin temperature at the current moment. Skin temperature at the start of the window The absolute value of the difference, combined with the sampling period and the smallest positive number Calculate the time scale of physiological changes To characterize the dynamic change cycle of the physiological parameters:

[0074]

[0075] Calculate the overall response lag time of the system. With respect to the time scale of the physiological changes The ratio of the fabric pore transport retention coefficient is then calculated. The transmission distortion correction factor is obtained by exponentiation. :

[0076] .

[0077] Preferably, the characteristic time scales of physiological parameter changes are first extracted using a sliding time window mechanism. A time window of fixed length is selected. This parameter determines the system's field of view for observing physiological trends, and the preferred value is 300 to 600 sampling points (corresponding to 300 to 600 seconds). This is because the significant changes in core body temperature or skin temperature during strenuous exercise typically occur within minutes. (Calculate the current skin temperature.) Skin temperature at the start of the window The absolute value of the difference reflects the magnitude of the physiological state shift during this period. To avoid division by zero errors and smooth the calculation results, a very small positive number is introduced. (Preferred) ), and combined with the sampling period Construct a time-scale formula for physiological change characteristics: Specifically, when skin temperature changes drastically (i.e., When the denominator term is significantly greater than 1, it makes the denominator term significantly greater than 1. A rapid decrease accurately reflects a state of rapid change in the current physiological process; conversely, when the temperature is stable, the denominator approaches 1. Approaching the window length This indicates that the physiological process is in a slow-change or steady state.

[0078] Next, a transmission distortion correction factor is constructed through dimensionless processing. This factor is used to correlate transmission lag with physiological dynamics. It utilizes the overall system response lag time. Time scale of physiological changes Calculate the ratio of the two. This ratio is the Deborah number, which in rheology describes the relative relationship between material response time and process time: when the ratio is much less than 1, the system response is fast enough and distortion is negligible; when the ratio is greater than 1, the system exhibits significant memory and hysteresis.

[0079] However, considering the nonlinear transport characteristics caused by the fractal structure of porous fabrics, a simple linear ratio is insufficient to describe the attenuation law of distortion. Therefore, by introducing the fabric pore transport retention coefficient... As an exponent, the power operation is performed to obtain the final transmission distortion correction factor: Here, It plays a key role in compressing or amplifying distortion weights: for Smaller fabric structures (i.e., with a stronger memory effect) show better retention even with a slight increase in the time ratio. This will also increase significantly, thus allowing for more severe penalties in risk assessments. This ensures the system can be tailored to different fabric structures (made from...). (characterization) and different physiological states (by) (Characteristics) Adaptively output distortion metric.

[0080] In a preferred embodiment, a porous media transport inversion model is established using the fabric pore transport retention coefficient, including:

[0081] Utilizing the fabric pore transport retention coefficient As the order of the fractional derivative, the fractional differential operator is constructed using the Grinwald-Letnikov discretization method. :

[0082]

[0083] in , For gamma function, The signal value at a historical moment; the biochemical signal of the sweat The result after processing by the fractional differential operator is multiplied by the system's overall response lag time. of The concentration is then raised to the power of the power and added to the original sweat biochemical signal to obtain the real-time concentration data at the skin end. :

[0084] .

[0085] Preferably, a porous medium transport inversion model is established using the pore transport retention coefficient of the fabric to eliminate the long-tail drag effect of the porous structure of electronic textiles and clothing on biochemical signals, thereby restoring the true physiological change trend of the skin surface.

[0086] Based on the ability of fractional calculus to describe the phenomenon of anomalous diffusion, the transport of sweat in the fractal porous network of fabrics does not follow classical Fick's law, but rather obeys a fractional relaxation equation. In order to obtain the measured output signal (Right now Inversely deduce the unknown input signal (Right now The processing module must construct and execute the discretized inverse operator of the equation.

[0087] In practice, the first step is to utilize the fabric pore transport retention coefficient. (The value is preferably between 0.3 and 0.8, representing the tortuosity and retention strength of the porous network) as the order of the fractional derivative. The processing module uses the Grünwald-Letnikov discretization method to construct the numerical fractional differential operator. This is achieved by approximating the fractional derivative as a weighted linear combination of the signal values ​​at the current and historical moments. The calculation formula is as follows: ,in represent Historical signal values ​​from one sampling period ago, The sampling period (preferably set to 1 second to ensure temporal resolution). The weighting coefficients in the formula... It is the core calculation term, utilizing the definition of the generalized binomial coefficients and the gamma function. Calculated as .because If it is a non-integer, the coefficient Follow The increase in gamma function exhibits power-law decay rather than truncation, corresponding to the long memory characteristics of the fabric medium. To balance accuracy and computational efficiency, the gamma function... The value is obtained through the Lanczos approximation algorithm, and the calculation precision needs to be maintained at more than 6 decimal places.

[0088] After constructing the differential operator, real-time skin concentration data is obtained based on the inverse transformation logic of the fractional relaxation model. It consists of the linear superposition of the measured signal itself and its fractional derivative terms. The specific calculation formula is as follows: In this formula, It is the apparent concentration measured by the ion-selective electrode at the current moment; It is obtained by convolving the historical measurement sequence using the Grinwald-Laitnikov operator. The fractional derivative at time step represents the combination of the rate of signal change and historical inertia. It is the overall system response lag time (in seconds). Power of 1 As a generalized time scale scaling factor (dimensions: ), used to balance fractional derivative terms (dimensions: The dimensions of the concentration term are determined, and the strength of the inversion compensation is adjusted. Specifically... The term estimates the high-frequency variation components that are swallowed up and smoothed out by the fabric structure, and compensates them back to the original measurements. This significantly sharpens the rising and falling edges of the signal, restoring the true physiological peak values ​​that are attenuated during transmission.

[0089] In a preferred embodiment, the sweat biochemical signal is restored to real-time concentration data at the skin end, and the transmission distortion correction factor is introduced as a confidence weighting term into the health risk assessment function. A warning signal is output when the assessment result exceeds a preset safety threshold, including:

[0090] Compensated instantaneous sweat rate With the real-time concentration data at the skin end Multiply to obtain the instantaneous loss rate, and within a preset time window The cumulative electrolyte loss is obtained by integrating and summing the instantaneous loss rates. :

[0091]

[0092] Calculate skin temperature Compared with reference temperature The difference and temperature scale parameters The ratio is used to perform an exponential operation on the ratio, then one is added, and finally the natural logarithm is calculated to obtain the heat load index. :

[0093]

[0094] Calculate the transmission distortion correction factor mentioned above. The natural logarithm yields a confidence penalty term. :

[0095] The cumulative electrolyte loss and the critical loss threshold The ratio, the heat burden index and temperature weight The product of the confidence penalty term and the distortion weight Adding the products together yields the overall risk value. :

[0096]

[0097] The combined risk value is mapped to a probability value using an S-shaped function. The probability value is then incremented by half and rounded to output a binarized warning signal. :

[0098]

[0099]

[0100] in is the kurtosis constant.

[0101] Preferably, the instantaneous electrolyte loss rate is first calculated, and the compensated instantaneous sweat rate is then... (Unit is) The real-time sodium ion concentration at the skin end, recovered via fractional-order inversion. (Unit is) The product of these two products represents the molar mass of electrolyte lost per unit time.

[0102] Subsequently, within the preset time window The loss rate is calculated by discrete integral summation, and the formula is as follows: ,in The preferred sampling point is 300 to 600 (corresponding to 5 to 10 minutes). This time length is determined based on the minimum significant onset cycle of acute dehydration or electrolyte imbalance in the human body. This quantifies the cumulative load loss in the short term.

[0103] Meanwhile, in order to assess the risk of thermal stress, the processing module calculates the thermal burden index. The design employs a variant of the Softplus function. ,in For real-time skin temperature, For reference to the safe threshold temperature, it is preferably set to 35.0. Up to 37.5 It is determined based on the critical point at which heat dissipation from human skin is blocked; For temperature-scale parameters, a value of 0.5 is preferred. Up to 2.0 This is used to adjust for the nonlinear growth rate of risk with increasing temperature. The advantage of this formula lies in the fact that it is not merely a simple difference, but provides a smooth, non-negative penalty: when... hour, Approaching 0; when Exceed At that time, the risk value increased exponentially and rapidly, which is consistent with the nonlinear biological characteristics of thermal damage.

[0104] By introducing a confidence penalty term The calculation formula is: Here, To transmit distortion correction factors. Based on the conservative early warning principle: when A larger reading indicates a severe retention effect of clothing or slow response due to biocontamination, in which case the sensor reading may underestimate the true rate of risk escalation. By introducing... As a positive penalty, the system proactively raises the overall risk score when signal reliability is low, thereby offsetting the possibility of missed detections caused by lag. Subsequently, the system calculates the comprehensive risk value. ,in The critical loss threshold constant is preferably 5 mmol to 20 mmol, and is determined based on individual body weight and tolerance. and The dimensionless weighting coefficient is preferably set between 0.5 and 1.5 to balance the relative importance of loss risk, thermal risk, and distortion risk. Finally, to output a deterministic binarized early warning signal... The system uses a sigmoid function for mapping: and calculate .in The steepness constant is preferably 10 to 20, such that when the comprehensive risk value... When it slightly exceeds 1.0, the probability It rapidly approaches 1, triggering (Warning status); otherwise, maintain. (Normal state). This sigmoid-based soft thresholding mechanism effectively avoids signal chattering near the critical point, ensuring the stability and robustness of the warning output.

[0105] like Figure 2The diagram illustrates an application scenario comprised of a smart garment body, a data acquisition and processing module, a mobile terminal, and a cloud-based analysis platform. The smart garment body is a flexible electronic fabric structure. A porous fabric / microfluidic guide layer is positioned on the skin side to collect sweat, allowing sweat secreted by the skin to be directionally transported along the fabric's pores to the sensing area via capillary action. The sensing area integrates a sweat ion sensing unit, a sweat rate sensing unit, and a microclimate sensing unit. The sweat ion sensing unit uses an ion-selective electrochemical sensing structure to acquire signals related to sodium and chloride ions in sweat. The sweat rate sensing unit uses an electrical impedance / capacitance method to acquire signals related to sweat wetting and flow changes. The microclimate sensing unit collects skin-side temperature, ambient temperature, and relative humidity to characterize the garment's microclimate state and for subsequent temperature and humidity corrections. Each sensing unit is electrically connected to the acquisition and processing module via flexible wires / fabric conductors. The acquisition and processing module integrates signal conditioning, electrochemical sampling, sweat rate measurement excitation and demodulation, microclimate sampling, processor, and wireless communication circuitry, and is powered by a battery. The acquisition and processing module synchronously samples the ion signal and sweat rate signal and performs noise reduction and temperature compensation. It further utilizes the ion correlation channel to decouple and compensate the sweat rate electrical channel to reduce the interference of sweat electrolyte concentration changes on sweat rate estimation. At the same time, it combines the skin temperature, ambient temperature, and relative humidity obtained from the microclimate sensor to estimate the contribution of sweat evaporation, thus obtaining the total fluid flux reflecting sweat injection and evaporation loss. During calibration and operation, the system can utilize tracer pulses to generate a residence time distribution response within the fabric's pore channels during initialization or periodic self-calibration phases. This identifies the fabric's transport and retention characteristics, and, together with the estimated result of the sensor interface sensitivity decaying over time, yields a comprehensive response hysteresis. Subsequently, based on the comparison between the comprehensive hysteresis and the rate of physiological change, a distortion correction intensity is constructed. This is used to compensate for the measured sweat ion signal through porous media transport inversion, outputting a real-time ion concentration sequence that more closely approximates the actual changes at the skin end. In the risk assessment phase, the acquisition and processing module calculates the electrolyte loss level based on the compensated sweat rate and ion concentration, and combines this with the heat burden represented by skin temperature and the distortion correction intensity to form a comprehensive risk value. When the comprehensive risk value exceeds a preset threshold, the acquisition and processing module sends a risk warning to a mobile device via wireless communication, while simultaneously uploading it to a cloud analysis platform for long-term trend analysis and individualized parameter updates. If the threshold is not exceeded, the system maintains continuous monitoring and cyclically updates the above calculation and assessment process, thereby achieving real-time monitoring and early warning of dehydration / electrolyte imbalance and heat-related risks while worn.

[0106] The embodiments of this example have been described above. However, this example is not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms based on the guidance of this example, and all of them are within the protection scope of this example.

Claims

1. A wearable intelligent clothing health monitoring system, characterized in that, include: A multimodal sensing unit integrated into the electronic textile and garment body is used to continuously collect sweat biochemical signals, sweat rate related signals, and garment microclimate signals; as well as The data processing module connected to the multimodal sensing unit is configured to perform the following operations: calculate the fluid transport flux of the clothing sweat pathway based on the microclimate signal and the sweat rate correlation signal, and identify the fabric pore transport retention coefficient, which characterizes the fluid resistance properties of the fabric pore structure. By combining the biopassivation time constant of the sensing interface with the fluid transport flux, a comprehensive system response lag time is constructed, and a transmission distortion correction factor is calculated based on this lag time and the dynamic change cycle of physiological parameters. A porous media transport inversion model is established using the fabric pore transport retention coefficient to restore the sweat biochemical signal to real-time concentration data at the skin end. At the same time, the transmission distortion correction factor is introduced as a confidence weighting term into the health risk assessment function, and an early warning signal is output when the assessment result exceeds a preset safety threshold.

2. The wearable intelligent clothing health monitoring system according to claim 1, characterized in that, A multimodal sensing unit integrated into the electronic textile garment body is used to continuously collect sweat biochemical signals, sweat rate-related signals, and garment microclimate signals, including: An ion-selective electrode assembly for acquiring sodium ion potential signals and chloride ion potential signals; a main perspiration rate sensing channel for acquiring impedance modulus readings or capacitance readings, and an ion interference compensation channel for acquiring background ion concentration data; and a skin temperature sensor deployed on the inner side of the garment in contact with the skin surface, an ambient temperature sensor deployed on the outer side of the garment, and a relative humidity sensor for monitoring the humidity inside the garment.

3. The wearable intelligent clothing health monitoring system according to claim 2, characterized in that, The fluid transport flux of the clothing sweat pathway is calculated based on the microclimate signal and the sweat rate correlation signal, including: The ion intensity compensation factor is calculated based on the difference between the reading of the ion interference compensation channel and the reference baseline. This ion intensity compensation factor is then used to correct the time change rate of the main perspiration rate sensing channel reading by division, resulting in the compensated instantaneous perspiration rate. Using the empirical exponential relationship between temperature and saturated vapor pressure, the skin-side saturated vapor pressure and the ambient-side saturated vapor pressure are calculated based on the readings of the skin temperature sensor and the ambient temperature sensor, respectively. The difference between the skin-side saturated vapor pressure and the ambient-side saturated vapor pressure after correction by the relative humidity sensor reading is calculated, and this difference is multiplied by a preset equivalent mass transfer coefficient and the effective gas-liquid interface area to obtain the evaporation equivalent volumetric flow rate. The compensated instantaneous perspiration rate and the evaporation equivalent volumetric flow rate together constitute the fluid transport flux.

4. The wearable intelligent clothing health monitoring system according to claim 1, characterized in that, The fabric pore transport retention coefficient, which characterizes the fluid retention properties of fabric pore structure, includes: A very small volume tracer pulse is released within the microfluidic module of the garment, and the tracer response concentration at the sensing site is recorded. The tracer response concentration is divided by the product of the total integral of the tracer response concentration and the sampling period to obtain a normalized residence time distribution curve. The tail time interval of the residence time distribution curve is selected, and the variance of the logarithm of time and the covariance of the logarithm of time and the logarithm of the distribution curve within this interval are calculated. The covariance is divided by the variance to obtain the logarithmic decay slope. The negative value of the logarithmic decay slope plus one is used as the pore transport retention coefficient of the fabric.

5. A wearable intelligent clothing health monitoring system according to claim 3, characterized in that, By combining the biopassivation time constant of the sensing interface with the fluid transport flux, the overall system response lag time is constructed, including: The hydraulic residence time, characterizing the fluid transport flux, is calculated by multiplying the sampling time by the residence time distribution curve and integrating and summing the results. The natural logarithm of the sensor sensitivity attenuation ratio corresponding to the sweat biochemical signal is calculated, and the negative of the ratio of the sampling period to the natural logarithm is used to calculate the bio-passivation time constant of the sensing interface. The ratio of the hydraulic residence time to the bio-passivation time constant of the sensing interface is added by one and then multiplied by the hydraulic residence time to obtain the overall system response lag time.

6. A wearable intelligent clothing health monitoring system according to claim 1, characterized in that, Based on this lag time and the dynamic change cycle of physiological parameters, the transmission distortion correction factor is calculated, including: A fixed-length time window is selected, and the absolute value of the difference between the skin temperature at the current moment and the skin temperature at the beginning of the window is calculated. This absolute value is divided by the window time length to obtain the temperature change rate. The physiological change characteristic time scale is obtained by dividing the window time length by the sum of one and the normalized term of the temperature change rate, which is used to characterize the dynamic change cycle of the physiological parameter. The ratio of the system's overall response lag time to the physiological change characteristic time scale is calculated, and this ratio is raised to the power of the fabric pore transport retention coefficient to obtain the transmission distortion correction factor.

7. A wearable intelligent clothing health monitoring system according to claim 4, characterized in that, A porous media transport inversion model is established using the fabric pore transport retention coefficient, including: Using the fabric pore transport retention coefficient as the order of the fractional derivative, a fractional differential operator is constructed using the Greenwald-Letnikov discretization method. This fractional differential operator is obtained by weighted summation of historical sweat biochemical signals, where the weights are determined by binomial coefficients based on gamma functions. The result of processing the sweat biochemical signal using the fractional differential operator is multiplied by a power term of the system's overall response lag time, and then added to the original sweat biochemical signal to obtain the real-time concentration data at the skin end.

8. A wearable intelligent clothing health monitoring system according to claim 1, characterized in that, The sweat biochemical signal is restored to real-time concentration data at the skin end. Simultaneously, the transmission distortion correction factor is introduced as a confidence weighting term into the health risk assessment function. When the assessment result exceeds a preset safety threshold, a warning signal is output, including: The instantaneous loss rate is obtained by multiplying the compensated instantaneous sweat rate by the real-time concentration data at the skin end, and the cumulative electrolyte loss is obtained by integrating and summing the instantaneous loss rate within a preset time window. The ratio of the difference between the skin temperature and the reference temperature to the temperature scale parameter is calculated, and the ratio is exponentially calculated, then incremented by one, and its natural logarithm is calculated to obtain the heat burden index. The confidence penalty term is obtained by adding the natural logarithm of the transmission distortion correction factor. The comprehensive risk value is obtained by adding the ratio of the cumulative electrolyte loss to the critical loss threshold, the product of the heat burden index and the temperature weight, and the product of the confidence penalty term and the distortion weight. The comprehensive risk value is mapped to a probability value using an sigmoid function, and the probability value is halved and rounded to output a binarized warning signal.

Images