A health management system based on dual regulation of body fat and endogenous fat

Abstract

The application is a health management system based on body fat and endogenous fat double regulation, and relates to the field of bioimpedance measurement technology, comprising: a multi-frequency impedance signal synchronous acquisition module; a time-frequency conversion and physiological noise elimination module; a body surface and internal impedance characteristic separation module; a fat distribution nonlinear mapping calculation module; a multi-dimensional parameter compensation and result output module. In the application, the multi-frequency composite excitation synchronous acquisition mechanism realizes concurrent synchronous acquisition of multi-frequency signals, greatly improving the synchronicity and signal-to-noise ratio of the original data; through an adaptive time-frequency domain joint denoising algorithm, respiratory baseline drift and transient high-frequency interference are accurately eliminated, ensuring the purity of the data source; through microscopic equivalent circuit modeling and frequency domain phase angle characteristic separation algorithm, the subcutaneous and internal fat signals are accurately separated, the internal fat quantitative index with clinical reference value can be output, and the error limitation of traditional empirical formula estimation is eliminated.

Description

Technical Field

[0001] This invention relates to the field of bioimpedance measurement technology, and in particular to a health management system based on the dual regulation of body fat and visceral fat. Background Technology

[0002] With the increasing awareness of health management among residents, accurate monitoring of body fat and visceral fat has become a core aspect of the prevention and control of obesity-related chronic diseases. Body fat measurement devices based on bioelectrical impedance analysis have become the mainstream solution for home health management due to their portability, low cost, and lack of radiation. However, existing technologies have the following drawbacks: Existing equipment mostly adopts a single-frequency sequential scanning excitation method. Multi-frequency signal acquisition is time-consuming, has poor synchronization, is easily affected by human physiological activities, and has a low signal-to-noise ratio of raw data. Some multi-frequency equipment has insufficient sampling rate and lacks an efficient buffer scheduling mechanism. High-frequency sampling is prone to data bus congestion, data packet loss, or stack overflow, which cannot meet the needs of continuous physiological monitoring. Fixed filtering or fast Fourier transform (FFT) is often used for denoising, which cannot effectively eliminate low-frequency nonlinear baseline drift caused by respiratory movements; moreover, FFT transform loses time-domain information and cannot accurately locate and remove transient high-frequency spikes caused by poor electrode contact, electromagnetic interference, and muscle electrical activity, which directly leads to a serious decrease in subsequent calculations. Unable to solve the problems of the "bypass escape effect" of current caused by low impedance tissues on the body surface and the "potential shielding" of deep visceral fat, it is impossible to accurately separate the impedance signals of subcutaneous fat and visceral fat. It can only estimate the visceral fat level through empirical formulas based on parameters such as BMI and waist circumference, and cannot output clinical-grade visceral fat cross-sectional area values. The measurement results have large errors and low clinical reference value.

[0003] Therefore, a health management system based on dual regulation of body fat and visceral fat is proposed to address the aforementioned problems. Summary of the Invention

[0004] The purpose of this invention is to provide a health management system based on the dual regulation of body fat and visceral fat in order to solve the above-mentioned problems.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A health management system based on dual regulation of body fat and visceral fat includes: The multi-frequency impedance signal synchronous acquisition module is configured to generate multi-frequency composite excitation signals and send them to the human body, synchronously acquire weak voltage drop signals of independent segments of the human body, and extract the original complex impedance matrix under multi-frequency bands. The time-frequency conversion and physiological noise cancellation module is configured to: acquire the original complex impedance matrix of the output, remove transient high-frequency glitches and noise, and reconstruct the output multi-frequency complex impedance array; The body surface and visceral impedance feature separation module is configured to: resolve the reference electrical characteristic parameters from the multi-frequency complex impedance array and extract the characteristic dispersion coefficient of visceral fat; The nonlinear mapping calculation module for fat distribution is configured to: calculate the initial fat-free mass and total fat percentage based on the reference electrical characteristic parameters, and calculate the initial visceral fat cross-sectional area by combining the multi-frequency impedance ratio and the characteristic dispersion coefficient. The multidimensional parameter compensation and result output module is configured to: obtain the dynamic conductivity compensation coefficient of skin and ambient temperature, calculate the water steady-state distribution index, calibrate the initial total fat percentage and the initial visceral fat cross-sectional area, calculate the first-order time derivative in combination with the timestamp, and output the high-dimensional tensor of the health status.

[0006] Preferably, the multi-frequency impedance signal synchronous acquisition module generates the multi-frequency composite excitation signal in the following specific way: The sinusoidal waves corresponding to a preset number of basic excitation frequency points are linearly superimposed to generate a composite current time-domain discrete equation. Meanwhile, the multi-frequency impedance signal synchronous acquisition module allocates a multi-channel ring buffer storage area in the system memory. When the filling difference of the buffer exceeds the preset safety level, a hardware interrupt is triggered and the data is calculated and read from the read pointer address.

[0007] Preferably, the expression for the time-domain discrete equation of the composite current is: ; in, The time-domain discrete equation is for the composite current. Index of the base excitation frequency point For the first The current amplitude at a basic excitation frequency point, limited to a safety threshold range. For the set discrete target frequency points, The initial phase of each frequency component; Furthermore, the initial phase difference between each frequency component is set to a multiple of a preset angle to avoid excessively large instantaneous peak values ​​in the multi-frequency signal.

[0008] Preferably, the step of the time-frequency conversion and physiological noise cancellation module using adaptive cubic spline interpolation to fit and eliminate low-frequency respiratory baseline drift includes: Local minima and their corresponding time indices of the original complex impedance real part sequence are extracted within a dynamic time sliding window and used as anchor points. A cubic spline interpolation polynomial is constructed using anchor points. A system of linear algebraic equations is established based on the function values ​​at the intersection of adjacent intervals and the curve smoothing boundary conditions with continuous derivatives. The coefficients of the interpolation polynomial are then calculated to obtain a continuous fitting function. Subtracting the continuous fitting function from the original complex impedance real part sequence yields the pure impedance time series after eliminating low-frequency respiratory baseline drift.

[0009] Preferably, the specific process for removing transient high-frequency glitch noise is as follows: A continuous wavelet transform is performed on the pure impedance time series using a mother wavelet function that meets preset conditions to generate a two-dimensional floating-point matrix of wavelet coefficients containing time-frequency states. The absolute median difference statistical algorithm is used to adaptively calculate the noise threshold limit of the current scale level; Wavelet coefficients with absolute values ​​below the noise threshold are forcibly cleared to zero, while wavelet coefficients with absolute values ​​above the noise threshold are retained. These coefficients are then substituted into the wavelet inverse transform restoration calculus equation that satisfies the allowable constant constraint to generate a reconstructed, pure multi-frequency complex impedance array.

[0010] Preferably, the complex function corresponding to the equivalent circuit topology model is: ; in, It is the overall complex impedance complex variable function. The excitation frequency point of the alternating current. To characterize the extreme values ​​of DC impedance for extracellular fluid resistance, The limiting impedance is used to characterize the overall conductivity of water. The characteristic relaxation time constant is The phase relaxation distribution coefficient is... The basic imaginary unit; The damping iterative optimization algorithm generates a residual column vector consisting of the complex difference between the actual measured complex impedance and the model predicted impedance in each iteration, and adaptively adjusts the convergence damping factor according to the residual descent rate to solve for the DC impedance extreme value, the limiting impedance extreme value, the characteristic relaxation time constant, and the phase relaxation distribution coefficient as reference electrical characteristic parameters.

[0011] Preferably, the process of extracting the characteristic dispersion coefficients of visceral fat by the body surface visceral impedance characteristic separation module is as follows: Extract the pure multi-frequency complex impedance array and calculate the phase angle distribution curve at discrete frequencies; The equal-interval linear interpolation resampling engine is invoked to map the phase angle distribution curve into a dense uniform array with equal intervals at fixed bandwidths; The central finite difference method is used to perform second-order numerical differentiation on a dense uniform array, and local maxima are searched in the generated second-order derivative array to extract the characteristic dispersion coefficients specific to visceral fat.

[0012] Preferably, the process by which the nonlinear mapping calculation module for fat distribution calculates the initial fat-free mass and total fat percentage is as follows: The total fluid volume of the human body is estimated by calling a nonlinear multivariate polynomial mapping equation. The nonlinear multivariate polynomial mapping equation introduces the ratio of the square of the height parameter to the high-frequency limiting impedance of the whole body as the core dominant term, and uses the total mass parameter and waist circumference parameter for system bias compensation. The initial fat-free mass is obtained by dividing the total body fluid volume by the physiological cell hydration constant; the absolute total fat mass is obtained by subtracting the initial fat-free mass from the total mass, and then the initial total fat percentage is calculated.

[0013] Preferably, the method further includes: To compensate for the differences in visceral volume caused by different user skeleton sizes, the system uses a series impedance voltage divider topology estimation formula to calculate the independent mass of the torso. Then, the system extracts the low-frequency extreme values ​​specific to the torso. and high frequency extrema Combined with the extracted specific dispersion coefficients Substitute this into the nonlinear exponential mapping equation for calculating the anatomical cross-sectional area of ​​visceral fat at the umbilicus plane:

[0014] in, This represents the initial cross-sectional area of ​​visceral fat. It is a natural exponential function. The characteristic relaxation time constant of the torso. , , , These are the system's preset weighted matrix constants.

[0015] Preferably, the multidimensional parameter compensation and result output module specifically includes: Real-time synchronous acquisition of absolute temperature sequence of skin surface at electrode contact point and ambient temperature sequence The system instantiates a two-factor linear conductivity compensation coefficient mapping function based on a variant of the nonlinear Arrhenius equation within the central processing unit. ; Call DC limiting impedance and ultra-high frequency limiting impedance Constructing a water steady-state distribution index in memory ; Real-time calculation of the current When the value changes beyond a preset threshold within a preset time period, adaptive penalty correction is activated to adjust the initial coarse calculation results for fat-free mass. Perform damped convergence compensation; Obtain real-time total body fat percentage based on the user's actual physiological state. and visceral fat cross-sectional area ; to compute and obtain a high-dimensional health state tensor with temporal and spatial coupling properties. .

[0016] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. This invention achieves concurrent synchronous acquisition of multi-frequency signals through a multi-frequency composite excitation synchronous acquisition mechanism, significantly improving the synchronization and signal-to-noise ratio of the raw data; through an adaptive time-frequency domain joint denoising algorithm, it accurately eliminates respiratory baseline drift and transient high-frequency interference, ensuring the purity of the data source; through microscopic equivalent circuit modeling and frequency domain phase angle feature separation algorithm, it achieves precise separation of subcutaneous and visceral fat signals, and can output visceral fat quantitative indicators with clinical reference value, overcoming the error limitations of traditional empirical formula estimation.

[0017] 2. This invention effectively eliminates measurement biases caused by non-fat factors through a dual-factor temperature dynamic compensation and moisture steady-state adaptive correction algorithm, ensuring measurement stability under complex physiological and environmental scenarios. At the same time, it breaks through the limitation of traditional devices that can only output static values. By quantifying the dynamic trend of fat changes, it constructs a high-dimensional health tensor that includes fat content, rate of change, and physiological state, providing complete quantitative data for the dual precise regulation of body fat and visceral fat. Attached Figure Description

[0018] Further details, features, and advantages of this application are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which: Figure 1 This is a system structure diagram of the present invention. Detailed Implementation

[0019] Several embodiments of this application will now be described in more detail with reference to the accompanying drawings to enable those skilled in the art to implement this application. This application may be embodied in many different forms and for various purposes and should not be limited to the embodiments set forth herein. These embodiments are provided to make this application thorough and complete, and to fully convey the scope of this application to those skilled in the art. The embodiments described do not limit this application.

[0020] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It will be further understood that terms such as those defined in commonly used dictionaries shall be interpreted as having a meaning consistent with their meaning in the relevant field and / or the context of this specification, and shall not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0021] Example 1

[0022] Its specific implementation method is combined with the appendix Figure 1 Please provide a detailed explanation.

[0023] Appendix Figure 1 This invention provides a structural block diagram of a health management system based on dual regulation of body fat and visceral fat, illustrating the connection relationship between the multi-frequency impedance signal synchronous acquisition module and the multi-dimensional parameter compensation and result output module, and marking the main functional interaction flow of each module.

[0024] In this embodiment, it includes: The multi-frequency impedance signal synchronous acquisition module is configured to generate multi-frequency composite excitation signals and send them to the human body, synchronously and concurrently acquire weak voltage drop signals of independent segments of the human body, and extract the original complex impedance matrix under multi-frequency bands through digital quadrature phasor demodulation. This module is the physical data source of the entire computing system. It is responsible for managing the interface between the microprocessor and the peripheral analog-to-digital / digital-to-analog converter hardware, and realizing high-precision, multi-band concurrent synchronous acquisition and buffering scheduling logic for electrophysiological signals.

[0025] The system needs to drive an external eight-electrode system to inject a safe, weak alternating current into the human body. To achieve "multi-band concurrency" and abandon the time-consuming single-frequency sequential scanning mechanism, the system constructs a discrete-time sequence of composite excitation signals in the underlying central processing unit.

[0026] The system statically defines four basic excitation frequency points in memory, covering low, medium, and high frequency bands (e.g., 5kHz, 50kHz, 250kHz, and 500kHz). The system generates discrete-time equations for composite currents through direct digital synthesis logic. Its mathematical model is a linear superposition of multiple sine waves: ; The parameters are strictly defined as follows: Represents the index of the frequency point registered in the system memory; Representing the The system uses a programmable gain amplifier to measure the current amplitude at each frequency. Strictly controlled within the human safety threshold range (less than 1 mA); Represents the set discrete target frequency points; This represents the initial phase of each frequency component. To avoid multiple sine waves overlapping at a certain time point, which could lead to excessively large instantaneous peaks that break through the hardware input range, the system executes an extremum minimization algorithm, setting the initial phase difference of each frequency to be [value missing]. Multiples of.

[0027] Continuous signal The data is digitized and discretized at an extremely high sampling rate (e.g., 2MHz, far exceeding the lower limit of the Nyquist sampling theorem) set by the system clock, generating a discrete digital array which is then pushed to the front-end data register of a high-speed digital-to-analog converter (DAC) for continuous output.

[0028] When the composite alternating current flows through different segments of the human body, the system controls a multi-channel high-speed analog-to-digital converter (ADC) to synchronously acquire the weak voltage drop signals generated by each independent segment of the human body.

[0029] Let the continuous voltage signal be The system performs discrete sampling to generate discrete-time voltage sequences. ,in This is a memory index for discrete time steps. To prevent data bus congestion and memory overflow caused by high-frequency sampling, this module allocates independent direct memory access (DMA) channels and multi-channel ring buffer storage areas in the system's random access memory (RAM).

[0030] The underlying state machine logic of the circular buffer is as follows: The total length of the buffer containing contiguous physical addresses is set to... The current data stream write pointer address is The current operation reads the pointer address as Each time a new sampling point is generated The hardware DMA controller writes it directly. The address pointed to, and the pointer offset updated to .

[0031] The system's main frequency timer dynamically monitors the pointer's catch-up status. When the absolute fill difference of the buffer exceeds a preset safety level (e.g., 80%), the system's underlying kernel immediately triggers a high-priority hardware interrupt, forcibly waking up the waiting digital signal processing thread and immediately starting from... The system begins reading data for calculations, thus avoiding data loss or stack overflow disasters that occur in high-frequency continuous physiological monitoring.

[0032] Obtaining stable composite voltage time series Then, the system executes a fully digital phase-sensitive detection algorithm to demodulate the time-domain mixed signal into complex impedance at independent frequencies.

[0033] For each target frequency The system generates two orthogonal digital local oscillator reference sequences in real time within the local arithmetic logic unit: In-phase reference sequence: ; Orthogonal reference sequence: ; in This is the sampling period of the system's ADC. Subsequently, the system performs a vector dot product operation, converting the acquired sequence... The signals are multiplied point-by-point by the two reference sequences mentioned above, and an accumulating average logic is performed within a strictly aligned, complete signal envelope integration period. Thus, the system accurately extracts the frequency-specific signals from the aliased voltage signals. In-phase components and orthogonal components .

[0034] By combining a constant transmit current vector, the system uses the complex division rule to calculate the real part (resistance) and imaginary part (capacitive reactance) at each frequency point, and constructs the original physical impedance matrix composed of multiple sets of complex elements in memory, thus announcing the completion of the comprehensive transformation of analog signals in the physical world into digital state variables of the computer.

[0035] The time-frequency conversion and physiological noise elimination module is configured to: acquire the original complex impedance matrix output by the multi-frequency impedance signal synchronous acquisition module, use adaptive cubic spline interpolation to fit and eliminate low-frequency respiratory baseline drift, and remove transient high-frequency glitches and noise through continuous wavelet transform combined with dynamic hard threshold absolute median difference algorithm, and reconstruct the output multi-frequency complex impedance array. This module receives the raw impedance matrix from the underlying buffer and uses purely digital frequency domain calculus and high-dimensional waveform transformation to filter out baseline drift caused by breathing motion and transient high-frequency glitches caused by poor electrode contact, ensuring the purity of subsequent modeling calculations.

[0036] Because human respiration causes periodic changes in thoracic cavity volume, low-frequency impedance data will be superimposed on a nonlinear drift curve with a frequency of approximately 0.15Hz to 0.35Hz. The system employs an adaptive time-domain sliding window strategy to eliminate this error.

[0037] First, the system is based on the original sequence of real parts of complex impedances. Set a width of A dynamic time sliding window is used to perform a local minimum traversal search within the window, extracting a series of time indices of the occurrence of local minimums. and the corresponding impedance amplitude A linked list of feature nodes is constructed in memory.

[0038] Subsequently, the system uses these minima as anchor points to construct a cubic spline interpolation function. At any two adjacent discrete anchor points and Within an interval, the interpolation polynomial is strictly defined as: ; To solve for these polynomial coefficients, the system constructs a system of linear algebraic equations with strict diagonal dominance, based on the curve smoothing boundary condition (i.e., the function values, first derivatives, and second derivatives at the intersection of adjacent intervals must be absolutely continuous). The system then uses a chasing method to perform a fast inversion operation on this matrix, calculating the coefficients for all intervals. .

[0039] Calculated continuous fitting function The nonlinear respiratory baseline was reproduced with extremely high accuracy. The system then performed a subtraction operation: ; This forces the original signal to flatten, obtaining a pure impedance time series that fluctuates around an absolute zero reference, thus eliminating the systematic measurement bias caused by macroscopic physiological motion.

[0040] For transient high-frequency broadband spikes caused by sudden electromagnetic interference or electromyography in the environment, the traditional Fast Fourier Transform (FFT) cannot perform time-domain localization because it loses the time dimension. This module introduces a Continuous Wavelet Transform (CWT) architecture.

[0041] The system selects "Morlet complex wavelet" as the mother wavelet function from the function library. Because its envelope shape is highly homologous to transient bioelectrical signals. The core calculus equation of wavelet transform is defined as: ; The physical meaning of this formula in a discrete computing environment is as follows: This represents the continuity of the impedance signal after baseline drift has been eliminated; An array of scale factors for the inverse frequency mapping; The translation factor step size on the time axis; To select the complex conjugate form of the mother wavelet function; The system uses nested double for loops to control the translation and scaling of the mother wavelet over the entire signal length, and calculates and outputs a two-dimensional floating-point matrix of wavelet coefficients containing the holographic time-frequency state. .

[0042] get After matrixing, the system performs non-linear threshold pruning to remove high-frequency spikes. This is done for specific scale levels. The corresponding row vector Instead of using fixed empirical values, the system adaptively calculates the current noise threshold limit using the median absolute difference (MAD) statistical algorithm. : ; in This represents the total number of sampling points at the current scale level. The noise standard deviation is estimated using the formula. The calculation yielded the result.

[0043] The system iterates through the matrix data, and once it finds that the absolute value is lower than... The wavelet coefficients are identified as Gaussian background noise and forcibly cleared to zero; isolated highlight coefficients with an absolute value greater than the dynamic threshold are identified as sudden system interference and are smoothed out; while effective coefficients in large areas are fully preserved.

[0044] After matrix cleaning, the system initiates a rigorous inverse wavelet transform to restore the differential equations: ; Among them, parameters Let be the admissibility constant of the mother wavelet, representing the strict mathematical constraint on the conservation of mother wavelet energy. It is defined as a frequency domain integral in Fourier space: ; Mother wavelet The frequency domain function expression under continuous Fourier transform, and the system enforces verification. The mathematical boundary conditions are used to ensure that the reconstructed integral equation converges absolutely in the real number field; This is the wavelet coefficient matrix after dynamic hard thresholding (denoising). The coefficients representing high-frequency interference (such as electromagnetic noise and muscle tremors) have been forcibly cleared to zero, leaving only the true physiological signal characteristics. The energy normalization factor ensures that the mother wavelet is normalized regardless of whether it is stretched or compressed (scaled). (Regardless of how it changes), its energy remains constant; Through repeated mathematical transformations from the time domain to the frequency domain and back to the time domain, the system officially outputs a high-fidelity complex impedance array of the body surface with extremely strong anti-interference capabilities, laying an unshakable data foundation for subsequent physiological feature separation.

[0045] The body surface and visceral impedance feature separation module is configured to connect with the time-frequency conversion and physiological noise cancellation module, and to parse the reference electrical characteristic parameters from the multi-frequency complex impedance array based on the equivalent circuit topology model and damping iterative optimization algorithm, and to extract the characteristic dispersion coefficients specific to visceral fat using the second derivative of the frequency domain phase angle after equal-interval resampling. Without relying on traditional medical imaging equipment (such as dual-energy X-ray absorptiometry (DEXA) or nuclear magnetic resonance MRI), this module constructs an equivalent circuit topology model of biological tissue at the system's bottom layer and combines it with impedance trajectory partial differential equations from computational physiology to transform the abstract discrete frequency-impedance matrix into macroscopic physical characteristic parameters that can directly characterize the microscopic cellular structure of the human body.

[0046] The human body is extremely complex. The system first calls and instantiates a generalized nonlinear Cole-Cole micro equivalent circuit model in the kernel database. This physical model abstracts human tissue as a combination of three main electrical components: extracellular fluid, which exhibits purely resistive properties; intracellular fluid, which also exhibits purely resistive properties; and the cell membrane, which exhibits polarization dispersion properties (a non-ideal capacitor).

[0047] Based on this physical topology, the system defines the global complex impedance complex variable function in the complex domain. for: ; The system allocates a strict memory register for each physical variable in the above equation and defines its physical meaning: This refers to the current alternating current excitation frequency point at which the system is currently being analyzed; This is the extreme value of DC impedance when the frequency is extremely close to zero (i.e., extremely low frequency DC state). Since DC current cannot break down the cell membrane capacitance at all, the current is forced to flow entirely through the intercellular space. Therefore, this parameter absolutely characterizes the total resistance of the extracellular fluid at the physical level. This represents the limiting impedance extreme value when the frequency approaches infinity (ultra-high frequency state). At this point, the high-frequency alternating current treats the capacitive reactance of the cell membrane as a short circuit, and the current simultaneously penetrates both intracellular and extracellular fluids. Therefore, this parameter characterizes the overall conductivity of the total water content in the human body. The relaxation time constant is a characteristic feature of a specific tissue, reflecting the time delay characteristic of the cell membrane as a biological capacitor undergoing charging and discharging operations. This applies to subcutaneous fat and dense visceral fat. The values ​​exhibit significant physical differences; It is a dimensionless phase relaxation distribution coefficient (its numerical range is strictly controlled between 0 and 1), used to compensate for and characterize the capacitance phase angle dispersion effect caused by the uneven size and shape of various cell communities in the human body. Based on the imaginary unit, satisfying Mathematical laws.

[0048] The system has received a set of true complex impedance measurements containing very little noise from the second module. In order to inversely solve for the four unknown core characteristic parameters from the above nonlinear complex function ( The system activates the Levenberg-Marquardt (LM) damping iterative optimization algorithm in the underlying algorithm library.

[0049] The system first establishes a cost function (global error energy function) to quantify the deviation between the theoretical model's predicted values ​​and the actual sampled values. : ; in This represents the total number of frequency points sampled. The core task of the system is to find a set of parameter combinations in the four-dimensional solution space that satisfy the energy function. It converges to the global minimum.

[0050] The system initializes a four-dimensional parameter vector matrix in memory. (subscript) (For the number of iterations), and calculate the Jacobian partial derivative matrix of the cost function with respect to these four parameters in real time. .

[0051] To ensure the computability of the nonlinear iteration, the system first generates an accurate residual column vector in each iteration loop. Residual vector It is The matrix whose first... element Strictly defined as at a specific frequency point Below, the actual measured complex impedance of the system is compared with the current iteration step. The complex difference between the predicted impedances of the physical model: ; The system updates the parameter vector with damping according to the following formula: ; in, It is the standard identity matrix; This is an adaptive convergence damping factor. The system kernel continuously monitors the norm of the residual vector after each iteration. When the residual decreases rapidly, the system reduces The value causes the algorithm matrix to degenerate into the Gauss-Newton method in pursuit of the ultimate convergence speed; when the residual descent shows signs of stagnation or divergence, the system immediately increases... The value forces the algorithm to be converted to gradient descent to ensure mathematically absolute convergence.

[0052] When the objective function of two consecutive iterations The difference is less than the system's preset minimum value. (like The loop is forcibly terminated at this point. The matrix... The locked and fixed values ​​represent the current actual microscopic physical properties of the user's tissue. .

[0053] Because visceral fat is located deep within the abdominal cavity between the organs and is tightly wrapped by the abdominal muscle group and thick subcutaneous fat, conventional high-frequency alternating current can easily pass through the low-impedance epidermal layer, resulting in a "bypass escape effect." The impedance signal deep in the abdomen presents an extremely weak "potential shielding" on the body surface.

[0054] To overcome this physical barrier on a purely mathematical level, the system introduces the "frequency domain phase angle second derivative mutation separation method." The system extracts the complex matrix and calculates the phase angle distribution curves at discrete frequencies. : ; in and Let represent the imaginary part of the capacitive reactance and the real part of the resistance.

[0055] Before performing differentiation, it's important to understand that the actual frequencies emitted by the hardware acquisition circuits are typically logarithmically distributed (e.g., 5kHz, 50kHz, 250kHz, 500kHz, with varying bandwidths). Direct differentiation would lead to catastrophic discrete errors. Therefore, the system forcibly introduces an "equal-interval linear interpolation" resampling engine to correct the non-uniform frequency distribution. Scattered data is remapped in computer memory to a format with a very small fixed bandwidth. (For example ) is a dense, uniform array with equal spacing.

[0056] Ensuring the step size of the denominator After achieving strict mathematical consistency, the system uses the central finite difference method to perform second-order numerical differentiation operations on the phase angle curves of the trunk segments: ; The physical distribution of subcutaneous fat tends to be homogeneous, and the first derivative of its phase angle with frequency exhibits a smooth constant, thus the second derivative approaches zero. However, visceral fat cell clusters exhibit a nonlinear phase angle abrupt response to specific mid-frequency bands (e.g., 50kHz to 100kHz). The system searches in memory... The local maxima (peak points) of the array were successfully used to separate and extract the characteristic dispersion coefficients specific to the capacitive properties of visceral fat. This achieves perfect separation of subcutaneous surface signals from deep visceral signals at the data source.

[0057] The nonlinear mapping calculation module for fat distribution is configured to connect to the body surface visceral impedance characteristic separation module, calculate the initial fat-free mass and total fat percentage based on the reference electrical characteristic parameters through multi-segment impedance volume integration, and combine the multi-frequency impedance ratio and characteristic dispersion coefficient to calculate the initial visceral fat cross-sectional area by substituting into the nonlinear exponential mapping equation. The system obtained characteristics of overall hydration in the previous module. and specific visceral indicators Multidimensional electrical parameters. The core function of this module is to use multi-voxel fluid dynamics theory to construct a set of deterministic nonlinear algebraic mapping equations to convert intangible electrical parameters into user-visible biological mass (total fat percentage) and geometric space area (cross-sectional area of ​​visceral fat).

[0058] When calculating total body fat, the system abandons the highly inaccurate single-channel body mass index (BMI) estimation method. The system constructs a "five-cylinder independent series geometric model" at the bottom layer, consisting of the left upper limb, right upper limb, trunk, left lower limb, and right lower limb.

[0059] First, the system needs to calculate the lean mass of the human body, which is an excellent conductor of electric current. The system utilizes a high-frequency limiting characteristic resistor. The impedance volume index is used to estimate the total body fluid volume (TBW). The system invokes a high-dimensional nonlinear multivariate polynomial mapping equation: ; The physical compensation logic for each dimension in this equation is as follows: This represents the height parameter in the user's basic profile, calculated by dividing the square of the height by the total body's maximum high-frequency impedance. This constitutes the cornerstone and core dominant factor in estimating the volume of conductive fluid in a cylinder; Provides baseline density compensation of absolute quality in a polynomial for the user's current total weight parameters; This represents the current actual waist circumference, obtained through auxiliary peripherals or user input. The system is used to correct geometric and topological deformation errors in cylindrical models caused by severe obesity bulging in the abdomen. These are the geometric morphological constants of a large sample of human bodies that have been pre-programmed into the system's flash memory and rigorously calibrated using medical gold standards (such as the heavy water isotope dilution method).

[0060] The calculations yielded extremely accurate results. Then, the system divides the constant physiological cell hydration rate (usually fixed at 0.732) of the human body in a healthy state by this constant to obtain the total fat-free mass: ; Subsequently, the system invokes the fundamental law of conservation of mass in macroscopic physics, directly subtracting the mass of a good conductor from the total mass to obtain the absolute total body fat mass (TBFM): ; Then, the overall body fat percentage is calculated: .

[0061] Visceral fat accumulates in the greater omentum and inter-organ spaces. Instead of proportionally increasing the conductive volume length of the trunk, it produces a non-linear effect similar to "sponge absorbing water" by changing the tortuosity and low / high frequency impedance ratio of the fluid conduction medium in the abdomen.

[0062] Before performing the index calculation, to compensate for the differences in visceral volume caused by different user skeleton sizes, the system must first accurately segment the independent weight of the torso. The system calls the series impedance voltage divider topology estimation formula to calculate the independent mass of the torso. : ; The system utilizes segmentally acquired high-frequency limiting impedance data of the torso. With total body impedance The resulting reciprocal of the ratio, combined with a preset empirical coefficient for trunk mass distribution. From the total weight The core area of ​​the abdomen is precisely separated and weighed independently.

[0063] In obtaining Then, the system extracts the low-frequency extreme values ​​specific to the torso. and high frequency extrema Combined with the extracted specific dispersion coefficients Substitute into the calculation of the visceral fat anatomical cross-sectional area at the umbilicus (VFA, unit) The nonlinear exponential mapping equation: ; In the formula: Ratio Term The severity of extracellular fluid being squeezed out by fat was monitored. Since visceral fat is a three-dimensional expansion, an external natural exponential function was applied. This amplifies the impedance ratio into a nonlinear area-increasing effect; Product term By physically anchoring the dispersion characteristics of pure electricity to the user's actual abdominal weight, the misjudgment that may occur in people with "large body weight and low body fat" is eliminated. Logarithmic terms We captured the small but deterministic logarithmic shift in cell membrane charge-discharge relaxation time caused by fatty degeneration.

[0064] Weighted matrix constants This ensures that the physical dimensions of each dimension are ultimately unified into clinically relevant area values. This purely mathematical deduction framework avoids the risk of uncontrollable divergence that comes with relying on AI black-box training.

[0065] The multidimensional parameter compensation and result output module is configured to connect with the fat distribution nonlinear mapping calculation module, obtain the dynamic conductivity compensation coefficient generated by skin and ambient temperature, calculate the water steady-state distribution index to remove the weight of free water, calibrate the initial total fat percentage and the initial visceral fat cross-sectional area, calculate the first time derivative in combination with the timestamp, and package and output the high-dimensional tensor of the health status. In real-world physical monitoring environments, the human body's bioelectrical impedance is highly vulnerable to environmental changes. Even minor shifts in the body's microclimate and temporary imbalances in cellular fluid can trigger drastic drifts in the impedance baseline at the underlying level. Before abstracting the final results and feeding them back to the user or intervention engine, this module performs multi-dimensional accuracy calibration at the system's underlying level by constructing a rigorous calculus compensation mechanism and a state-adaptive bias algorithm.

[0066] The electrical conductivity of human soft tissue exhibits a strong positive correlation with temperature. This is essentially because increased temperature leads to increased Brownian motion of ions in the interstitial fluid and a higher ion mobility. If a user has just completed strenuous aerobic exercise or is in a hot / cold environment, the raw impedance data collected by the system will be significantly shifted, resulting in unacceptable "spurious changes" in the calculated fat volume.

[0067] To ensure the underlying data is resistant to temperature interference, the system must synchronously drive an external high-precision NTC thermistor or infrared temperature sensor to collect the absolute temperature sequence of the skin surface at the electrode contact points in real time. and ambient temperature sequence ; The system instantiates a two-factor linear conductivity compensation coefficient mapping function based on a variant of the nonlinear Arrhenius equation within the central processing unit. : ; In this compensation topology, the strict definitions of each parameter are as follows: The ideal room temperature test reference physical temperature (e.g., factory-calibrated 25.0 degrees Celsius) is used to harden the system's configuration in the underlying registers. The slope coefficient for skin temperature sensitivity feedback compensation represents the response rate of the conductivity of the superficial subcutaneous fat and capillary network to the gradient change of local temperature. The environmental thermal radiation hysteresis bias compensation coefficient is used to compensate for the physical offset caused by the change in the smooth muscle relaxation state of deep visceral blood vessels due to the overall environmental cold and heat stimulation, thereby affecting the penetration path of the low-frequency detection current.

[0068] The core electrical characteristic reference parameters output by the system call include and dispersion coefficient Compare them with the dynamic compensation coefficient respectively. Perform scalar multiplication. Through this mathematically forced mapping, the system perfectly shifts and calibrates all impedance physical quantities under current non-standard thermodynamic states back to their equivalent pure impedance values ​​under the standard reference plane.

[0069] A user's instantaneous hydration state (such as drinking a lot of water before a test, or dehydration due to excessive sweating from high-intensity training) is a huge source of error that is extremely difficult to eliminate with a single sensor. This system does not rely on the user manually inputting any diet or water intake status on the front-end UI interface, but instead directly calculates and detects the cell membrane penetration characteristics through underlying physical frequencies.

[0070] The system calls the calculated DC limiting impedance in parallel. (Its physical nature fully characterizes the absolute volume of extracellular fluid) and extremely high frequency limiting impedance (Its physical essence represents the total water volume including intracellular and extracellular fluids). The system constructs a water steady-state distribution index in memory. : ; In a healthy and homeostatic human physiological environment, this The values ​​are strictly controlled within an extremely narrow range of biological constants. During long-term monitoring, the system maintains a baseline average of each user's historical water content, calculated using an exponential smoothing algorithm, within a secure storage area. .

[0071] When the system calculates and discovers the current situation When the value changes beyond a preset threshold within a preset time period (e.g., after a user drinks a large amount of water, water remains in the digestive tract and extracellular fluid, leading to...), A precipitous drop occurred, which in turn triggered... (If the value is relatively small compared to the baseline), the system immediately activates the adaptive penalty correction equation to adjust the initial coarse calculation results for fat-free mass. (Obtained in the nonlinear mapping calculation module for fat distribution) This is a theoretical value that has not yet been calibrated for environmental and intravenous instantaneous water fluctuations; it is recorded as such in this module for the purpose of differentiation. (In reality, it's a single value) Perform damped convergence compensation: ; Among them, parameters This is a preset weighted constant for the moisture anomaly ratio bias penalty. When the system detects a sudden surge in instantaneous free moisture, the penalty term in this feedback equation will generate a negative bias. The system automatically uses this mathematical relationship to... The system proportionally removes excess "free water" and false weight. This purely computational mechanism ensures that fluctuations in a user's overall weight are accurately attributed to water metabolism, rather than being misinterpreted by the system as transient increases in visceral fat or muscle and bone mass.

[0072] After rigorous cross-compensation across all physical dimensions and microclimate conditions, the system finally achieved an extremely high-precision real-time overall body fat percentage that closely approximates the user's actual physiological state. and visceral fat cross-sectional area .

[0073] In order to enable the entire system to not only "measure" but also provide data with dynamic trend prediction capabilities for subsequent "dual intervention and regulation", the system calls the historical timestamp data stored in the buffer memory and uses the backward finite difference calculus method to calculate the time first derivative (i.e., the acceleration of change) of the total fat percentage and visceral fat area in real time.

[0074] The differential equations for finite difference calculations are defined as follows: ; ; In the two derivative equations above, This represents the absolute timestamp at which the system completed its data measurement and calculation task. This represents the absolute timestamp of the last valid measurement sampling point recorded in the system memory. The difference between the denominators is the time interval between the two measurements, and the numerator is the difference in the compensated true physiological data. By extracting this first derivative, the system can mathematically determine whether the user's current fat status is in an upward, plateau, or downward phase.

[0075] Finally, this computation module formally constructs and instantiates a high-dimensional health state tensor with strong temporal and spatial coupling properties in the computer's internal bus register. : ; The system's underlying hardware triggers a direct memory access (DMA) interrupt, pushing the high-dimensional tensor data block to the system's application layer and rule scheduling engine with zero latency and no loss. This completes the process from capturing multi-frequency weak voltage signals, cleaning high-dimensional time-frequency noise, separating microscopic physical topologies, mapping fluid dynamics, to multi-dimensional anti-interference compensation.

[0076] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0077] It should be noted that, in this document, the use of relational terms such as "first" and "second" is merely for distinguishing one entity or operation from another, and does not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0078] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0079] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0080] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A health management system based on dual regulation of body fat and visceral fat, characterized in that, include: The multi-frequency impedance signal synchronous acquisition module is configured to generate multi-frequency composite excitation signals and send them to the human body, synchronously acquire weak voltage drop signals of independent segments of the human body, and extract the original complex impedance matrix under multi-frequency bands. The time-frequency conversion and physiological noise cancellation module is configured to: acquire the original complex impedance matrix of the output, remove transient high-frequency glitches and noise, and reconstruct the output multi-frequency complex impedance array; The body surface and visceral impedance feature separation module is configured to: resolve the reference electrical characteristic parameters from the multi-frequency complex impedance array and extract the characteristic dispersion coefficient of visceral fat; The nonlinear mapping calculation module for fat distribution is configured to: calculate the initial fat-free mass and total fat percentage based on the reference electrical characteristic parameters, and calculate the initial visceral fat cross-sectional area by combining the multi-frequency impedance ratio and the characteristic dispersion coefficient. The multidimensional parameter compensation and result output module is configured to: obtain the dynamic conductivity compensation coefficient of skin and ambient temperature, calculate the water steady-state distribution index, calibrate the initial total fat percentage and the initial visceral fat cross-sectional area, calculate the first-order time derivative in combination with the timestamp, and output the high-dimensional tensor of the health status.

2. The health management system based on dual regulation of body fat and visceral fat according to claim 1, characterized in that, The specific method by which the multi-frequency impedance signal synchronous acquisition module generates multi-frequency composite excitation signals is as follows: The sinusoidal waves corresponding to a preset number of basic excitation frequency points are linearly superimposed to generate a composite current time-domain discrete equation. Meanwhile, the multi-frequency impedance signal synchronous acquisition module allocates a multi-channel ring buffer storage area in the system memory. When the filling difference of the buffer exceeds the preset safety level, a hardware interrupt is triggered and the data is calculated and read from the read pointer address.

3. The health management system based on dual regulation of body fat and visceral fat according to claim 2, characterized in that, The expression for the time-domain discrete equation of the composite current is: ； in, The time-domain discrete equation is for the composite current. Index of the base excitation frequency point For the first The current amplitude at a basic excitation frequency point, limited to a safety threshold range. For the set discrete target frequency points, The initial phase of each frequency component; Furthermore, the initial phase difference between each frequency component is set to a multiple of a preset angle to avoid excessively large instantaneous peak values ​​in the multi-frequency signal.

4. The health management system based on dual regulation of body fat and visceral fat according to claim 1, characterized in that, The time-frequency conversion and physiological noise cancellation module uses adaptive cubic spline interpolation to fit and eliminate low-frequency respiratory baseline drift. The steps include: Local minima and their corresponding time indices of the original complex impedance real part sequence are extracted within a dynamic time sliding window and used as anchor points. A cubic spline interpolation polynomial is constructed using anchor points. A system of linear algebraic equations is established based on the function values ​​at the intersection of adjacent intervals and the curve smoothing boundary conditions with continuous derivatives. The coefficients of the interpolation polynomial are then calculated to obtain a continuous fitting function. Subtracting the continuous fitting function from the original complex impedance real part sequence yields the pure impedance time series after eliminating low-frequency respiratory baseline drift.

5. A health management system based on dual regulation of body fat and visceral fat according to claim 4, characterized in that, The specific process for removing transient high-frequency glitches and noise is as follows: A continuous wavelet transform is performed on the pure impedance time series using a mother wavelet function that meets preset conditions to generate a two-dimensional floating-point matrix of wavelet coefficients containing time-frequency states. The absolute median difference statistical algorithm is used to adaptively calculate the noise threshold limit of the current scale level; Wavelet coefficients with absolute values ​​below the noise threshold are forcibly cleared to zero, while wavelet coefficients with absolute values ​​above the noise threshold are retained. These coefficients are then substituted into the wavelet inverse transform restoration calculus equation that satisfies the allowable constant constraint to generate a reconstructed, pure multi-frequency complex impedance array.

6. A health management system based on dual regulation of body fat and visceral fat according to claim 1, characterized in that, The complex function corresponding to the equivalent circuit topology model is: ； in, It is the overall complex impedance complex variable function. The excitation frequency point of the alternating current. To characterize the extreme values ​​of DC impedance for extracellular fluid resistance, The limiting impedance is used to characterize the overall conductivity of water. The characteristic relaxation time constant is The phase relaxation distribution coefficient is... The basic imaginary unit; The damping iterative optimization algorithm generates a residual column vector consisting of the complex difference between the actual measured complex impedance and the model predicted impedance in each iteration, and adaptively adjusts the convergence damping factor according to the residual descent rate to solve for the DC impedance extreme value, the limiting impedance extreme value, the characteristic relaxation time constant, and the phase relaxation distribution coefficient as reference electrical characteristic parameters.

7. A health management system based on dual regulation of body fat and visceral fat according to claim 1, characterized in that, The process of extracting the characteristic dispersion coefficients of visceral fat using the body surface visceral impedance feature separation module is as follows: Extract the pure multi-frequency complex impedance array and calculate the phase angle distribution curve at discrete frequencies; The equal-interval linear interpolation resampling engine is invoked to map the phase angle distribution curve into a dense uniform array with equal intervals at fixed bandwidths; The central finite difference method is used to perform second-order numerical differentiation on a dense uniform array, and local maxima are searched in the generated second-order derivative array to extract the characteristic dispersion coefficients specific to visceral fat.

8. A health management system based on dual regulation of body fat and visceral fat according to claim 1, characterized in that, The process by which the nonlinear mapping calculation module for fat distribution calculates the initial fat-free mass and total fat percentage is as follows: The total fluid volume of the human body is estimated by calling a nonlinear multivariate polynomial mapping equation. The nonlinear multivariate polynomial mapping equation introduces the ratio of the square of the height parameter to the high-frequency limiting impedance of the whole body as the core dominant term, and uses the total mass parameter and waist circumference parameter for system bias compensation. The initial fat-free mass is obtained by dividing the total body fluid volume by the physiological cell hydration constant. The absolute total fat mass is obtained by subtracting the initial fat-free mass from the total mass, and then the initial total fat percentage is calculated.

9. A health management system based on dual regulation of body fat and visceral fat according to claim 8, characterized in that, Also includes: To compensate for the differences in visceral volume caused by different user skeleton sizes, the system uses a series impedance voltage divider topology estimation formula to calculate the independent mass of the torso. Then, the system extracts the low-frequency extreme values ​​specific to the torso. and high frequency extrema Combined with the extracted specific dispersion coefficients Substitute this into the nonlinear exponential mapping equation for calculating the anatomical cross-sectional area of ​​visceral fat at the umbilicus plane: in, This represents the initial cross-sectional area of ​​visceral fat. It is a natural exponential function. The characteristic relaxation time constant of the torso. , , , These are the system's preset weighted matrix constants.

10. A health management system based on dual regulation of body fat and visceral fat according to claim 1, characterized in that, The multidimensional parameter compensation and result output module specifically includes: Real-time synchronous acquisition of absolute temperature sequence of skin surface at electrode contact point and ambient temperature sequence The system instantiates a two-factor linear conductivity compensation coefficient mapping function based on a variant of the nonlinear Arrhenius equation within the central processing unit. ; Call DC limiting impedance and ultra-high frequency limiting impedance Constructing a water steady-state distribution index in memory ; Real-time calculation of the current When the value changes beyond a preset threshold within a preset time period, adaptive penalty correction is activated to adjust the initial coarse calculation results for fat-free mass. Perform damped convergence compensation; Obtain real-time total body fat percentage based on the user's actual physiological state. and visceral fat cross-sectional area ; to compute and obtain a high-dimensional health state tensor with temporal and spatial coupling properties. .

Images