Continuous blood pressure monitoring method, system, terminal and storage medium based on elastic cavity model and pulse wave velocity equation

Abstract

The application relates to the technical field of blood pressure monitoring, and discloses a continuous blood pressure monitoring method and system based on an elastic cavity model and a pulse wave velocity equation, a terminal and a storage medium, the method comprising the following steps: utilizing a two-element elastic cavity model to reflect the dynamic process among cardiac ejection, arterial compliance and peripheral resistance from a system level, establishing a time-domain coupling relationship between blood pressure and blood flow, and introducing a pulse wave velocity equation to calculate the arterial compliance C parameter of different objects, calibrate the model compliance C parameter, and obtain targeted blood pressure monitoring results. The application introduces individual physiological information, and improves the accuracy and interpretability of blood pressure prediction for different objects.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of blood pressure monitoring, and particularly relates to a continuous blood pressure monitoring method, system, terminal and computer readable storage medium based on an elastic cavity model and a pulse wave velocity equation. BACKGROUND

[0002] Blood pressure is an important physiological indicator reflecting the function state of the human circulatory system and plays a core role in maintaining the homeostasis of the cardiovascular system.

[0003] Existing non-invasive blood pressure measurement techniques are mostly based on the pulse wave velocity equation or the elastic cavity model to establish a blood pressure estimation model alone, but the former is difficult to depict the dynamic process between heart pumping and peripheral resistance, and the latter can describe the transient response of pressure-flow, but is insufficiently sensitive to the spatial distribution of blood vessel elasticity changes.

[0004] Therefore, the prior art still needs to be improved and developed. SUMMARY

[0005] The main purpose of the present application is to provide a continuous blood pressure monitoring method, system, terminal and computer readable storage medium based on an elastic cavity model and a pulse wave velocity equation, aiming to solve the problem of inaccurate blood pressure measurement caused by the incomplete physical mechanism, insufficient model universality and significant individual difference in existing non-invasive continuous blood pressure measurement techniques.

[0006] To achieve the above-mentioned purpose, the present application provides a continuous blood pressure monitoring method based on an elastic cavity model and a pulse wave velocity equation, which comprises the following steps:

[0007] Collecting a plurality of physiological signals of a target object, pre-processing all the physiological signals to obtain corresponding standard physiological signals, and obtaining a plurality of conventional parameters from all the standard physiological signals;

[0008] Constructing an elastic cavity model, adding all the conventional parameters to the elastic cavity model to construct a state equation, extracting features from all the standard physiological signals to obtain morphological features, statistical features and sequence features;

[0009] Mapping the morphological features, the statistical features and the sequence features to a plurality of parameters in the state equation to obtain a training sample set, and training the elastic cavity model using the training sample set to obtain a prediction model;

[0010] The pulse wave velocity is obtained from the standard physiological signals, the arterial wall Young's modulus is calculated according to the pulse wave velocity, the pulse wave velocity equation is constructed, the arterial wall Young's modulus is added to the pulse wave velocity equation, and a specific expression of the arterial compliance value C parameter is obtained;

[0011] The arterial compliance value C parameter of the target object is calculated according to the specific expression, the model compliance C parameter in the prediction model is adjusted, the continuous blood pressure estimation value is obtained by monitoring the blood pressure of the target object by using the adjusted prediction model.

[0012] Optionally, the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation, wherein the regular parameters include a cardiac systolic ejection time, a cardiac diastolic period time, a cardiac cycle time and a blood flow parameter;

[0013] The multiple physiological signals of the target object are collected, all the physiological signals are preprocessed to obtain corresponding standard physiological signals, and multiple regular parameters are obtained from all the standard physiological signals, specifically including:

[0014] The data of the target object is collected by using multiple sensors to obtain a photoplethysmogram, an electrocardiogram and a thoracic impedance graph;

[0015] The first noise in the photoplethysmogram, the electrocardiogram and the thoracic impedance graph is filtered by using a first filter, and the second noise in the photoplethysmogram, the electrocardiogram and the thoracic impedance graph is filtered by using a second filter to obtain a standard photoplethysmogram, a standard electrocardiogram and a standard thoracic impedance graph;

[0016] The cardiac systolic ejection time, the cardiac diastolic period time and the cardiac cycle time of the target object are extracted from the standard photoplethysmogram;

[0017] The blood flow parameter of the target object is extracted from the standard thoracic impedance graph:

[0018] ;

[0019] Wherein, represents a blood flow parameter, represents a blood density, and respectively represent the time of blood flowing from a point and a point, and respectively represent the arterial wall thickness at a point and a point, ​​​​​represents the length of the blood vessel, represents the monitoring start point.

[0020] Optionally, the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation, wherein the elastic cavity model is constructed, all the conventional parameters are added to the elastic cavity model to construct a state equation, feature extraction is performed on all standard physiological signals to obtain morphological features, statistical features and sequence features, and specifically includes:

[0021] The elastic cavity model of the target object is constructed, and the cardiac systolic ejection time, the cardiac diastolic period time and the cardiac cycle time are added to the elastic cavity model:

[0022] ;

[0023] ;

[0024] wherein, represents the current input at the moment, represents the cardiac output, represents the cardiac cycle time, represents the cardiac systolic ejection time, represents the time index of the entire cardiac cycle time, represents the number of cardiac cycles, represents the model compliance C parameter, represents the voltage at the moment, represents the resistance;

[0025] Based on the cardiac cycle time, the standard photoplethysmogram, the standard electrocardiogram and the standard thoracic impedance are statistically obtained to obtain statistical features, and the statistical features are averaged to obtain morphological features.

[0026] Obtain a plurality of cardiac cycle photoplethysmogram signals or thoracic impedance signals, calculate the variation signal of the photoplethysmogram signal or the thoracic impedance signal, and obtain sequence features.

[0027] Optionally, the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation, wherein the mapping between the morphological features, the statistical features and the sequence features and the plurality of parameters in the state equation is obtained to obtain a training sample set, and the elastic cavity model is trained using the training sample set to obtain a prediction model, and specifically includes:

[0028] For the cardiac cycle time, a preset algorithm is used to construct a mapping from the morphological features, statistical features, and sequence features to the resistance and initial model compliance C parameter in the state equation, resulting in multiple feature-parameter pairs;

[0029] All the feature-parameter pairs are divided into a training sample set, a validation sample set, and a test sample set, wherein the validation sample set is used to validate the prediction model, and the test sample set is used to test the prediction model.

[0030] The training sample set is input into the elastic cavity model for prediction. A weighted loss term is constructed based on the prediction results and each actual value in the training sample set. The elastic cavity model is iteratively optimized based on the weighted loss term to obtain the prediction model.

[0031] Optionally, the continuous blood pressure monitoring method based on the elastic cavity model and pulse wave velocity equation, wherein the step of constructing a weighted loss term based on the prediction result and each actual value in the training sample set, and iteratively optimizing the elastic cavity model based on the weighted loss term to obtain the prediction model, specifically includes:

[0032] A weighted loss term is constructed based on the actual value of each training sample in the training sample set and the prediction result:

[0033] ;

[0034] ;

[0035] in, Indicates the weighted loss term. This represents the total number of training samples. Indicates the first Each sample weight, Indicates the first One actual value, Indicates the prediction result. Indicates the first training samples, Indicates the first Weights for the next iteration Indicates the first In the next iteration, the training samples are input into the output value of the weak regressor. This represents the input training samples. Indicates the number of iterations. This represents the median;

[0036] The sample weights are initialized, and the initial pseudo-residuals are calculated to train the weak regressor, obtaining the minimized weighted squared error:

[0037] ;

[0038] wherein, denotes the sample weight in the th iteration;

[0039] ;

[0040] wherein, denotes the pseudo-residual of the th training sample in the th iteration, denotes the prediction result in the th iteration;

[0041] ;

[0042] wherein, denotes the weighted squared error minimized in the th iteration training, denotes the sample weight of the th sample in the th iteration training, denotes the output value of the th training sample input weak regressor in the

[0043] th iteration;

[0044] ;

[0045] ;

[0046] wherein, denotes the th absolute error in the th iteration, denotes the average weighted loss in the th iteration, denotes the sample weight in the th iteration;

[0047] For the current round, the learner weight is calculated by using the average weighted loss, and the sample weight is updated by using the learner weight to obtain the sample weight of the next round, until the updated parameter that minimizes the weighted loss term is obtained, and a prediction model is constructed by using the updated parameter:

[0048] ;

[0049] ;

[0050] wherein, denotes the learner weight in the th iteration, denotes the th sample weight in the th iteration.

[0051] Optionally, the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation, wherein the pulse wave velocity is obtained from the plurality of standard physiological signals, the arterial wall Young's modulus is calculated according to the pulse wave velocity, the pulse wave velocity equation is constructed, the arterial wall Young's modulus is added to the pulse wave velocity equation, and a specific expression of the arterial compliance value C parameter is obtained, specifically comprising:

[0052] According to the peak time difference in the standard photoplethysmogram or the pulse transmission time in the standard electrocardiogram, the pulse wave velocity is obtained, and the arterial wall Young's modulus is calculated according to the calculation expression of the pulse wave velocity:

[0053] ;

[0054] ;

[0055] wherein, denotes the pulse wave velocity, denotes the arterial wall Young's modulus, denotes the arterial wall thickness, denotes the blood density, denotes the arterial radius;

[0056] The pulse wave velocity equation is constructed:

[0057] ;

[0058] ;

[0059] ;

[0060] wherein, denotes the arterial compliance value C parameter, denotes the volume change, denotes the pressure change, denotes the blood vessel length, denotes the change of arterial radius;

[0061] The arterial wall Young's modulus is added to the pulse wave velocity equation to construct a specific expression of the arterial compliance value C parameter:

[0062] ;

[0063] wherein, represents the volume of the cavity tube.

[0064] Optionally, the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation, wherein the arterial compliance value C parameter of the target object is calculated according to the specific expression, the model compliance C parameter in the prediction model is adjusted, the blood pressure of the target object is monitored by using the adjusted prediction model, and a continuous blood pressure estimation value is obtained, and specifically comprises:

[0065] The arterial compliance value C parameter of the target object is calculated according to the blood density, the pulse wave velocity and the volume of the cavity tube of the target object.

[0066] The model compliance C parameter in the prediction model is adjusted by using the arterial compliance value C parameter, and a final prediction model is obtained.

[0067] The blood pressure of the target object is continuously monitored by using the final prediction model, and a continuous blood pressure estimation value is obtained.

[0068] In addition, in order to achieve the above object, the application further provides a continuous blood pressure monitoring system based on an elastic cavity model and a pulse wave velocity equation, wherein the continuous blood pressure monitoring system based on the elastic cavity model and the pulse wave velocity equation comprises:

[0069] A parameter acquisition module is configured to acquire a plurality of physiological signals of a target object, pre-process all the physiological signals to obtain corresponding standard physiological signals, and obtain a plurality of conventional parameters from all the standard physiological signals.

[0070] A feature extraction module is configured to construct an elastic cavity model, add all the conventional parameters to the elastic cavity model to construct a state equation, and extract features from all the standard physiological signals to obtain morphological features, statistical features and sequence features.

[0071] A model training module is configured to construct a mapping between the morphological features, the statistical features, the sequence features and a plurality of parameters in the state equation, obtain a training sample set, and train the elastic cavity model by using the training sample set to obtain a prediction model.

[0072] A constraint parameter construction module is configured to obtain a pulse wave velocity from a plurality of the standard physiological signals, calculate an arterial wall Young's modulus according to the pulse wave velocity, construct a pulse wave velocity equation, add the arterial wall Young's modulus to the pulse wave velocity equation, and obtain a specific expression of an arterial compliance value C parameter.

[0073] A blood pressure monitoring module is configured to calculate an arterial compliance value C parameter of the target object according to the specific expression, adjust a model compliance C parameter in the prediction model, monitor blood pressure of the target object by using the adjusted prediction model, and obtain continuous blood pressure estimation values.

[0074] In addition, to achieve the above object, the application further provides a terminal, wherein the terminal comprises a memory, a processor, and a continuous blood pressure monitoring program based on an elastic cavity model and a pulse wave velocity equation stored in the memory and executable on the processor, and the continuous blood pressure monitoring program based on the elastic cavity model and the pulse wave velocity equation implements the steps of the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation when executed by the processor.

[0075] In addition, to achieve the above object, the application further provides a computer readable storage medium, wherein the computer readable storage medium stores a continuous blood pressure monitoring program based on an elastic cavity model and a pulse wave velocity equation, and the continuous blood pressure monitoring program based on the elastic cavity model and the pulse wave velocity equation implements the steps of the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation when executed by a processor.

[0076] In the application, a plurality of physiological signals of a target object are collected, all the physiological signals are preprocessed to obtain corresponding standard physiological signals, and a plurality of conventional parameters are obtained from all the standard physiological signals; an elastic cavity model is constructed, all the conventional parameters are added to the elastic cavity model to construct a state equation, and feature extraction is performed on all the standard physiological signals to obtain morphological features, statistical features and sequence features; a mapping between the morphological features, the statistical features, the sequence features and a plurality of parameters in the state equation is constructed to obtain a training sample set, and the elastic cavity model is trained by using the training sample set to obtain a prediction model; a pulse wave velocity is obtained from the plurality of standard physiological signals, an arterial wall Young's modulus is calculated according to the pulse wave velocity, a pulse wave velocity equation is constructed, and the arterial wall Young's modulus is added to the pulse wave velocity equation to obtain a specific expression of an arterial compliance value C parameter; the arterial compliance value C parameter of the target object is calculated according to the specific expression, a model compliance C parameter in the prediction model is adjusted, the blood pressure of the target object is monitored by using the adjusted prediction model, and continuous blood pressure estimation values are obtained. The application introduces individual physiological information, and improves the accuracy and biological authenticity of blood pressure prediction for different objects. BRIEF DESCRIPTION OF DRAWINGS

[0077] Figure 1 is a method flowchart of a preferred embodiment of the continuous blood pressure monitoring method based on an elastic cavity model and a pulse wave velocity equation of the application;

[0078] Figure 2 is a multi-physiological signal acquisition and processing flowchart of a preferred embodiment of the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation of the present application;

[0079] Figure 3 is a whole flowchart of a preferred embodiment of the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation of the present application;

[0080] Figure 4 is a schematic diagram of the elastic cavity model of a preferred embodiment of the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation of the present application;

[0081] Figure 5 is a structure diagram of a preferred embodiment of the continuous blood pressure monitoring system based on the elastic cavity model and the pulse wave velocity equation of the present application;

[0082] Figure 6 is a structure diagram of a preferred embodiment of the terminal of the present application. DETAILED DESCRIPTION

[0083] In order to make the objects, technical solutions and advantages of the present application clearer and more explicit, the present application will be further described in detail below with reference to the drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0084] The continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation described in the preferred embodiment of the present application, as shown in Figure 1 The continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation includes the following steps:

[0085] Step S10, acquiring a plurality of physiological signals of a target object, pre-processing all the physiological signals to obtain corresponding standard physiological signals, and acquiring a plurality of conventional parameters from all the standard physiological signals.

[0086] For the existing non-invasive blood pressure monitoring method, it mostly relies on experience regression or data-driven model (such as machine learning, deep learning, etc.) to directly fit the blood pressure value. Although these methods can achieve higher accuracy on a specific data set, they lack direct physiological mechanism support and are difficult to explain the physiological source of blood pressure changes. Therefore, the present application first acquires a plurality of physiological signals of a target object, and extracts a plurality of conventional parameters therefrom, which are used to construct a WindKessel model (i.e. an elastic cavity model, which is replaced by WK model hereinafter), to reflect the time-domain process of the target object in hemodynamics, so as to describe the transient response of pressure-flow and improve the monitoring accuracy of blood pressure.

[0087] The physiological signals include: photoplethysmography (PPG signal), electrocardiogram (ECG signal), and impedance cardiogram (ICG signal). The corresponding standard physiological signals include: standard photoplethysmography (PPG signal), standard electrocardiogram (ECG signal), and standard impedance cardiogram (ICG signal). The routine parameters include: cardiac systolic ejection time, cardiac diastolic time, cardiac cycle time, and blood flow parameters.

[0088] Specifically, multiple sensors are used to collect data from the target object, obtaining photoplethysmography (PPG), electrocardiogram (ECG), and cardiac impedance cardiogram (IAC). A first filter is used to filter out first noise in the PPG, ECG, and IAC, and a second filter is used to filter out second noise in the PPG, ECG, and IAC, resulting in standard PPG, standard ECG, and standard IAC. The cardiac systolic ejection time, diastolic time, and cardiac cycle time of the target object are extracted from the standard PPG. The blood flow parameters of the target object are extracted from the standard IAC.

[0089] ;

[0090] in, This represents blood flow parameters. Indicates blood density, and These respectively represent blood from Flowing to Point and Time at point, and They represent Point and The arterial wall is thick at the point. Indicates the length of the blood vessel. Indicates the starting point of the monitoring.

[0091] In order to achieve continuous blood pressure monitoring, this invention collects physiological signals of the target object through a variety of sensors, including photoplethysmography (PPG) sensors, electrocardiogram (ECG) sensors, and impedance cardiography (ICP) sensors. These sensors are placed in wearable devices (such as wristbands or chest straps) and fit non-invasively against the human skin, improving the convenience of blood pressure monitoring.

[0092] Further, when collecting, a certain frequency can be set to capture fine waveform changes in the cardiac cycle and improve data integrity. It should be noted that when extracting physiological signals, noise is often accompanied, so the pre-processing process of the physiological signals is extremely important. Noise is mainly divided into low-frequency and high-frequency interference: low-frequency interference mainly manifests as baseline drift of the waveform, which is mainly derived from low-frequency signals generated by human respiration; high-frequency interference mainly manifests as small fluctuations and burrs on the waveform signal, which is derived from myoelectric noise generated by normal physiological activities of human tissues and electrical signal interference of the collection device itself.

[0093] Further, in the pre-processing of the physiological signals, the original signal waveform needs to be preserved while the baseline drift and power frequency interference are reduced as much as possible, so as to ensure that the quality of the physiological signals is not affected. As shown in FIG. 2, in the embodiment disclosed in the present application, for low-frequency noise, a Butterworth high-pass filter with a cutoff frequency of 0.5 Hz and an order of 5 can be used to remove the baseline drift; for high-frequency noise, a 4th-order Butterworth filter with a cutoff frequency of 10 Hz can be used for processing to eliminate high-frequency noise in the signal, and finally standard physiological signals (standard photoplethysmogram, standard electrocardiogram and standard thoracic impedance) are obtained. Figure 2

[0094] Further, a plurality of conventional parameters need to be extracted from the standard physiological signals for constructing a WK model suitable for the target object and improving the accuracy of blood pressure monitoring. As shown in FIG. 3, the conventional parameters (such as Figure 3 ) that need to be obtained in the WK model include cardiac systolic ejection time , cardiac diastolic period time , cardiac cycle time (HR equivalent to heart rate), cardiac output and stroke volume (SV, which refers to the amount of blood ejected by one side of the heart chamber in one heartbeat).

[0095] Among them, the cardiac systolic ejection time , the cardiac diastolic period time , and the cardiac cycle time are obtained from the standard PPG signal, and each trough of the standard PPG signal is used to divide the cardiac cycle and serve as a starting point. In the same cardiac cycle, the time from the starting point to the peak is , the time from the peak to the starting point of the next cardiac cycle is , and the time of the entire cardiac cycle is ​For SV, it is obtained from standard ICG signal, according to C (the starting point of c wave in cardio impedance diagram), B (the starting mark of ventricular ejection) and X point (the termination mark of ventricular ejection) in ICG signal, cardiac function is analyzed and blood flow parameters are calculated.

[0096] Step S20, constructing an elastic cavity model, adding all the conventional parameters to the elastic cavity model to construct a state equation, and performing feature extraction on all standard physiological signals to obtain morphological features, statistical features and sequence features.

[0097] Among them, the existing continuous blood pressure monitoring method has insufficient signal feature utilization and insufficient dynamic correlation modeling. Most of them only use PPG signal or pulse arrival time (PTT, Pulse Transit Time) and other simple waveform features to estimate blood pressure, ignoring the coupling relationship between pulse wave propagation characteristics, vascular compliance and hemodynamic parameters. This shallow feature-based modeling method is difficult to fully excavate the deep vascular dynamic information contained in multiple physiological signals, resulting in a decrease in the stability and accuracy of the model under complex physiological changes or motion state. Therefore, the present application discloses a WK model, and adds multiple standard physiological signals of a target object to the WK model, reflects the dynamic process between cardiac ejection, arterial compliance and peripheral resistance from a system level, establishes the time domain coupling relationship between blood pressure and blood flow, and thus realizes the interpretable monitoring of blood pressure.

[0098] Specifically, an elastic cavity model of the target object is constructed, and the cardiac systolic ejection time, the cardiac diastolic period time and the cardiac cycle time are added to the elastic cavity model:

[0099] ;

[0100] ;

[0101] Among them, represents the current input at the time, represents cardiac output, represents the cardiac cycle time, represents the cardiac systolic ejection time, represents the time index of the entire cardiac cycle time, represents the number of cardiac cycles, represents the model compliance C parameter, represents the voltage at the time, represents a resistor; based on the cardiac cycle time, statistical features are obtained by performing statistics on the standard photoplethysmogram, the standard electrocardiogram and the standard impedance cardiogram, and morphological features are obtained by performing average processing on the statistical features; a plurality of cardiac cycle photoplethysmogram signals or impedance cardiogram signals are obtained, and a variation signal of the photoplethysmogram signal or the impedance cardiogram signal is calculated to obtain sequence features.

[0102] The WK model is designed by using a resistor and a capacitor, as shown in the following formula: Figure 4 The resistor R is connected in parallel with the capacitor (i.e., a model compliance C parameter), and then an input current is obtained, so as to obtain the WK model, and a state equation can be derived according to the model.

[0103] In the embodiments disclosed in the present application, the standard physiological signals (standard PPG signals, standard ECG signals and standard ICG signals) that have been obtained are subjected to feature extraction, so as to obtain waveform features (i.e., morphological features, statistical features and sequence features) as shown in the following formula: Figure 3 These features are used for subsequent parameter prediction of the WK model, and can ensure comprehensiveness and physiological correlation of model input.

[0104] Further, the morphological features are extracted based on a single cardiac cycle, the ECG signal is subjected to R-peak monitoring, so as to divide the cardiac cycle, and then the features extracted for each cardiac cycle are subjected to average processing, so as to obtain final feature values; these morphological features have been widely defined in existing research, and specifically include time parameters, signal amplitudes, area parameters and proportional coefficients between these features. The statistical features are similar to the morphological features, and are extracted based on a single cardiac cycle and then average values are obtained; specifically, these statistical features include mean, median, standard deviation, variance, maximum value, minimum value, difference between maximum and minimum values, skewness and kurtosis of the PPG signal and the ICG signal. Different from the morphological features and the statistical features, the sequence features are not extracted based on a single cardiac cycle, but are obtained based on a PPG signal or an ICG signal; in the embodiments disclosed in the present application, a heart rate analysis toolkit can be used to calculate variability of a signal segment, so as to describe a change mode of a signal sequence, including variation of a heartbeat interval, short-term and long-term scatter diagram standard deviations in a Poincare plot analysis, and in addition, sample entropy and approximate entropy of the signal sequence are calculated, so as to capture nonlinear dynamic characteristics of the physiological signal.

[0105] In step S30, a mapping between the morphological features, the statistical features and the sequence features and a plurality of parameters in the state equation is constructed, a training sample set is obtained, and the elastic cavity model is trained by using the training sample set, so as to obtain a prediction model.

[0106] Wherein, based on the feature extraction of step S20, further using a specific algorithm to realize the mapping from the multi-waveform features to the WK model parameters (such as Figure 4 The resistance R, the initial model compliance C parameters ) shown in the figure.

[0107] Specifically, for the cardiac cycle time, a preset algorithm is used to construct a mapping from the shape features, statistical features and sequence features to the resistance and initial model compliance C parameters in the state equation, obtaining a plurality of feature-parameter pairs; all the feature-parameter pairs are divided into a training sample set, a verification sample set and a test sample set, wherein the verification sample set is used to verify the prediction model, and the test sample set is used to test the prediction model; the training sample set is input into the elastic cavity model for prediction, a weighted loss term is constructed according to the prediction result and each actual value in the training sample set, and the elastic cavity model is iteratively optimized according to the weighted loss term, obtaining a prediction model.

[0108] Wherein, in the embodiments disclosed in the present application, the collected multi-physiological signals and their corresponding WK model parameters are first divided at the sample level: training set (80%), verification set (10%), test set (10%), and each sample is constructed into a feature-parameter pair of a cardiac cycle.

[0109] Further, a weighted loss term is constructed according to the actual value of each training sample in the training sample set and the prediction result:

[0110] ;

[0111] ;

[0112] Wherein, Indicates the weighted loss term, Indicates the total number of training samples, Indicates the weight of the th sample, Indicates the th actual value, Indicates the prediction result, Indicates the th training sample, Indicates the th iteration weight, Indicates the output value of the training sample input into the weak regressor in the th iteration, Indicates the input training sample, Indicates the number of iterations, denotes the median; the sample weights are initialized and initialized pseudo-residuals are calculated to train the weak regressor to minimize the weighted squared error:

[0113] ;

[0114] wherein, denotes the sample weight of the i-th sample in the j-th iteration, denotes the i-th sample weight in the j-th iteration,

[0115] ;

[0116] wherein, denotes the pseudo-residual of the i-th training sample in the j-th iteration, denotes the prediction result in the j-th iteration, ;

[0117] ;

[0118] wherein, denotes the weighted squared error minimized in the j-th iteration training, denotes the sample weight of the i-th sample in the j-th iteration training, denotes the output value of the i-th training sample inputted into the weak regressor in the j-th iteration, ;

[0119] ;

[0120] ;

[0121] wherein, denotes the i-th absolute error in the j-th iteration, denotes the average weighted loss in the j-th iteration, denotes the i-th sample weight in the j-th iteration, ;

[0122] ​​​​​​​​​​​ ;

[0123] ;

[0124] in, Indicates the first Learner weights in the next iteration Indicates the first The th iteration Individual sample weights.

[0125] Among them, the prediction results The weighted median of the weak learner sequence is used. After constructing the weighted loss, iterative optimization can be performed, starting from the initial prediction. The process begins with T iterations: For each iteration, pseudo-residuals are first calculated to train a weak learner, thereby minimizing the weighted squared error. Based on this, the normalized absolute error is calculated, and the average weighted loss is calculated to obtain the learner weights. Finally, the sample weights for the next iteration are updated based on the learner weights. The parameters are continuously updated during iterations to continuously reduce the loss function, thus achieving the final model construction. By minimizing the prediction residuals and physiological biases, the convergence of the iterative calibration process is accelerated and its robustness is improved, avoiding the parameter drift and overfitting risks caused by purely empirical calibration. Ultimately, this significantly improves the overall model's accuracy and generalization ability in individualized blood pressure prediction.

[0126] Step S40: Obtain pulse wave velocity from multiple standard physiological signals, calculate the Young's modulus of the arterial wall based on the pulse wave velocity, construct a pulse wave velocity equation, add the Young's modulus of the arterial wall to the pulse wave velocity equation, and obtain a specific expression for the arterial compliance value C parameter.

[0127] Existing technologies for blood pressure monitoring rely solely on either the MK equation or the WK model. The former focuses on describing the relationship between pulse wave velocity and vascular elasticity, while the latter focuses on the transient dynamic response of pressure and flow. Neither can simultaneously reflect the spatial distribution characteristics of vascular elasticity changes nor the temporal characteristics of cardiac pumping and peripheral resistance, resulting in limited accuracy in blood pressure estimation and an inability to comprehensively characterize the true physiological processes of the cardiovascular system. Therefore, this invention, based on the construction and optimization of the WK model to obtain a predictive model, further introduces the MK equation, achieving a fusion of the MK equation and the WK model. This not only overcomes the insufficient predictive accuracy of the arterial compliance value C parameter in the single WK model but also establishes the relationship between vascular elasticity, vascular geometry, and pulse wave velocity, enabling a description of vascular compliance and achieving high-precision, stable, continuous, and interpretable monitoring of blood pressure.

[0128] Specifically, according to the peak time difference in the standard photoplethysmogram or the pulse transmission time in the standard electrocardiogram, the pulse wave velocity is obtained, and according to the calculation expression of the pulse wave velocity, the arterial wall Young's modulus is calculated:

[0129] ;

[0130] ;

[0131] wherein, represents the pulse wave velocity, represents the arterial wall Young's modulus, represents the arterial wall thickness, represents the blood density, represents the arterial radius; the pulse wave velocity equation is constructed:

[0132] ;

[0133] ;

[0134] ;

[0135] wherein, represents the arterial compliance value C parameter, represents the volume change, represents the pressure change, represents the blood vessel length, represents the arterial radius change; the arterial wall Young's modulus is added to the pulse wave velocity equation, and the specificity expression of the arterial compliance value C parameter is constructed:

[0136] ;

[0137] wherein, represents the lumen volume.

[0138] According to the MK equation, the expression for calculating the pulse wave velocity can be constructed, so that the arterial wall Young's modulus can be inversely deduced; and the pulse wave velocity is the reciprocal of PTT, which can be obtained from the standard ECG signal and the standard PPG signal. Specifically, according to the R peak of the standard ECG signal, the cardiac cycle is divided, and then the standard ECG signal and the standard PPG signal are segmented by the cardiac cycle. The PPT is the event from the R peak to the peak value of the standard PPG signal. After obtaining the PPT, the pulse wave velocity can be obtained.

[0139] Furthermore, in the MK equation, the unit of pulse wave velocity is m / s; the unit of Young's modulus of the arterial wall is Pa, reflecting the stiffness characteristics of the arterial wall; blood density is approximately 1050 kg / m³; and the unit of arterial radius is m. After deriving the Young's modulus of the arterial wall, an expression for calculating the arterial compliance C-parameter (i.e., the ratio of volume change to pressure change) is further constructed. For an artery simulated as a circular tube, the change in arterial radius is imported into this expression to obtain the volume change. Combined with the expression for calculating the arterial compliance C-parameter, the arterial compliance C-parameter can be calculated. This provides a calibration method for the arterial compliance C-parameter based on the physical mechanism of pulse wave propagation. Moreover, individual differences are considered in the calculation, enabling individualized adjustment of the arterial compliance C-parameter to ensure its physiological rationality.

[0140] Step S50: Calculate the arterial compliance value C parameter of the target object according to the specificity expression, adjust the model compliance C parameter in the prediction model, and use the adjusted prediction model to monitor the blood pressure of the target object to obtain a continuous blood pressure estimate.

[0141] Specifically, based on the target object's blood density, pulse wave velocity, and lumen volume, the arterial compliance value C parameter of the target object is calculated; the model compliance C parameter in the prediction model is adjusted using the arterial compliance value C parameter to obtain the final prediction model; the blood pressure of the target object is continuously monitored using the final prediction model to obtain a continuous blood pressure estimate.

[0142] Specifically, the arterial compliance value C parameter calculated from the MK equation is used as a soft physiological constraint on the model compliance C parameter to correct it. By constraining the arterial compliance value C parameter, the model's predicted pressure and clinically observed pressure are jointly optimized, rather than through unconstrained fitting or a single loss function as in existing technologies. This ensures that the model compliance C parameter is always anchored to non-invasive physiological measurements, thereby avoiding the problem of parameters being out of sync with actual arterial elasticity in existing technologies and improving the biomechanical realism of the model.

[0143] Furthermore, such as Figure 3 As shown, besides the C parameter of model compliance ( Figure 3 In addition to correcting the C parameter in the model, the resistance R predicted by waveform features is also used as input to monitor the blood pressure of the target object together with the model compliance C parameter, thereby improving the accuracy and biological authenticity of blood pressure prediction for different objects.

[0144] Furthermore, such as Figure 5As shown, based on the above-mentioned continuous blood pressure monitoring method based on the elastic cavity model and pulse wave velocity equation, the present invention also provides a continuous blood pressure monitoring system based on the elastic cavity model and pulse wave velocity equation, wherein the continuous blood pressure monitoring system based on the elastic cavity model and pulse wave velocity equation includes:

[0145] The parameter acquisition module 51 is used to acquire multiple physiological signals of the target object, preprocess all the physiological signals to obtain corresponding standard physiological signals, and obtain multiple conventional parameters from all the standard physiological signals.

[0146] The feature extraction module 52 is used to construct an elastic cavity model, add all the conventional parameters to the elastic cavity model to construct a state equation, and extract features from all standard physiological signals to obtain morphological features, statistical features and sequence features.

[0147] The model training module 53 is used to construct the mapping between the morphological features, the statistical features, and the sequence features and multiple parameters in the state equation, to obtain a training sample set, and to use the training sample set to train the elastic cavity model to obtain a prediction model.

[0148] The constraint parameter construction module 54 is used to obtain pulse wave velocity from multiple standard physiological signals, calculate the Young's modulus of the arterial wall based on the pulse wave velocity, construct a pulse wave velocity equation, add the Young's modulus of the arterial wall to the pulse wave velocity equation, and obtain a specific expression for the arterial compliance value C parameter.

[0149] The blood pressure monitoring module 55 is used to calculate the arterial compliance value C parameter of the target object according to the specific expression, adjust the model compliance C parameter in the prediction model, and use the adjusted prediction model to monitor the blood pressure of the target object to obtain a continuous blood pressure estimate.

[0150] Furthermore, such as Figure 6 As shown, based on the above-mentioned continuous blood pressure monitoring method and system based on the elastic cavity model and pulse wave velocity equation, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 6 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0151] The memory 20 can be an internal storage unit of the terminal in some embodiments, such as a hard disk or a memory of the terminal. The memory 20 can also be an external storage device of the terminal in other embodiments, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. equipped on the terminal. Further, the memory 20 can include both an internal storage unit and an external storage device of the terminal. The memory 20 is used to store application software and various data installed on the terminal, such as program codes of the terminal, etc. The memory 20 can also be used to temporarily store data that has been output or will be output. In an embodiment, the memory 20 stores a continuous blood pressure monitoring program based on the elastic cavity model and the pulse wave velocity equation, which can be executed by the processor 10 to implement the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation in the present application.

[0152] The processor 10 can be a Central Processing Unit (CPU), a microprocessor or other data processing chip in some embodiments, which is used to run program codes or process data stored in the memory 20, such as to execute the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation, etc.

[0153] The display 30 can be an LED display, a liquid crystal display, a touch liquid crystal display, an OLED (Organic Light-Emitting Diode) touch, etc. in some embodiments. The display 30 is used to display information of the terminal and to display a visualized user interface. The components of the terminal communicate with each other through a system bus.

[0154] In an embodiment, the steps of the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation as described above are implemented when the processor 10 executes the continuous blood pressure monitoring program based on the elastic cavity model and the pulse wave velocity equation in the memory 20.

[0155] The present application also provides a computer readable storage medium, wherein the computer readable storage medium stores a continuous blood pressure monitoring program based on the elastic cavity model and the pulse wave velocity equation, which, when executed by a processor, implements the steps of the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation as described above.

[0156] In summary, the present application provides a continuous blood pressure monitoring method and related device based on an elastic cavity model and a pulse wave velocity equation, the method comprising: collecting a plurality of physiological signals of a target object, preprocessing all the physiological signals to obtain corresponding standard physiological signals, and obtaining a plurality of conventional parameters from all the standard physiological signals; constructing an elastic cavity model, adding all the conventional parameters to the elastic cavity model to construct a state equation, and extracting features from all the standard physiological signals to obtain morphological features, statistical features, and sequence features; constructing a mapping between the morphological features, the statistical features, the sequence features, and a plurality of parameters in the state equation to obtain a training sample set, and training the elastic cavity model using the training sample set to obtain a prediction model; obtaining a pulse wave velocity from a plurality of the standard physiological signals, calculating an arterial wall Young's modulus according to the pulse wave velocity, constructing a pulse wave velocity equation, adding the arterial wall Young's modulus to the pulse wave velocity equation to obtain a specific expression of an arterial compliance value C parameter; calculating the arterial compliance value C parameter of the target object according to the specific expression, adjusting a model compliance C parameter in the prediction model, and monitoring blood pressure of the target object using the adjusted prediction model to obtain a continuous blood pressure estimate. The present application introduces individualized physiological information, and improves the accuracy and biological authenticity of blood pressure prediction for different objects.

[0157] It should be noted that in this document, the terms "comprising", "containing", or any other similar term are intended to encompass non-exclusive inclusion, such that a process, method, article, or terminal including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or further includes elements inherent to such a process, method, article, or terminal. Without more limitations, the element defined by the phrase "comprising a" does not exclude the presence of additional identical elements in the process, method, article, or terminal including the element.

[0158] Of course, those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing relevant hardware (such as a processor, a controller, etc.) to complete, and the program can be stored in a computer-readable computer-readable storage medium, and the program can include the processes of the above-mentioned method embodiments when executed. The computer-readable storage medium can be a memory, a magnetic disc, an optical disc, etc.

[0159] It should be understood that the application of the present application is not limited to the above examples, and those skilled in the art can make improvements or changes according to the above description, and all such improvements and changes shall fall within the scope of protection of the claims of the present application.

Claims

1. A continuous blood pressure monitoring method based on an elastic cavity model and a pulse wave velocity equation, characterized in that, The continuous blood pressure monitoring method based on the elastic cavity model and pulse wave velocity equation includes: Multiple physiological signals from the target object are collected, all of which are preprocessed to obtain corresponding standard physiological signals, and multiple routine parameters are obtained from all of the standard physiological signals, specifically including: The conventional parameters include: cardiac contraction and ejection time, cardiac diastolic time, cardiac cycle time, and blood flow parameters; Data is collected from the target object using multiple sensors to obtain photoplethysmography (PPG), electrocardiogram (ECG), and cardiac impedance cardiogram (CIK). The first noise in the photoplethysmography (PPG), electrocardiogram (ECG), and impedance cardiogram is filtered out using a first filter, and the second noise in the PPG, ECG, and impedance cardiogram is filtered out using a second filter, resulting in a standard PPG, standard ECG, and standard impedance cardiogram. Extract the cardiac systolic ejection time, cardiac diastolic time, and cardiac cycle time of the target object from the standard photoplethysmography pulse wave; Extract the blood flow parameters of the target object from the standard impedance cardiogram: ； in, This represents blood flow parameters. Indicates blood density, and These respectively represent blood from Flowing to Point and Time at point, and They represent Point and The arterial wall is thick at the point. Indicates the length of the blood vessel. Indicates the monitoring starting point; An elastic cavity model is constructed, and all the aforementioned conventional parameters are added to the elastic cavity model to construct a state equation. Feature extraction is performed on all standard physiological signals to obtain morphological features, statistical features, and sequence features, specifically including: Construct an elastic cavity model of the target object, and add the cardiac systolic ejection time, the cardiac diastolic time, and the cardiac cycle time to the elastic cavity model: ； ； in, Indicates the first Current input at any given time, Indicates cardiac output. Indicates the duration of the cardiac cycle. It indicates the duration of cardiac contraction and ejection. A time index representing the entire cardiac cycle. Indicates the number of cardiac cycles. The C-parameter represents the model's compliance. Indicates the first Voltage at time, Indicates resistance; Based on the cardiac cycle time, statistical features are obtained by statistically analyzing the standard photoplethysmography pulse wave, the standard electrocardiogram, and the standard impedance cardiogram. The morphological features are then obtained by averaging the statistical features. Acquire photoplethysmography (PPG) signals or impedance cardiogram (ETC) signals from multiple cardiac cycles, calculate the variation signals of the PPG signals or ETC signals, and obtain sequence characteristics. A mapping is constructed between the morphological features, statistical features, and sequence features and multiple parameters in the state equation to obtain a training sample set. The training sample set is then used to train the elastic cavity model to obtain a prediction model. Pulse wave velocity is obtained from multiple standard physiological signals, the Young's modulus of the arterial wall is calculated based on the pulse wave velocity, a pulse wave velocity equation is constructed, and the Young's modulus of the arterial wall is added to the pulse wave velocity equation to obtain a specific expression for the arterial compliance value C parameter. The arterial compliance value C parameter of the target object is calculated based on the specific expression, and the model compliance C parameter in the prediction model is adjusted. The blood pressure of the target object is monitored using the adjusted prediction model to obtain a continuous blood pressure estimate.

2. The continuous blood pressure monitoring method based on the elastic cavity model and pulse wave velocity equation according to claim 1, characterized in that, The process of constructing a mapping between the morphological features, statistical features, and sequence features and multiple parameters in the state equation to obtain a training sample set, and then using the training sample set to train the elastic cavity model to obtain a prediction model, specifically includes: For the cardiac cycle time, a preset algorithm is used to construct a mapping from the morphological features, statistical features, and sequence features to the resistance and initial model compliance C parameter in the state equation, resulting in multiple feature-parameter pairs; All the feature-parameter pairs are divided into a training sample set, a validation sample set, and a test sample set, wherein the validation sample set is used to validate the prediction model, and the test sample set is used to test the prediction model. The training sample set is input into the elastic cavity model for prediction. A weighted loss term is constructed based on the prediction results and each actual value in the training sample set. The elastic cavity model is iteratively optimized based on the weighted loss term to obtain the prediction model.

3. The continuous blood pressure monitoring method based on the elastic cavity model and pulse wave velocity equation according to claim 2, characterized in that, The step of constructing a weighted loss term based on the prediction results and each actual value in the training sample set, and iteratively optimizing the elastic cavity model based on the weighted loss term to obtain the prediction model, specifically includes: A weighted loss term is constructed based on the actual value of each training sample in the training sample set and the prediction result: ； ； in, Indicates the weighted loss term. This represents the total number of training samples. Indicates the first Each sample weight, Indicates the first One actual value, Indicates the prediction result. Indicates the first training samples, Indicates the first Weights for the next iteration Indicates the first In the next iteration, the training samples are input into the output value of the weak regressor. This represents the input training samples. Indicates the number of iterations. This represents the median; The sample weights are initialized, and the initial pseudo-residuals are calculated to train the weak regressor, obtaining the minimized weighted squared error: ； in, Indicates the first The first iteration in the round Individual sample weights; ； in, Indicates the first The first iteration in the round The pseudo residuals of each training sample Indicates the first Prediction results in round iteration; ； in, Indicates the first Minimize the weighted squared error during rounds of iterative training. Indicates the first The first round of iterative training The sample weights of each sample. Indicates the first In the nth iteration Each training sample is input into the output value of the weak regressor; An absolute error is constructed based on the output value and the corresponding pseudo-residual, and the average weighted loss is calculated using the absolute error and the sample weights of the current round: ； ； in, Indicates the first In the nth iteration One absolute error, Indicates the first The average weighted loss of the next iteration Indicates the first In the nth iteration Individual sample weights; For the current round, the learner weights are calculated using the average weighted loss, and the sample weights are updated using these learner weights to obtain the sample weights for the next round. This process continues until the updated parameters that minimize the weighted loss term are obtained, and the prediction model is constructed using these updated parameters. ； ； in, Indicates the first Learner weights in the next iteration Indicates the first The th iteration Individual sample weights.

4. The continuous blood pressure monitoring method based on the elastic cavity model and pulse wave velocity equation according to claim 1, characterized in that, The process of obtaining pulse wave velocities from multiple standard physiological signals, calculating the Young's modulus of the arterial wall based on the pulse wave velocities, constructing a pulse wave velocity equation, and adding the Young's modulus of the arterial wall to the pulse wave velocity equation to obtain a specific expression for the arterial compliance value C parameter specifically includes: Based on the peak time difference in the standard photoplethysmography pulse wave or the pulse conduction time in the standard electrocardiogram, the pulse wave velocity is obtained, and the Young's modulus of the arterial wall is calculated according to the calculation expression of the pulse wave velocity. ； ； in, Indicates pulse wave velocity, This represents the Young's modulus of the arterial wall. This indicates thickened arterial walls. Indicates blood density, Indicates the radius of the artery; Constructing the pulse wave velocity equation: ； ； ； in, The C parameter represents the arterial compliance value. Indicates volume change, Indicates pressure changes. Indicates the length of the blood vessel. Indicates changes in arterial radius; By adding the Young's modulus of the arterial wall to the pulse velocity equation, a specific expression for the arterial compliance value C parameter is constructed: ； in, This indicates the volume of the cavity.

5. The continuous blood pressure monitoring method based on the elastic cavity model and pulse wave velocity equation according to claim 1, characterized in that, The step of calculating the arterial compliance value C parameter of the target object based on the specificity expression, adjusting the model compliance C parameter in the prediction model, and using the adjusted prediction model to monitor the blood pressure of the target object to obtain continuous blood pressure estimates specifically includes: Calculate the arterial compliance value C parameter of the target object based on the target object's blood density, pulse wave velocity, and lumen volume; The model compliance C parameter in the prediction model is adjusted using the arterial compliance value C parameter to obtain the final prediction model; The blood pressure of the target object is continuously monitored using the final prediction model to obtain continuous blood pressure estimates.

6. A continuous blood pressure monitoring system based on an elastic cavity model and a pulse wave velocity equation, characterized in that, The continuous blood pressure monitoring system based on the elastic cavity model and pulse wave velocity equation is used to implement the continuous blood pressure monitoring method based on the elastic cavity model and pulse wave velocity equation as described in any one of claims 1-5, wherein the continuous blood pressure monitoring system based on the elastic cavity model and pulse wave velocity equation comprises: The parameter acquisition module is used to acquire multiple physiological signals of the target object, preprocess all the physiological signals to obtain corresponding standard physiological signals, and obtain multiple conventional parameters from all the standard physiological signals. The feature extraction module is used to construct an elastic cavity model, add all the conventional parameters to the elastic cavity model to construct a state equation, and extract features from all standard physiological signals to obtain morphological features, statistical features and sequence features. The model training module is used to construct the mapping between the morphological features, the statistical features, and the sequence features and multiple parameters in the state equation, to obtain a training sample set, and to train the elastic cavity model using the training sample set to obtain a prediction model. The constraint parameter construction module is used to obtain pulse wave velocity from multiple standard physiological signals, calculate the Young's modulus of the arterial wall based on the pulse wave velocity, construct the pulse wave velocity equation, add the Young's modulus of the arterial wall to the pulse wave velocity equation, and obtain a specific expression for the arterial compliance value C parameter. The blood pressure monitoring module is used to calculate the arterial compliance value C parameter of the target object according to the specific expression, adjust the model compliance C parameter in the prediction model, and use the adjusted prediction model to monitor the blood pressure of the target object to obtain continuous blood pressure estimates.

7. A terminal, characterized in that, The terminal includes: a memory, a processor, and a continuous blood pressure monitoring program based on an elastic cavity model and a pulse wave velocity equation stored in the memory and executable on the processor. When the continuous blood pressure monitoring program based on an elastic cavity model and a pulse wave velocity equation is executed by the processor, it implements the steps of the continuous blood pressure monitoring method based on an elastic cavity model and a pulse wave velocity equation as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a continuous blood pressure monitoring program based on an elastic cavity model and a pulse wave velocity equation. When the continuous blood pressure monitoring program based on the elastic cavity model and the pulse wave velocity equation is executed by a processor, it implements the steps of the continuous blood pressure monitoring method based on the elastic cavity model and the pulse wave velocity equation as described in any one of claims 1-5.

Images