Blood pressure monitoring method, device and system and electronic equipment

Abstract

The invention provides a blood pressure monitoring method, device and system and electronic equipment. A heartbeat signal and a pulse wave signal of a target object at the current time are obtained. And the pulse transmission time is calculated by using the two signals. The pulse transmission time is used to estimate blood pressure in a first blood pressure prediction model. The model is established based on pulse transmission time data of a target object in past time. Feature parameters are extracted from the pulse wave signals. The characteristic parameters are used for a second blood pressure prediction model, and the model is based on pulse wave characteristic data of the target object in the past time. And combining the output values of the two models through a fusion algorithm to obtain a comprehensive blood pressure measurement value. The heartbeat signal and the pulse wave signal of the target object are collected, the pulse transmission time and the pulse wave characteristic parameters are used for conducting composite blood pressure prediction and fusion, the blood pressure of the target object is accurately predicted, and therefore continuous monitoring of the blood pressure can be achieved through the heartbeat signal and the pulse wave signal of the target object.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of blood pressure monitoring, and in particular to a blood pressure monitoring method, device, system and electronic equipment. BACKGROUND

[0002] Currently, common blood pressure measurement methods mainly include mercury sphygmomanometer based on Korotkoff sound auscultation method and electronic sphygmomanometer based on oscillometric method. However, these blood pressure measurement methods can only provide instantaneous blood pressure data, and cannot capture systolic pressure and diastolic pressure in the same cardiac cycle. At the same time, such instantaneous measurement method may not accurately reflect the true blood pressure level of the object due to "white coat effect".

[0003] On the other hand, the accuracy of blood pressure prediction is also a key problem. When using biological radar to detect blood pressure alone, the measurement result is often inaccurate due to signal quality fluctuations, which cannot meet the needs of actual application.

[0004] Therefore, how to accurately and continuously monitor the blood pressure of the object has become a technical problem that technicians in the field urgently need to solve. SUMMARY

[0005] In view of the above problems, the present application provides a blood pressure monitoring method, device, system and electronic equipment which can overcome the above problems or at least partially solve the above problems, and the technical solutions are as follows:

[0006] A blood pressure monitoring method comprises:

[0007] obtaining a first heartbeat signal and a first pulse wave signal of a target object at a current time;

[0008] obtaining a first pulse transit time of the target object by using the first heartbeat signal and the first pulse wave signal;

[0009] inputting the first pulse transit time into a first blood pressure prediction model to obtain a blood pressure prediction value output by the first blood pressure prediction model, wherein the first blood pressure prediction model is a blood pressure prediction model constructed based on pulse transit times of the target object at historical times, and the historical times are times before the current time;

[0010] obtaining a pulse wave feature parameter of the first pulse wave signal;

[0011] inputting the pulse wave feature parameter into a second blood pressure prediction model to obtain a blood pressure state value output by the second blood pressure prediction model, wherein the second blood pressure prediction model is a blood pressure prediction model constructed based on pulse wave features of the target object at the historical times;

[0012] The blood pressure fusion estimation algorithm is used for blood pressure fusion processing on the blood pressure prediction value and the blood pressure state value, to obtain a target blood pressure value of the target object.

[0013] A blood pressure monitoring device comprises a target object signal obtaining unit, a pulse transit time obtaining unit, a blood pressure prediction value obtaining unit, a pulse wave feature parameter obtaining unit, a blood pressure state value obtaining unit, and a target blood pressure value obtaining unit,

[0014] The target object signal obtaining unit is configured to obtain a first heartbeat signal and a first pulse wave signal of a target object at a current time.

[0015] The pulse transit time obtaining unit is configured to obtain a first pulse transit time of the target object by using the first heartbeat signal and the first pulse wave signal.

[0016] The blood pressure prediction value obtaining unit is configured to input the first pulse transit time into a first blood pressure prediction model to obtain a blood pressure prediction value output by the first blood pressure prediction model, wherein the first blood pressure prediction model is a blood pressure prediction model constructed based on pulse transit times of the target object at historical times, and the historical times are times before the current time.

[0017] The pulse wave feature parameter obtaining unit is configured to obtain a pulse wave feature parameter of the first pulse wave signal.

[0018] The blood pressure state value obtaining unit is configured to input the pulse wave feature parameter into a second blood pressure prediction model to obtain a blood pressure state value output by the second blood pressure prediction model, wherein the second blood pressure prediction model is a blood pressure prediction model constructed based on pulse wave features of the target object at the historical times.

[0019] The target blood pressure value obtaining unit is configured to use a preset blood pressure fusion estimation algorithm to perform blood pressure fusion processing on the blood pressure prediction value and the blood pressure state value, to obtain a target blood pressure value of the target object.

[0020] A blood pressure monitoring system comprises a biological radar, a pulse bracelet, a cloud server, and an upper computer,

[0021] The upper computer is configured to send a signal acquisition instruction to the biological radar and the pulse bracelet through the cloud server, and store and display a target blood pressure value fed back by the cloud server.

[0022] The biological radar is configured to continuously acquire a heartbeat signal of a target object in response to the signal acquisition instruction, and feed back the heartbeat signal to the cloud server.

[0023] The pulse bracelet is configured to continuously collect the pulse wave signal of the target object in response to the signal collection instruction and feed back the pulse wave signal to the cloud server.

[0024] The cloud server is configured to construct a first blood pressure prediction model and a second blood pressure prediction model of the target object based on the heartbeat signal and the pulse wave signal, perform blood pressure fusion processing on a blood pressure prediction value output by the first blood pressure prediction model and a blood pressure state value output by the second blood pressure prediction model, and continuously output the target blood pressure value of the target object to the upper computer.

[0025] An electronic device includes at least one processor and at least one memory connected to the processor via a bus; the processor, the memory and the bus complete mutual communication; the processor is configured to call program instructions in the memory to execute the blood pressure monitoring method.

[0026] By the above technical solution, the blood pressure monitoring method, device, system and electronic device provided by the application obtain a first heartbeat signal and a first pulse wave signal of a target object at a current time; a first pulse transit time of the target object is obtained by using the first heartbeat signal and the first pulse wave signal; the first pulse transit time is input into a first blood pressure prediction model to obtain a blood pressure prediction value output by the first blood pressure prediction model, wherein the first blood pressure prediction model is a blood pressure prediction model constructed based on the pulse transit time of the target object at a historical time, and the historical time is a time before the current time; a pulse wave feature parameter of the first pulse wave signal is obtained; the pulse wave feature parameter is input into a second blood pressure prediction model to obtain a blood pressure state value output by the second blood pressure prediction model, wherein the second blood pressure prediction model is a blood pressure prediction model constructed based on the pulse wave feature of the target object at the historical time; a target blood pressure value of the target object is obtained by using a preset blood pressure fusion estimation algorithm to perform blood pressure fusion processing on the blood pressure prediction value and the blood pressure state value. The application predicts the blood pressure of the target object accurately by collecting the heartbeat signal and the pulse wave signal of the target object, using the pulse transit time and the pulse wave feature parameter for composite blood pressure prediction and fusion, so that the continuous monitoring of the blood pressure can be realized by continuously collecting the heartbeat signal and the pulse wave signal of the target object.

[0027] The above description is only a summary of the technical solutions of the application. In order to make the technical means of the application more clear, the application can be implemented according to the content of the specification, and in order to make the above and other purposes, features and advantages of the application more obvious and easy to understand, the following specific embodiments of the application are described. BRIEF DESCRIPTION OF DRAWINGS

[0028] Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments with reference made to the accompanying drawings. The drawings provided herein are for illustrative purposes only and, therefore, should not be considered to be limiting in any way. Like reference characters often refer to parts throughout the drawings. In the drawings:

[0029] Figure 1 A flow chart illustrating one embodiment of the blood pressure monitoring method provided by the embodiments of the present application is shown;

[0030] Figure 2 A schematic diagram illustrating the process of the pulse wave time domain feature recognition algorithm provided by the embodiments of the present application is shown;

[0031] Figure 3 A schematic diagram illustrating the process of the pulse wave shape classification provided by the embodiments of the present application is shown;

[0032] Figure 4 A schematic diagram illustrating one embodiment of the tongue site division provided by the embodiments of the present application is shown;

[0033] Figure 5 A flow chart illustrating one embodiment of the step S100 of the blood pressure monitoring method provided by the embodiments of the present application is shown;

[0034] Figure 6 A flow chart illustrating the heartbeat signal blind source separation algorithm provided by the embodiments of the present application is shown;

[0035] Figure 7 A flow chart illustrating one embodiment of the step S200 of the blood pressure monitoring method provided by the embodiments of the present application is shown;

[0036] Figure 8 A flow chart illustrating the optimal peak point algorithm based on envelope detection provided by the embodiments of the present application is shown;

[0037] Figure 9 A structural schematic diagram of the blood pressure monitoring device provided by the embodiments of the present application is shown;

[0038] Figure 10 A schematic diagram of the blood pressure monitoring system provided by the embodiments of the present application is shown. DETAILED DESCRIPTION

[0039] Exemplary embodiments of the present application will be described herein below with reference to the accompanying drawings. Although exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the application to those skilled in the art.

[0040] With the increasing demand for blood pressure monitoring, traditional blood pressure measurement methods such as Korotkoff auscultation and oscillometric electronic sphygmomanometers have limitations in clinical use. These methods usually only provide instantaneous blood pressure values, and some people experience temporary blood pressure elevation when measuring blood pressure in front of a doctor or in a clinical environment due to stress, anxiety, and other emotional factors. However, when measuring at home or in a non-medical environment, blood pressure returns to normal or relatively low "white coat effect", which may not accurately reflect the user's true blood pressure level.

[0041] In terms of hardware devices for blood pressure measurement, current blood pressure measurement methods such as volume compensation and arterial tension method rely on air bag cuffs, which may cause user discomfort and hinder blood flow when inflated. In addition, blood pressure estimation methods based on pulse transit time (PTT) require the attachment of electrodes on the skin, which are not very comfortable to wear.

[0042] In terms of blood pressure prediction accuracy, existing methods have obvious shortcomings. For example, using biological radar alone to measure blood pressure may affect accuracy due to fluctuations in signal quality. At the same time, models that simply assume a linear relationship between PTT and systolic blood pressure (SBP) and diastolic blood pressure (DBP) ignore the actual complex nonlinear relationship in the body, which can all lead to large errors in blood pressure prediction results.

[0043] In summary, existing blood pressure monitoring technology cannot simultaneously meet the needs of accurate continuous monitoring and user comfort, and there is an urgent need to develop a blood pressure monitoring method that can overcome the limitations of instantaneous blood pressure measurement, achieve continuous, accurate, and comfortable blood pressure monitoring, and provide valuable data support for long-term health management of users.

[0044] Based on this, in the embodiment of the present application, a blood pressure monitoring method is provided, first, the heartbeat signal and the pulse wave signal of the target object at the current time are obtained. The two signals are used to calculate the pulse transit time. The pulse transit time is used to estimate the blood pressure in the first blood pressure prediction model. This model is established based on the pulse transit time data of the target object at the past time. At the same time, the feature parameters are extracted from the pulse wave signal. These feature parameters are used for the second blood pressure prediction model, which is based on the pulse wave feature data of the target object at the past time. Finally, the output values of the two models are combined by a fusion algorithm to obtain the comprehensive blood pressure measurement value. The present application can accurately predict the blood pressure of the target object by collecting the heartbeat signal and the pulse wave signal of the target object, using the pulse transit time and the pulse wave feature parameters for composite blood pressure prediction and fusion, so as to realize the continuous monitoring of the blood pressure by continuously collecting the heartbeat signal and the pulse wave signal of the target object.

[0045] As shown in Figure 1 The flowchart of an embodiment of the blood pressure monitoring method provided by the present application, which can include:

[0046] S100, obtaining the first heartbeat signal and the first pulse wave signal of the target object at the current time.

[0047] Wherein, the target object is an individual who needs blood pressure monitoring.

[0048] Wherein, the heartbeat signal usually refers to the bioelectric signal generated by the heart beat. The heartbeat signal mainly reflects the electrical physiological activity of the heart in each beat cycle. The heartbeat signal of the target object can be collected by the commonly used electrocardiogram device in the embodiment of the present application. For example, the heartbeat signal of the target object can be collected by the electrode patch in the embodiment of the present application. Optionally, the heartbeat signal of the target object can also be collected by the biological radar in the embodiment of the present application. The biological radar can obtain the original heartbeat signal from the echo signal by non-contact way through the millimeter wave radar, which can reduce the discomfort of the target object when collecting the heartbeat signal.

[0049] The pulse wave signal is a physiological signal reflecting heart activity and blood vessel state through blood flow dynamics change. The pulse wave signal of the target object can be collected by the pulse wave sensor. Optionally, the pulse wave signal at the radial artery of the target object can be acquired by using a pressure sensor, a photoelectric sensor, a capacitive sensor or an ultrasonic sensor. The pulse wave signal at the fingertip of the target object can also be acquired by using a finger clip sensor. The pulse wave signal of the measured part of the target object can be obtained by using a photoelectric sensor to detect blood volume change of the measured part of the target object through photo plethysmography (PPG).

[0050] The first heartbeat signal and the first pulse wave signal of the target object can be collected synchronously, so that the electrocardiogram signal and the pulse wave signal of the same cardiac cycle are recorded synchronously, and the first pulse transit time of the target object can be calculated accurately subsequently.

[0051] S200, the first pulse transit time of the target object is obtained by using the first heartbeat signal and the first pulse wave signal.

[0052] Specifically, the heartbeat peak point in the first heartbeat signal and the pulse wave peak point in the first pulse wave signal in the same cardiac cycle can be captured, and then the time difference between the heartbeat peak point and the pulse wave peak point is calculated, and the time difference is confirmed as the first pulse transit time of the target object.

[0053] S300, the first pulse transit time is input into the first blood pressure prediction model to obtain a blood pressure prediction value output by the first blood pressure prediction model, wherein the first blood pressure prediction model is a blood pressure prediction model constructed based on the pulse transit time of the target object at a historical time, and the historical time is a time before the current time.

[0054] The blood pressure prediction value can include a systolic pressure prediction value and a diastolic pressure prediction value.

[0055] The historical time can be a time before the current time and continuous with the current time.

[0056] The first blood pressure prediction model can be constructed based on the Moens-Korteweg equation and the Hughes equation and the individual-specific parameter estimated based on the pulse transit time of the target object.

[0057] The Moens-Korteweg equation is used to reflect the relationship between pulse propagation time and arterial wall elasticity and arterial wall distensibility. The pulse transit time calculation formula in the Moens-Korteweg equation for an artery with an equivalent length of L is:

[0058]

[0059] where E is the Young's modulus of the vessel wall, h is the wall thickness, D is the internal diameter of the elastic tube at equilibrium, p is the fluid density, and K is the so-called Moens constant, an empirical dimensionless value, K = 0.8 for the human aorta.

[0060] where the Hughes equation is used to reflect the relationship between the blood pressure P and the Young's modulus of the vessel wall E. The Hughes equation is:

[0061] E = E0 x e αp

[0062] where E0 is the modulus of elasticity of the vessel when the pressure is zero, p is the blood pressure, and a is a quantity representing the characteristics of the vessel, with a value of 0.016-0.018 (mmHg -1 ).

[0063] Combining the Moens-Korteweg equation and the Hughes equation, we have:

[0064]

[0065] Through the derivation of the pulse transit time, we have:

[0066]

[0067] Thus, the linear relationship between the systolic blood pressure and the pulse transit time is:

[0068] SBP = a + b x PTT

[0069] where a and b are undetermined coefficients, which are different for different individuals, but for the same individual, the values of the coefficients do not change within a certain period of time.

[0070] Based on the E.C.Y. Poon equation, the linear relationship between the mean blood pressure (MBP) and the pulse transit time is:

[0071]

[0072] where MBP0 and PTT0 are reference calibration values. The pressure difference PP and the mean blood pressure can be obtained from the corresponding systolic and diastolic blood pressures:

[0073] PP = SBP - DBP

[0074]

[0075] In combination with the above formula, it can be known that the linear relationship between diastolic pressure and pulse transit time is:

[0076]

[0077] wherein SBP0 and DBP0 are reference calibration values. Therefore, diastolic pressure can be calculated by measuring pulse transit time.

[0078] In combination with the above analysis, the linear relationship between systolic pressure and pulse transit time and the linear relationship between diastolic pressure and pulse transit time are simplified as the linear relationship between blood pressure and pulse transit time:

[0079] SBP = a + b x PTT

[0080]

[0081] wherein a, b, A, B and C are individual-specific parameters that can be obtained by regression fitting.

[0082] In the construction phase of the first blood pressure prediction model, the second heartbeat signal, the second pulse wave signal and the actual blood pressure value of the target object in the same cardiac cycle can be collected. The second pulse transit time of the target object is calculated by using the second heartbeat signal and the second pulse wave signal. The individual-specific parameters of the target object are obtained by regression fitting of the second pulse transit time and the actual blood pressure value. The individual-specific parameters are substituted into the linear relationship between blood pressure and pulse transit time to obtain the first blood pressure prediction model of the target object. The first pulse transit time calculated by the first heartbeat signal and the first pulse wave signal of the target object obtained synchronously in the subsequent process is used to continuously output the blood pressure prediction value of the target object.

[0083] Optionally, in the above Figure 1 Based on one or more embodiments, in another optional embodiment of the present embodiment, the first pulse transit time is input into the first blood pressure prediction model to obtain the blood pressure prediction value output by the first blood pressure prediction model. Specifically, the method can include:

[0084] The first pulse transit time is input into the first blood pressure prediction model so that the first blood pressure prediction model uses the first pulse transit time and the individual-specific parameters of the target object to obtain the blood pressure prediction value.

[0085] Specifically, the first pulse transit time can be substituted into the linear relationship between blood pressure and pulse transit time of the first blood pressure prediction model. The individual-specific parameters of the target object provided by the linear relationship between blood pressure and pulse transit time are used for calculation to obtain the systolic pressure prediction value and the diastolic pressure prediction value of the target object.

[0086] The embodiment of the present application can use a more personalized blood pressure prediction model to predict the blood pressure of the target object by combining the first pulse transit time and the individual-specific parameter, thereby improving the accuracy of the blood pressure prediction value.

[0087] S400, obtain a pulse wave feature parameter of the first pulse wave signal.

[0088] The pulse wave feature parameter can include a feature parameter related to blood pressure in the pulse wave signal.

[0089] Optionally, the pulse wave feature parameter includes a pulse wave time domain feature and a pulse wave frequency domain feature of the first pulse wave signal.

[0090] The pulse wave time domain feature refers to a feature of the pulse wave form observed in the time sequence.

[0091] Optionally, the pulse wave time domain feature can include one or more features of the average intensity of the pulse wave signal, the pulse wave peak value, the peak area, the peak-valley difference, the slope of the peak-valley connecting line of the falling branch, and the pulse width.

[0092] The pulse wave frequency domain feature refers to a feature obtained by analyzing the pulse wave signal after converting the pulse wave signal from the time domain to the frequency domain. This conversion is usually realized through Fourier Transform, which can decompose the pulse wave signal in the time domain into a set of sinusoidal waves of different frequencies.

[0093] The embodiment of the present application can extract the time domain feature of the pulse wave signal by analyzing the form of the pulse wave signal. Specifically, the embodiment of the present application can use Fourier Transform and wavelet transform to process the pulse wave signal based on the pulse wave time domain feature recognition algorithm of the pulse wave time domain form classification, extract the feature parameters of the pulse wave signal form feature from the overall and local angles respectively, and then classify the pulse wave signal according to the form feature of the pulse wave signal by using the support vector machine (SVM) algorithm, to identify the pulse wave form features of different categories.

[0094] Optionally, Figure 2 The figure shows the pulse wave time domain feature recognition process of the pulse wave time domain feature recognition algorithm provided by the embodiment of the present application. The input of the pulse wave time domain feature recognition algorithm based on the pulse wave time domain form classification is the pulse wave of different forms, and the output is the pulse wave time domain feature. The pulse wave time domain feature is embodied as the corresponding time domain feature coordinates in the pulse wave signal. The key steps of the pulse wave time domain feature recognition algorithm based on the pulse wave time domain form classification are: classifying the pulse wave form feature based on the support vector machine and extracting the pulse wave time domain feature based on the angle method and the difference method.

[0095] For the classification of pulse waveform morphology based on SVM, the main components of a single cycle pulse wave include the main wave, the pre-systolic wave and the systolic wave. However, due to physiological and pathological differences among individuals, these waveform features may vary, such as the pre-systolic wave and the systolic wave may become unobvious or feature fusion, or even disappear. In order to effectively identify the time domain features, the pulse wave is first classified according to the existing prior knowledge. By applying Fourier transform, the frequency domain features of the pulse wave can be obtained. Then, according to the frequency range of the main components of the pulse wave (i.e. the main wave, the pre-systolic wave and the systolic wave), the pulse wave signal is further analyzed using wavelet decomposition, and the wavelet coefficients are extracted at a specific decomposition level. These wavelet coefficients are combined with the frequency domain features of the pulse wave to form the feature parameters of the pulse waveform morphology. In order to reduce the data redundancy of the feature parameters of the morphology features, singular value decomposition technique can also be applied. Finally, support vector machine is used as a classifier to classify the pulse waveform morphology.

[0096] The process of classifying pulse waveform morphology by support vector machine can be as shown in Figure 3 In the data preparation stage, the single cycle pulse wave signal is first processed to generate a feature vector, and the corresponding label is assigned to the data set, thereby forming a sample data set. 80% of the data in this data set is selected as the training sample set to construct the SVM classifier, and the remaining 20% is used as the test sample set to evaluate the performance of the classifier. In the algorithm calibration stage, Gaussian radial basis function (RBF) is selected as the kernel function, and the values of the penalty factor and the radial basis parameter are determined. The hyperplane of the classifier is solved using the sequential minimal optimization (SMO) method to obtain the required decision function, and a binary SVM model is constructed. In the classification prediction stage, the test sample set divided in the previous stage is used to verify the classification effect of the SVM model, and the evaluation indexes are obtained, which are used to measure the performance of the classification.

[0097] For the pulse wave time domain feature extraction based on the angle method and the difference method, the angle method and the difference method are optional pulse wave time domain feature extraction algorithms, and the selection principle mainly depends on the obviousness of the pulse wave morphology feature. The difference method mainly extracts the extreme points of the wave by calculating the difference between any point on the pulse wave and the two points before and after it. Simply speaking, the difference method directly identifies the maximum and minimum points in the pulse wave, and can be used for time domain feature extraction of the pulse wave with obvious morphology feature (for example: the pulse wave with three peaks or two peaks). The angle method is suitable for the pulse wave with less obvious time domain feature (for example: the pulse wave with less obvious pre-rebound wave). In this case, the angle method represents the feature points by calculating the curvature of the feature points. By angle calculation, the less obvious feature points are converted into wave peaks or wave troughs with larger curvature, and then the difference method is used to determine the peak values of these points, so as to accurately extract the feature values of the feature points. In summary, the angle method and the difference method are two time domain feature parameter extraction techniques selected according to the different pulse wave morphology features. The difference method is suitable for the case with obvious feature, while the angle method is suitable for the pulse wave with less obvious time domain feature, and both of them can effectively extract the time domain features of the pulse wave for subsequent analysis.

[0098] Optionally, the obtaining process of the pulse wave time domain feature provided by the embodiment of the present application can comprise: analyzing the morphology feature of the first pulse wave signal, and determining the pulse wave category corresponding to the first pulse wave signal. The pulse wave time domain feature of the first pulse wave signal is obtained by using the pulse wave time domain feature extraction algorithm corresponding to the pulse wave category.

[0099] The pulse wave category is used to reflect the morphology feature of the pulse wave signal. The pulse wave category of the pulse wave signal can be three peaks, two peaks or less obvious pre-rebound wave.

[0100] Specifically, the embodiment of the present application can input the first pulse wave signal into the pulse wave time domain feature recognition algorithm based on the pulse wave time domain morphology classification, determine the pulse wave category corresponding to the first pulse wave signal by using the trained vector machine, and extract the pulse wave time domain feature of the first pulse wave signal according to the pulse wave time domain feature extraction algorithm corresponding to the pulse wave category.

[0101] The embodiment of the present application can increase the relevance and accuracy of the extracted pulse wave time domain feature by analyzing the morphology feature of the first pulse wave signal and determining the corresponding category, and then accurately extracting the pulse wave time domain feature by using the pulse wave time domain feature extraction algorithm designed for the category, thereby improving the effectiveness of subsequent blood pressure analysis and prediction.

[0102] S500, input the pulse wave feature parameter into the second blood pressure prediction model to obtain the blood pressure state value output by the second blood pressure prediction model, wherein the second blood pressure prediction model is a blood pressure prediction model constructed based on the pulse wave feature of the target object at the historical time.

[0103] wherein the blood pressure status value can include a systolic blood pressure status value and a diastolic blood pressure status value.

[0104] The embodiment of the present application can construct a second blood pressure prediction model according to the pulse wave characteristic parameters and individual characteristic parameters collected at historical time.

[0105] wherein the individual characteristic parameter is various indexes describing the unique physiological characteristics of the individual. In order to improve the accuracy of blood pressure prediction, considering that the differences in physiological characteristics between different individuals can have different effects on blood pressure values, it is necessary to combine the individual characteristic parameters and the pulse wave characteristic parameters to construct a blood pressure prediction model. Optionally, the individual characteristic parameters can include the age, gender, height, weight, body mass index (BMI), heart rate and tongue color of the target object. As shown in Figure 4 When considering the tongue color as a parameter, the tongue is divided into four regions: the root of the tongue, the middle of the tongue, the tip of the tongue and the edge of the tongue. The tongue color of each region can be classified into categories such as pale white, pale red, red, pale purple and purple. By measuring the average RGB value (a color representation method representing the mixing degree of red, green and blue) of the four regions, the average RGB value is taken as one of the individual characteristic parameters in the blood pressure prediction model.

[0106] In order to establish an accurate blood pressure prediction model, first, the true systolic pressure measurement value and diastolic pressure measurement value of the target object are obtained by the mercury sphygmomanometer, which are recorded in the systolic pressure vector ySBP and the diastolic pressure vector yDBP. Then, the correlation analysis is performed on the systolic pressure vector ySBP and the diastolic pressure vector yDBP, and the Pearson correlation coefficient between different characteristic parameters and the systolic pressure vector ySBP and the diastolic pressure vector yDBP is calculated. The Pearson correlation coefficient measures the degree of linear correlation between each characteristic parameter and the systolic pressure and diastolic pressure.

[0107] In the analysis process, in order to improve the effectiveness of the model, the characteristic parameters with an absolute value of the Pearson correlation coefficient less than 0.2 in the systolic pressure vector ySBP and the diastolic pressure vector yDBP can be excluded, so as to exclude these parameters with weak linear relationship with blood pressure, thereby improving the accuracy of the model.

[0108] For the systolic pressure matrix R PCAS and the diastolic pressure matrix R PCAD , the systolic pressure matrix R PCAS and the diastolic pressure matrix R PCADPrincipal Component Analysis (PCA) is performed. Principal Component Analysis is a statistical method that can convert multiple variables into a few independent variables (principal components) while retaining as much information as possible from the original data. In this process, principal components with eigenvalues greater than or equal to 1 can be extracted from each analysis, resulting in two m-dimensional principal component matrices: systolic blood pressure principal component matrix SPS and diastolic blood pressure principal component matrix SPD.

[0109] Systolic blood pressure principal component matrix SPS:

[0110]

[0111] Diastolic blood pressure principal component matrix SPD:

[0112]

[0113] The embodiments of the present application can perform multiple linear regression analysis on the systolic blood pressure principal component matrix SPS and the diastolic blood pressure principal component matrix SPD as independent variables, and the systolic blood pressure vector ySBP and the diastolic blood pressure vector yDBP as dependent variables, to obtain blood pressure regression coefficients: systolic blood pressure regression coefficient B SBP and diastolic blood pressure regression coefficient B DBP , both of which are (m+1) x 1 matrices.

[0114] Systolic blood pressure regression coefficient B SBP :

[0115]

[0116] Diastolic blood pressure regression coefficient B DBP :

[0117]

[0118] The embodiments of the present application can combine the pulse wave feature parameters and individual feature parameters of the target object into a principal component parameter vector W:

[0119] W = [w1, w2, w3,... w m ]

[0120] The embodiments of the present application are based on the principal component parameter vector and the systolic blood pressure regression coefficient B SBP and the diastolic blood pressure regression coefficient B DBP , respectively, to obtain the systolic blood pressure state value BP S and the diastolic blood pressure state value BP D .

[0121] BP S = [1, W] · B SBP

[0122] BP D =[1,W]·B DBP

[0123] Optionally, in the above Figure 1 Based on one or more embodiments corresponding, another optional embodiment provided by the embodiment of the application is that the pulse wave feature parameter is input into the second blood pressure prediction model to obtain a blood pressure state value output by the second blood pressure prediction model, which can specifically include:

[0124] The pulse wave feature parameter is input into the second blood pressure prediction model, so that the second blood pressure prediction model combines the pulse wave feature parameter with individual feature parameters of the target object to obtain a principal component parameter vector of the target object, and the blood pressure state value is obtained by using the principal component parameter vector and a blood pressure regression coefficient of the target object.

[0125] Specifically, after obtaining the pulse wave feature parameter of the first pulse wave signal of the target object at the current time, the embodiment of the application can combine the individual feature parameters of the target object collected in advance into a principal component parameter vector, and calculate the systolic pressure state value BP SBP and the diastolic pressure state value BP DBP of the target object at the current time according to the principal component vector and the systolic pressure regression coefficient B S and the diastolic pressure regression coefficient B D .

[0126] The embodiment of the application inputs the pulse wave feature parameter into the second blood pressure prediction model, integrates multi-dimensional data by combining the individual feature parameters of the target object collected in advance, effectively converts high-dimensional data into key information capable of representing the individual blood pressure condition, thereby improving the performance and practical value of the blood pressure prediction model and enhancing the accuracy of the blood pressure state value.

[0127] S600, using a preset blood pressure fusion estimation algorithm, performing blood pressure fusion processing on the blood pressure prediction value and the blood pressure state value to obtain a target blood pressure value of the target object.

[0128] The preset blood pressure fusion estimation algorithm can be a blood pressure fusion estimation algorithm based on Kalman filter.

[0129] In the blood pressure fusion estimation algorithm based on Kalman filter, it can be assumed that the blood pressure state value X k (X k ∈R n ) and the blood pressure prediction value z k (z k ∈R n ) satisfy the discrete-time process difference equation respectively:

[0130] X k = AX k-1 + Bu k + w k

[0131] z k = HX k + v k

[0132] where the process excitation noise w and the measurement noise v are mutually independent, normally distributed white noises with mean 0 and covariance Q and R respectively, i.e. P(w) ~ N(0, Q) and P(v) ~ N(0, R), A, B and H are state transition coefficient matrices, and u is an optional control input variable.

[0133] Definition is the prior state estimate of the blood pressure at the kth step given the state before the kth step.

[0134] Definition is the posterior state estimate after the blood pressure prediction at the kth step.

[0135] The prior estimation error e and the posterior estimation error e k are defined as:

[0136]

[0137] The covariance of the prior estimation error e

[0138]

[0139] The covariance of the posterior estimation error e

[0140]

[0141] The expression of the Kalman filter is:

[0142]

[0143] i.e. the prior state estimate x is a linear combination of the weighted blood pressure prediction z k and the difference between the prior state estimate x and the weighted blood pressure prediction z

[0144] The residual of the measurement process of the Kalman filter is:

[0145]

[0146] The residual reflects the degree of inconsistency between the blood pressure prediction value and the blood pressure state value. k is a gain coefficient, or a mixing factor, for the residual. K k The covariance P k .

[0147] K k One form of representation can be:

[0148]

[0149] The embodiments of the present application can substitute the blood pressure prediction value and the blood pressure state value into the above formula to solve, and the calculated posterior state estimation as the target blood pressure value of the target object.

[0150] The blood pressure monitoring method provided by the present application comprises the following steps: acquiring a first heartbeat signal and a first pulse wave signal of a target object at a current time; obtaining a first pulse transit time of the target object by using the first heartbeat signal and the first pulse wave signal; inputting the first pulse transit time into a first blood pressure prediction model to obtain a blood pressure prediction value output by the first blood pressure prediction model, wherein the first blood pressure prediction model is a blood pressure prediction model constructed based on pulse transit times of the target object at historical times, and the historical times are times before the current time; obtaining a pulse wave feature parameter of the first pulse wave signal; inputting the pulse wave feature parameter into a second blood pressure prediction model to obtain a blood pressure state value output by the second blood pressure prediction model, wherein the second blood pressure prediction model is a blood pressure prediction model constructed based on pulse wave features of the target object at the historical times; and performing blood pressure fusion processing on the blood pressure prediction value and the blood pressure state value by using a preset blood pressure fusion estimation algorithm to obtain a target blood pressure value of the target object. The present application can accurately predict the blood pressure of the target object by collecting the heartbeat signal and the pulse wave signal of the target object, and using the pulse transit time and the pulse wave feature parameter for composite blood pressure prediction and fusion, so that the continuous monitoring of the blood pressure of the target object can be realized by continuously collecting the heartbeat signal and the pulse wave signal of the target object.

[0151] Optionally, based on the method shown in FIG. 1, the embodiments of the present application provide a flowchart of a specific implementation of the blood pressure monitoring method of step S100, which can comprise the following steps: Figure 1 Figure 5

[0152] S110, collecting a first pulse wave signal of a target object.

[0153] S120, collecting a thoracic aorta pulse echo signal of the target object.

[0154] ​​Specifically, in this embodiment of the invention, the thoracic aortic pulse echo signal of the target object can be collected by bio-radar.

[0155] S130. Perform blind source separation processing on the thoracic aortic pulse echo signal to extract the first heartbeat signal of the target object from the thoracic aortic pulse echo signal.

[0156] The thoracic aortic pulse echo signal can be affected by various physiological activities. The main influencing factors include chest wall vibrations generated during cardiac systole and diastole. In addition, respiratory movements and changes in venous blood flow also produce vibrations in the chest, further affecting the pulse echo signal. To extract a complete and well-defined heartbeat signal from the aliased thoracic aortic pulse echo signal, a blind source separation algorithm based on Empirical Mode Decomposition (EMD) can be used to perform blind source separation processing on the thoracic aortic pulse echo signal, separating the heartbeat signal SCG (Seismocardiogram) from the thoracic aortic pulse echo signal.

[0157] like Figure 6 As shown in the figure, the heartbeat signal blind source separation algorithm provided in this embodiment of the invention includes a raw signal decomposition stage, a mask signal construction stage, a new signal construction stage, and a flooding signal cancellation stage.

[0158] In the original signal decomposition stage: the thoracic aortic pulse echo signal is used as the original signal x(n), and the Empirical Mode Decomposition (EMD) algorithm is used to process the original signal to decompose it into multiple intrinsic mode functions (IMFs). Among the multiple decomposed IMFs, the first IMF (IMF1) is selected, and a Hilbert transform is performed on it to calculate the instantaneous amplitude a1(n) and instantaneous frequency f1(n) of IMF1.

[0159] During the mask signal construction stage: first calculate the energy-weighted average of the instantaneous frequencies of the IMF1 component:

[0160]

[0161] Next, the mean amplitude of IMF1 was calculated:

[0162]

[0163] The constructed mask signal is:

[0164]

[0165] The mask signal is used to load onto the original signal.

[0166] In the phase of building a new signal: Building a new signal:

[0167] x + (n) = x(n) + s(n)

[0168] x - (n) = x(n) - s(n)

[0169] For x respectively + (n) and x - (n) Perform empirical mode decomposition to obtain a series of intrinsic mode components (IMFs), the i-th IMF... i The components are represented as z i+ (n) and z i- (n).

[0170] During the phase of offsetting flooding signals:

[0171] z i (n)=(z i+ (n)+z i- (n)) / 2

[0172] When processing heartbeat signals affected by noise, a mask signal is first superimposed on the original signal to be analyzed, and then empirical mode decomposition is used to decompose the composite signal. By subtracting the original signal from the first intrinsic mode function (perturbation component) obtained from the decomposition, noise can be effectively removed, thus obtaining a near-pure heartbeat signal.

[0173] This invention, through the acquisition of pulse wave signals and thoracic aortic pulse echo signals of the target object and the blind source separation processing of the thoracic aortic pulse echo signals, can effectively separate the heartbeat signal of the target object from the thoracic aortic pulse echo signals, thereby helping to reduce non-heartbeat-related interference noise and improving the clarity and quality of the heartbeat signal.

[0174] Optional, based on Figure 1 The method shown is as follows: Figure 7 The diagram shows a specific implementation of step S200 in the blood pressure monitoring method provided by this invention. Step S200 may include:

[0175] S210. Capture the optimal peak points of the first heartbeat signal and the first pulse wave signal respectively to obtain the optimal peak point of the first heartbeat signal and the optimal peak point of the first pulse wave signal.

[0176] Specifically, in this embodiment of the invention, the optimal peak point can be captured for the first heartbeat signal and the first pulse wave signal based on the optimal peak point algorithm of envelope detection.

[0177] The envelope detection is a signal processing technique for extracting the envelope of a complex modulated signal.

[0178] Optionally, as shown in FIG. 1, the envelope detection-based optimal peak point algorithm provided by the embodiment of the present application comprises three algorithm stages, i.e., envelope extraction based on Viola integral and Shannon energy, initial peak point capture based on fundamental frequency and envelope wave peak mean value, and adaptive optimal peak searching based on dynamic interval threshold segmentation. Figure 8

[0179] In the algorithm stage of envelope extraction based on Viola integral and Shannon energy, the Viola integral and the Shannon energy are used to extract the envelope of the heartbeat signal or the pulse signal. First, the Viola integral is applied to the preprocessed signal to extract the envelope, and then the Shannon energy algorithm is used to highlight the low-amplitude peak value. Compared with other envelope algorithms (for example, Hilbert envelope), the envelope obtained by the algorithm is smoother.

[0180] The embodiment of the present application can determine the time scale L T wherein L T is related to the sampling frequency of the signal. The Viola integral is used to obtain the mean value sequence of the heartbeat signal or the pulse signal, i.e., to extract the envelope E

[0181]

[0182] Then, the Viola integral feature envelope E T (m) is obtained:

[0183]

[0184] m=L T , L T +1,..., M-1-L T

[0185] The new envelope E(m) is obtained by normalization processing:

[0186]

[0187] In order to better reflect the peak value of the low-intensity amplitude, the Shannon energy is emphasized in the middle-intensity component and weakened in other intensity components, and the first-order Shannon energy is selected: when X(t)=1, logX(t)=0. In order to avoid the double-peak problem that may occur in the analysis of the signal with periodic fluctuations or multiple main frequency components, the new envelope E(m) is multiplied by an empirical parameter parameter to obtain the sequence X(t):

[0188] ​X(t) = parameter x E(m)

[0189] E(t) = -X(t) n logX(t) n

[0190] wherein the empirical parameter parameter can be 0.7.

[0191] In order to improve the smoothness of the feature sequence, the signal is then smoothed:

[0192]

[0193] In order to solve the problem that small peaks around the main peak make it difficult to detect the main peak when calculating the Shannon energy, by introducing the normalized average Shannon energy (Shannon Energy Envelope, SEE), the small peaks can be eliminated, so that the main peak is easier to detect and analyze. After normalization, we get:

[0194]

[0195] In the algorithm stage based on the fundamental frequency and envelope peak mean to capture the initial peak point, first, the envelope extracted from the signal is processed by the empirical mode decomposition method to obtain the processed signal X(n). Since the heartbeat signal or pulse wave signal has a limited cycle length in the time domain, the fundamental frequency range of the heartbeat signal or pulse wave signal is estimated according to the range of heart rate. Then, the envelope of the heartbeat signal or pulse wave signal is extracted from the fundamental frequency range by using Fast Fourier Transform (FFT), so as to determine the fundamental frequency and the initial step T initial In addition, the extracted envelope is again processed by the empirical mode decomposition method to extract its upper and lower envelopes, and the average of the two is calculated to obtain the dynamic threshold Thre. The minimum distance between two peaks is set to T min , the peak points greater than the dynamic threshold are selected, and the relatively small envelope peak values (false positive points) are removed.

[0196] After selecting the dynamic threshold based on the amplitude, further peak searching is performed by the time dynamic segmentation method. Based on the data of the last peak point, the time interval T dynamic is obtained by differentiating the peak point abscissa, and the dynamic time segmentation offset T offset is calculated. In this quasi-periodic segmentation interval, the maximum peak point is searched, which represents the true peak point of the heartbeat signal or pulse signal.

[0197] In the adaptive optimal peak seeking algorithm stage based on dynamic interval threshold segmentation, in order to deal with the problem of missing envelope peak points in the heartbeat signal or pulse wave signal due to human jitter or low-frequency noise, an algorithm needs to be used to detect these missing points. The first step of the algorithm is to calculate the average interval between the last two detected peak points Next, an upper threshold is set If the interval between two envelope peaks exceeds this upper threshold, it is considered that there is a missing point; similarly, a lower threshold is set If the interval between two envelope peaks is less than this lower threshold, it is considered that there is a redundant point. In this way, the third iteration is completed, and then the update is continued, and iteration is performed until the average interval between the peak points obtained by two consecutive iterations is the same, at which time a set of optimal envelope peak points is confirmed.

[0198] S220, using the first heartbeat optimal peak point and the first pulse wave optimal peak point, obtaining the first pulse transit time of the target object.

[0199] Specifically, the embodiment of the present application can calculate the time difference between the first heartbeat optimal peak point and the first pulse wave optimal peak point, and confirm the time difference as the first pulse transit time of the target object.

[0200] The embodiment of the present application can obtain the optimal peak points of the heartbeat signal and the pulse wave signal by capturing the optimal peak points of the heartbeat signal and the pulse wave signal. Using the optimal peak points of the two signals, the pulse transit time of the target object can be accurately calculated, thereby solving the problem of incorrect judgment of the traditional threshold peak seeking algorithm for weak non-steady-state signals such as heartbeat signals and pulse wave signals, and ensuring accurate measurement of the blood pressure of the target object.

[0201] Although the operations are described in a particular order, this should not be understood as requiring the operations to be performed in the particular order shown or in sequential order. Under certain circumstances, multitasking and parallel processing can be advantageous.

[0202] It should be understood that each of the steps described in the method embodiments of the present application can be performed in a different order and / or in parallel. In addition, the method embodiments can include additional steps and / or omit the execution of the steps shown. The scope of the present application is not limited in this respect.

[0203] Corresponding to the above-mentioned method embodiments, the embodiment of the present application also provides a blood pressure monitoring device, the structure of which is as follows Figure 9As shown, the device can include a target object signal obtaining unit 10, a pulse transit time obtaining unit 20, a blood pressure prediction value obtaining unit 30, a pulse wave characteristic parameter obtaining unit 40, a blood pressure state value obtaining unit 50, and a target blood pressure value obtaining unit 60.

[0204] The target object signal obtaining unit 10 is configured to obtain a first heartbeat signal and a first pulse wave signal of a target object at a current time.

[0205] The pulse transit time obtaining unit 20 is configured to obtain a first pulse transit time of the target object by using the first heartbeat signal and the first pulse wave signal.

[0206] The blood pressure prediction value obtaining unit 30 is configured to input the first pulse transit time into a first blood pressure prediction model to obtain a blood pressure prediction value output by the first blood pressure prediction model, wherein the first blood pressure prediction model is a blood pressure prediction model constructed based on pulse transit times of the target object at historical times, and the historical times are times before the current time.

[0207] The pulse wave characteristic parameter obtaining unit 40 is configured to obtain a pulse wave characteristic parameter of the first pulse wave signal.

[0208] The blood pressure state value obtaining unit 50 is configured to input the pulse wave characteristic parameter into a second blood pressure prediction model to obtain a blood pressure state value output by the second blood pressure prediction model, wherein the second blood pressure prediction model is a blood pressure prediction model constructed based on pulse wave characteristics of the target object at the historical times.

[0209] The target blood pressure value obtaining unit 60 is configured to perform blood pressure fusion processing on the blood pressure prediction value and the blood pressure state value by using a preset blood pressure fusion estimation algorithm to obtain a target blood pressure value of the target object.

[0210] Optionally, the target object signal obtaining unit 10 can include a first pulse wave signal obtaining subunit, a thoracic aorta pulse echo signal obtaining subunit, and a first heartbeat signal obtaining subunit.

[0211] The first pulse wave signal obtaining subunit is configured to collect a first pulse wave signal of the target object.

[0212] The thoracic aorta pulse echo signal obtaining subunit is configured to collect a thoracic aorta pulse echo signal of the target object.

[0213] The first heartbeat signal obtaining subunit is configured to perform blind source separation processing on the thoracic aorta pulse echo signal to extract a first heartbeat signal of the target object from the thoracic aorta pulse echo signal.

[0214] Optionally, the pulse transit time obtaining unit 20 can include an optimal peak point obtaining subunit and a pulse transit time obtaining subunit.

[0215] The optimal peak point obtaining subunit is configured to capture optimal peak points of the first heartbeat signal and the first pulse wave signal respectively, and obtain a first heartbeat optimal peak point of the first heartbeat signal and a first pulse wave optimal peak point of the first pulse wave signal.

[0216] The pulse transit time obtaining subunit is configured to obtain the first pulse transit time of the target object by using the first heartbeat optimal peak point and the first pulse wave optimal peak point.

[0217] Optionally, the blood pressure state value obtaining unit 50 can be specifically configured to input the pulse wave characteristic parameter into the second blood pressure prediction model, so that the second blood pressure prediction model combines the pulse wave characteristic parameter with the individual characteristic parameter of the target object to obtain a principal component parameter vector of the target object, and obtains the blood pressure state value by using the principal component parameter vector and the blood pressure regression coefficient of the target object.

[0218] Optionally, the pulse wave characteristic parameter includes a pulse wave time domain characteristic and a pulse wave frequency domain characteristic of the first pulse wave signal.

[0219] Optionally, the pulse wave time domain characteristic can include one or more of an average intensity, a pulse wave peak value, a peak area, a peak-valley difference, a descending branch peak-valley connecting line slope, and a pulse width of the pulse wave signal.

[0220] Optionally, the blood pressure monitoring device can further include a pulse wave time domain characteristic obtaining unit.

[0221] The pulse wave time domain characteristic obtaining unit is configured to analyze a shape characteristic of the first pulse wave signal, determine a pulse wave category corresponding to the first pulse wave signal, and obtain a pulse wave time domain characteristic of the first pulse wave signal by using a pulse wave time domain characteristic extraction algorithm corresponding to the pulse wave category.

[0222] Optionally, the blood pressure prediction value obtaining unit 30 is specifically configured to input the first pulse transit time into the first blood pressure prediction model, so that the first blood pressure prediction model obtains the blood pressure prediction value by using the first pulse transit time and the individual specificity parameter of the target object.

[0223] The blood pressure monitoring device provided by the application obtains a first heartbeat signal and a first pulse wave signal of a target object at a current time; uses the first heartbeat signal and the first pulse wave signal to obtain a first pulse transit time of the target object; inputs the first pulse transit time into a first blood pressure prediction model to obtain a blood pressure prediction value output by the first blood pressure prediction model, wherein the first blood pressure prediction model is a blood pressure prediction model constructed based on a pulse transit time of the target object at a historical time, and the historical time is a time before the current time; obtains a pulse wave feature parameter of the first pulse wave signal; inputs the pulse wave feature parameter into a second blood pressure prediction model to obtain a blood pressure state value output by the second blood pressure prediction model, wherein the second blood pressure prediction model is a blood pressure prediction model constructed based on a pulse wave feature of the target object at the historical time; and uses a preset blood pressure fusion estimation algorithm to perform blood pressure fusion processing on the blood pressure prediction value and the blood pressure state value to obtain a target blood pressure value of the target object. The heartbeat signal and the pulse wave signal of the target object are collected, the pulse transit time and the pulse wave feature parameter are used for composite blood pressure prediction and fusion, and the blood pressure of the target object is accurately predicted, so that the continuous monitoring of the blood pressure can be realized by continuously collecting the heartbeat signal and the pulse wave signal of the target object.

[0224] As to the device in the above-mentioned embodiments, the specific manner in which each unit performs an operation has been described in detail in the embodiments related to the method, and will not be described in detail here.

[0225] The blood pressure monitoring device includes a processor and a memory, and the above-mentioned target object signal obtaining unit 10, pulse transit time obtaining unit 20, blood pressure prediction value obtaining unit 30, pulse wave feature parameter obtaining unit 40, blood pressure state value obtaining unit 50 and target blood pressure value obtaining unit 60 are all stored in the memory as program units, and the corresponding functions are realized by the processor executing the above-mentioned program units stored in the memory.

[0226] The processor includes a core, and the core calls the corresponding program units in the memory. One or more than one core can be set, and the core parameters are adjusted to collect the heartbeat signal and the pulse wave signal of the target object, use the pulse transit time and the pulse wave feature parameter for composite blood pressure prediction and fusion, accurately predict the blood pressure of the target object, and thus realize the continuous monitoring of the blood pressure through the heartbeat signal and the pulse wave signal of the target object.

[0227] The embodiment of the application provides a computer readable storage medium, which has a program stored thereon, and the program is executed by a processor to realize the blood pressure monitoring method.

[0228] The embodiment of the application provides a processor, which is used for running a program, and the program is executed to perform the blood pressure monitoring method.

[0229] As Figure 10 shown, the blood pressure monitoring system provided by the embodiment of the application can include a biological radar 1000, a pulse bracelet 2000, a cloud server 3000 and an upper computer 4000.

[0230] The upper computer 4000 is configured to send a signal acquisition instruction to the biological radar 1000 and the pulse bracelet 2000 through the cloud server 3000, and store and display a target blood pressure value fed back by the cloud server 3000.

[0231] The biological radar 1000 is configured to continuously acquire a heartbeat signal of a target object in response to the signal acquisition instruction, and feed back the heartbeat signal to the cloud server 3000.

[0232] The pulse bracelet 2000 is configured to continuously acquire a pulse wave signal of the target object in response to the signal acquisition instruction, and feed back the pulse wave signal to the cloud server 3000.

[0233] The cloud server 3000 is configured to construct a first blood pressure prediction model and a second blood pressure prediction model of the target object based on the heartbeat signal and the pulse wave signal, and perform blood pressure fusion processing on a blood pressure prediction value output by the first blood pressure prediction model and a blood pressure state value output by the second blood pressure prediction model, and continuously output a target blood pressure value of the target object to the upper computer 4000.

[0234] The target object can lie on a bed, and wear the pulse bracelet 2000 at the radial artery of the wrist. A biological radar 1000 is placed slightly downward at an angle opposite to the heart position above the head of the target object. The control software of the upper computer 4000 is used to send a control instruction, and the instruction is transmitted to the biological radar 1000 and the pulse bracelet 2000 through the cloud server 3000. After receiving the control command, the biological radar 1000 and the pulse bracelet 2000 start to collect heartbeat and pulse data, and send the real-time collected original data to the cloud server 3000. The cloud server 3000 processes the original heartbeat and pulse data, including a non-invasive continuous blood pressure sensing algorithm, blind source separation of the heartbeat signal, optimal peak capture of the heartbeat signal and the pulse wave signal, and identification of the time domain characteristics of the pulse wave. Through multi-dimensional blood pressure modeling of the pulse wave transmission time and the time domain characteristics, the blood pressure value is predicted, and the processed heartbeat, pulse wave data and model predicted blood pressure value are transmitted to the upper computer 4000. The upper computer 4000 stores and displays the heartbeat data, pulse wave data and blood pressure value in real time, as a visual monitor, to monitor the continuous blood pressure state of the target object.

[0235] The electronic device can be a server, a PC, a PAD, a mobile phone, etc.

[0236] The application further provides a computer program product adapted to execute the program initialized with the steps of the blood pressure monitoring method when executed on the electronic device.

[0237] The application is described with reference to flowcharts and / or block diagrams of the method, device, system, electronic device and computer program product according to the embodiments of the application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a general-purpose computer, a special-purpose computer, an embedded processor or other programmable devices to produce a machine, so that the instructions executed by the computer or other programmable devices produce a machine for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 The device for implementing the functions specified in one flow or multiple flows and / or blocks. Figure 1 The device for implementing the functions specified in one flow or multiple flows and / or blocks.

[0238] In a typical configuration, the electronic device includes one or more processors (CPU), memories and buses. The electronic device can further include input / output interfaces, network interfaces, etc.

[0239] The memory can include a non-permanent memory in a computer readable medium, random access memory (RAM) and / or non-volatile memory such as read-only memory (ROM) or flash memory (flash RAM), and the memory includes at least one memory chip. The memory is an example of the computer readable medium.

[0240] Computer-readable media includes permanent and non-permanent, removable and non-removable media implemented in any method or technology for storage of information such as computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer-readable media does not include transitory media such as modulated data signals and carriers.

[0241] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0242] In the description of the present application, it should be understood that if the orientation or positional relationship indicated by the terms such as "upper", "lower", "front", "back", "left" and "right" is based on the orientation or positional relationship shown in the drawings, it is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the indicated position or element must have a particular orientation, be constructed and operated in a particular orientation, therefore it cannot be understood as a limitation of the present application.

[0243] It should be noted that in this paper, the relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. It should also be noted that the terms "include", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, commodity or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or includes elements inherent to such process, method, commodity or equipment. Without more limitation, the element defined by the sentence "including a…" does not exclude the presence of other identical elements in the process, method, commodity or equipment including the element.

[0244] Those skilled in the art will appreciate that embodiments of the present application can be devised for a method, a system or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer readable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer readable program code thereon for use by or in connection with an instruction execution system. In this particular case, computer storage media can include a computer program product that is in use when being read by a computer; therefore, such computer program product, non-transitory, also falls under the present application.

[0245] The above merely provides an embodiment of the present application and is not intended to limit the present application. The present application can have various modifications and changes for those skilled in the art. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application shall be included in the scope of claims of the present application.

Claims

1. A method of blood pressure monitoring, characterized by, The method comprises the following steps: obtaining a first heartbeat signal and a first pulse wave signal of a target object at a current time; obtaining a first pulse transit time of the target object by using the first heartbeat signal and the first pulse wave signal; inputting the first pulse transit time into a first blood pressure prediction model to obtain a blood pressure prediction value output by the first blood pressure prediction model, wherein the first blood pressure prediction model is a blood pressure prediction model constructed based on pulse transit times of the target object at historical times, and the historical times are times before the current time; obtaining a pulse wave feature parameter of the first pulse wave signal; inputting the pulse wave feature parameter into a second blood pressure prediction model to obtain a blood pressure state value output by the second blood pressure prediction model, wherein the second blood pressure prediction model is a blood pressure prediction model constructed based on pulse wave features of the target object at the historical times; performing blood pressure fusion processing on the blood pressure prediction value and the blood pressure state value by using a preset blood pressure fusion estimation algorithm to obtain a target blood pressure value of the target object.

2. The method of claim 1, wherein, The method comprises the following steps: collecting a first pulse wave signal of the target object; collecting a thoracic aorta pulse echo signal of the target object; performing blind source separation processing on the thoracic aorta pulse echo signal to extract a first heartbeat signal of the target object from the thoracic aorta pulse echo signal.

3. The method of claim 1, wherein, The method comprises the following steps: respectively performing optimal peak point capturing on the first heartbeat signal and the first pulse wave signal to obtain a first heartbeat optimal peak point of the first heartbeat signal and a first pulse wave optimal peak point of the first pulse wave signal; obtaining a first pulse transit time of the target object by using the first heartbeat optimal peak point and the first pulse wave optimal peak point.

4. The method of claim 1, wherein, The method comprises the following steps: inputting the pulse wave feature parameter into the second blood pressure prediction model to make the second blood pressure prediction model combine the pulse wave feature parameter with individual feature parameters of the target object to obtain a principal component parameter vector of the target object, and obtaining a blood pressure state value by using the principal component parameter vector and a blood pressure regression coefficient of the target object.

5. The method according to claim 1 or 4, characterized in that, The pulse wave feature parameter comprises pulse wave time domain features and pulse wave frequency domain features of the first pulse wave signal.

6. The method of claim 5, wherein, The method comprises the following steps: analyzing a shape feature of the first pulse wave signal to determine a pulse wave category corresponding to the first pulse wave signal; obtaining the pulse wave time domain features of the first pulse wave signal by using a pulse wave time domain feature extraction algorithm corresponding to the pulse wave category.

7. The method of claim 1, wherein, The method comprises the following steps: input the first pulse transit time into a first blood pressure prediction model, so that the first blood pressure prediction model obtains a blood pressure prediction value by using the first pulse transit time and individual-specific parameters of the target object.

8. A blood pressure monitoring device, characterized by Comprise: A target object signal obtaining unit, a pulse transit time obtaining unit, a blood pressure prediction value obtaining unit, a pulse wave characteristic parameter obtaining unit, a blood pressure state value obtaining unit, and a target blood pressure value obtaining unit, The target object signal obtaining unit is configured to obtain a first heartbeat signal and a first pulse wave signal of a target object at a current time. The pulse transit time obtaining unit is configured to obtain a first pulse transit time of the target object by using the first heartbeat signal and the first pulse wave signal. The blood pressure prediction value obtaining unit is configured to input the first pulse transit time into a first blood pressure prediction model to obtain a blood pressure prediction value output by the first blood pressure prediction model, wherein the first blood pressure prediction model is a blood pressure prediction model constructed based on pulse transit times of the target object at historical times, and the historical times are times before the current time. The pulse wave characteristic parameter obtaining unit is configured to obtain a pulse wave characteristic parameter of the first pulse wave signal. The blood pressure state value obtaining unit is configured to input the pulse wave characteristic parameter into a second blood pressure prediction model to obtain a blood pressure state value output by the second blood pressure prediction model, wherein the second blood pressure prediction model is a blood pressure prediction model constructed based on pulse wave characteristics of the target object at the historical times. The target blood pressure value obtaining unit is configured to perform blood pressure fusion processing on the blood pressure prediction value and the blood pressure state value by using a preset blood pressure fusion estimation algorithm to obtain a target blood pressure value of the target object.

9. A blood pressure monitoring system characterized by, Comprise: A biological radar, a pulse bracelet, a cloud server, and an upper computer, The upper computer is configured to send a signal acquisition instruction to the biological radar and the pulse bracelet through the cloud server, and store and display a target blood pressure value fed back by the cloud server. The biological radar is configured to continuously collect a heartbeat signal of a target object in response to the signal acquisition instruction, and feed back the heartbeat signal to the cloud server. The pulse bracelet is configured to continuously collect a pulse wave signal of the target object in response to the signal acquisition instruction, and feed back the pulse wave signal to the cloud server. The cloud server is configured to construct a first blood pressure prediction model and a second blood pressure prediction model of the target object based on the heartbeat signal and the pulse wave signal, and perform blood pressure fusion processing on a blood pressure prediction value output by the first blood pressure prediction model and a blood pressure state value output by the second blood pressure prediction model to continuously output the target blood pressure value of the target object to the upper computer.

10. An electronic device, comprising at least one processor, and at least one memory connected to the processor via a bus; wherein, The processor, the memory, and the bus complete communication with each other; The processor is configured to invoke program instructions in the memory to execute the blood pressure monitoring method in any one of claims 1 to 7.

Images