Cardiovascular monitoring and evaluating method and system based on multi-mode non-invasive signals

Abstract

The invention belongs to the technical field of health monitoring, and particularly discloses a cardiovascular monitoring and evaluating method and system based on multi-mode non-invasive signals, and the method comprises the following steps: collecting clinical detection data and basic information of a testee and 32 body signals related to the cardiovascular system, and carrying out the preprocessing, and obtaining initial parameters; performing feature selection and dimension reduction on the initial parameters to obtain secondary parameters; correlation analysis is conducted on the secondary parameters, correlation coefficients of the secondary parameters and cardiovascular disease results are ranked from large to small, and the secondary parameters are selected from front to back according to the total proportion of set samples to serve as tertiary parameters; screening for eliminating multivariate colinearity is carried out on the cubic parameters, and evaluation parameters are obtained; and based on a machine learning algorithm, constructing an evaluation model, inputting the evaluation parameters into the evaluation model, and outputting an evaluation result. By adopting the technical scheme, more cardiovascular information is obtained, and tracking monitoring and comprehensive evaluation of cardiovascular health of a human body are realized.

Description

Technical Field

[0001] This invention belongs to the field of health monitoring technology and relates to a cardiovascular monitoring and assessment method and system based on multimodal non-invasive signals. Background Technology

[0002] Cardiovascular diseases, as a group of diseases that seriously threaten human health, encompass a wide range, including coronary heart disease, myocardial infarction, heart failure, arrhythmia, hypertension, and stroke. These diseases have enormous and far-reaching consequences, not only severely impacting individual health and reducing quality of life, but also potentially imposing a heavy economic burden on families due to long-term medical expenses. Furthermore, they have a significant impact on the allocation of medical resources and the supply of labor in society as a whole.

[0003] In my country and many developed countries, cardiovascular disease has become the leading cause of death, the "number one killer" threatening human health. According to authoritative statistics, the number of people suffering from cardiovascular disease in China is as high as 330 million. Studies have shown that in the early stages of certain cardiovascular diseases such as hypertension and arteriosclerosis, patients often have no obvious symptoms. However, at this time, a series of cardiovascular parameters, such as blood pressure, vascular resistance, blood vessel elasticity, and blood viscosity, have already quietly changed. Through proactive interventions, such as changing unhealthy lifestyles and adjusting dietary habits, the occurrence of cardiovascular disease can be effectively prevented, and cardiovascular health can be significantly improved. Therefore, the monitoring of cardiovascular-related parameters is particularly important.

[0004] On the one hand, early detection of potential cardiovascular disease risks can buy patients valuable treatment time and prevent their condition from worsening. On the other hand, test results can provide strong evidence for cardiovascular disease monitoring, guide doctors in rational drug use, and scientifically evaluate treatment effectiveness. Therefore, early risk screening, early warning, and assessment of cardiovascular disease are undoubtedly key measures for maintaining human health.

[0005] Numerous studies and practices have demonstrated a close link between many cardiovascular parameters and cardiovascular diseases. These parameters can be effectively used for early risk screening, warning, and assessment of cardiovascular diseases. Common cardiovascular parameters include blood pressure, heart rate, vascular resistance, vascular wall elasticity, and blood viscosity. Many instruments can be used to detect these parameters, such as ultrasound, MRI, and X-ray CT, which can detect and diagnose the morphology and some functions of the heart and blood vessels; and 12-channel electrocardiographs and impedance cardiographs, which can obtain information on cardiac function. However, these instruments are complex to operate and not suitable for cardiovascular health testing, screening, long-term monitoring, or other scenarios such as in-vehicle use, home settings, private clinics, and communities.

[0006] While a few devices (such as smart bracelets and smartwatches) can be used for cardiovascular function testing or monitoring in homes, private clinics, and communities, most are simply monitors or multichannel physiological instruments used to detect, display, or record relevant signals. They can obtain fewer cardiovascular parameters, have very limited role in evaluating cardiovascular function, and their accuracy is also low. Summary of the Invention

[0007] The purpose of this invention is to provide a cardiovascular monitoring and assessment method and system based on multimodal non-invasive signals, so as to achieve more accurate comprehensive assessment and tracking of human cardiovascular health.

[0008] To achieve the above objectives, the basic solution of this invention is: a cardiovascular monitoring and assessment method based on multimodal non-invasive signals, comprising the following steps:

[0009] The basic information, clinical test results, and 32 cardiovascular-related bodily signals of the sample subjects were collected. The collection process of the 32 bodily signals was divided into: no pressure process, slow deflation or inflation process, personalized blood pressure process, and blood flow occlusion process. The basic information, clinical test results, and 32 cardiovascular-related bodily signals of the subjects were preprocessed to obtain initial parameters.

[0010] By using the minimum absolute value shrinkage and selection operator Lasso function, feature selection and dimensionality reduction are performed on the initial parameters to obtain quadratic parameters;

[0011] Correlation analysis was performed on the quadratic parameters, and the correlation coefficients between the quadratic parameters and cardiovascular disease outcomes were sorted from largest to smallest. The quadratic parameters were then selected from the beginning to the end according to the set proportion of the total sample size, and used as the cubic parameters.

[0012] The cubic parameters are screened to exclude multivariate collinearity, and the evaluation parameters are obtained.

[0013] An evaluation model is constructed based on machine learning algorithms and evaluation parameters.

[0014] The study acquires 32 cardiovascular-related bodily signals from the subject. The acquisition process of these 32 bodily signals is divided into: no-pressure process, slow deflation or inflation process, personalized blood pressure process, and blood flow occlusion process. The acquired data is input into the evaluation model, and the evaluation results are output.

[0015] The working principle and beneficial effects of this basic solution are as follows: This technical solution non-invasively and synchronously acquires 32 multimodal cardiovascular-related signals, uses a specific parameter extraction method to obtain cardiovascular-related characteristic parameters, and performs tracking monitoring and comprehensive assessment of cardiovascular health status based on the assessment model. This achieves efficient, high-frequency, low-cost, convenient, and non-invasive detection of human cardiovascular health, improves assessment accuracy, and provides strong protection for human health.

[0016] Furthermore, based on the sub-process, 32 cardiovascular-related bodily signals were collected from the subjects and test subjects, specifically:

[0017] Air bands were placed on both wrists or arms and ankles of the subject and the test subject.

[0018] Pressureless process: Without applying air pressure to all air zones, collect three standard electrocardiogram signals, a single heart sound and two lung sound signals for a period of time, as well as dual-wavelength signals of blood oxygen saturation from both earlobes, temples, a single finger and a single toe. The acquisition frequency of each channel can be set individually.

[0019] Slow deflation or inflation process: Simultaneously inflate the air belts at both arms or wrists and ankles to the set pressure value, and then slowly deflate them at a constant deflation rate;

[0020] Alternatively, each air bag can be slowly inflated until it exceeds the systolic pressure of the subject at the corresponding position by 20 mmHg, and then quickly deflated.

[0021] Obtain blood pressure values ​​at both arms or wrists and ankles, including systolic blood pressure, diastolic blood pressure, and mean blood pressure;

[0022] Personalized blood pressure sub-process: While applying personalized average blood pressure to corresponding sites on both arms or wrists and ankles with air cuffs, 32 signals are collected simultaneously within a first preset time period.

[0023] Blood flow occlusion process: When applying personalized maximum blood pressure to the corresponding parts of both arms or wrists and ankles, 32 signals are simultaneously collected within a second preset time period to obtain the changes in dual-wavelength signals of blood oxygen saturation in the fingers and toes under the condition of basic blood flow occlusion.

[0024] Based on the sub-process, 32 cardiovascular-related bodily signals were collected from the subjects and test subjects. The collected signals are richer and more beneficial for subsequent use.

[0025] Furthermore, the 32 cardiovascular-related bodily signals of the subjects and test subjects were preprocessed to obtain initial parameters. The steps were as follows:

[0026] The ECG signal feature points of 32 cardiovascular-related body signals of subjects and test subjects and their positions in the ECG signal data sequence were extracted. Based on the position of the ECG signal feature points, the feature parameters of the ECG signal were extracted and their mean and standard deviation were calculated.

[0027] By utilizing the relationship between electrocardiogram (ECG) signals and other signals, feature points and feature parameters of other signals are extracted, and the mean and standard deviation of the feature parameters of other signals are calculated.

[0028] Calculate the ratio of the same feature parameters obtained in different sub-processes, and extract the ratio of feature parameters of the left and right symmetrical parts of the subject and test subject in the same sub-process.

[0029] Based on the pressure and pressure pulse signals in 32 body signals, the systolic pressure, diastolic pressure and mean blood pressure of the corresponding parts were extracted using the normalized oscilloscope method.

[0030] Calculate the dual-wavelength correlation parameters of blood oxygen saturation at the fingers and toes during the blood flow occlusion subprocess.

[0031] The study preprocessed 32 cardiovascular-related bodily signals of subjects and test subjects to obtain assessment parameters, streamlined the data, and made it easier to use.

[0032] Furthermore, the method for extracting ECG signal feature points from 32 cardiovascular-related bodily signals of subjects and test subjects, along with their positions in the ECG signal data sequence, and then extracting feature parameters of the ECG signals based on the positions of these feature points and calculating their mean and standard deviation is as follows:

[0033] By utilizing the characteristics of the electrocardiogram (ECG) signal itself, we can extract the feature points in the ECG signal and their positions in the ECG signal data sequence.

[0034] Based on the location of each characteristic point of the ECG signal, various characteristic parameters of the ECG signal are extracted, including the RR interval, the time between QRS complexes, and the amplitudes of QR waves, SR waves, ST waves, and SP waves and their ratios. The characteristic parameter values ​​are sorted according to the acquisition sub-process, and the mean and standard deviation of each parameter are calculated from the middle percentage of the data.

[0035] We extract the ECG signal feature points of 32 cardiovascular-related bodily signals from subjects and test subjects and their positions in the ECG signal data sequence. Based on the positions of the ECG signal feature points, we extract the feature parameters of the ECG signals and calculate their mean and standard deviation to obtain the required data features for evaluation and analysis.

[0036] Furthermore, other signal characteristic points include the maximum value point, minimum value point, maximum rate of increase point, and maximum rate of decrease point in each cardiac cycle;

[0037] Other characteristic parameters of the signals include the difference between the maximum and minimum values ​​within a cardiac cycle, the maximum rate of ascent, the maximum rate of descent, the ratio of the maximum rate of ascent to the maximum rate of descent, the time taken for the ECG R wave to reach the maximum rate of ascent and the maximum rate of descent and their ratio, the ratio of the time taken for the ECG R wave to reach its maximum value to the cardiac cycle, and the ratio of the mean, maximum rate of ascent, and maximum rate of descent of each signal within a cardiac cycle to the difference between the maximum and minimum values, respectively.

[0038] The characteristic parameter values ​​of other signals are sorted by size according to the acquisition sub-process, and the mean and standard deviation of the characteristic parameters of other signals in each sub-process are calculated by taking the middle percentage of the data.

[0039] Obtain the characteristic parameters of the signals required in each sub-process for analysis.

[0040] Furthermore, based on the pressure and pressure pulse signals from 32 body signals, the normalized oscilloscope method was used to extract the systolic blood pressure, diastolic blood pressure, and mean blood pressure at the corresponding sites, as follows:

[0041] Using dual-wavelength signals of blood oxygen saturation from the forehead, bilateral earlobes, temples, a single finger, and a single toe, and based on the blood oxygen saturation obtained from a blood oxygen simulator, the average blood oxygen saturation values ​​at various points during the pressure-free process and the personalized average blood pressure process are calculated, along with the ratio between the two processes. Furthermore, the ratios of the average blood pressure saturation values ​​at the forehead, bilateral earlobes, temples, and a single toe during the pressure-free process and the personalized average blood pressure process to the blood oxygen saturation of a single finger on the same side during the same process are calculated, as are the ratios of the average blood oxygen saturation values ​​on the left and right sides of the pressure-free process and the personalized average blood pressure process at the left and right earlobes, temples, a single toe, and a single finger.

[0042]

[0043] Wherein, SpO2 represents blood oxygen saturation, R represents the mean of the ratio of AC to DC of the 660nm photoelectric signal and the ratio of AC to DC of the 940nm photoelectric signal in the same cardiac cycle in the blood oxygen saturation dual-wavelength signal in the corresponding sub-process, and a, b, and c represent the fitting coefficients obtained by the blood oxygen simulator.

[0044] Based on the pressure and pressure pulse signals from 32 body signals, the systolic pressure, diastolic pressure and mean blood pressure of the corresponding sites were extracted using the normalized oscillometric method.

[0045] Furthermore, the dual-wavelength correlation parameters of blood oxygen saturation at the fingers and toes during the blood flow occlusion sub-process were calculated as follows:

[0046] D ikλ =±(Max ikλ -Minikλ )

[0047] DR ik =

[0048] DRR k =

[0049] DRR i =

[0050] Among them, Max ikλ Min ikλ Let represent the maximum and minimum values ​​of the photoelectric signal with wavelength λ at point k on side i of the blood flow occlusion subprocess, respectively. The difference is positive when the maximum value is earlier and negative when the maximum value is later, used to distinguish the trend of change; D ikλ This represents the difference between the maximum and minimum values ​​of the photoelectric signal with wavelength λ at point k on side i of the blood flow blocking subprocess;

[0051] DR ik D represents the ratio of the difference between the maximum and minimum values ​​of the photoelectric signal at wavelength 660nm to the difference between the maximum and minimum values ​​of the photoelectric signal at wavelength 940nm at site k on side i of the blood flow occlusion subprocess; ik940 and D ik660 These represent the differences between the maximum and minimum values ​​of the photoelectric signals at wavelengths of 940 nm and 660 nm at location k on side i of the blood flow occlusion subprocess, respectively.

[0052] DRR k This indicates the DR at k points on the left and right sides of the blood flow blocking subprocess. ik The ratio; This represents the ratio of the maximum and minimum difference of the 660nm wavelength photoelectric signal at the left k site of the blood flow blocking subprocess to the maximum and minimum difference of the 940nm wavelength photoelectric signal. This represents the ratio of the maximum and minimum difference between the 660nm wavelength photoelectric signal and the maximum and minimum difference between the 940nm wavelength photoelectric signal at the k site on the right side of the blood flow blocking subprocess.

[0053] DRR i This indicates the blood flow blocking subprocess at the i-th toe and finger, on both sides of the DR. ik The ratio; This represents the ratio of the maximum and minimum difference of the 660nm wavelength photoelectric signal at the toe location on the i-th side of the blood flow blocking subprocess to the maximum and minimum difference of the 940nm wavelength photoelectric signal. This represents the ratio of the maximum and minimum difference between the 660nm wavelength photoelectric signal and the 940nm wavelength photoelectric signal at the finger site on the i-th side of the blood flow blocking subprocess.

[0054] The blood flow blocking subprocess uses dual-wavelength correlation parameters of blood oxygen saturation at the fingers and toes, which is beneficial for use.

[0055] Furthermore, the steps for screening the cubic parameters to exclude multivariate collinearity are as follows:

[0056] The cubic parameters were sorted from largest to smallest according to their correlation coefficients, and the variance inflation factor collinearity analysis was performed using the forward selection method, specifically as follows:

[0057] If the maximum variance inflation factor (VIF) exceeds the threshold, delete the cubic parameter corresponding to the maximum VIF value, add the next cubic parameter for collinearity analysis, and continue until the required number of feature parameters are selected.

[0058] If collinearity cannot select the required number of cubic parameters, then the remaining cubic parameters after VIF multicollinearity processing are used as the final evaluation parameters.

[0059] Enables parameter selection and simplifies data.

[0060] The present invention also provides a cardiovascular monitoring and assessment system based on the method described in the present invention, including a non-invasive signal acquisition terminal, a cloud server, and a smart terminal;

[0061] The non-invasive signal acquisition terminal includes a sensor module, a signal processing module, and a microprocessor module. The sensor module is used to acquire body temperature, ambient temperature, and 32 cardiovascular-related non-invasive signals. The input end of the signal processing module is connected to the output end of the sensor module, and the output end of the signal processing module is connected to the microprocessor. The microprocessor controls the signal processing module to communicate with the cloud server and smart terminal.

[0062] The cloud server is bidirectionally connected to both the non-invasive signal acquisition terminal and the smart terminal, and the non-invasive signal acquisition terminal is bidirectionally connected to the smart terminal.

[0063] This system measures body temperature and ambient temperature, and simultaneously acquires 32 non-invasive cardiovascular-related signals. It enables accurate prediction of some blood parameters that can only be obtained invasively (including whole blood high, low, and medium shear rates, mean corpuscular hemoglobin content, mean corpuscular hemoglobin concentration, red blood cell distribution width standard deviation, mean corpuscular volume, etc.), as well as some results from echocardiography (such as ejection fraction, valvular regurgitation, and reduced left ventricular diastolic function), and some results from carotid artery echocardiography (such as abnormalities in the carotid and vertebral arteries), thus achieving the predictive assessment of cardiovascular diseases.

[0064] Furthermore, the sensor module includes a human body temperature sensor, an ambient temperature sensor, an electrocardiogram electrode, a pressure sensor, a sound sensor, and a photoelectric sensor;

[0065] The human body temperature sensor is used to collect the body temperature of the subject and the person to be tested.

[0066] The ambient temperature sensor is mounted on the outer casing of the non-invasive signal acquisition terminal;

[0067] The ECG electrodes can be attached to the left sternal border, right sternal border, lower left abdomen, and right leg of the passenger.

[0068] The pressure sensor is mounted on the air belt and can come into contact with the occupant's body. The air belt is positioned at the occupant's wrist or arm and ankle on the vehicle seat and can be connected to the occupant's wrist and ankle. The pressure sensor is used to collect pressure and pressure pulse signals.

[0069] The sound sensor is used to collect lung sound signals and heart sound signals from the passenger;

[0070] The photoelectric sensor is integrated into the blood oxygen probe, which has at least nine sensors corresponding to the forehead area, left and right temple areas, left and right earlobes, left and right fingers, and left and right toes.

[0071] In addition to collecting body temperature and ambient temperature, the sensor module also simultaneously collects 32 multimodal cardiovascular-related signals throughout the signal acquisition process, namely standard three-channel electrocardiogram, single-channel heart sound, double-channel lung sound, and nine channels (forehead area, left and right temple areas, left and right earlobes, left and right fingers, and left and right toes) blood oxygen saturation dual-wavelength signals (i.e., 18 photoelectric signals). The air belt is used to obtain pressure and pressure pulse at the left and right arms (or wrists) and ankles (i.e., 8 signals). Attached Figure Description

[0072] Figure 1 This is a flowchart illustrating the cardiovascular monitoring and assessment method based on multimodal non-invasive signals of the present invention.

[0073] Figure 2 This is a schematic diagram of the blood oxygen saturation dual-wavelength correlation parameter waveform of the blood flow blocking subprocess in the cardiovascular monitoring and assessment method based on multimodal non-invasive signals of this invention. Detailed Implementation

[0074] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0075] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0076] In the description of this invention, unless otherwise specified and limited, it should be noted that the terms "installation", "connection" and "linking" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two components. They can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0077] This invention discloses a cardiovascular monitoring and assessment method based on multimodal non-invasive signals, which can realize human body temperature detection and simultaneous acquisition of 32 multimodal non-invasive human signals, enabling tracking, monitoring, and comprehensive assessment of human cardiovascular health. Figure 1 As shown, the cardiovascular monitoring and assessment method based on multimodal noninvasive signals includes the following steps:

[0078] Clinical test results (including clinical biochemical data, imaging data, and various cardiovascular assessment results, including electrocardiogram), basic information (including height, weight, gender, age, distance between wrist and fingers, distance between ankle and toes, distance between arm and wrist, etc.) and 32 cardiovascular-related body signals were collected from the subjects. The cardiovascular information collection process was divided into: no-pressure process, slow depressurization or inflation process, personalized blood pressure process, and blood flow occlusion process. Based on each sub-process, 32 cardiovascular-related body signals (i.e., standard three-channel electrocardiogram, single-channel heart sound, double-channel lung sound, 9 channels (forehead area, left and right temple areas, left and right earlobes, left and right fingers, left and right toes) blood oxygen saturation dual-wavelength signals (i.e., 18 photoelectric signals) were collected from the subjects. Pressure and pressure pulse at the left and right arms (or wrists) and ankles (i.e., 8 signals) were obtained using an air belt. The basic information, clinical test results, and 32 cardiovascular-related body signals of the subjects were preprocessed to obtain initial parameters.

[0079] The Lasso function (Least Absolute Shrinkage and Selection Operator) is used to perform feature selection and dimensionality reduction on the initial parameters to obtain quadratic parameters. By adjusting the regularization parameter α in the Lasso function, the selection of feature parameters is achieved.

[0080] Correlation analysis was performed on the quadratic parameters, and the correlation coefficients between the quadratic parameters and cardiovascular disease outcomes were sorted from largest to smallest. The quadratic parameters were then selected from the beginning to the end according to the set proportion of the total sample size, and used as the cubic parameters.

[0081] The cubic parameters are screened to exclude multivariate collinearity, and the evaluation parameters are obtained.

[0082] For prediction problems (i.e. regression problems) where the output is a continuously changing quantity, an evaluation model is constructed using evaluation parameters based on machine learning algorithms (such as multiple linear regression, support vector regression, decision tree regression, random forest regression, neural network regression, gradient boosting regression tree, etc.).

[0083] Thirty-two cardiovascular-related bodily signals were acquired from the subjects. The acquisition process of these 32 bodily signals was divided into: no-pressure process, slow depressurization or inflation process, personalized blood pressure process, and blood flow occlusion process. The acquired data were input into the evaluation model, and the cardiovascular-related evaluation results of the subjects were output (such as the probability of carotid artery abnormality, the probability of heart valve abnormality, the probability of reduced left ventricular diastolic function, ejection fraction, vascular elasticity, etc.).

[0084] It can clearly identify the presence of premature beats, arrhythmia, atrial fibrillation, and abnormalities in the blood vessels of the upper and lower limbs, as well as accurately measure heart rate, blood pressure in the left and right wrists or arms and ankles, blood oxygen saturation in the fingers and toes, and use non-invasive methods to accurately predict blood parameters that can only be obtained invasively (including high, low, and medium shear rates of whole blood, mean corpuscular hemoglobin content, mean corpuscular hemoglobin concentration, standard deviation of red blood cell distribution width, mean corpuscular volume, etc.). Combined with cloud storage and cloud servers, it can conveniently achieve non-invasive tracking, monitoring, and comprehensive assessment of human cardiovascular health.

[0085] This invention features low cost, high efficiency, non-invasiveness, harmlessness, and comprehensiveness in monitoring and assessing human cardiovascular health. It can be widely used in hospitals, communities, private clinics, nursing homes, and even constitutes a vehicle-mounted personal cardiovascular health tracking, monitoring, and comprehensive assessment system.

[0086] In a preferred embodiment of the present invention, 32 cardiovascular-related bodily signals of the subject and test subject are collected based on a subprocess, specifically as follows:

[0087] Air bands were placed on both wrists or arms and ankles of the subject and the test subject.

[0088] 1. Air-belt pressure-free process: Without applying air pressure to all air belts, collect three standard electrocardiogram signals, one channel heart sound and two channels lung sound signals for a period of time (this time can be set by the host computer, the default is 20 seconds, used for comparison with pressurization to obtain more cardiovascular information), as well as blood oxygen saturation dual-wavelength signals from both earlobes, temple area, single finger (such as index finger, middle finger or ring finger) and single toe (such as big toe, middle toe). The acquisition frequency of each channel can be set individually.

[0089] 2. Slow deflation or inflation process: Simultaneously inflate the air belts at both arms or wrists and ankles to the set pressure value (20 mmHg greater than the corresponding systolic pressure of the subject), and then slowly deflate at a constant deflation rate (this value can be set, the default is 5 mmHg / s).

[0090] Alternatively, each air bag can be slowly inflated simultaneously (this value can be set, the default is 5 mmHg / s) until it exceeds the systolic pressure of the subject at the corresponding position by 20 mmHg, and then the air bag can be quickly deflated.

[0091] Obtain blood pressure values ​​at both arms or wrists and ankles, including systolic blood pressure, diastolic blood pressure, and mean blood pressure;

[0092] 3. Personalized Blood Pressure Sub-process: When applying a cuff to the corresponding sites on both arms or wrists and ankles, the personalized average blood pressure (referring to the average blood pressure of each site, which may vary from person to person and has personalized characteristics, which can be obtained through sub-process 2) is simultaneously collected for 32 signals within a first preset time period; the collection period can be set by the host computer, with a default value of 20 seconds, and is used to compare with the pressure without a cuff to obtain more cardiovascular information.

[0093] 4. Blood Flow Interruption Process: While applying personalized maximum blood pressure (greater than the corresponding systolic blood pressure by 20 mmHg to essentially block blood flow) to the corresponding sites on both arms or wrists and ankles via air bags, 32 signals are simultaneously collected within a second preset time period to obtain changes in dual-wavelength signals of blood oxygen saturation in the fingers and toes under conditions of basic blood flow interruption. This second preset time period can be set via a host computer, with a default value of 40 seconds. It is used to obtain changes in dual-wavelength signals of blood oxygen saturation in the fingers and toes under conditions of basic blood flow interruption, thereby extracting information on the metabolism and oxygen consumption of the fingers and toes, and comparing it with the previous data to extract more cardiovascular information.

[0094] By comparing and analyzing signals acquired from symmetrical sites on both sides, the ratio and difference of the same parameters on the left and right sides, as well as the ratio and difference of the same parameters from the same sites acquired in different acquisition sub-processes, can be extracted. This can effectively assess cardiovascular health status, especially in detecting unilateral abnormalities and assessing the degree of abnormality.

[0095] In a preferred embodiment of the present invention, the step of preprocessing 32 cardiovascular-related bodily signals of the subject and test subject to obtain initial parameters is as follows:

[0096] Utilizing the inherent characteristics of the electrocardiogram (ECG) signal (such as the large rate of change of the QR wave, the R wave being extracted using the differential thresholding method, and the position and amplitude relationship between the P, Q, R, S, and T waves and the R wave), ECG signal feature points (such as the P, Q, R, S, and T waves in the ECG signal) and their positions in the ECG signal data sequence of the subjects and test subjects are extracted from 32 cardiovascular-related body signals. Based on the position of the ECG signal feature points, ECG signal feature parameters (such as the RR interval, the time between QRS complexes, and the amplitudes of the QR wave, SR wave, ST wave, and SP wave and their ratios) are extracted. The feature parameter values ​​are sorted according to the acquisition sub-process, and the mean and standard deviation of each parameter are calculated from the middle given percentages (such as 50%, 75%, and 80%) of the data throughout the acquisition process.

[0097] By utilizing the relationship between ECG signals and other signals, feature points (such as the maximum value, minimum value, maximum rate of rise, and maximum rate of fall in each cardiac cycle) and feature parameters (such as the difference between the maximum and minimum values ​​within a cardiac cycle, the maximum rate of rise, the maximum rate of fall, the ratio of the maximum rate of rise to the maximum rate of fall, the time taken for the ECG R wave to reach the maximum rate of rise and the maximum rate of fall, and their ratio, the ratio of the time taken for the ECG R wave to reach the maximum value to the cardiac cycle, and the ratio of the mean, maximum rate of rise, and maximum rate of fall of each signal to the difference between the maximum and minimum values ​​within a cardiac cycle, etc.) are extracted. The mean and standard deviation of the feature parameters of other signals are calculated (the feature parameter values ​​are sorted according to the size of the acquisition sub-process, and the mean and standard deviation of each parameter in each sub-process are calculated from the middle given percentage (such as 50%, 75%, 80%)).

[0098] Calculate the ratio of the same characteristic parameters obtained in different sub-processes, and extract the ratio of characteristic parameters of the left and right symmetrical parts of the subject and the test subject in the same sub-process (such as the ratio of the average maximum rise rate of the 940nm photoelectric signal of the left finger in the personalized average blood pressure sub-process to the average maximum rise rate of the air belt without pressure sub-process).

[0099] Based on the pressure and pressure pulse signals in 32 body signals, the systolic pressure, diastolic pressure and mean blood pressure of the corresponding parts were extracted using the normalized oscilloscope method.

[0100] Calculate the dual-wavelength correlation parameters of blood oxygen saturation at the fingers and toes during the blood flow occlusion subprocess.

[0101] To improve the accuracy of the prediction model in this patent, incomplete data records are directly deleted. The interquartile range (IQR) criterion and the deviation principle of the three means are used to judge outliers, and data records judged as outliers are also deleted.

[0102] In a preferred embodiment of the present invention, the method for extracting ECG signal feature points of 32 cardiovascular-related body signals from subjects and test subjects and their positions in the ECG signal data sequence, and extracting feature parameters of the ECG signal based on the positions of the ECG signal feature points and calculating their mean and standard deviation is as follows:

[0103] By utilizing the characteristics of the electrocardiogram (ECG) signal itself, we can extract the feature points in the ECG signal and their positions in the ECG signal data sequence.

[0104] Based on the location of each characteristic point of the ECG signal, various characteristic parameters of the ECG signal are extracted, including the RR interval, the time between QRS complexes, and the amplitudes of QR waves, SR waves, ST waves, and SP waves and their ratios. The characteristic parameter values ​​are sorted according to the acquisition sub-process, and the mean and standard deviation of each parameter are calculated from the middle percentage of the data.

[0105] In a preferred embodiment of the present invention, the feature points of other signals include the maximum value point, minimum value point, maximum rate of increase point, and maximum rate of decrease point in each cardiac cycle;

[0106] Other characteristic parameters of the signals include the difference between the maximum and minimum values ​​within a cardiac cycle, the maximum rate of ascent, the maximum rate of descent, the ratio of the maximum rate of ascent to the maximum rate of descent, the time taken for the ECG R wave to reach the maximum rate of ascent and the maximum rate of descent and their ratio, the ratio of the time taken for the ECG R wave to reach its maximum value to the cardiac cycle, and the ratio of the mean, maximum rate of ascent, and maximum rate of descent of each signal within a cardiac cycle to the difference between the maximum and minimum values, respectively.

[0107] The characteristic parameter values ​​of other signals are sorted by size according to the acquisition sub-process, and the mean and standard deviation of the characteristic parameters of other signals in each sub-process are calculated by taking the middle percentage of the data.

[0108] In a preferred embodiment of the present invention, the method for extracting systolic blood pressure, diastolic blood pressure, and mean blood pressure at corresponding sites using normalized oscillometric method based on pressure and pressure pulse signals from 32 body signals is as follows:

[0109] Using dual-wavelength signals of blood oxygen saturation from the forehead, bilateral earlobes, temples, a single finger, and a single toe, and based on the blood oxygen saturation obtained from a blood oxygen simulator, the average blood oxygen saturation values ​​at various points during the pressure-free process and the personalized average blood pressure process are calculated, along with the ratio between the two processes. Furthermore, the ratios of the average blood pressure saturation values ​​at the forehead, bilateral earlobes, temples, and a single toe during the pressure-free process and the personalized average blood pressure process to the blood oxygen saturation of a single finger on the same side during the same process are calculated, as are the ratios of the average blood oxygen saturation values ​​on the left and right sides of the pressure-free process and the personalized average blood pressure process at the left and right earlobes, temples, a single toe, and a single finger.

[0110]

[0111] Wherein, SpO2 represents blood oxygen saturation, R represents the mean of the ratio of AC to DC of the 660nm photoelectric signal and the ratio of AC to DC of the 940nm photoelectric signal in the same cardiac cycle in the blood oxygen saturation dual-wavelength signal in the corresponding sub-process, and a, b, and c represent the fitting coefficients obtained by the blood oxygen simulator.

[0112] In a preferred embodiment of the present invention, such as Figure 2 As shown, Figure 2 The waveforms in (a) represent the pressure and pressure pulse signal waveforms at the arm (or wrist) or ankle during the blood flow occlusion subprocess, respectively. Figure 2 (b) and Figure 2 The two waveforms in (c) represent the photoelectric signal waveforms obtained when the blood oxygen saturation at the finger or toe is irradiated with light at two wavelengths of 940nm and 660nm in the dual-wavelength signal of blood oxygen saturation at the finger or toe during the blood flow blocking subprocess. Figure 2 (b) and Figure 2 (c) represents two typical different trends of change, in order to better describe the characteristics of their changes.

[0113] The dual-wavelength correlation parameters of blood oxygen saturation at the fingers and toes during the blood flow occlusion subprocess are calculated as follows:

[0114] D ikλ =±(Max ikλ -Min ikλ )

[0115] DR ik =

[0116] DRR k =

[0117] DRR i =

[0118] Among them, Maxikλ Min ikλ Let represent the maximum and minimum values ​​of the photoelectric signal with wavelength λ at point k on side i of the blood flow occlusion subprocess, respectively. The difference is positive when the maximum value is earlier and negative when the maximum value is later, used to distinguish the trend of change; D ikλ This represents the difference between the maximum and minimum values ​​of the photoelectric signal with wavelength λ at point k on side i of the blood flow occlusion subprocess; such as... Figure 2 (b) indicates D at point k on side i. ik940 and D ik660 The differences are all positive, and Figure 2 (c) represents D at point k on side i. ik940 and D ik660 The differences are all negative.

[0119] DR ik D represents the ratio of the difference between the maximum and minimum values ​​of the photoelectric signal at wavelength 660nm to the difference between the maximum and minimum values ​​of the photoelectric signal at wavelength 940nm at site k on side i of the blood flow occlusion subprocess; ik940 and D ik660 These represent the differences between the maximum and minimum values ​​of the photoelectric signals at wavelengths of 940 nm and 660 nm at location k on side i of the blood flow occlusion subprocess, respectively.

[0120] DRR k This indicates the DR at k points on the left and right sides of the blood flow blocking subprocess. ik The ratio; This represents the ratio of the maximum and minimum difference of the 660nm wavelength photoelectric signal at the left k site of the blood flow blocking subprocess to the maximum and minimum difference of the 940nm wavelength photoelectric signal. This represents the ratio of the maximum and minimum difference between the 660nm wavelength photoelectric signal and the maximum and minimum difference between the 940nm wavelength photoelectric signal at the k site on the right side of the blood flow blocking subprocess.

[0121] DRR i This indicates the blood flow blocking subprocess at the i-th toe and finger, on both sides of the DR. ik The ratio; This represents the ratio of the maximum and minimum difference of the 660nm wavelength photoelectric signal at the toe location on the i-th side of the blood flow blocking subprocess to the maximum and minimum difference of the 940nm wavelength photoelectric signal. This represents the ratio of the maximum and minimum difference between the 660nm wavelength photoelectric signal and the 940nm wavelength photoelectric signal at the finger site on the i-th side of the blood flow blocking subprocess.

[0122] In a preferred embodiment of the present invention, the steps for screening the cubic parameters to exclude multivariate collinearity and obtaining the evaluation parameters are as follows:

[0123] The cubic parameters were sorted from largest to smallest according to their correlation coefficients, and the variance inflation factor collinearity analysis was performed using the forward selection method, specifically as follows:

[0124] If the maximum variance inflation factor (VIF) exceeds the threshold, delete the cubic parameter corresponding to the maximum VIF value, add the next cubic parameter for collinearity analysis, and continue until the required number of feature parameters are selected.

[0125] If collinearity cannot select the required number of cubic parameters, then the remaining cubic parameters after VIF multicollinearity processing are used as the final evaluation parameters.

[0126] The parameters extracted by the parameter extraction method proposed in this patent, and the model established by the machine learning algorithm, can be used non-invasively to accurately predict some blood indicators that can only be obtained invasively (including whole blood high, low, and medium shear, mean corpuscular hemoglobin content, mean corpuscular hemoglobin concentration, red blood cell distribution width standard deviation, mean corpuscular volume, etc.), as well as some results of cardiac ultrasound detection (such as ejection fraction, valvular regurgitation, reduced left ventricular diastolic function, etc.) and some results of carotid ultrasound detection (such as abnormalities of the carotid and vertebral arteries), with a percentage error expected to be less than 10%. (5) This patented system can also realize the prediction and assessment of cardiovascular diseases (such as coronary heart disease, carotid artery abnormalities, arteriosclerosis, premature beats, abnormal valvular regurgitation, etc.).

[0127] The present invention also provides a cardiovascular monitoring and assessment system based on the method described in the present invention, including a non-invasive signal acquisition terminal, a cloud server, and a smart terminal.

[0128] The non-invasive signal acquisition terminal includes a sensor module (which uses temperature sensors, blood oxygen probes, pressure sensors, sound sensors, etc. to obtain corresponding signals from the human body), a signal processing module, and a microprocessor module. The sensor module is used to collect body temperature, ambient temperature, and 32 cardiovascular-related non-invasive signals. The input end of the signal processing module is electrically connected to the output end of the sensor module, and the output end of the signal processing module is connected to the microprocessor. The microprocessor controls the signal processing module to communicate with the cloud server and smart terminal.

[0129] Preferably, the microprocessor is mainly composed of a high-performance microprocessor (such as the GD32F470VG) to realize various control and data transmission. The GD32F470VG microprocessor can be selected. It is a 32-bit microprocessor from GigaDevice, with a maximum operating frequency of 240MHz. It has 768KB Flash program memory, 512KB static data storage SRAM, 14 timers, 8 serial ports, 3 I2C communication ports, and 5 SPI communication ports, which is sufficient to meet the requirements of this patented system.

[0130] The cloud server is bidirectionally electrically connected to both the non-invasive signal acquisition terminal and the smart terminal, and the non-invasive signal acquisition terminal is bidirectionally electrically connected to the smart terminal. The cloud server can utilize cloud servers and cloud databases provided by vendors such as Huawei, Alibaba, and Tencent.

[0131] This invention enables the tracking and monitoring (including remote tracking and detection) and comprehensive assessment of human cardiovascular health. It can effectively conduct early risk screening and early warning of cardiovascular diseases, as well as follow-up monitoring, medication guidance and efficacy evaluation of cardiovascular diseases.

[0132] This invention can not only form in-vehicle intelligent systems (such as in-vehicle intelligent seat systems), but also intelligent cardiovascular health tracking, monitoring and comprehensive assessment systems (such as intelligent seat systems and intelligent testing beds) for use in homes, communities, private clinics, hospitals and other occasions, and has very broad application prospects and market value.

[0133] Preferably, a personal identification module can also be set up, which can be implemented using technologies such as fingerprint recognition, facial recognition or voice recognition. It is mainly used for automatic identification of personal identity, and on the one hand, it can automatically retrieve basic personal information (including name, gender, height, weight, age, arm-finger distance, etc.) for software analysis.

[0134] In a preferred embodiment of the present invention, the sensor module includes a human body temperature sensor, an ambient temperature sensor, an electrocardiogram electrode, a pressure sensor, a sound sensor, and a photoelectric sensor.

[0135] Human body temperature sensors (such as infrared temperature sensors (GD609141 with an accuracy of 0.1℃)) are used to collect the body temperature of subjects and those to be tested, while ambient temperature sensors (such as thermistors) are installed on the outer shell of the non-invasive signal acquisition terminal.

[0136] ECG electrodes can be attached to the left sternal border, right sternal border, lower left abdomen, and right leg of the passenger.

[0137] The pressure sensor is mounted on the air belt and can come into contact with the occupant's body. The air belt is positioned on the vehicle seat corresponding to the occupant's wrist or arm and ankle, and can be connected to the occupant's wrist and ankle. The pressure sensor is used to collect pressure and pressure pulse signals.

[0138] Sound sensors are used to collect lung sound and heart sound signals from passengers.

[0139] The photoelectric sensor is integrated into the blood oxygen probe, which has at least nine sensors corresponding to the forehead area, left and right temple areas, left and right earlobes, left and right fingers, and left and right toes.

[0140] In addition to collecting body temperature and ambient temperature, the sensor module also simultaneously collects 32 multimodal cardiovascular-related signals throughout the signal acquisition process, namely standard three-channel electrocardiogram, single-channel heart sound, double-channel lung sound, and nine channels (forehead area, left and right temple areas, left and right earlobes, left and right fingers, and left and right toes) blood oxygen saturation dual-wavelength signals (i.e., 18 photoelectric signals). The air belt is used to obtain pressure and pressure pulse at the left and right arms (or wrists) and ankles (i.e., 8 signals).

[0141] The 32 multimodal cardiovascular-related signals obtained through non-invasive synchronous acquisition are used to extract cardiovascular-related characteristic parameters using specific parameter extraction methods. The cardiovascular health status can be tracked, monitored, and comprehensively assessed through specific assessment models built into APP software or cloud software.

[0142] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0143] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A cardiovascular monitoring and assessment method based on multimodal noninvasive signals, characterized in that, Includes the following steps: The basic information, clinical test results, and 32 cardiovascular-related bodily signals of the sample subjects were collected. The collection process of the 32 bodily signals was divided into: no pressure process, slow deflation or inflation process, personalized blood pressure process, and blood flow occlusion process. The basic information, clinical test results, and 32 cardiovascular-related bodily signals of the subjects were preprocessed to obtain initial parameters. By using the minimum absolute value shrinkage and selection operator Lasso function, feature selection and dimensionality reduction are performed on the initial parameters to obtain quadratic parameters; Correlation analysis was performed on the quadratic parameters, and the correlation coefficients between the quadratic parameters and cardiovascular disease outcomes were sorted from largest to smallest. The quadratic parameters were then selected from the beginning to the end according to the set proportion of the total sample size, and used as the cubic parameters. The cubic parameters are screened to exclude multivariate collinearity, and the evaluation parameters are obtained. An evaluation model is constructed based on machine learning algorithms and evaluation parameters. The study acquires 32 cardiovascular-related bodily signals from the subject. The acquisition process of these 32 bodily signals is divided into: no-pressure process, slow deflation or inflation process, personalized blood pressure process, and blood flow occlusion process. The acquired data is input into the evaluation model, and the evaluation results are output.

2. The cardiovascular monitoring and assessment method based on multimodal noninvasive signals as described in claim 1, characterized in that, Based on the subprocess, 32 cardiovascular-related bodily signals were collected from the subjects and test subjects, specifically: Air bands were placed on both wrists or arms and ankles of the subject and the test subject. Pressureless process: Without applying air pressure to all air zones, collect three standard electrocardiogram signals, a single heart sound and two lung sound signals for a period of time, as well as dual-wavelength signals of blood oxygen saturation from both earlobes, temples, a single finger and a single toe. The acquisition frequency of each channel can be set individually. Slow deflation or inflation process: Simultaneously inflate the air belts at both arms or wrists and ankles to the set pressure value, and then slowly deflate them at a constant deflation rate; Alternatively, each air bag can be slowly inflated until it exceeds the systolic pressure of the subject at the corresponding position by 20 mmHg, and then quickly deflated. Obtain blood pressure values ​​at both arms or wrists and ankles, including systolic blood pressure, diastolic blood pressure, and mean blood pressure; Personalized blood pressure sub-process: While applying personalized average blood pressure to corresponding sites on both arms or wrists and ankles with air cuffs, 32 signals are collected simultaneously within a first preset time period. Blood flow occlusion process: When applying personalized maximum blood pressure to the corresponding parts of both arms or wrists and ankles, 32 signals are simultaneously collected within a second preset time period to obtain the changes in dual-wavelength signals of blood oxygen saturation in the fingers and toes under the condition of basic blood flow occlusion.

3. The cardiovascular monitoring and assessment method based on multimodal noninvasive signals as described in claim 1, characterized in that, The steps for preprocessing 32 cardiovascular-related bodily signals from subjects and test subjects to obtain initial parameters are as follows: The ECG signal feature points of 32 cardiovascular-related body signals of subjects and test subjects and their positions in the ECG signal data sequence were extracted. Based on the position of the ECG signal feature points, the feature parameters of the ECG signal were extracted and their mean and standard deviation were calculated. By utilizing the relationship between electrocardiogram (ECG) signals and other signals, feature points and feature parameters of other signals are extracted, and the mean and standard deviation of the feature parameters of other signals are calculated. Calculate the ratio of the same feature parameters obtained in different sub-processes, and extract the ratio of feature parameters of the left and right symmetrical parts of the subject and test subject in the same sub-process. Based on the pressure and pressure pulse signals in 32 body signals, the systolic pressure, diastolic pressure and mean blood pressure of the corresponding parts were extracted using the normalized oscilloscope method. Calculate the dual-wavelength correlation parameters of blood oxygen saturation at the fingers and toes during the blood flow occlusion subprocess.

4. The cardiovascular monitoring and assessment method based on multimodal noninvasive signals as described in claim 3, characterized in that, The method for extracting ECG feature points from 32 cardiovascular-related bodily signals of subjects and participants, along with their positions in the ECG signal data sequence, and then extracting feature parameters of the ECG signals based on the positions of these feature points and calculating their mean and standard deviation is as follows: By utilizing the characteristics of the electrocardiogram (ECG) signal itself, we can extract the feature points in the ECG signal and their positions in the ECG signal data sequence. Based on the location of each characteristic point of the ECG signal, various characteristic parameters of the ECG signal are extracted, including the RR interval, the time between QRS complexes, and the amplitudes of QR waves, SR waves, ST waves, and SP waves and their ratios. The characteristic parameter values ​​are sorted according to the acquisition sub-process, and the mean and standard deviation of each parameter are calculated from the middle percentage of the data.

5. The cardiovascular monitoring and assessment method based on multimodal noninvasive signals as described in claim 3, characterized in that, Other signal characteristic points include the maximum value point, minimum value point, maximum rate of increase point, and maximum rate of decrease point in each cardiac cycle; Other characteristic parameters of the signals include the difference between the maximum and minimum values ​​within a cardiac cycle, the maximum rate of ascent, the maximum rate of descent, the ratio of the maximum rate of ascent to the maximum rate of descent, the time taken for the ECG R wave to reach the maximum rate of ascent and the maximum rate of descent and their ratio, the ratio of the time taken for the ECG R wave to reach its maximum value to the cardiac cycle, and the ratio of the mean, maximum rate of ascent, and maximum rate of descent of each signal within a cardiac cycle to the difference between the maximum and minimum values, respectively. The characteristic parameter values ​​of other signals are sorted by size according to the acquisition sub-process, and the mean and standard deviation of the characteristic parameters of other signals in each sub-process are calculated by taking the middle percentage of the data.

6. The cardiovascular monitoring and assessment method based on multimodal noninvasive signals as described in claim 3, characterized in that, Based on pressure and pressure pulse signals from 32 body signals, the following method is used to extract systolic blood pressure, diastolic blood pressure, and mean blood pressure at corresponding sites using normalized oscillometric analysis: Using dual-wavelength signals of blood oxygen saturation from the forehead, bilateral earlobes, temples, a single finger, and a single toe, and based on the blood oxygen saturation obtained from a blood oxygen simulator, the average blood oxygen saturation values ​​at various points during the pressure-free process and the personalized average blood pressure process are calculated, along with the ratio between the two processes. Furthermore, the ratios of the average blood pressure saturation values ​​at the forehead, bilateral earlobes, temples, and a single toe during the pressure-free process and the personalized average blood pressure process to the blood oxygen saturation of a single finger on the same side during the same process are calculated, as are the ratios of the average blood oxygen saturation values ​​on the left and right sides of the pressure-free process and the personalized average blood pressure process at the left and right earlobes, temples, a single toe, and a single finger. ， Wherein, SpO2 represents blood oxygen saturation, R represents the mean of the ratio of AC to DC of the 660nm photoelectric signal and the ratio of AC to DC of the 940nm photoelectric signal in the same cardiac cycle in the blood oxygen saturation dual-wavelength signal in the corresponding sub-process, and a, b, and c represent the fitting coefficients obtained by the blood oxygen simulator.

7. The cardiovascular monitoring and assessment method based on multimodal noninvasive signals as described in claim 3, characterized in that, The dual-wavelength correlation parameters of blood oxygen saturation at the fingers and toes during the blood flow occlusion subprocess are calculated as follows: D ikλ =±(Max ikλ -Min ikλ ), DR ik = , DRR k = , DRR i = , Among them, Max ikλ Min ikλ Let represent the maximum and minimum values ​​of the photoelectric signal with wavelength λ at point k on side i of the blood flow occlusion subprocess, respectively. The difference is positive when the maximum value is earlier and negative when the maximum value is later, used to distinguish the trend of change; D ikλ This represents the difference between the maximum and minimum values ​​of the photoelectric signal with wavelength λ at point k on side i of the blood flow blocking subprocess; DR ik D represents the ratio of the difference between the maximum and minimum values ​​of the photoelectric signal at wavelength 660nm to the difference between the maximum and minimum values ​​of the photoelectric signal at wavelength 940nm at site k on side i of the blood flow occlusion subprocess; ik940 and D ik660 These represent the differences between the maximum and minimum values ​​of the photoelectric signals at wavelengths of 940 nm and 660 nm at location k on side i of the blood flow occlusion subprocess, respectively. DRR k This indicates the DR at k points on the left and right sides of the blood flow blocking subprocess. ik The ratio; This represents the ratio of the maximum and minimum difference of the 660nm wavelength photoelectric signal at the left k site of the blood flow blocking subprocess to the maximum and minimum difference of the 940nm wavelength photoelectric signal. This represents the ratio of the maximum and minimum difference between the 660nm wavelength photoelectric signal and the maximum and minimum difference between the 940nm wavelength photoelectric signal at the k site on the right side of the blood flow blocking subprocess. DRR i This indicates the blood flow blocking subprocess at the i-th toe and finger, on both sides of the DR. ik The ratio; This represents the ratio of the maximum and minimum difference of the 660nm wavelength photoelectric signal at the toe location on the i-th side of the blood flow blocking subprocess to the maximum and minimum difference of the 940nm wavelength photoelectric signal. This represents the ratio of the maximum and minimum difference between the 660nm wavelength photoelectric signal and the 940nm wavelength photoelectric signal at the finger site on the i-th side of the blood flow blocking subprocess.

8. The cardiovascular monitoring and assessment method based on multimodal noninvasive signals as described in claim 1, characterized in that, The steps for screening cubic parameters to exclude multivariate collinearity and obtaining the evaluation parameters are as follows: The cubic parameters were sorted from largest to smallest according to their correlation coefficients, and the variance inflation factor collinearity analysis was performed using the forward selection method, specifically as follows: If the maximum variance inflation factor (VIF) exceeds the threshold, delete the cubic parameter corresponding to the maximum VIF value, add the next cubic parameter for collinearity analysis, and continue until the required number of feature parameters are selected. If collinearity cannot select the required number of cubic parameters, then the remaining cubic parameters after VIF multicollinearity processing are used as the final evaluation parameters.

9. A cardiovascular monitoring and assessment system based on the method of any one of claims 1-8, characterized in that, This includes non-invasive signal acquisition terminals, cloud servers, and smart terminals; The non-invasive signal acquisition terminal includes a sensor module, a signal processing module, and a microprocessor module. The sensor module is used to acquire body temperature, ambient temperature, and 32 cardiovascular-related non-invasive signals. The input end of the signal processing module is connected to the output end of the sensor module, and the output end of the signal processing module is connected to the microprocessor. The microprocessor controls the signal processing module to communicate with the cloud server and smart terminal. The cloud server is bidirectionally connected to both the non-invasive signal acquisition terminal and the smart terminal, and the non-invasive signal acquisition terminal is bidirectionally connected to the smart terminal.

10. The cardiovascular monitoring and assessment system as described in claim 9, characterized in that, The sensor module includes a human body temperature sensor, an ambient temperature sensor, an electrocardiogram electrode, a pressure sensor, a sound sensor, and a photoelectric sensor. The human body temperature sensor is used to collect the body temperature of the subject and the person to be tested. The ambient temperature sensor is mounted on the outer casing of the non-invasive signal acquisition terminal; The ECG electrodes can be attached to the left sternal border, right sternal border, lower left abdomen, and right leg of the passenger. The pressure sensor is mounted on the air belt and can come into contact with the occupant's body. The air belt is positioned at the occupant's wrist or arm and ankle on the vehicle seat and can be connected to the occupant's wrist and ankle. The pressure sensor is used to collect pressure and pressure pulse signals. The sound sensor is used to collect lung sound signals and heart sound signals from the passenger; The photoelectric sensor is integrated into the blood oxygen probe, which has at least nine sensors corresponding to the forehead area, left and right temple areas, left and right earlobes, left and right fingers, and left and right toes.

Images