Cuffless continuous blood pressure measurement method and system

Abstract

The present invention relates to a cuffless continuous blood pressure measurement method, comprising: synchronously acquiring BCG signals and PPG signals of a subject; removing noise from the BCG signals and the PPG signals, and assessing beat by beat the signal quality of the BCG signals and the PPG signals; for the BCG signals and the PPG signals that meet signal quality requirements, detecting relevant feature point locations, and extracting PTT features, BCG signal features and PPG signal features from the relevant feature point locations; and on the basis of the extracted PTT features, BCG signal features and PPG signal features and statistical data and calibrated blood pressure of a user, calculating a blood pressure value by using a PTT model and a multi-parameter model. The present invention further relates to a cuffless continuous blood pressure measurement system. The present invention can implement cuffless, low-load and high-precision blood pressure measurement.

Description

Cuffless continuous blood pressure measurement method and system Technical Field

[0001] The present invention relates to the technical field of medical equipment, and in particular to a cuffless continuous blood pressure measurement method and system. Background Art

[0002] Cardiovascular disease is a major disease that seriously endangers the world's population and is the leading cause of death among urban and rural residents in my country. According to the "China Cardiovascular Health and Disease Report 2021", there are 330 million people suffering from cardiovascular disease in the country, of which hypertension accounts for as high as 72.5%. As an important physiological parameter of the human body, blood pressure and its variability are important bases for assessing the health status of heart and vascular function. Continuous measurement of blood pressure can reflect the trend of blood pressure changes and instantaneous changes, which is of great significance for timely detection of cardiovascular disease emergencies and reducing cardiovascular disease mortality. Therefore, continuous blood pressure measurement plays an important role in cardiovascular health management.

[0003] Existing blood pressure measurement methods include direct and indirect methods. The direct method mainly includes arterial catheterization, while the indirect methods mainly include oscillometric method, auscultation method, arterial tension method, volume compensation method and pulse wave method.

[0004] Arterial catheterization involves inserting a catheter with a pressure sensor directly into a human artery to measure blood pressure. This method is the gold standard for continuous blood pressure measurement, but it is invasive and requires professional operation. It is usually only used in intensive care units and is not suitable for daily monitoring environments.

[0005] The oscillometric method and auscultatory method are the two most commonly used blood pressure measurement methods in outpatient practice. Both methods require an inflated cuff to block the blood vessels. The oscillometric method determines blood pressure by detecting the pulsation (i.e., oscillation wave) generated by the arteries during cuff inflation and deflation, while the auscultatory method determines blood pressure by detecting the pulsating sound produced by blood flow reopening the blood vessels during cuff deflation. Due to the restraint of the inflated cuff, these two methods are less comfortable. In addition, both methods can only perform single blood pressure measurements, not continuous blood pressure measurements, making it difficult to capture abnormal blood pressure variations in a timely manner.

[0006] Arterial tension and volume compensation are two of the more established noninvasive continuous blood pressure measurement technologies. Arterial tension measures blood pressure by flattening the artery, but this method places stringent demands on sensor positioning. Volume compensation requires external force, such as an inflated cuff, to maintain a constant arterial volume. However, prolonged measurements can lead to venous congestion, affecting accuracy. Furthermore, the complex equipment required for both methods makes them unsuitable for prolonged use in everyday settings.

[0007] The pulse wave method collects surface-measurable physiological signals, such as electrocardiogram (ECG), pulse generator (PPG), and blood pressure (BCG), and then uses pulse wave conduction theory and artificial intelligence techniques to map these surface signals into blood pressure values, achieving indirect blood pressure measurement. This method, which eliminates the constraints of an inflatable cuff and enables comfortable and non-invasive continuous blood pressure monitoring, has garnered widespread attention from both academia and industry in recent years and is currently the mainstream technology in the field of continuous blood pressure measurement. Existing continuous blood pressure measurement methods based on the pulse wave method primarily estimate blood pressure indirectly by measuring ECG and PPG signals and calculating pulse transit time (PTT). However, this method has the following disadvantages: 1) it requires surface-mounted ECG electrodes, making it uncomfortable to wear; and 2) the calculated PTT includes the pre-ejection period (PEP), which can introduce errors in blood pressure measurement.

[0008] To provide a more comfortable continuous blood pressure measurement method and avoid interference from PEP, researchers have proposed a continuous blood pressure measurement method based on BCG signals. BCG signals can be measured contactlessly, significantly improving the comfort of continuous blood pressure measurement. However, existing technologies fail to fully utilize the blood pressure characteristics contained in PPG signals. Compared to BCG signals, PPG signals are generated by arterial pulsation and therefore contain richer blood pressure information.

[0009] Currently, existing technologies related to continuous blood pressure measurement have poor comfort and insufficient accuracy, and are unable to extract information from PPG signals that contain rich blood pressure information to achieve low-load, accurate blood pressure calculation.

[0010] Summary of the Invention

[0011] In view of this, it is necessary to provide a cuffless continuous blood pressure measurement method and system.

[0012] The present invention provides a cuffless continuous blood pressure measurement method, which includes the following steps: a. synchronously collecting a subject's BCG signal and PPG signal; b. removing noise from the BCG signal and PPG signal, and evaluating the signal quality of the BCG signal and PPG signal beat by beat; c. for BCG signals and PPG signals that meet the signal quality requirements, detecting the positions of relevant feature points and extracting PTT features, BCG signal features, and PPG signal features from them; d. calculating the blood pressure value using a PTT model and a multi-parameter model based on the extracted PTT features, BCG signal features, and PPG signal features, as well as the user's statistical data and calibrated blood pressure.

[0013] Preferably, the step b comprises:

[0014] First, a bandpass filter is used to remove high-frequency and low-frequency noise from the BCG and PPG signals, and then the signal quality of the denoised BCG and PPG signals is evaluated beat by beat.

[0015] Preferably, the signal quality assessment includes:

[0016] The BCG and PPG signals are segmented beat by beat based on the H wave position of the BCG signal and the trough position of the PPG signal;

[0017] Calculate the kurtosis and skewness of each heartbeat of the BCG signal and PPG signal for ease of description. In this embodiment, the kurtosis and skewness of the BCG signal of the i-th heartbeat are respectively and The kurtosis and skewness of the PPG signal of the i-th heart beat are and

[0018] For the BCG signal and PPG signal of the i-th heartbeat, if their kurtosis and skewness are both within the preset range, the BCG signal and PPG signal of the heartbeat meet the signal quality requirements; otherwise, the BCG signal and PPG signal of the heartbeat do not meet the signal quality requirements; repeat the above judgment to complete the signal quality evaluation of each heartbeat.

[0019] Preferably, the step c comprises:

[0020] For heartbeats whose signal quality meets the requirements, the characteristic points of the BCG signal and the characteristic points of the PPG signal are detected; based on the detection results, the PTT features, BCG signal features and PPG signal features are calculated and extracted; wherein the BCG signal features include: HI interval, HI amplitude difference, IJ interval, IJ amplitude difference, JK interval, JK amplitude difference and JJ interval; the PPG signal features include: FP interval, FP slope, area under the curve of FP package, PN interval, PF interval, PF slope, area under the curve of PF package and FF interval.

[0021] Preferably, the step d comprises:

[0022] Step S41: Input the extracted PTT characteristics and the user's calibrated blood pressure into the PTT equation to obtain an estimated SBP value SBP est1 and the estimated value of DBP est1 , the PTT equation is:

[0023] In the above formula, SBP0 and DBP0 are the calibrated systolic blood pressure and calibrated diastolic blood pressure respectively measured by a cuff-type blood pressure monitor; PTT0 is the initial PTT; γ is a vascular-related constant;

[0024] Step S42: The extracted BCG signal features, PPG signal features, PTT features, user statistics and calibrated blood pressure are input into a multi-parameter model to obtain an estimated SBP value SBP. est2 and the estimated value of DBP est2 ; Wherein, the user's statistical information includes: gender, age, height and weight; the multi-parameter model is a data model trained using a machine learning algorithm;

[0025] Step S43: Fusing the blood pressure estimation results of the PTT model and the multi-parameter model to obtain the final blood pressure meter estimation value. The fusion method is as follows: DBP = α·DBP est1 +(1-a)·DBP est2 SBP=β·SBP est1 +(1-β)·SBP est2

[0026] In the above formula, α and β are constants determined based on experiments.

[0027] The present invention provides a cuffless continuous blood pressure measurement system, which includes a signal acquisition module, a signal preprocessing module, a signal feature extraction module, and a blood pressure calculation module, wherein: the signal acquisition module is used to synchronously acquire a subject's BCG signal and PPG signal; the signal preprocessing module is used to remove noise from the BCG signal and PPG signal, and evaluate the signal quality of the BCG signal and PPG signal beat by beat; the signal feature extraction module is used to detect the positions of relevant feature points of BCG signals and PPG signals that meet the signal quality requirements and extract PTT features, BCG signal features, and PPG signal features from them; the signal feature extraction module is used to calculate the blood pressure value using a PTT model and a multi-parameter model based on the extracted PTT features, BCG signal features, and PPG signal features, as well as the user's statistical data and calibrated blood pressure.

[0028] Preferably, the signal preprocessing module is specifically used for:

[0029] First, a bandpass filter is used to remove high-frequency and low-frequency noise from the BCG and PPG signals, and then the signal quality of the denoised BCG and PPG signals is evaluated beat by beat.

[0030] Preferably, the signal quality assessment includes:

[0031] The BCG and PPG signals are segmented beat by beat based on the H wave position of the BCG signal and the trough position of the PPG signal;

[0032] Calculate the kurtosis and skewness of each heartbeat of the BCG signal and PPG signal for ease of description. In this embodiment, the kurtosis and skewness of the BCG signal of the i-th heartbeat are respectively and The kurtosis and skewness of the PPG signal of the i-th heart beat are and

[0033] For the BCG signal and PPG signal of the i-th heartbeat, if their kurtosis and skewness are both within the preset range, the BCG signal and PPG signal of the heartbeat meet the signal quality requirements; otherwise, the BCG signal and PPG signal of the heartbeat do not meet the signal quality requirements; repeat the above judgment to complete the signal quality evaluation of each heartbeat.

[0034] Preferably, the signal feature extraction module is specifically used to:

[0035] For heartbeats whose signal quality meets the requirements, the characteristic points of the BCG signal and the characteristic points of the PPG signal are detected; based on the detection results, the PTT features, BCG signal features and PPG signal features are calculated and extracted; wherein the BCG signal features include: HI interval, HI amplitude difference, IJ interval, IJ amplitude difference, JK interval, JK amplitude difference and JJ interval; the PPG signal features include: FP interval, FP slope, area under the curve of FP package, PN interval, PF interval, PF slope, area under the curve of PF package and FF interval.

[0036] Preferably, the blood pressure calculation module is specifically used to:

[0037] The extracted PTT features and the user's calibrated blood pressure are input into the PTT equation to obtain the estimated SBP value SBP est1 and the estimated value of DBP est1 , the PTT equation is:

[0038] In the above formula, SBP0 and DBP0 are the calibrated systolic blood pressure and calibrated diastolic blood pressure respectively measured by a cuff-type blood pressure monitor; PTT0 is the initial PTT; γ is a vascular-related constant;

[0039] The extracted BCG signal features, PPG signal features, PTT features, user statistics and calibrated blood pressure are input into the multi-parameter model to obtain the estimated SBP value SBP est2 and the estimated value of DBP est2 ; Wherein, the user's statistical information includes: gender, age, height and weight; the multi-parameter model is a data model trained using a machine learning algorithm;

[0040] The blood pressure estimation results of the PTT model and the multi-parameter model are integrated to obtain the final blood pressure meter estimation value. The integration method is as follows: DBP = α·DBP est1 +(1-a)·DBP est2 SBP=β·SBP est1 +(1-β)·SBP est2

[0041] In the above formula, α and β are constants determined based on experiments.

[0042] The present invention extracts multi-parameter features from BCG and PPG signals and combines them with a pulse wave transit time model and a multi-parameter fusion model to achieve cuff-free, low-load, and highly accurate blood pressure measurement. Specifically:

[0043] Compared with existing technologies based on ECG signals and PPG signals, the BCG signal used in the present invention can be measured in a non-contact manner and can avoid the interference of PEP on the results, so it is more comfortable and more accurate; compared with existing technologies based on BCG signals and PPG signals, the present invention takes into account the advantage of PPG signals containing rich blood pressure information, extracts multiple features from PPG signals for blood pressure calculation, and further integrates the blood pressure calculation results of the PTT model and the multi-parameter model, so it is more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] FIG1 is a flow chart of a cuffless continuous blood pressure measurement method according to the present invention;

[0045] FIG2 is a schematic diagram of beat-by-beat segmentation of BCG signals and PPG signals according to an embodiment of the present invention;

[0046] FIG3 is a schematic diagram of extracting PTT features, BCG signal features, and PPG signal features according to an embodiment of the present invention;

[0047] FIG4 is a schematic diagram of calculating blood pressure values ​​according to an embodiment of the present invention;

[0048] FIG5 is a hardware architecture diagram of the cuffless continuous blood pressure measurement system of the present invention. DETAILED DESCRIPTION

[0049] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0050] 1 , which is a flowchart of a preferred embodiment of a cuffless continuous blood pressure measurement method of the present invention.

[0051] Step S1: synchronously collect the BCG signal and PPG signal of the subject. Specifically:

[0052] The BCG signal and PPG signal of the subject are collected synchronously, and the collection positions of the PPG signal include: fingers, wrists, ears, and toes.

[0053] Step S2: remove noise from the BCG signal and the PPG signal, and evaluate the signal quality of the BCG signal and the PPG signal beat by beat. Specifically:

[0054] First, a bandpass filter is used to remove high-frequency and low-frequency noise from the BCG and PPG signals. Then, the signal quality of the denoised BCG and PPG signals is evaluated beat by beat. The specific steps of signal quality evaluation include:

[0055] Step S21, segmenting the BCG signal and the PPG signal beat by beat (i.e., segmenting by heartbeat cycle) based on the position of the H wave of the BCG signal and the position of the trough of the PPG signal (i.e., F in FIG2 ). FIG2 shows segmenting the BCG signal and the PPG signal into four heartbeats;

[0056] Step S22, calculate the kurtosis and skewness of each heartbeat of the BCG signal and the PPG signal. For ease of description, in this embodiment, the kurtosis and skewness of the BCG signal of the i-th heartbeat are respectively and The kurtosis and skewness of the PPG signal of the i-th heart beat are and

[0057] Step S23: For the BCG signal and PPG signal of the i-th heartbeat, if their kurtosis and skewness values ​​are within the preset range, that is, if and If both are satisfied, it means that the BCG signal and PPG signal of the heartbeat meet the signal quality requirements, otherwise they are not satisfied. Repeat this step to complete the signal quality judgment of each heartbeat. It should be noted that bcgK th1 、bcgK th2 、bcgS th1 、bcgS th2 ,ppgK th1 ,ppgK th2 ,ppgS th1 and ppgS th2 There are 8 preset experience thresholds.

[0058] Step S3: For BCG signals and PPG signals that meet the signal quality requirements, detect the locations of relevant feature points and extract signal features from them. Specifically:

[0059] Refer to Figure 3. For heartbeats with satisfactory signal quality, detect the characteristic points of the BCG signal (including the H, I, J, and K waves) and the characteristic points of the PPG signal (including the trough (F), peak (P), and mid-fall (N)). Based on the detection results, calculate and extract the PTT, BCG, and PPG signal features. BCG signal features include: HI interval, HI amplitude difference, IJ interval, IJ amplitude difference, JK interval, JK amplitude difference, and JJ interval; PPG signal features include: FP interval, FP slope, area under the curve of the FP package, PN interval, PF interval, PF slope, area under the curve of the PF package, and FF interval.

[0060] Step S4, based on the extracted signal features and the user's statistical data and calibrated blood pressure, the blood pressure value is calculated using the PTT model and the multi-parameter model. Specifically:

[0061] PTT describes the conduction time of the pulse wave in the artery, reflects the degree of arteriosclerosis and thus reflects the size of blood pressure. Based on the association mechanism, the PTT model can track rapid changes in blood pressure. The multi-parameter model integrates a variety of factors that affect blood pressure changes and can reflect the overall trend of blood pressure changes. This application combines the advantages of the two and integrates the PTT model and the multi-parameter model to calculate systolic blood pressure (SBP) and diastolic blood pressure (DBP) (as shown in Figure 4), specifically including:

[0062] Step S41: Input the extracted PTT characteristics and the user's calibrated blood pressure into the PTT equation to obtain an estimated SBP value SBP est1 and the estimated value of DBP est1 , the PTT equation is:

[0063] In the above formula, SBP0 and DBP0 are the calibrated systolic blood pressure and the calibrated diastolic blood pressure, respectively, which are measured by a cuff-type sphygmomanometer; PTT0 is the initial PTT; and γ is a vascular-related constant.

[0064] Step S42: The extracted BCG signal features, PPG signal features, PTT features, user statistics and calibrated blood pressure are input into a multi-parameter model to obtain an estimated SBP value SBP. est2 and the estimated value of DBP est2 The user's statistical information includes gender, age, height and weight; and the multi-parameter model is a data model trained using a machine learning algorithm.

[0065] Step S43: Fusing the blood pressure estimation results of the PTT model and the multi-parameter model to obtain the final blood pressure meter estimation value. The fusion method is as follows: DBP = α·DBP est1+(1-a)·DBP est2 SBP=β·SBP est1 +(1-β)·SBP ets2

[0066] In the above formula, α and β are constants determined experimentally.

[0067] Refer to FIG5 , which is a hardware architecture diagram of the cuffless continuous blood pressure measurement system 10 of the present invention. The system includes: a signal acquisition module 101, a signal preprocessing module 102, a signal feature extraction module 103, and a blood pressure calculation module 104. Among them:

[0068] The signal acquisition module 101 is used to synchronously acquire the BCG signal and PPG signal of the subject. Specifically:

[0069] The signal acquisition module 101 synchronously acquires BCG signals and PPG signals of the subject, and the acquisition locations of the PPG signals include: fingers, wrists, ears, and toes.

[0070] The signal preprocessing module 102 is used to remove noise from the BCG signal and the PPG signal, and to evaluate the signal quality of the BCG signal and the PPG signal beat by beat. Specifically:

[0071] The signal preprocessing module 102 first uses a bandpass filter to remove high-frequency and low-frequency noise from the BCG signal and the PPG signal, and then performs a beat-by-beat signal quality assessment on the denoised BCG signal and the PPG signal. The signal quality assessment specifically includes:

[0072] First, the signal preprocessing module 102 segments the BCG signal and the PPG signal beat by beat (i.e., segment by heartbeat cycle) according to the position of the H wave of the BCG signal and the position of the trough of the PPG signal (i.e., F in FIG2 ). FIG2 shows the segmentation of the BCG signal and the PPG signal into four heartbeats.

[0073] Next, the signal preprocessing module 102 calculates the kurtosis and skewness of each heartbeat of the BCG signal and the PPG signal. For ease of description, in this embodiment, the kurtosis and skewness of the BCG signal of the i-th heartbeat are respectively and The kurtosis and skewness of the PPG signal of the i-th heart beat are and

[0074] Then, for the BCG signal and PPG signal of the i-th heartbeat, if their kurtosis and skewness values ​​are within the preset range, that is, if and If both are satisfied, it means that the BCG signal and PPG signal of the heartbeat meet the signal quality requirements, otherwise they are not satisfied. The signal preprocessing module 102 repeats the above judgment to complete the signal quality evaluation of each heartbeat. It should be noted that bcgK th1 、bcgK th2 、bcgS th1 、bcgS th2 ,ppgK th1 ,ppgK th2 ,ppgS th1 and ppgS th2 There are 8 preset experience thresholds.

[0075] The signal feature extraction module 103 is used to detect the positions of relevant feature points of BCG signals and PPG signals that meet the signal quality requirements and extract signal features from them. Specifically:

[0076] The signal feature extraction module 103 detects characteristic points of the BCG signal (including H, I, J, and K waves) and characteristic points of the PPG signal (including trough (F), peak (P point), and mid-fall (N point)) for heartbeats whose signal quality meets the requirements; based on the detection results, the module calculates and extracts PTT features, BCG signal features, and PPG signal features (please refer to FIG3 ). The BCG signal features include: HI interval, HI amplitude difference, IJ interval, IJ amplitude difference, JK interval, JK amplitude difference, and JJ interval; the PPG signal features include: FP interval, FP slope, area under the curve of the FP package, PN interval, PF interval, PF slope, area under the curve of the PF package, and FF interval.

[0077] The blood pressure calculation module 104 is used to calculate the blood pressure value using the PTT model and the multi-parameter model based on the extracted signal features, the user's statistical data and the calibrated blood pressure. Specifically:

[0078] PTT describes the conduction time of the pulse wave in the artery, reflects the degree of arteriosclerosis and thus reflects the size of blood pressure. Based on the association mechanism, the PTT model can track rapid changes in blood pressure. The multi-parameter model integrates a variety of factors that affect blood pressure changes and can reflect the overall trend of blood pressure changes. This application combines the advantages of the two and integrates the PTT model and the multi-parameter model to calculate systolic blood pressure (SBP) and diastolic blood pressure (DBP) (as shown in Figure 4), specifically including:

[0079] The blood pressure calculation module 104 inputs the extracted PTT features and the user's calibrated blood pressure into the PTT equation to obtain an estimated SBP value SBP est1 and the estimated value of DBP est1 , the PTT equation is:

[0080] In the above formula, SBP0 and DBP0 are the calibrated systolic blood pressure and the calibrated diastolic blood pressure, respectively, which are measured by a cuff-type sphygmomanometer; PTT0 is the initial PTT; and γ is a vascular-related constant.

[0081] The blood pressure calculation module 104 inputs the extracted BCG signal features, PPG signal features, PTT features, user statistics and calibrated blood pressure into a multi-parameter model to obtain an estimated SBP value SBP est2 and the estimated value of DBP est2 The user's statistical information includes gender, age, height and weight; and the multi-parameter model is a data model trained using a machine learning algorithm.

[0082] The blood pressure calculation module 104 fuses the blood pressure estimation results of the PTT model and the multi-parameter model to obtain the final blood pressure meter estimation value. The fusion method is as follows: DBP = α·DBP est1 +(1-a)·DBP est2 SBP=β·SBP est1 +(1-β)·SBP est2

[0083] In the above formula, α and β are constants determined experimentally.

[0084] Although the present invention has been described with reference to the current preferred embodiments, those skilled in the art should understand that the above-mentioned preferred embodiments are only used to illustrate the present invention and are not used to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A cuffless continuous blood pressure measurement method, characterized in that: The method comprises the following steps: a. Synchronously collect the BCG signal and PPG signal of the subject; b. removing noise from the BCG signal and the PPG signal, and evaluating the signal quality of the BCG signal and the PPG signal beat by beat; c. For BCG signals and PPG signals that meet the signal quality requirements, detect the positions of relevant feature points and extract PTT features, BCG signal features and PPG signal features from them; d. Based on the extracted PTT features, BCG signal features and PPG signal features, as well as the user's statistical data and calibrated blood pressure, the blood pressure value is calculated using the PTT model and the multi-parameter model.

2. The cuffless continuous blood pressure measurement method according to claim 1, characterized in that: The step b comprises: Firstly, a bandpass filter is used to remove high-frequency and low-frequency noise from the BCG signal and the PPG signal, and then the signal quality of the denoised BCG signal and the PPG signal is evaluated beat by beat.

3. The cuffless continuous blood pressure measurement method according to claim 2, characterized in that: The signal quality assessment includes: The BCG signal and the PPG signal are segmented beat by beat according to the H wave position of the BCG signal and the trough position of the PPG signal; The kurtosis and skewness of each heartbeat of the BCG signal and the PPG signal are calculated. For ease of description, in this embodiment, the kurtosis and skewness of the BCG signal of the i-th heartbeat are recorded as and The kurtosis and skewness of the PPG signal of the i-th heart beat are and For the BCG signal and PPG signal of the i-th heart beat, if their kurtosis and skewness are within the preset range, the BCG signal and PPG signal of the heart beat meet the signal quality requirements, otherwise the BCG signal and PPG signal of the heart beat do not meet the signal quality requirements; repeat the above judgment to complete the signal quality evaluation of each heart beat.

4. The cuffless continuous blood pressure measurement method according to claim 3, characterized in that: The step c comprises: For heartbeats whose signal quality meets the requirements, the characteristic points of the BCG signal and the characteristic points of the PPG signal are detected; based on the detection results, the PTT features, BCG signal features and PPG signal features are calculated and extracted; wherein the BCG signal features include: HI interval, HI amplitude difference, IJ interval, IJ amplitude difference, JK interval, JK amplitude difference and JJ interval; the PPG signal features include: FP interval, FP slope, area under the curve of FP package, PN interval, PF interval, PF slope, area under the curve of PF package and FF interval.

5. The cuffless continuous blood pressure measurement method according to claim 4, characterized in that: The step d comprises: Step S41, input the extracted PTT feature and the user's calibrated blood pressure into the PTT equation to obtain an estimated value of SBP SBP est1 and the estimated value of DBP est1 , the PTT equation is: In the above formula, SBP0 and DBP0 are the calibrated systolic pressure and calibrated diastolic pressure measured by a cuff sphygmomanometer, respectively; PTT0 is the initial PTT; γ is a vascular-related constant; Step S42: Input the extracted BCG signal features, PPG signal features, PTT features, user statistics and calibrated blood pressure into a multi-parameter model to obtain an estimated value of SBP SBP est2 and the estimated value of DBP est2 ; Wherein, the user's statistical information includes: gender, age, height and weight; the multi-parameter model is a data model trained using a machine learning algorithm; Step S43, fusing the blood pressure estimation results of the PTT model and the multi-parameter model to obtain the final sphygmomanometer estimation value, and the fusing method is as follows: DBP=α·DBP est1 +(1-a)·DBP est2 SBP=β·SBP est1 +(1-β)·SBP est2 In the above formula, α and β are constants determined experimentally.

6. A cuffless continuous blood pressure measurement system, characterized in that: The system includes a signal acquisition module, a signal preprocessing module, a signal feature extraction module, and a blood pressure calculation module, wherein: The signal acquisition module is used to synchronously acquire the BCG signal and PPG signal of the subject; The signal preprocessing module is used to remove noise from the BCG signal and the PPG signal, and evaluate the signal quality of the BCG signal and the PPG signal beat by beat; The signal feature extraction module is used to detect the positions of relevant feature points of BCG signals and PPG signals that meet the signal quality requirements and extract PTT features, BCG signal features and PPG signal features from them; The signal feature extraction module is used to calculate the blood pressure value using the PTT model and the multi-parameter model based on the extracted PTT features, BCG signal features and PPG signal features, as well as the user's statistical data and calibrated blood pressure.

7. The cuffless continuous blood pressure measurement system according to claim 6, characterized in that: The signal preprocessing module is specifically used for: Firstly, a bandpass filter is used to remove high-frequency and low-frequency noise from the BCG signal and the PPG signal, and then the signal quality of the denoised BCG signal and the PPG signal is evaluated beat by beat.

8. The cuffless continuous blood pressure measurement system according to claim 7, characterized in that: The signal quality assessment includes: The BCG signal and the PPG signal are segmented beat by beat according to the H wave position of the BCG signal and the trough position of the PPG signal; The kurtosis and skewness of each heartbeat of the BCG signal and the PPG signal are calculated. For ease of description, in this embodiment, the kurtosis and skewness of the BCG signal of the i-th heartbeat are recorded as and The kurtosis and skewness of the PPG signal of the i-th heart beat are and For the BCG signal and PPG signal of the i-th heart beat, if their kurtosis and skewness are within the preset range, the BCG signal and PPG signal of the heart beat meet the signal quality requirements, otherwise the BCG signal and PPG signal of the heart beat do not meet the signal quality requirements; repeat the above judgment to complete the signal quality evaluation of each heart beat.

9. The cuffless continuous blood pressure measurement system according to claim 8, characterized in that: The signal feature extraction module is specifically used for: For heartbeats that meet the signal quality requirements, detect the characteristic points of the BCG signal and the characteristic points of the PPG signal; calculate and extract the PTT feature, the BCG signal feature and the PPG signal feature based on the detection results; The BCG signal characteristics include: HI interval, HI amplitude difference, IJ interval, IJ amplitude difference, JK interval, JK amplitude difference and JJ interval; the PPG signal characteristics include: FP interval, FP slope, area under the curve of FP package, PN interval, PF interval, PF slope, area under the curve of PF package and FF interval.

10. The cuffless continuous blood pressure measurement system according to claim 9, characterized in that: The blood pressure calculation module is specifically used for: The extracted PTT features and the user's calibrated blood pressure are input into the PTT equation to obtain the estimated SBP value SBP est1 and the estimated value of DBP est1 , the PTT equation is: In the above formula, SBP0 and DBP0 are the calibrated systolic pressure and calibrated diastolic pressure measured by a cuff sphygmomanometer, respectively; PTT0 is the initial PTT; γ is a vascular-related constant; The extracted BCG signal features, PPG signal features, PTT features, user statistics and calibrated blood pressure are input into the multi-parameter model to obtain the estimated SBP value SBP est2 and the estimated value of DBP est2 ; Wherein, the user's statistical information includes: gender, age, height and weight; the multi-parameter model is a data model trained using a machine learning algorithm; The blood pressure estimation results of the PTT model and the multi-parameter model are fused to obtain the final sphygmomanometer estimation value. The fusion method is as follows: DBP=α·DBP est1 +(1-a)·DBP est2 SBP=β·SBP est1 +(1-β)·SBP est2 In the above formula, α and β are constants determined experimentally.

Images