Blood pressure dynamic measurement method based on multi-sensor fusion and smart wearable device

Abstract

The application provides a blood pressure dynamic measurement method based on multi-sensor fusion and a smart wearable device, and the method comprises the following steps: collecting an arterial pulsation signal, a motion state signal and a photoelectric pulse wave signal of a wearing user based on an arterial pulsation sensor, an acceleration sensor and a photoelectric pulse wave sensor respectively; performing interference optimization on an initial peak value feature sequence in the arterial pulsation signal based on a motion amplitude change sequence of the motion state signal, to obtain an optimized peak value feature sequence; performing arterial pulsation and pulse wave feature correlation matching based on the optimized peak value feature sequence and a waveform feature sequence of the photoelectric pulse wave signal, to obtain a feature correlation result; analyzing a systolic pressure corresponding feature matching pair and a diastolic pressure corresponding feature matching pair based on the feature correlation result, to obtain a systolic pressure and diastolic pressure representation amplitude, and performing blood pressure measurement on the wearing user based on the systolic pressure and diastolic pressure representation amplitude, to obtain a dynamic blood pressure measurement result. The application improves the reliability and accuracy of dynamic blood pressure monitoring.

Description

Technical Field

[0001] This invention relates to the field of smart wearable technology, and in particular to a method for dynamic blood pressure measurement based on multi-sensor fusion and a smart wearable device. Background Technology

[0002] Blood pressure is an important physiological parameter for the human body, and its accurate measurement is of great significance for the prevention, diagnosis, and treatment of cardiovascular diseases. Currently, the most widely used blood pressure measurement method in clinical and home settings is the oscillometric blood pressure measurement method. This method uses an inflatable cuff to compress the artery in the arm, and a pressure sensor detects the pressure fluctuation signal generated by the arterial pulsation during the deflation of the cuff. Based on preset feature extraction rules, it identifies the feature points corresponding to systolic and diastolic blood pressure, and then calculates the blood pressure value.

[0003] However, since this method relies on a single pressure sensor to acquire cuff pressure fluctuation signals, interference factors such as changes in the human arm posture, muscle tension, and external environmental vibrations can directly affect the detection accuracy of the pressure sensor. This results in a large amount of interference noise in the acquired pressure fluctuation signals, which cannot accurately reflect the true characteristics of arterial pulsation. Ultimately, this leads to a significant deviation in blood pressure measurement results, especially in dynamic scenarios (such as walking or light activity), where this deviation problem is more prominent and cannot meet the accuracy requirements of ambulatory blood pressure monitoring. Summary of the Invention

[0004] This invention provides a method for dynamic blood pressure measurement based on multi-sensor fusion and an intelligent wearable device to solve the problem of large measurement deviations in dynamic scenarios caused by relying on a single sensor, thereby achieving accurate blood pressure measurement in dynamic scenarios and improving the reliability and accuracy of dynamic blood pressure monitoring.

[0005] In a first aspect, the present invention provides a dynamic blood pressure measurement method based on multi-sensor fusion, applicable to a smart wearable device; the smart wearable device integrates an arterial pulsation sensor, an acceleration sensor, and a photoelectric pulse wave sensor; the method includes:

[0006] The arterial pulsation sensor, the acceleration sensor, and the photoelectric pulse wave sensor respectively collect the wearer's arterial pulsation signal, motion state signal, and photoelectric pulse wave signal;

[0007] Based on the motion amplitude change sequence of the motion state signal, the initial peak feature sequence in the arterial pulsation signal is subjected to interference optimization to obtain the optimized peak feature sequence.

[0008] Based on the optimized peak feature sequence and the waveform feature sequence of the photoelectric pulse wave signal, feature association matching between arterial pulsation and pulse wave is performed to obtain feature association results;

[0009] Based on the feature association results, the feature matching pairs corresponding to systolic blood pressure and diastolic blood pressure are analyzed to obtain the systolic and diastolic blood pressure characterization amplitude. Based on the systolic and diastolic blood pressure characterization amplitude, the blood pressure of the wearer is measured to obtain the dynamic blood pressure measurement results.

[0010] In a second aspect, the present invention also provides a smart wearable device for use in the multi-sensor fusion-based dynamic blood pressure measurement method as described in the first aspect; the smart wearable device integrates an arterial pulsation sensor, an acceleration sensor, and a photoelectric pulse wave sensor; the smart wearable device includes:

[0011] The signal acquisition module is used to acquire the wearer's arterial pulsation signal, motion state signal, and photoelectric pulse wave signal based on the arterial pulsation sensor, the acceleration sensor, and the photoelectric pulse wave sensor, respectively.

[0012] An interference optimization module is used to perform interference optimization on the initial peak feature sequence in the arterial pulsation signal based on the motion amplitude change sequence of the motion state signal, so as to obtain an optimized peak feature sequence.

[0013] The feature matching module is used to perform feature association matching between arterial pulsation and pulse wave based on the optimized peak feature sequence and the waveform feature sequence of the photoelectric pulse wave signal, and obtain feature association results;

[0014] The blood pressure measurement module is used to analyze the feature matching pairs corresponding to systolic blood pressure and diastolic blood pressure based on the feature association results, obtain the systolic-diastolic blood pressure characteristic amplitude, and measure the blood pressure of the wearer based on the systolic-diastolic blood pressure characteristic amplitude to obtain the dynamic blood pressure measurement result.

[0015] Thirdly, the present invention also provides an electronic device, comprising: a memory for storing computer software programs; and a processor for reading and executing the computer software programs, thereby realizing the dynamic blood pressure measurement method based on multi-sensor fusion as described above.

[0016] Fourthly, the present invention also provides a non-transitory computer-readable storage medium storing a computer software program, which, when executed by a processor, implements the dynamic blood pressure measurement method based on multi-sensor fusion as described above.

[0017] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the dynamic blood pressure measurement method based on multi-sensor fusion as described above.

[0018] The blood pressure dynamic measurement method based on multi-sensor fusion provided in this invention collects the wearer's arterial pulsation signal, motion state signal, and photoelectric pulse wave signal using an arterial pulsation sensor, an accelerometer, and a photoelectric pulse wave sensor, respectively, overcoming the limitation of relying solely on a single pressure sensor. Based on the motion amplitude change sequence of the motion state signal, interference optimization is performed on the initial peak feature sequence in the arterial pulsation signal to obtain an optimized peak feature sequence. The motion state signal is used to accurately identify motion interference components in the arterial pulsation signal, and interference optimization effectively removes interference noise, solving the core problem that a single pressure sensor signal is easily distorted by motion interference, thus ensuring the accuracy of the arterial pulsation peak value. The features accurately reflect the actual state of arterial pulsation. Based on the optimized peak feature sequence and the waveform feature sequence of the photoelectric pulse wave signal, feature association matching between arterial pulsation and pulse wave is performed to obtain feature association results. By leveraging the synergistic verification of features from two different types of physiological signals, the reliability of blood pressure-related features is improved, avoiding potential biases in single-signal feature extraction. Based on the feature association results, feature matching pairs corresponding to systolic and diastolic blood pressure are analyzed to obtain the systolic-diastolic pressure amplitude. Blood pressure is then measured on the wearer based on this amplitude, yielding dynamic blood pressure measurement results. By precisely locating feature pairs matching the physiological mechanisms of systolic and diastolic blood pressure, the core basis for blood pressure measurement is ensured to be accurate and reliable. Therefore, this embodiment of the invention achieves simultaneous acquisition of interference signals and physiological signals through multi-source sensor signal acquisition, extracts real arterial pulsation features through motion interference optimization, enhances feature reliability through multi-signal feature association matching, and ultimately achieves blood pressure measurement based on precise features. This solves the problem of large measurement deviations in dynamic scenarios caused by relying solely on a single sensor, achieving accurate blood pressure measurement in dynamic scenarios and improving the reliability and accuracy of dynamic blood pressure monitoring. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the dynamic blood pressure measurement method based on multi-sensor fusion provided in an embodiment of the present invention.

[0020] Figure 2 This is a schematic diagram of the structure of the smart wearable device provided in an embodiment of the present invention;

[0021] Figure 3 An embodiment diagram of the electronic device provided in this invention;

[0022] Figure 4 An embodiment diagram of a computer-readable storage medium provided in accordance with the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0025] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0026] See Figure 1 , Figure 1 This is a flowchart illustrating the dynamic blood pressure measurement method based on multi-sensor fusion provided by the present invention. In this embodiment, the main body executing the dynamic blood pressure measurement method based on multi-sensor fusion is a smart wearable device, which integrates an arterial pulsation sensor, an accelerometer, and a photoelectric pulse wave sensor. The pulsation sensor can sense the pulsating motion of the wearer's arteries as the heart contracts and relaxes. The accelerometer can capture the wearer's body movement status in real time (including amplitude, direction, and frequency of movement). The photoelectric pulse wave sensor senses changes in the blood volume within the wearer's blood vessels with the pulse using photoplethysmography.

[0027] Therefore, dynamic blood pressure measurement methods based on multi-sensor fusion include:

[0028] Step 10: Collect the wearer's arterial pulsation signal, motion state signal, and photoelectric pulse wave signal based on the arterial pulsation sensor, accelerometer, and photoelectric pulse wave sensor, respectively.

[0029] Optionally, the smart wearable device can complete the initial configuration of each sensor, including setting parameters such as the signal acquisition frequency, signal amplification factor, and data sampling accuracy of each sensor, to ensure that the signals acquired by each sensor are consistent and effective, and to avoid the inability to process signals in subsequent collaborative processing due to parameter mismatch.

[0030] Furthermore, after initialization, the smart wearable device can control the arterial pulsation sensor, accelerometer, and photoelectric pulse wave sensor to simultaneously start signal acquisition. The arterial pulsation sensor, through close contact with the wearer's skin (the area where arteries are distributed, such as the radial artery in the wrist), senses the periodic expansion and contraction of arteries under the pumping action of the heart, converting this mechanical action into a corresponding electrical signal, i.e., the arterial pulsation signal. Therefore, the arterial pulsation signal reflects the wearer's arterial pulsation pattern. The accelerometer, through its built-in inertial measurement unit, captures the wearer's body movements in real time during the measurement process, including changes in amplitude and direction of movement in different motion scenarios such as stationary, walking, and running states. It converts this motion information into a corresponding electrical signal, i.e., the motion state signal. Therefore, the motion state signal reflects the wearer's body movement characteristics. The photoelectric pulse wave sensor emits light of a specific wavelength (such as green or red light) that penetrates the wearer's skin tissue. When the light shines on the blood vessels, the change in blood volume causes periodic changes in the amount of light absorbed and reflected. The photoelectric pulse wave sensor captures this light change information and converts it into a corresponding electrical signal, namely the photoelectric pulse wave signal. Therefore, the photoelectric pulse wave signal reflects the periodic changes in the blood volume in the wearer's blood vessels with the pulse.

[0031] Furthermore, the smart wearable device synchronously stores and pre-processes the arterial pulsation signal, motion state signal, and photoelectric pulse wave signal collected by the three sensors. The pre-processing only includes signal filtering and noise reduction (removing noise generated by interference from the sensor's own circuitry) and format standardization (converting different format electrical signals output by different sensors into a unified data format), without changing the core characteristics of the signal.

[0032] In one embodiment, a wrist-worn smart wearable device (such as a smartwatch) is selected and worn on the user's left wrist where the radial artery pulsation is most prominent, ensuring that the arterial pulsation sensor, accelerometer, and photoelectric pulse wave sensor within the device maintain good contact with the skin. The smart wearable device initializes the parameters of each sensor: the arterial pulsation sensor is set to a sampling frequency of 100 Hz, a signal amplification factor of 1000 times, and a sampling accuracy of 16 bits; the accelerometer is set to a sampling frequency of 100 Hz, a sampling accuracy of 16 bits, and a motion range of ±8 gravitational acceleration; the photoelectric pulse wave sensor is set to emit green light, with a sampling frequency of 100 Hz and a sampling accuracy of 16 bits. After synchronous acquisition is initiated, when the user is walking, the arterial pulsation sensor continuously senses the pulsation of the radial artery and outputs the corresponding arterial pulsation signal; the accelerometer captures motion information such as the amplitude and frequency of the wrist's up-and-down swing during walking and outputs the corresponding motion state signal; the photoelectric pulse wave sensor transmits green light through the wrist skin, capturing changes in light reflection as blood volume changes with the pulse, and outputs the corresponding photoelectric pulse wave signal. The smart wearable device stores the three signals synchronously and removes noise with frequencies higher than 20 Hz through a built-in low-pass filter, converting them into a unified digital signal format before storing them in the device's memory.

[0033] Step 20: Based on the motion amplitude change sequence of the motion state signal, the initial peak feature sequence in the arterial pulsation signal is subjected to interference optimization to obtain the optimized peak feature sequence.

[0034] Optionally, the smart wearable device can extract features from motion state signals to generate a motion amplitude change sequence, clarifying the motion amplitude of the wearer at each time point; at the same time, it can perform peak detection on arterial pulsation signals and extract an initial peak feature sequence, which reflects the sequence data of the change law of the peak position and peak size of arterial pulsation signals over time.

[0035] Furthermore, since the user's movements can interfere with the arterial pulsation sensor, resulting in interference peaks in the initial peak feature sequence that do not correspond to the actual arterial pulsation, the smart wearable device can determine the intensity of motion interference at each time point based on the motion amplitude change sequence. For time points where the interference intensity exceeds a set threshold, the corresponding peaks in the initial peak feature sequence are removed or corrected, ultimately obtaining an optimized peak feature sequence that can accurately reflect the characteristics of the actual arterial pulsation, as described in steps 201 to 204.

[0036] Step 30: Based on the optimized peak feature sequence and the waveform feature sequence of the photoelectric pulse wave signal, perform feature association matching between arterial pulsation and pulse wave to obtain feature association results.

[0037] Optionally, the smart wearable device can extract features from the photoelectric pulse wave signal to generate a waveform feature sequence, clarifying the waveform morphology characteristics of the photoelectric pulse wave signal at each time point. The waveform feature sequence reflects the sequence data that reflects the time-varying pattern of the photoelectric pulse wave signal waveform morphology (such as rising edge slope, falling edge slope, number of peaks, and trough positions).

[0038] Furthermore, the smart wearable device uses the timestamps of each sensor signal acquisition as a synchronization benchmark to time-align the optimized peak feature sequence with the extracted waveform feature sequence, ensuring that the time dimension of the two sequences remains consistent.

[0039] Furthermore, the intelligent wearable device establishes matching rules for each feature parameter in the optimized peak feature sequence and waveform feature sequence based on the physiological correlation between arterial pulsation and pulse wave (both arterial pulsation and pulse wave are driven by cardiac contraction and relaxation and have the same periodic characteristics). Through the matching rules, feature parameter combinations with corresponding physiological significance are selected to obtain feature correlation results, thereby achieving accurate correlation between arterial pulsation features and pulse wave features, as detailed in steps 301 to 304.

[0040] Step 40: Based on the feature association results, analyze the feature matching pairs corresponding to systolic blood pressure and diastolic blood pressure to obtain the systolic-diastolic blood pressure amplitude. Then, measure the blood pressure of the wearer based on the systolic-diastolic blood pressure amplitude to obtain the dynamic blood pressure measurement results.

[0041] Optionally, the smart wearable device analyzes the feature matching pairs corresponding to systolic blood pressure and diastolic blood pressure based on the feature association results to obtain the systolic-diastolic blood pressure amplitude, specifically as described in steps 401 to 405. The feature matching pair corresponding to systolic blood pressure refers to the corresponding combination of optimized peak features and waveform features related to the wearer's physiological mechanism of systolic blood pressure in the feature association results. The feature matching pair corresponding to diastolic blood pressure refers to the corresponding combination of optimized peak features and waveform features related to the wearer's physiological mechanism of diastolic blood pressure in the feature association results. The systolic-diastolic blood pressure amplitude refers to the amplitude data that characterizes the magnitude of systolic and diastolic blood pressure, obtained by analyzing the feature matching pairs corresponding to systolic and diastolic blood pressure.

[0042] Furthermore, based on the systolic and diastolic blood pressure amplitude, the smart wearable device, combined with a preset blood pressure calibration model (which is built based on a large amount of clinical data and is used to convert the systolic and diastolic blood pressure amplitude into accurate blood pressure values), calculates the wearer's systolic and diastolic blood pressure values, thus completing the dynamic blood pressure measurement and finally outputting the dynamic blood pressure measurement results.

[0043] This invention achieves simultaneous acquisition of interference signals and physiological signals through multi-source sensor signal acquisition, extracts real arterial pulsation features by optimizing motion interference, enhances feature reliability by relying on multi-signal feature association matching, and finally achieves blood pressure measurement based on accurate features. This solves the problem of large measurement deviation in dynamic scenarios caused by relying on a single sensor, realizes accurate blood pressure measurement in dynamic scenarios, and improves the reliability and accuracy of dynamic blood pressure monitoring.

[0044] Optionally, the processes of steps 201 to 204 include:

[0045] Step 201: Identify the motion interference time intervals in which the continuous amplitudes exceed the preset motion interference threshold based on the motion amplitude change sequence.

[0046] Optionally, the preset motion interference threshold refers to a pre-stored critical amplitude value used to determine whether motion amplitude will effectively interfere with the arterial pulsation signal. This value is calibrated based on a large amount of experimental data, specifically the minimum motion amplitude that can cause distortion in the signal acquired by the arterial pulsation sensor. Continuous amplitude refers to the set of amplitudes in a motion amplitude change sequence where the motion amplitudes corresponding to multiple adjacent time points all satisfy the same judgment condition. The motion interference time interval refers to the continuous time range formed by the time points in the motion amplitude change sequence where the motion amplitude continuously exceeds the preset motion interference threshold.

[0047] Optionally, the smart wearable device can traverse the sequence of motion amplitude changes node by node in chronological order, and determine whether the motion amplitude corresponding to each time node exceeds the preset motion interference threshold.

[0048] During the traversal, the smart wearable device records the starting time node when the motion amplitude first exceeds the preset motion interference threshold, and continues traversing until the ending time node when the motion amplitude falls below the preset motion interference threshold is reached. The continuous time range from the starting time node to the ending time node is marked as a motion interference time interval. If multiple discontinuous cases of "continuous amplitude exceeding the preset motion interference threshold" occur during the traversal, each independent motion interference time interval is marked sequentially using the above method, ultimately obtaining a set of all motion interference time intervals.

[0049] Step 202: Determine whether the time position of each peak point in the initial peak feature sequence falls within any motion interference time interval, and obtain the subset of peak points affected by motion interference.

[0050] Optionally, a peak point refers to a feature point in the initial peak feature sequence that reflects the local maximum value of the arterial pulsation signal, and each peak point corresponds to a unique time position. The time position refers to the specific moment corresponding to the peak point on the arterial pulsation signal acquisition time axis. The motion-disturbed peak point subset refers to the set of all peak points in the initial peak feature sequence whose time positions fall within any motion-disturbed time interval.

[0051] Optionally, the smart wearable device extracts the time position corresponding to each peak point in the initial peak feature sequence, and simultaneously retrieves all motion interference time intervals obtained in step 201. For the time position of each peak point, the smart wearable device compares whether the time position is between the start and end time nodes of a certain motion interference time interval.

[0052] If the time position of a peak point falls within a motion interference time interval, then the peak point is determined to be a motion-interferenced peak point; if the time position of a peak point does not fall within any motion interference time interval, then the peak point is determined to be an uninterrupted peak point. The smart wearable device can collect all peak points determined to be motion-interferenced and construct a subset of motion-interferenced peak points.

[0053] Step 203: For each disturbed peak point in the subset of peak points affected by motion disturbance, a local time neighborhood window is constructed on the time axis with reference to its adjacent undisturbed peak points.

[0054] Optionally, adjacent undisturbed peak points refer to the two closest undisturbed peak points located before and after a disturbed peak point on the time axis of the initial peak feature sequence. A local time neighborhood window refers to a specific time range defined on the time axis, centered on the disturbed peak point and bounded by the time positions of its adjacent undisturbed peak points. This range covers the local waveform of the arterial pulsation signal corresponding to the disturbed peak point and the normal waveform corresponding to the adjacent undisturbed peak points.

[0055] Optionally, the smart wearable device can, for each disturbed peak point in the subset of disturbed peak points, search for its forward adjacent undisturbed peak point and backward adjacent undisturbed peak point on the time axis in the initial peak feature sequence. The forward adjacent undisturbed peak point is the nearest undisturbed peak point before the disturbed peak point, and the backward adjacent undisturbed peak point is the nearest undisturbed peak point after the disturbed peak point.

[0056] After identifying the two adjacent undisturbed peak points, the smart wearable device extracts the corresponding time positions of these two peak points. The time position of the forward-adjacent undisturbed peak point is used as the starting boundary of the window, and the time position of the backward-adjacent undisturbed peak point is used as the ending boundary of the window. A local time neighborhood window is constructed on the time axis with the current undisturbed peak point as its core. If the undisturbed peak point is located at the beginning of the initial peak feature sequence, and only backward-adjacent undisturbed peak points exist, then the starting time of the arterial pulsation signal acquisition is used as the starting boundary of the window, and the time position of the backward-adjacent undisturbed peak point is used as the ending boundary of the window. If the undisturbed peak point is located at the end of the initial peak feature sequence, and only forward-adjacent undisturbed peak points exist, then the time position of the forward-adjacent undisturbed peak point is used as the starting boundary of the window, and the ending time of the arterial pulsation signal acquisition is used as the ending boundary of the window.

[0057] Step 204: Based on the arterial pulsation signal waveform within each local time neighborhood window, interference optimization is performed to obtain the optimized peak feature sequence.

[0058] Optionally, the interference optimization refers to removing interference components and restoring the true arterial pulsation waveform by using preset waveform repair logic for arterial pulsation signal waveforms that are subject to motion interference within a local time neighborhood window.

[0059] Optionally, the smart wearable device can perform interference optimization based on the arterial pulsation signal waveform within each local time neighborhood window to obtain an optimized peak feature sequence, as described in steps 2041 to 2043.

[0060] This invention uses motion amplitude change sequences as the basis for interference judgment. By accurately identifying the time interval of motion interference, it achieves precise positioning of the motion interference range, avoiding ineffective processing of areas without motion interference and ensuring the targeted and efficient optimization process. Based on time position comparison, it filters a subset of motion-interferenced peak points, achieving accurate differentiation between interference peak points and normal peak points. By constructing a local time neighborhood window with adjacent undisturbed peak points as boundaries, it can retain the normal arterial pulsation waveform characteristics around the interference area to the maximum extent, avoiding optimization distortion caused by deviating from the normal waveform background. Finally, through interference optimization and peak point integration within the local window, it effectively eliminates motion interference components in the initial peak feature sequence, corrects the deviation of interference peak points, and enables the optimized peak feature sequence to truly and accurately reflect the wearer's arterial pulsation physiological characteristics, improving the accuracy and stability of blood pressure measurement.

[0061] Optionally, the processes of steps 2041 to 2043 include:

[0062] Step 2041: Perform morphological consistency verification on the arterial pulsation signal waveform within the local time neighborhood window to identify whether there are non-physiological waveform distortions caused by motion artifacts, and obtain the distortion judgment result.

[0063] Optionally, morphological consistency verification refers to the process by which a smart wearable device compares the arterial pulsation signal waveform within a local time neighborhood window with preset normal arterial pulsation waveform characteristics to determine whether the waveform within the window conforms to the physiological laws of human arterial pulsation. Motion artifacts refer to interference signal components generated when the arterial pulsation sensor collects signals due to the user's movement, which are not corresponding to human physiological activities. Non-physiological waveform distortion refers to abnormal changes in the arterial pulsation signal waveform that deviate from the normal physiological waveform shape due to the influence of motion artifacts, such as abrupt increases or decreases in waveform peak value, waveform period disorder, or abnormal abrupt changes in the slope of the rising or falling edge of the waveform.

[0064] Optionally, the smart wearable device can acquire preset normal arterial pulsation waveform features. These features are constructed based on a large amount of physiological data on arterial pulsation in healthy individuals and include key physiological parameters such as the peak range, period range, rising edge slope range, falling edge slope range, and waveform smoothness threshold of the normal arterial pulsation waveform. These parameters serve as a benchmark for verifying morphological consistency.

[0065] Furthermore, the smart wearable device extracts the corresponding arterial pulsation signal waveform within each local time neighborhood window. For each extracted waveform, its waveform feature parameters are extracted, including the peak value, waveform period, rising edge slope, falling edge slope, and waveform smoothness (waveform smoothness refers to the stability of the continuous change of the waveform curve, which is achieved by judging whether the amplitude change between adjacent data points of the waveform is within a preset stable range).

[0066] Furthermore, the intelligent wearable device compares the extracted waveform feature parameters within the window with preset normal arterial pulsation waveform feature parameters one by one to determine whether each feature parameter is within its corresponding normal range. If at least one feature parameter exceeds the normal range, and this exceedance cannot be explained by normal physiological fluctuations (the normal physiological fluctuation range is ±5% of the preset normal feature parameter range), then it is determined that the arterial pulsation signal waveform within the local time neighborhood window has non-physiological waveform distortion caused by motion artifacts, and a distortion determination result indicating the presence of non-physiological waveform distortion is output; if all feature parameters are within their corresponding normal ranges, or the exceedance is within the normal physiological fluctuation range, then it is determined that there is no non-physiological waveform distortion, and a distortion determination result indicating the absence of non-physiological waveform distortion is output.

[0067] Step 2042: Based on the distortion determination results, the disturbed peak points that are determined to have non-physiological waveform distortion are removed from the initial peak feature sequence to obtain the purified peak feature sequence.

[0068] Optionally, the smart wearable device can associate the distortion determination results of each local time neighborhood window with each disturbed peak point in the subset of motion-disturbed peak points, and clarify the distortion determination results of the local time neighborhood window corresponding to each disturbed peak point.

[0069] Furthermore, for the interference peak points whose distortion determination results indicate the presence of non-physiological waveform distortion, the smart wearable device can perform a removal operation to remove such interference peak points from the initial peak feature sequence; for the interference peak points whose distortion determination results indicate the absence of non-physiological waveform distortion, it is determined that their peak features are not substantially affected by motion artifacts, and the smart wearable device retains their position in the initial peak feature sequence; at the same time, the peak points that were originally retained in the initial peak feature sequence and were not affected by motion interference are retained.

[0070] Furthermore, the smart wearable device reorders and integrates all the remaining peak points (including peak points that are not affected by motion and peak points whose corresponding waveforms are not distorted) according to their corresponding time positions to obtain a purified peak feature sequence.

[0071] Step 2043: Based on the purified peak feature sequence, the peak missing intervals caused by the removal operation are interpolated at equal intervals according to the time interval pattern between adjacent retained peak points to obtain the optimized peak feature sequence.

[0072] Optionally, a peak missing interval refers to a continuous time interval in the purified peak feature sequence where, due to a removal operation, a peak point gap occurs within the time range where the peak points originally existed, and the time span of this gap exceeds a preset normal peak interval threshold. Adjacent retained peak points refer to the two nearest peak points in the purified peak feature sequence before and after the peak missing interval, including peak points unaffected by motion interference and interference peak points whose corresponding waveforms are undistorted. The time interval regularity refers to the periodic regularity of the time intervals between adjacent retained peak points, based on the periodic physiological characteristics of human arterial pulsation. Equal-interval interpolation refers to the operation of evenly distributing the time positions of interpolated peak points within the peak missing interval according to the time interval regularity between adjacent retained peak points, and determining the characteristic parameters of the interpolated peak points based on the peak feature parameters of the adjacent retained peak points.

[0073] Optionally, the smart wearable device can traverse the purification peak feature sequence, calculate the time interval between two adjacent peak points, and compare each time interval with a preset normal peak interval threshold (the normal peak interval threshold is set based on the arterial pulsation cycle corresponding to a normal human heart rate, and is 1.5 times the normal arterial pulsation cycle). If the time interval between a certain adjacent peak point exceeds the preset normal peak interval threshold, it is determined that there is a peak missing interval between the adjacent peak points, and the start time (time position of the previous retained peak point) and end time (time position of the next retained peak point) of the peak missing interval are recorded.

[0074] Furthermore, for each peak missing interval, the smart wearable device can extract the adjacent retained peak points before and after it, calculate the time interval between these two adjacent retained peak points, and based on the time interval and the periodic physiological characteristics of human arterial pulsation, determine the number of peak points that should be supplemented in the peak missing interval (the logic for determining the number of supplemented peak points is: divide the time interval between adjacent retained peak points by the normal arterial pulsation cycle, and take the integer part as the number of supplemented peak points).

[0075] Based on the determined number of supplementary peak points, the time position of each interpolated peak point is allocated at equal time intervals between the start and end times of the peak missing interval. At the same time, feature parameters such as peak size and peak slope of adjacent retained peak points are extracted. Based on the changing trend of the feature parameters of adjacent retained peak points, the feature parameters of each interpolated peak point are determined by linear fitting (linear fitting refers to calculating the feature parameters of the interpolated peak point according to the uniform changing trend based on the feature parameters of adjacent retained peak points).

[0076] Furthermore, the smart wearable device inserts all interpolated peak points into the peak missing intervals corresponding to the purified peak feature sequence according to their time positions, and then re-integrates all the inserted peak points in chronological order to obtain the optimized peak feature sequence.

[0077] This invention uses the arterial pulsation signal waveform within a local time neighborhood window as the analysis object. Through morphological consistency verification, it accurately distinguishes between non-physiological waveform distortions caused by motion artifacts and normal physiological fluctuations, avoiding the misjudgment and rejection of normal physiological fluctuations as interference. This ensures that distorted peak points substantially affected by motion artifacts can be accurately identified. Based on the distortion determination results, targeted removal of interfering peak points effectively eliminates interference components from the initial peak feature sequence, resulting in a preliminarily purified peak feature sequence. For the peak missing intervals generated by the removal operation, equal-interval interpolation is performed based on the time interval pattern of adjacent retained peak points. This ensures that the interpolated peak points match the periodic physiological characteristics of normal arterial pulsation and compensates for the incomplete sequence caused by peak missing points, avoiding the impact of peak missing points on the accuracy of feature association matching and ensuring the accuracy and stability of blood pressure measurement results.

[0078] Optionally, the processes of steps 301 to 304 include:

[0079] Step 301: Determine the time of arterial pulsation event based on the timestamps of each peak point in the optimized peak feature sequence, and determine the time of photoelectric pulse wave event based on the timestamps of each feature point in the waveform feature sequence.

[0080] Optionally, the peak point timestamp refers to the acquisition time stamp corresponding to each peak point in the optimized peak feature sequence. The arterial pulsation event time refers to the time point that can characterize the occurrence of a complete arterial pulsation physiological event, directly determined by the timestamps of each peak point in the optimized peak feature sequence, because the peak point corresponds to the moment of strongest arterial pulsation and can serve as the core characterization moment of the arterial pulsation event. The feature point timestamp refers to the acquisition time stamp corresponding to each feature point (such as peak point, trough point, rising edge inflection point, falling edge inflection point, etc.) in the waveform feature sequence. The photoelectric pulse wave event time refers to the time point that can characterize the occurrence of a complete photoelectric pulse wave physiological event, determined by the timestamps of each feature point in the waveform feature sequence. The timestamp corresponding to the feature point directly related to the arterial pulsation physiological process (such as the peak point of the photoelectric pulse wave) is selected as the photoelectric pulse wave event time.

[0081] Optionally, the smart wearable device can extract the peak point timestamp corresponding to each peak point from the optimized peak feature sequence. Since each peak point corresponds to the strongest pulsation moment of the artery caused by the heart pumping blood, each peak point timestamp can be directly determined as the arterial pulsation event time corresponding to an arterial pulsation event.

[0082] Subsequently, the feature point timestamps corresponding to all feature points are extracted from the waveform feature sequence. Simultaneously, a preset feature point filtering rule is retrieved (this rule is based on the physiological correlation mechanism between arterial pulsation and photoelectric pulse waves, explicitly selecting feature points in the photoelectric pulse wave waveform that have a synchronous physiological correlation with the peak time of arterial pulsation, i.e., the photoelectric pulse wave peak point, because the photoelectric pulse wave peak point corresponds to the moment when the blood volume in the blood vessel is largest, and has a direct physiological conduction correlation with the moment when the arterial pulsation is strongest). Based on the preset feature point filtering rule, the feature point timestamps corresponding to the photoelectric pulse wave peak points are selected from all feature point timestamps, and these feature point timestamps are determined as the photoelectric pulse wave event time corresponding to a single photoelectric pulse wave event.

[0083] Step 302: Construct a time difference absolute value matrix based on the absolute values ​​of the time differences between all time point pairs between the arterial pulsation event time and the photoelectric pulse wave event time.

[0084] Optionally, a time point pair refers to a combination of time points consisting of an arterial pulsation event time from the set of arterial pulsation event time points and an optoelectronic pulse wave event time from the set of optoelectronic pulse wave event time points.

[0085] Optionally, the smart wearable device can determine the number of elements in the arterial pulsation event time set and the photoelectric pulse wave event time set, using each arterial pulsation event time in the arterial pulsation event time set as a row of a matrix and each photoelectric pulse wave event time in the photoelectric pulse wave event time set as a column of a matrix, and determine the number of rows and columns of the absolute value of time difference matrix.

[0086] Furthermore, for each element in the matrix, the smart wearable device extracts the arterial pulsation event time of the corresponding row and the photoelectric pulse wave event time of the corresponding column, calculates the time difference between these two time points (by subtracting the values ​​of the two time points), and then takes the absolute value of the obtained time difference to obtain the absolute value of the time difference corresponding to that element position. This process is repeated to calculate the absolute values ​​of the time differences for all elements in the matrix. After filling all elements, a time difference absolute value matrix is ​​obtained. Each element in the time difference absolute value matrix represents the absolute value of the time difference between the arterial pulsation event time of the corresponding row and the photoelectric pulse wave event time of the corresponding column. This matrix characterizes the relationship between the time intervals of all arterial pulsation event times and all photoelectric pulse wave event times.

[0087] Step 303: Based on the absolute value matrix of time difference, select point pairs whose absolute value of time difference is less than the preset upper limit of physiological propagation delay to obtain candidate feature matching pairs.

[0088] Optionally, the preset upper limit of physiological propagation delay refers to the maximum time delay threshold set based on the physiological conduction mechanism of human arterial pulsation and photoelectric pulse wave. It represents the maximum reasonable range of time difference between the arterial pulsation signal traveling from the heart to the acquisition site and the photoelectric pulse wave signal at the same acquisition site. Time point pairs with time differences exceeding this range are not physiologically correlated. This threshold is calibrated based on a large amount of clinical physiological data. Candidate feature matching pairs refer to time point pairs with an absolute time difference less than the preset upper limit of physiological propagation delay. These time point pairs correspond to arterial pulsation events and photoelectric pulse wave events with a reasonable physiological conduction correlation.

[0089] Optionally, the smart wearable device can iterate through each element in the absolute time difference matrix and compare the absolute time difference of each element with a preset upper limit for physiological propagation delay. If the absolute time difference of an element is less than the preset upper limit for physiological propagation delay, the time point pair corresponding to that element is determined to have a physiological correlation and is marked as a candidate feature matching pair; if the absolute time difference of an element is greater than or equal to the preset upper limit for physiological propagation delay, the time point pair corresponding to that element is determined to have no physiological correlation and is directly removed from the process.

[0090] Furthermore, all time point pairs marked as candidate feature matching pairs are collected to form a candidate feature matching pair set. At the same time, the arterial pulsation event time and photoelectric pulse wave event time corresponding to each candidate feature matching pair are retained, as well as the corresponding local waveforms of the arterial pulsation signal and the photoelectric pulse wave signal (the local waveforms are signal waveform segments centered on the event time and within a preset time length).

[0091] Step 304: Based on candidate feature matching, perform feature association matching between the corresponding local waveforms of arterial pulsation signal and photoelectric pulse wave signal to obtain feature association results.

[0092] Optionally, the smart wearable device can perform feature association matching between the corresponding local waveform of the arterial pulsation signal and the local waveform of the photoelectric pulse wave signal based on each candidate feature matching to obtain the feature association result, as specifically in steps 3041 to 3044.

[0093] This invention is based on the physiological correlation mechanism between arterial pulsation and photoelectric pulse wave. By accurately determining the event time, it achieves the preliminary positioning of the two types of physiological events in the time dimension. The constructed absolute value matrix of time difference comprehensively covers all combinations of the time of the two types of events, ensuring that no potentially related time point pairs are missed. Based on the screening based on the preset upper limit of physiological propagation delay, time point pairs without physiological correlation are efficiently eliminated, which greatly reduces the amount of computation for subsequent matching, while ensuring the physiological rationality of candidate feature matching pairs. Based on the fine matching of local waveforms, the true correlation of candidate feature matching pairs is verified, thereby ensuring the accuracy of dynamic blood pressure measurement.

[0094] Optionally, the processes of steps 3041 to 3044 include:

[0095] Step 3041: Based on each candidate feature matching, perform morphological similarity determination on the local waveform of the corresponding arterial pulsation signal and the local waveform of the photoelectric pulse wave signal, determine whether the two have the same physiological rhythm characteristics, and obtain the morphological consistency determination result.

[0096] Optionally, morphological similarity determination refers to the process of extracting morphological feature parameters from two sets of local waveforms, comparing the degree of fit between the two sets of parameters, and determining whether the two sets of waveforms originate from the same physiological rhythm driven by the heart's pumping action. Homologous physiological rhythm characteristics refer to the common physiological features, such as periodicity and waveform change trends, shared by the local waveforms of arterial pulsation signals and photoelectric pulse wave signals, because both are driven by blood flow fluctuations generated by the contraction and relaxation of the human heart.

[0097] Optionally, the smart wearable device can acquire preset morphological feature parameter extraction rules. These rules specify the core parameters that need to be extracted to characterize the physiological rhythm of the waveform, including waveform period, peak amplitude variation trend, rising edge duration, falling edge duration, number of troughs, and trough spacing. These parameters can accurately reflect the core physiological rhythm characteristics of the waveform. For each candidate feature matching pair in the candidate feature matching pair set, the smart wearable device extracts the morphological feature parameters of the corresponding local waveforms of arterial pulsation signal and photoelectric pulse wave signal. During the extraction process, each morphological feature parameter of each group of waveforms is determined one by one according to preset parameter calculation standards. The waveform period is determined by identifying the time interval between two adjacent peak points in the waveform; the peak amplitude variation trend is determined by fitting the curve of the peak point amplitude changing over time; the rising edge duration is determined by calculating the time span from the trough to the peak; the falling edge duration is determined by calculating the time span from the peak to the next trough; the number of troughs is determined by traversing the waveform to identify the number of local minimum points below the preset amplitude threshold; and the trough spacing is determined by calculating the time interval between two adjacent trough points.

[0098] Furthermore, the morphological feature parameters of the two extracted local waveforms are compared one by one according to their corresponding types, and the matching degree of each set of corresponding parameters is calculated (the matching degree is calculated by taking the absolute value of the difference between the two sets of parameters and the preset parameter tolerance threshold; if the ratio is less than 1, the parameters of that set are considered to be matching). If the matching degree of all morphological feature parameters meets the preset matching condition (the preset matching condition is that more than 80% of the corresponding parameter types match), then the two sets of local waveforms are determined to have homologous physiological rhythm characteristics, and the morphological consistency judgment result of homologous physiological rhythm characteristics is output; if the preset matching condition is not met, then they are determined not to have homologous physiological rhythm characteristics, and the morphological consistency judgment result of not having homologous physiological rhythm characteristics is output.

[0099] Step 3042: Based on the morphological consistency determination result, retain the candidate feature matching pairs that are determined to have homologous physiological rhythm characteristics, and remove the remaining candidate feature matching pairs to obtain the initial screening feature matching pairs.

[0100] Optionally, the smart wearable device associates the morphological consistency determination result of each candidate feature matching pair with each candidate feature matching pair in the candidate feature matching pair set, clarifying the morphological consistency determination conclusion corresponding to each candidate feature matching pair. For candidate feature matching pairs whose corresponding morphological consistency determination result is that they have homologous physiological rhythm features, the smart wearable device retains them and includes them in the subsequent processing scope; for candidate feature matching pairs whose corresponding morphological consistency determination result is that they do not have homologous physiological rhythm features, the smart wearable device performs a rejection operation, removing them from the candidate feature matching pair set, to avoid such interfering matching pairs without morphological association affecting the accuracy of subsequent association results.

[0101] All retained candidate feature matching pairs are collected, sorted and arranged according to the time sequence of their corresponding arterial pulsation events to form a set of initial screening feature matching pairs, and the morphological feature parameters of the arterial pulsation peak point, photoelectric pulse wave feature point and two sets of local waveforms corresponding to each initial screening feature matching pair are retained.

[0102] Step 3043: Based on the sequential relationship between the peak point of arterial pulsation and the photoelectric pulse wave feature point in each initial screening feature matching pair, match pairs that violate the temporal logic of arterial blood flow propagation direction are excluded, and target feature matching pairs are obtained.

[0103] Optionally, the arterial blood flow propagation direction timing logic refers to the timing pattern formed based on the human arterial blood flow propagation mechanism, that is, arterial pulsation is conducted from the heart to the blood vessels throughout the body. At the same acquisition site (such as the wrist), the arterial pulsation event should occur before the photoelectric pulse wave event, and the corresponding arterial pulsation event time should be earlier than the photoelectric pulse wave event time. Therefore, the smart wearable device can retrieve the preset arterial blood flow propagation direction timing logic judgment standard. This standard clearly states that at the acquisition site of the smart wearable device (such as the wrist), the timing relationship that conforms to physiological laws is that the arterial pulsation event time is earlier than the photoelectric pulse wave event time, and the time difference between the two is within a preset reasonable timing interval range (this range is based on the conduction velocity of arterial blood flow at the acquisition site and is a sub-range within the preset upper limit of physiological propagation delay).

[0104] For each initial screening feature matching pair in the initial screening feature matching pair set, the smart wearable device extracts its corresponding arterial pulsation event time and photoelectric pulse wave event time, compares the chronological order of the two times on the time axis, and calculates the time difference between them. If the arterial pulsation event time is earlier than the photoelectric pulse wave event time, and the time difference between the two is within a preset reasonable time interval, then the initial screening feature matching pair is determined to conform to the arterial blood flow propagation direction timing logic, and the matching pair is retained; if the arterial pulsation event time is later than or equal to the photoelectric pulse wave event time, or the time difference between the two exceeds the preset reasonable time interval, then the initial screening feature matching pair is determined to violate the arterial blood flow propagation direction timing logic, and a removal operation is performed to remove it from the initial screening feature matching pair set.

[0105] Collect all initial feature matching pairs that conform to the temporal logic of arterial blood flow propagation direction, sort them in chronological order, and form a target feature matching pair set.

[0106] Step 3044: Each time-series compliant feature matching pair is labeled as an associated event of arterial pulsation and photoelectric pulse wave, and all associated events are integrated to form a structured dataset to obtain the feature association results.

[0107] Optionally, an associated event refers to marking each target feature matching pair as an independent event that characterizes the synchronous physiological process of arterial pulsation and photoelectric pulse wave. Each associated event includes core information such as the arterial pulsation event time, photoelectric pulse wave event time, arterial pulsation peak characteristic parameters, and photoelectric pulse wave characteristic parameters corresponding to the target feature matching pair. Therefore, the smart wearable device can retrieve preset associated event marking rules and structured dataset field specifications. The associated event marking rules specify a unique associated event number for each target feature matching pair, and the structured dataset field specifications specify the core fields to be included and the data format of each field. For each target feature matching pair in the target feature matching pair set, the smart wearable device assigns a unique associated event number according to the associated event marking rules and extracts the core information corresponding to the target feature matching pair, including arterial pulsation event time, arterial pulsation peak amplitude, arterial pulsation waveform period, photoelectric pulse wave event time, photoelectric pulse wave peak amplitude, photoelectric pulse wave waveform period, and the time difference between the two. This information is then filled into the corresponding fields according to the structured dataset field specifications to form a data record for a single associated event. By integrating the data records of all individual related events in chronological order according to the related event numbers, a structured dataset is obtained, which is the feature association result.

[0108] This invention is based on the common physiological mechanism and temporal propagation law of arterial pulsation and photoelectric pulse wave. Through morphological similarity judgment, it accurately eliminates interfering candidate matching pairs with no morphological correlation, ensuring that the retained matching pairs have the commonality of core physiological rhythms. Based on the verification of the temporal logic of arterial blood flow propagation direction, matching pairs that violate physiological laws are eliminated from the time dimension, ensuring the temporal rationality of the matching pairs. This allows the feature association results to truly and comprehensively reflect the synchronous physiological correlation between arterial pulsation and photoelectric pulse wave, ensuring the accuracy and stability of dynamic blood pressure measurement results.

[0109] Optionally, the processes of steps 401 to 405 include:

[0110] Step 401: Based on the timestamp of the peak point of the arterial pulsation signal and the timestamp of the feature point of the photoelectric pulse wave signal corresponding to each feature matching pair in the feature association results, a time-aligned event sequence is obtained.

[0111] Optionally, the smart wearable device extracts the core time information corresponding to each feature matching pair from the feature association results (structured dataset), namely the timestamp of the peak point of the arterial pulsation signal and the timestamp of the feature point of the photoelectric pulse wave signal. Subsequently, the smart wearable device marks each feature matching pair as an independent alignment event according to a preset time alignment benchmark rule. The time alignment benchmark rule explicitly uses the timestamp of the peak point of the arterial pulsation signal as the unified time alignment benchmark (because the peak point of the arterial pulsation signal corresponds to the moment when the arterial pulsation is strongest, which is the starting moment of arterial blood flow propagation and has more stable time benchmark characteristics). The core information contained in each alignment event is: the alignment benchmark time (i.e., the timestamp of the peak point of the arterial pulsation signal), the timestamp of the associated photoelectric pulse wave signal feature point, and other associated information of the corresponding feature matching pair (such as the peak amplitude of the arterial pulsation and the peak amplitude of the photoelectric pulse wave).

[0112] Furthermore, all the marked alignment events are sorted and organized according to the alignment reference time (time stamp of the peak point of the arterial pulsation signal) to form a time-aligned event sequence, ensuring that the time order of adjacent alignment events in the sequence is completely consistent with the physiological time sequence of human arterial pulsation.

[0113] Step 402: Calculate the time delay value between the peak point of the arterial pulsation signal and the feature point of the photoelectric pulse wave signal in each feature matching pair based on the time-aligned event sequence, and obtain the time delay value sequence.

[0114] Optionally, the time delay value refers to the difference between the timestamp of the photoelectric pulse wave signal feature point and the timestamp of the arterial pulsation signal peak point in the same feature matching pair. This value characterizes the time interval from the time of arterial pulsation signal acquisition to the time of photoelectric pulse wave signal generation, reflecting the propagation speed-related characteristics of arterial blood flow at the acquisition site. Therefore, the smart wearable device can obtain a preset time delay value calculation rule. The time delay value calculation rule specifies that the time delay value is calculated as follows: subtract the timestamp of the arterial pulsation signal peak point from the timestamp of the photoelectric pulse wave signal feature point corresponding to the same feature matching pair (because it conforms to the timing logic of arterial blood flow propagation direction, the calculation result is positive and within the preset reasonable timing interval range).

[0115] For each alignment event in the time alignment event sequence, the smart wearable device can extract the corresponding photoelectric pulse wave signal feature point timestamp and arterial pulsation signal peak point timestamp, and calculate the time delay value corresponding to each alignment event one by one according to the preset calculation rules.

[0116] During the calculation process, the validity of each time delay value is verified to ensure that the calculation result is positive and within the preset reasonable time interval range (if a negative value or a value exceeding the reasonable range occurs, it is determined to be an invalid time delay value, and the corresponding alignment event is removed to avoid invalid data affecting subsequent analysis).

[0117] Furthermore, the smart wearable device arranges all valid time delay values ​​that have passed verification according to the order of their corresponding alignment events in the time alignment event sequence to form a time delay value sequence, while retaining the association between each time delay value and its corresponding alignment event.

[0118] Step 403: Based on the time delay value sequence, identify feature matching pairs whose time delay values ​​fall within a preset first physiological interval to obtain systolic candidate matching pairs. The preset first physiological interval corresponds to the typical propagation delay range of arterial blood flow during systole.

[0119] Optionally, the preset first physiological interval refers to the time delay value interval set based on the propagation characteristics of human arterial blood flow during the systolic phase. This interval corresponds to the typical propagation time delay range of arterial blood flow during the systolic phase (cardiac contraction, high-pressure blood jet propagation stage). The interval range is calibrated through a large amount of clinical systolic blood flow propagation data.

[0120] Optionally, the smart wearable device iterates through each time delay value in the time delay value sequence, comparing each time delay value with a preset first physiological interval one by one. During the comparison, it determines whether the current time delay value is greater than or equal to the lower limit of the preset first physiological interval and less than or equal to the upper limit of the preset first physiological interval. If a time delay value is within the preset first physiological interval, the feature matching pair corresponding to the time delay value is determined to be a systolic phase-related matching pair and is marked as a systolic phase candidate matching pair; if a time delay value is not within the preset first physiological interval, the feature matching pair corresponding to the time delay value is determined to be unrelated to the systolic phase physiological process and is not marked as a systolic phase candidate matching pair. All feature matching pairs marked as systolic phase candidate matching pairs are collected to form a systolic phase candidate matching pair set, and the time delay value corresponding to each systolic phase candidate matching pair and the associated arterial pulsation and photoelectric pulse wave feature parameters are retained.

[0121] Step 404: Based on the time delay value sequence, identify feature matching pairs whose time delay values ​​fall within a preset second physiological interval to obtain diastolic candidate matching pairs. The preset second physiological interval corresponds to the typical propagation delay range of arterial blood flow during diastole.

[0122] Optionally, the preset second physiological interval refers to the time delay value interval set based on the propagation characteristics of arterial blood flow during diastole, corresponding to the typical propagation time delay range of arterial blood flow during diastole (cardiac diastole, low-pressure blood return phase), which is calibrated through a large amount of clinical diastolic blood flow propagation data.

[0123] Optionally, the smart wearable device can iterate through each time delay value in the time delay value sequence and compare each time delay value with a preset second physiological interval one by one. During the comparison process, it is determined whether the current time delay value is greater than or equal to the lower limit of the preset second physiological interval and less than or equal to the upper limit of the preset second physiological interval. If a time delay value is within the preset second physiological interval, the feature matching pair corresponding to the time delay value is determined to be a diastolic-related matching pair and is marked as a diastolic candidate matching pair; if a time delay value is not within the preset second physiological interval, the feature matching pair corresponding to the time delay value is determined to be unrelated to the diastolic physiological process and is not marked as a diastolic candidate matching pair. All feature matching pairs marked as diastolic candidate matching pairs are collected to form a diastolic candidate matching pair set, and the time delay value corresponding to each diastolic candidate matching pair and the associated arterial pulsation and photoelectric pulse wave feature parameters are retained.

[0124] Step 405: Determine the amplitude of systolic and diastolic pressure based on candidate matching pairs during systole and diastole.

[0125] Optionally, the smart wearable device can determine the systolic and diastolic pressure characterization amplitude based on the candidate matching pairs during systole and diastole, as described in steps 4051 to 4053.

[0126] This invention is based on the different propagation characteristics of arterial blood flow during systole and diastole, and the physiological mechanism of blood pressure formation. Through time alignment processing, it achieves unified calibration of all feature matching pairs in the time dimension. By standardizing calculations to obtain a time delay value sequence, it accurately captures the temporal variation pattern of arterial blood flow propagation delay. Based on preset physiological intervals, it achieves accurate classification of candidate matching pairs during systole and diastole, effectively isolating matching pairs unrelated to systolic and diastolic blood pressure, ensuring the relevance of subsequent parameter derivation. Based on the classified candidate matching pairs, it derives the systolic and diastolic blood pressure amplitude, achieving an effective correlation between feature matching pairs and blood pressure characteristic parameters. This allows the final systolic and diastolic blood pressure amplitude to accurately map the actual magnitudes of systolic and diastolic blood pressure, thereby improving the accuracy and reliability of dynamic blood pressure measurement.

[0127] Optionally, the process of steps 4051 to 4053 includes:

[0128] Step 4051: Determine the systolic arterial pulsation amplitude sequence based on the amplitude corresponding to the peak point of the arterial pulsation signal in each candidate matching pair during systole.

[0129] Optionally, the amplitude corresponding to the peak point of the arterial pulsation signal refers to the signal intensity value corresponding to the peak point in the local waveform of the arterial pulsation signal. This value is directly related to the degree of expansion of the arterial blood vessels during contraction, and thus reflects the pressure state in the arterial blood vessels during the systolic phase.

[0130] Optionally, the smart wearable device can extract the core feature parameter corresponding to the peak point of the arterial pulsation signal for each candidate systolic pair from the set of candidate systolic pairs. During the extraction process, based on the structured data stored in the feature association results above, the amplitude information of the peak point of the arterial pulsation signal associated with each candidate systolic pair is accurately located, ensuring the uniqueness and accuracy of the extracted data and avoiding confusion with the amplitude information of the candidate diastolic pairs.

[0131] Furthermore, the smart wearable device obtains the alignment reference time (timestamp of the peak point of the arterial pulsation signal) corresponding to each candidate systolic pair based on the time alignment event sequence. Based on the order of the timestamps, the amplitude of the peak point of the arterial pulsation signal corresponding to all extracted candidate systolic pairs is sorted and arranged to ensure that the arrangement order of the amplitudes in the sequence is consistent with the physiological time sequence of arterial pulsation during human systole.

[0132] The peak amplitudes of the sorted systolic arterial pulsation signals are integrated sequentially to form a systolic arterial pulsation amplitude sequence, while preserving the association between each amplitude in the sequence and the corresponding candidate systolic pulsation pair.

[0133] Step 4052: Determine the diastolic arterial pulsation amplitude sequence based on the amplitude corresponding to the peak point of the arterial pulsation signal in each diastolic candidate matching pair.

[0134] Optionally, the smart wearable device can extract the amplitude corresponding to the peak point of the arterial pulsation signal for each diastolic candidate matching pair from the diastolic candidate matching pair set one by one. During the extraction process, relying on the structured data of the feature association results mentioned above, the amplitude information of diastolic and systolic candidate matching pairs is accurately distinguished, and only the amplitude data associated with diastolic candidate matching pairs is extracted to ensure the relevance of the extracted data.

[0135] Furthermore, the smart wearable device obtains the alignment reference time (timestamp of the peak point of the arterial pulsation signal) corresponding to each diastolic candidate matching pair based on the time-aligned event sequence. According to the order of these timestamps, the amplitude of the peak point of the arterial pulsation signal corresponding to all extracted diastolic candidate matching pairs is sorted to ensure that the sequence timing is consistent with the physiological process of diastolic arterial pulsation in the human body.

[0136] Furthermore, the smart wearable device integrates the sorted peak amplitudes of the diastolic arterial pulsation signal in sequence to form a diastolic arterial pulsation amplitude sequence, and retains the association between each amplitude in the sequence and the corresponding diastolic candidate matching pair.

[0137] Step 4053: Based on the systolic arterial pulsation amplitude sequence and the diastolic arterial pulsation amplitude sequence, determine the maximum stable amplitude in the systolic arterial pulsation amplitude sequence and the minimum stable amplitude in the diastolic arterial pulsation amplitude sequence, respectively, to obtain the systolic blood pressure characteristic amplitude and the diastolic blood pressure characteristic amplitude.

[0138] Optionally, the maximum stable amplitude refers to the median value of the amplitude interval with the highest frequency and longest duration in the systolic arterial pulsation amplitude sequence after excluding transient abnormal fluctuations. This amplitude stably reflects the pulsation intensity corresponding to the maximum pressure in the systolic artery. The minimum stable amplitude refers to the median value of the amplitude interval with the highest frequency and longest duration in the diastolic arterial pulsation amplitude sequence after excluding transient abnormal fluctuations. This amplitude stably reflects the pulsation intensity corresponding to the minimum pressure in the diastolic artery. The systolic blood pressure characterization amplitude refers to the characteristic amplitude parameter determined by the maximum stable amplitude in the systolic arterial pulsation amplitude sequence, which can map the actual magnitude of human systolic blood pressure. The diastolic blood pressure characterization amplitude refers to the characteristic amplitude parameter determined by the minimum stable amplitude in the diastolic arterial pulsation amplitude sequence, which can map the actual magnitude of human diastolic blood pressure.

[0139] Optionally, the smart wearable device can obtain a preset rule for eliminating abnormal fluctuation amplitudes. The rule specifies that the criteria for judging abnormal fluctuation amplitudes are: the absolute value of the difference between the amplitude and multiple adjacent amplitudes in the sequence exceeds the preset amplitude fluctuation threshold (the threshold is calibrated based on the normal physiological fluctuation range of human arterial pulsation amplitude). Such amplitudes are instantaneous interference amplitudes caused by non-physiological factors and need to be eliminated.

[0140] Furthermore, the intelligent wearable device performs anomaly removal processing on the systolic arterial pulsation amplitude sequence and the diastolic arterial pulsation amplitude sequence according to the abnormal fluctuation amplitude removal rules: it traverses the systolic arterial pulsation amplitude sequence and removes the instantaneous abnormal fluctuation amplitude to obtain the purified systolic arterial pulsation amplitude sequence; similarly, it traverses the diastolic arterial pulsation amplitude sequence and removes the instantaneous abnormal fluctuation amplitude to obtain the purified diastolic arterial pulsation amplitude sequence.

[0141] Furthermore, the smart wearable device can acquire a preset stable amplitude recognition rule. The rule specifies that the stable amplitude is identified as follows: the amplitude interval of the purified amplitude sequence is statistically analyzed, the frequency of occurrence and duration of amplitude in each amplitude interval are counted (duration refers to the time span corresponding to the continuous amplitude in the amplitude interval), the amplitude interval with the highest frequency and the longest duration is selected as the stable amplitude interval, and the median value of the interval is taken as the stable amplitude.

[0142] Furthermore, the intelligent wearable device identifies the stable amplitude of the purified systolic arterial pulsation amplitude sequence according to the stable amplitude identification rules, determines the maximum stable amplitude, and uses the maximum stable amplitude as the systolic blood pressure characteristic amplitude; it also identifies the stable amplitude of the purified diastolic arterial pulsation amplitude sequence, determines the minimum stable amplitude, and uses the minimum stable amplitude as the diastolic blood pressure characteristic amplitude.

[0143] Furthermore, the smart wearable device integrates the systolic blood pressure amplitude and the diastolic blood pressure amplitude to obtain a complete systolic and diastolic blood pressure amplitude.

[0144] This invention is based on the physiological correlation mechanism between systolic and diastolic arterial pulsation amplitude and blood pressure. By extracting and sorting the systolic and diastolic arterial pulsation amplitude sequences respectively, it achieves precise separation and temporal regularization of systolic and diastolic amplitude information. Through abnormal amplitude elimination and stable amplitude identification, it effectively eliminates amplitude fluctuations caused by non-physiological interference factors, accurately captures stable amplitudes that can truly reflect the maximum systolic pressure and minimum diastolic pressure, and finally determines the systolic and diastolic blood pressure characteristic amplitudes. This ensures that the final systolic and diastolic blood pressure characteristic amplitudes can accurately and stably map the actual systolic and diastolic blood pressure of the human body, guaranteeing the accuracy and stability of dynamic blood pressure measurement.

[0145] Furthermore, the intelligent wearable device provided by the present invention will be described in detail below. The intelligent wearable device described below can be referred to in correspondence with the blood pressure dynamic measurement method based on multi-sensor fusion described above.

[0146] Optional, refer to Figure 2 , Figure 2 This is a schematic diagram of the structure of the smart wearable device provided by the present invention. The smart wearable device includes:

[0147] The signal acquisition module 210 is used to acquire the wearer's arterial pulsation signal, motion state signal and photoelectric pulse wave signal based on the arterial pulsation sensor, acceleration sensor and photoelectric pulse wave sensor respectively;

[0148] The interference optimization module 220 is used to perform interference optimization on the initial peak feature sequence in the arterial pulsation signal based on the motion amplitude change sequence of the motion state signal, so as to obtain the optimized peak feature sequence.

[0149] The feature matching module 230 is used to perform feature association matching between arterial pulsation and pulse wave based on the optimized peak feature sequence and the waveform feature sequence of photoelectric pulse wave signal, and obtain feature association results;

[0150] The blood pressure measurement module 240 is used to analyze the feature matching pairs corresponding to systolic blood pressure and diastolic blood pressure based on the feature association results, obtain the systolic and diastolic blood pressure characteristic amplitude, and measure the blood pressure of the wearer based on the systolic and diastolic blood pressure characteristic amplitude to obtain the dynamic blood pressure measurement results.

[0151] This invention achieves simultaneous acquisition of interference signals and physiological signals through multi-source sensor signal acquisition, extracts real arterial pulsation features by optimizing motion interference, enhances feature reliability by relying on multi-signal feature association matching, and finally achieves blood pressure measurement based on accurate features. This solves the problem of large measurement deviation in dynamic scenarios caused by relying on a single sensor, realizes accurate blood pressure measurement in dynamic scenarios, and improves the reliability and accuracy of dynamic blood pressure monitoring.

[0152] Please see Figure 3 , Figure 3 An embodiment diagram of an electronic device provided in accordance with the present invention. For example... Figure 3 As shown, this embodiment of the invention provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor 320. When the processor 320 executes the computer program 311, it performs the following steps:

[0153] Arterial pulsation signals, motion state signals, and photoelectric pulse wave signals of the wearer are collected based on an arterial pulsation sensor, an accelerometer, and a photoelectric pulse wave sensor, respectively.

[0154] The initial peak feature sequence in the arterial pulsation signal is optimized by using the motion amplitude change sequence of the motion state signal to obtain the optimized peak feature sequence.

[0155] Based on the optimized peak feature sequence and the waveform feature sequence of the photoelectric pulse wave signal, feature association matching between arterial pulsation and pulse wave is performed to obtain feature association results;

[0156] Based on the feature association results analysis, the feature matching pairs corresponding to systolic blood pressure and diastolic blood pressure are analyzed to obtain the systolic and diastolic blood pressure characteristic amplitude. Based on the systolic and diastolic blood pressure characteristic amplitude, the blood pressure of the wearer is measured to obtain the dynamic blood pressure measurement results.

[0157] Please see Figure 4 , Figure 4 An embodiment diagram of a computer-readable storage medium provided in accordance with an embodiment of the present invention is shown. Figure 4As shown, this embodiment provides a computer-readable storage medium 400 on which a computer program 311 is stored. When the computer program 311 is executed by a processor, it performs the following steps:

[0158] Arterial pulsation signals, motion state signals, and photoelectric pulse wave signals of the wearer are collected based on an arterial pulsation sensor, an accelerometer, and a photoelectric pulse wave sensor, respectively.

[0159] The initial peak feature sequence in the arterial pulsation signal is optimized by using the motion amplitude change sequence of the motion state signal to obtain the optimized peak feature sequence.

[0160] Based on the optimized peak feature sequence and the waveform feature sequence of the photoelectric pulse wave signal, feature association matching between arterial pulsation and pulse wave is performed to obtain feature association results;

[0161] Based on the feature association results analysis, the feature matching pairs corresponding to systolic blood pressure and diastolic blood pressure are analyzed to obtain the systolic and diastolic blood pressure characteristic amplitude. Based on the systolic and diastolic blood pressure characteristic amplitude, the blood pressure of the wearer is measured to obtain the dynamic blood pressure measurement results.

[0162] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the multi-sensor fusion-based dynamic blood pressure measurement method provided by the above methods, the method comprising:

[0163] Arterial pulsation signals, motion state signals, and photoelectric pulse wave signals of the wearer are collected based on an arterial pulsation sensor, an accelerometer, and a photoelectric pulse wave sensor, respectively.

[0164] The initial peak feature sequence in the arterial pulsation signal is optimized by using the motion amplitude change sequence of the motion state signal to obtain the optimized peak feature sequence.

[0165] Based on the optimized peak feature sequence and the waveform feature sequence of the photoelectric pulse wave signal, feature association matching between arterial pulsation and pulse wave is performed to obtain feature association results;

[0166] Based on the feature association results analysis, the feature matching pairs corresponding to systolic blood pressure and diastolic blood pressure are analyzed to obtain the systolic and diastolic blood pressure characteristic amplitude. Based on the systolic and diastolic blood pressure characteristic amplitude, the blood pressure of the wearer is measured to obtain the dynamic blood pressure measurement results.

[0167] The above-described embodiments of the smart wearable device are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0168] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0169] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for dynamic blood pressure measurement based on multi-sensor fusion, characterized in that, Applications in smart wearable devices; The smart wearable device integrates an arterial pulsation sensor, an acceleration sensor, and a photoelectric pulse wave sensor; the method includes: The arterial pulsation sensor, the acceleration sensor, and the photoelectric pulse wave sensor respectively collect the wearer's arterial pulsation signal, motion state signal, and photoelectric pulse wave signal; Based on the motion amplitude change sequence of the motion state signal, the initial peak feature sequence in the arterial pulsation signal is subjected to interference optimization to obtain the optimized peak feature sequence. Based on the timestamps of each peak point in the optimized peak feature sequence, the time of the arterial pulsation event is determined, and based on the timestamps of each feature point in the waveform feature sequence of the photoelectric pulse wave signal, the time of the photoelectric pulse wave event is determined. A time difference absolute value matrix is ​​constructed based on the absolute values ​​of the time differences between all time point pairs between the arterial pulsation event time and the photoelectric pulse wave event time. Based on the absolute value matrix of time difference, point pairs with an absolute value of time difference less than the preset upper limit of physiological propagation delay are selected to obtain candidate feature matching pairs; Based on each candidate feature matching, feature association matching of arterial pulsation signal local waveform and photoelectric pulse wave signal local waveform is performed to obtain feature association results; Based on the timestamp of the peak point of the arterial pulsation signal and the timestamp of the feature point of the photoelectric pulse wave signal corresponding to each feature matching pair in the feature association results, a time-aligned event sequence is obtained; Based on the time-aligned event sequence, the time delay value between the peak point of the arterial pulsation signal and the feature point of the photoelectric pulse wave signal in each feature matching pair is calculated to obtain a time delay value sequence. Based on the time delay value sequence, feature matching pairs in which the time delay value is within a preset first physiological interval are identified to obtain systolic candidate matching pairs; the preset first physiological interval corresponds to the typical propagation delay range of arterial blood flow during systole. Based on the time delay value sequence, feature matching pairs in which the time delay value is within a preset second physiological interval are identified to obtain diastolic candidate matching pairs; the preset second physiological interval corresponds to the typical propagation delay range of arterial blood flow during diastole; Based on candidate matching pairs during systole and diastole, the amplitude of systolic and diastolic pressure is determined. Based on the amplitude of the systolic and diastolic pressure characterization, the blood pressure of the wearer is measured to obtain the dynamic blood pressure measurement result.

2. The method for dynamic blood pressure measurement based on multi-sensor fusion according to claim 1, characterized in that, The step of filtering point pairs whose absolute time difference is less than a preset upper limit for physiological propagation delay based on the absolute time difference matrix yields candidate feature matching pairs, including: Based on the matching of each candidate feature, the local waveform of the corresponding arterial pulsation signal and the local waveform of the photoelectric pulse wave signal are judged to determine whether they have the same physiological rhythm features, and the morphological consistency judgment result is obtained. Based on the morphological consistency determination results, candidate feature matching pairs that are determined to have homologous physiological rhythm features are retained, and the remaining candidate feature matching pairs are eliminated to obtain the initial screening feature matching pairs. Based on the sequential relationship between the peak point of arterial pulsation and the feature point of photoelectric pulse wave in each initial screening feature matching pair, matching pairs that violate the temporal logic of arterial blood flow propagation direction are excluded, and target feature matching pairs are obtained. Each time-series compliant feature matching pair is labeled as an associated event of arterial pulsation and photoelectric pulse wave, and all associated events are integrated to form a structured dataset to obtain the feature association results.

3. The method for dynamic blood pressure measurement based on multi-sensor fusion according to claim 1, characterized in that, The determination of the systolic and diastolic pressure characterization amplitude based on systolic candidate matching pairs and diastolic candidate matching pairs includes: Based on the amplitude corresponding to the peak point of the arterial pulsation signal in each candidate matching pair during systole, the systolic arterial pulsation amplitude sequence is determined; Based on the amplitude corresponding to the peak point of the arterial pulsation signal in each candidate matching pair during diastole, the diastolic arterial pulsation amplitude sequence is determined; Based on the systolic arterial pulsation amplitude sequence and the diastolic arterial pulsation amplitude sequence, the maximum stable amplitude in the systolic arterial pulsation amplitude sequence and the minimum stable amplitude in the diastolic arterial pulsation amplitude sequence are determined respectively, to obtain the systolic blood pressure characteristic amplitude and the diastolic blood pressure characteristic amplitude.

4. The method for dynamic blood pressure measurement based on multi-sensor fusion according to claim 1, characterized in that, The motion amplitude change sequence based on the motion state signal is used to optimize the initial peak feature sequence in the arterial pulsation signal to obtain an optimized peak feature sequence, including: Based on the motion amplitude change sequence, identify the motion interference time interval in which consecutive amplitudes exceed a preset motion interference threshold; Determine whether the time position of each peak point in the initial peak feature sequence falls within any motion interference time interval to obtain a subset of motion interference peak points; For each disturbed peak point in the subset of peak points affected by motion disturbance, a local time neighborhood window is constructed on the time axis with reference to its adjacent undisturbed peak points. Interference optimization is performed on the arterial pulsation signal waveform within each local time neighborhood window to obtain the optimized peak feature sequence.

5. The method for dynamic blood pressure measurement based on multi-sensor fusion according to claim 4, characterized in that, The interference optimization based on the arterial pulsation signal waveform within each local time neighborhood window yields an optimized peak feature sequence, including: The morphological consistency of the arterial pulsation signal waveform within each local time neighborhood window is checked to identify whether there are non-physiological waveform distortions caused by motion artifacts, and the distortion judgment result is obtained. Based on the distortion determination results, the disturbed peak points that are determined to have non-physiological waveform distortion will be removed from the initial peak feature sequence to obtain the purified peak feature sequence. Based on the purified peak feature sequence, the peak missing intervals caused by the removal operation are interpolated at equal intervals according to the time interval pattern between adjacent retained peak points to obtain the optimized peak feature sequence.

6. A smart wearable device, characterized in that, The method for dynamic blood pressure measurement based on multi-sensor fusion as described in any one of claims 1 to 5 is applied; the intelligent wearable device integrates an arterial pulsation sensor, an acceleration sensor, and a photoelectric pulse wave sensor; the intelligent wearable device comprises: The signal acquisition module is used to acquire the wearer's arterial pulsation signal, motion state signal, and photoelectric pulse wave signal based on the arterial pulsation sensor, the acceleration sensor, and the photoelectric pulse wave sensor, respectively. An interference optimization module is used to perform interference optimization on the initial peak feature sequence in the arterial pulsation signal based on the motion amplitude change sequence of the motion state signal, so as to obtain an optimized peak feature sequence. The feature matching module is used to perform feature association matching between arterial pulsation and pulse wave based on the optimized peak feature sequence and the waveform feature sequence of the photoelectric pulse wave signal, and obtain feature association results; The blood pressure measurement module is used to analyze the feature matching pairs corresponding to systolic blood pressure and diastolic blood pressure based on the feature association results, obtain the systolic-diastolic blood pressure characteristic amplitude, and measure the blood pressure of the wearer based on the systolic-diastolic blood pressure characteristic amplitude to obtain the dynamic blood pressure measurement result.

7. An electronic device, comprising: Memory, used to store computer software programs; A processor for reading and executing the computer software program, characterized in that, when the processor executes the computer software program, it implements the dynamic blood pressure measurement method based on multi-sensor fusion as described in any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium, wherein a computer software program is stored therein, characterized in that, When the computer software program is executed by the processor, it implements the dynamic blood pressure measurement method based on multi-sensor fusion as described in any one of claims 1 to 5.

Images