A pulse wave processing method based on a wearable intelligent pulse condition instrument

Abstract

The present application relates to medical detection technology, it discloses a kind of based on wearable smart pulse instrument pulse wave processing method, solve the interference of invalid signal generated by pressure switching difficult to eliminate in the conventional pulse wave processing method, resulting in the problem of low accuracy of pulse condition analysis result.In the present application scheme, first, using the wearable smart pulse instrument integrated pulse wave acquisition sensor and pressure detection element, from the beginning of applying buoyancy, pulse wave signal and pressure signal are synchronously collected until the pressure completes buoyancy, medium, and sinking stage switching;Then, the collected pressure data is determined by sudden change determination, and the pressure mutation point is determined;Then, according to the determined pressure mutation point, the collected whole pulse wave is divided into effective pulse wave subsection corresponding to buoyancy, medium force, and sinking force stable period;Finally, the segmented effective pulse wave subsection corresponding to buoyancy, medium force, and sinking force stable period is spliced, and the time sequence and signal amplitude are calibrated to obtain pure pulse wave data.

Description

Technical Field

[0001] This invention relates to medical testing technology, specifically to a pulse wave processing method based on a wearable smart pulse wave analyzer. Background Technology

[0002] In the field of medical testing technology, pulse diagnosis in traditional Chinese medicine is the core method of the four diagnostic methods of "inspection, auscultation, inquiry and palpation". Its diagnostic logic relies on the doctor applying three different levels of force (superficial, medium and deep) to the radial artery, and using the fingers to perceive the rhythm, strength and shape of the pulse, thereby judging the physiological and pathological state of the human body.

[0003] With the development of modern sensing technology and wearable devices, wearable smart pulse diagnostic instruments have emerged. These devices aim to simulate the pressure application process of traditional pulse diagnosis. By integrating pulse wave acquisition sensors and pressure detection elements, they can achieve automated acquisition of multi-dimensional pulse wave data, providing technical support for the objectification and standardization of pulse diagnosis, and demonstrating important application value in the field of auxiliary medical diagnosis.

[0004] However, existing wearable smart pulse wave devices face a key technical problem in actual pulse wave acquisition: invalid signals are generated due to pressure switching. Specifically, to simulate the superficial, medium, and deep pressure changes in traditional pulse diagnosis, the device needs to gradually adjust the pressure applied to the wrist during the acquisition process, transitioning from a buoyant phase to a medium pressure phase, and then from a medium pressure phase to a deep pressure phase. During these two pressure switching transitions, the pressure value undergoes non-stationary dynamic changes. These changes cause the pulse wave acquisition sensor to simultaneously capture a large number of interference signals unrelated to the actual pulse, such as sensor vibration noise caused by pressure fluctuations and instantaneous abnormal pulsation of blood vessels due to sudden pressure changes. These invalid or erroneous signals are directly mixed into the normal pulse wave data, forming a continuous pulse wave sequence that simultaneously contains valid signals and interference signals.

[0005] In existing technologies, conventional pulse wave processing methods mostly focus on simple filtering of the entire pulse wave after acquisition in an attempt to remove high-frequency noise. However, such methods cannot distinguish between interference signals generated by pressure switching and real pulse wave signals. Since the interference signals and real pulse wave signals have some overlap in frequency, amplitude and other characteristics, simple filtering methods not only fail to completely eliminate interference, but may also destroy the detailed features of the real pulse wave.

[0006] Therefore, existing technologies struggle to effectively handle interference signals generated by pressure switching in pulse wave signals, resulting in poor accuracy of analysis results and an inability to provide reliable evidence for medical diagnosis. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to provide a pulse wave processing method based on a wearable smart pulse wave analyzer, which solves the problem that conventional pulse wave processing methods have difficulty in eliminating interference from invalid signals generated by pressure switching, resulting in low accuracy of pulse wave analysis results.

[0008] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows:

[0009] A pulse wave processing method based on a wearable smart pulse oximeter includes the following steps:

[0010] S1. A wearable smart pulse meter that integrates a pulse wave acquisition sensor and a pressure detection element can simultaneously acquire pulse wave signals and pressure signals from the moment buoyancy is applied until the pressure completes the switching between floating, middle and sinking stages.

[0011] S2. Perform abrupt change detection on the collected pressure data to determine the pressure abrupt change points;

[0012] S3. Based on the determined pressure change point, the collected pulse wave is divided into effective pulse wave segments corresponding to the buoyancy, medium force, and sinking stable periods;

[0013] S4. The effective pulse wave segments corresponding to the stable periods of buoyancy, medium force, and sinking are spliced ​​together to obtain pure pulse wave data.

[0014] Furthermore, in step S1, when simultaneously acquiring pulse wave signals and pressure signals, the acquisition time for each of the floating, middle, and sinking stages shall not be less than 10 seconds.

[0015] Furthermore, in step S2, the step of determining abrupt changes in the collected pressure data and identifying pressure abrupt change points includes:

[0016] The collected pressure data is differentiated to obtain the slope data of pressure change. The starting point where the slope changes significantly from a stable state close to 0 is determined as the pre-pressure segmentation mutation point, and the regression point where the slope returns to a stable state from a significant change is determined as the post-pressure segmentation mutation point.

[0017] Furthermore, in step S3, the step of dividing the acquired pulse wave into effective pulse wave segments corresponding to the buoyancy, medium force, and sinking stability periods based on the determined pressure abrupt change points includes:

[0018] The buoyancy stabilization period segment corresponds to the point from the start of data acquisition to the abrupt transition point before the transition from buoyancy to neutral force.

[0019] The segment in the stable period of medium force corresponds to the abrupt change point after the transition from buoyancy to medium force to the abrupt change point before the transition from medium force to sinking force;

[0020] The stable period segment corresponds to the point of abrupt change in the transition from medium force to sinking force until the end of the data collection.

[0021] Furthermore, in step S4, the effective pulse wave segments corresponding to the stable periods of buoyancy, medium force, and sinking are spliced ​​together in the order of buoyancy, medium force, and sinking stages, and the time series and signal amplitude are calibrated during splicing.

[0022] Furthermore, the time series calibration employs a linear interpolation algorithm to insert new data points within the data time interval of adjacent segments to achieve time continuity; the signal amplitude calibration employs a normalization algorithm to map the signal amplitude of each segment to a unified standard range.

[0023] Furthermore, the linear interpolation algorithm includes:

[0024] The pulse wave data interpolation between two adjacent segments is calculated using the following formula:

[0025] ;

[0026] in, This refers to the acquisition time of the last data point in the preceding sub-segment of two adjacent sub-segments. for Pulse wave data collected at all times; This refers to the acquisition time of the first data point in the latter of two adjacent sub-segments. for Pulse wave data collected at all times; for arrive At any point in time between, For calculation Interpolation of pulse wave data corresponding to a given time.

[0027] Furthermore, the normalization algorithm includes:

[0028] Assuming the maximum amplitude of each segment of the pulse wave is The minimum value is The standard amplitude range is For any pulse wave data point amplitude value within each segment Normalization is performed using the following formula:

[0029] ;

[0030] in, for The mapping value.

[0031] The beneficial effects of this invention are:

[0032] (1) Improve the quality of pulse wave data:

[0033] This invention achieves precise segmentation of pulse waves by determining pressure abrupt change points, effectively eliminating invalid interference signals generated during pressure switching transitions and retaining only effective pulse wave segments during the stable periods of floating, middle, and deep pulses. After time series calibration and signal amplitude calibration, the generated pure pulse wave data is free from mixed interference, accurately reflecting the true pulse characteristics and providing a high-quality data foundation for subsequent analysis.

[0034] (2) Improve the accuracy of pulse analysis results:

[0035] Pulse modeling and analysis based on pure pulse wave data can avoid interference from invalid signals in the model learning process, enabling the model to accurately capture pulse characteristics under different pressure stages, reducing the risk of misjudgment caused by mixed data, and the output pulse analysis results are more consistent with the actual physiological and pathological state of the human body, providing a reliable basis for medical diagnosis.

[0036] (3) Achieve deep integration of traditional pulse diagnosis with modern technology:

[0037] This invention simulates the core logic of traditional Chinese medicine pulse diagnosis, which involves sensing pulse characteristics in stages: "superficial, middle, and deep." By synchronously acquiring signals and processing data in stages using a wearable smart pulse diagnostic device, it transforms the subjective experience of traditional pulse diagnosis into an objective and standardized technical process. This process retains the diagnostic essence of traditional pulse diagnosis while overcoming its subjective limitations with the help of modern sensing and data processing technologies, laying the foundation for the widespread application of wearable pulse diagnostic devices in clinical settings. Attached Figure Description

[0038] Figure 1 This is a flowchart of the pulse wave processing method based on a wearable smart pulse oximeter in an embodiment of the present invention. Detailed Implementation

[0039] This invention aims to provide a pulse wave processing method based on a wearable smart pulse wave analyzer, solving the problem of conventional pulse wave processing methods failing to eliminate interference from invalid signals generated by pressure switching, leading to low accuracy in pulse wave analysis results. Its core idea is to address the critical issue of interference signals generated by pressure switching during pulse wave acquisition using a wearable smart pulse wave analyzer. It employs a technical approach that combines simulated traditional pulse diagnosis logic with precise signal processing. Through a closed-loop process of synchronous acquisition, mutation detection, segmented extraction, and recombinant calibration, it achieves interference signal elimination and the generation of pure pulse waves, providing high-quality data support for pulse wave analysis.

[0040] Specifically, firstly, relying on wearable devices that integrate pulse wave acquisition sensors and pressure detection elements, pulse and pressure signals are simultaneously acquired to reproduce the traditional pulse diagnosis process of applying pressure at the superficial, middle, and deep levels, ensuring data integrity. Next, by analyzing the slope changes through differentiation of the pressure data, the abrupt change points before and after pressure switching are accurately determined, clarifying the time boundaries of each stable pressure stage. Then, based on the abrupt change points, the entire pulse wave is segmented, and effective segments of the stable periods of buoyancy, middle force, and sinking force are extracted to completely eliminate invalid interference during the pressure transition stage. Finally, the effective segments are spliced ​​together according to the traditional pulse diagnosis sequence, and a pure pulse wave is generated through time series calibration and signal amplitude calibration.

[0041] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0042] This embodiment provides a pulse wave processing method based on a wearable smart pulse oximeter. See [link to relevant documentation]. Figure 1 It includes the following implementation steps:

[0043] S1. Synchronously acquire pulse wave signals and pressure signals.

[0044] In this step, a wearable smart pulse oximeter with integrated pulse wave acquisition sensor and pressure detection element is used to simultaneously acquire pulse wave signals and pressure signals from the moment buoyancy is applied until the pressure completes the switching between floating, middle and sinking stages.

[0045] In one exemplary implementation, once the user correctly wears the smart pulse oximeter and starts the detection program, the detection process begins with the initial application of buoyancy. A pulse wave acquisition sensor captures the vibrations generated by the pulse at high frequency: the piezoelectric material generates weak charge changes under the influence of pulse vibrations, which are amplified and filtered by a signal conditioning circuit and converted into a stable electrical signal. This signal is then converted into a digital signal in a time sequence by a high-precision A / D conversion chip and stored in the instrument's built-in high-speed cache. Simultaneously, a pressure detection element senses the pressure on the wrist in real time at the same high frequency: the piezoresistive pressure detection element converts the pressure value into an electrical signal based on the characteristic that the resistance value changes with pressure. This signal is then converted into a digital signal by an A / D conversion module and stored synchronously with the pulse wave digital signal. This synchronous acquisition method can comprehensively and meticulously acquire pulse wave and pressure data at different pressure stages, laying a solid data foundation for subsequent in-depth data analysis. Both types of signals are continuously acquired as the pressure gradually adjusts from floating to medium and then to sinking. To obtain sufficient and stable data, each pressure stage is acquired for a relatively long time, for example, approximately 15 seconds per stage.

[0046] S2. Perform abrupt change detection on the collected pressure data to determine the pressure abrupt change point.

[0047] In this step, the pressure abrupt change point is the point where the pressure changes significantly. Identifying the pressure abrupt change point helps in the accurate segmentation of the pulse wave and can effectively eliminate invalid interference signals generated during the pressure switching transition phase.

[0048] In one exemplary implementation, the acquired pressure signal is subjected to abrupt change detection using a digital signal processing algorithm, and the pressure data sequence is differentiated point by point to obtain the slope data of pressure change. In the relatively stable stages of traditional pulse diagnosis (superficial, middle, and deep pulses), the pressure data shows a gentle trend of change, and after differentiation, the slope value approaches 0.

[0049] When the slope begins to change significantly from this stable state (value approaching 0), this initial point of change is identified as the pre-pressure segmentation abrupt change point. Taking the transition from buoyancy to neutral force as an example, during the stable buoyancy phase, the slope remains in a stable state approaching 0. When the force is gradually applied to transition to neutral force, once the slope deviates from 0 and shows a continuous changing trend, this point of deviation is the pre-pressure segmentation abrupt change point. The pulse waves corresponding to the pressure data before this point are all valid buoyancy pulse wave data.

[0050] As pressure continues to change, the slope is in a dynamic adjustment process. When the slope returns from a state of significant change to a stable state (the slope approaches 0 again), this point of return is determined as the pressure-related breakpoint. Continuing with the example of the transition from buoyancy to medium force, during the pressure adjustment process, the slope continues to change until the pressure stabilizes in the medium force stage, and the slope approaches 0 again. The point of return at this time is the pressure-related breakpoint, and the pulse wave corresponding to the pressure data after this point is the valid pulse wave data for medium force.

[0051] Similarly, the pressure transition from neutral force to sinking force is also segmented in the same way. Each time the pulse wave corresponding to the two forces is segmented, it contains two key pressure abrupt change points: one is the starting point where the slope changes from 0, and the other is the return point where the slope tends back to 0. Through the dynamic monitoring and analysis of the pressure signal slope, the start and end positions of different stable pressure stages can be accurately identified, providing a crucial basis for the subsequent accurate segmentation of the pulse wave.

[0052] Those skilled in the art will understand that "approaching 0" means that the difference between the slope value and 0 is less than a preset value, and "significant change" means that the difference between the slope value and 0 is greater than a preset value.

[0053] S3. Segment the pulse wave according to the pressure change point.

[0054] In this step, based on the determined pressure change point, the collected pulse wave is divided into effective pulse wave segments corresponding to the buoyancy, medium force, and sinking stable periods.

[0055] In one exemplary implementation, after determining the pre-pressure segmentation abrupt change point and the post-pressure segmentation abrupt change point, the continuously acquired pulse wave is segmented according to the determined pressure abrupt change point timestamp.

[0056] During the transition from buoyancy to medium force, once the timestamp of the pre-pressure abrupt change point is detected, the pulse wave data continuously collected before this time point is defined as the valid data segment for buoyancy pulse waves. This data segment reflects the pulse wave characteristics under buoyancy stability, eliminating interference from subsequent pressure changes. As the pressure continues to adjust, when the timestamp of the post-pressure abrupt change point is detected, the newly collected pulse wave data after this time point is identified as the valid data segment for medium force pulse waves. This part of the data corresponds to the pulse condition during the medium force stable phase.

[0057] Similarly, during the pressure transition from moderate to heavy force, the data is segmented based on the timestamps of the two pressure abrupt change points. When the pressure abrupt change point before the transition from moderate to heavy force is detected, the pulse wave data before that point is classified as valid data for the moderate force stage; when the pressure abrupt change point is detected after the transition, the pulse wave data collected subsequently belongs to the valid data for the heavy force stage.

[0058] During the aforementioned segmentation process, when transitioning from buoyancy to neutral force, the pulse wave data following the timestamp of the pressure-following abrupt change point constitutes the beginning of the effective data segment for the neutral force pulse wave; when transitioning from neutral force to sinking force, the pulse wave data preceding the timestamp of the pressure-following abrupt change point constitutes the end of the effective data segment for the neutral force pulse wave. Therefore, the complete effective data segment for the neutral force phase is the entire data segment from the timestamp of the pressure-following abrupt change point during the transition from buoyancy to neutral force to the timestamp of the pressure-following abrupt change point during the transition from neutral force to sinking force.

[0059] S4. The effective pulse wave segments obtained from the segmentation are spliced ​​together.

[0060] In this step, the effective pulse wave segments corresponding to the stable periods of buoyancy, medium force, and sinking are spliced ​​together to obtain pure pulse wave data.

[0061] In one exemplary implementation, the pulse wave segments during the stable phases of buoyancy, middle force, and sinking force are spliced ​​together, following the traditional pulse diagnosis sequence of floating, middle, and sinking. During splicing, time series calibration and signal amplitude calibration are also performed.

[0062] First, time series calibration is performed using a linear interpolation algorithm:

[0063] Taking adjacent buoyancy-stable pulse wave segments and medium-force-stable pulse wave segments as examples, if the acquisition time of the last data point in the buoyancy-stable pulse wave segment is... The corresponding pulse wave data value is The acquisition time for the first data point of the pulse wave segment during the stable phase of the medium force is... The corresponding pulse wave data value is .exist arrive Between them, according to the linear interpolation formula:

[0064] ;

[0065] in, for arrive At any point in time between, several new data points are inserted. In this way, adjacent segments are closely connected in time, ensuring the continuity of the reconstructed pulse wave time series, making the changes in the pulse wave smoother in time, and truly reflecting the dynamic changes of the pulse.

[0066] Then perform signal amplitude calibration:

[0067] The amplitude data of each sub-segment of the pulse wave are statistically analyzed to determine a unified amplitude standard range. An amplitude normalization algorithm is then used to adjust the amplitude of each sub-segment signal to this standard range.

[0068] Assuming the maximum amplitude of each segment of the pulse wave is The minimum value is The standard amplitude range is For any pulse wave data point amplitude value within each segment Normalization is performed using a linear transformation formula:

[0069] ;

[0070] in, for The mapping value represents the amplitude value of any pulse wave data point within each segment. Mapping to the standard range ensures the consistency of the reconstructed pulse wave signal amplitude, truly reflects the pulse characteristics, and generates a pure pulse wave.

[0071] In pulse wave modeling, the effective processing of the collected pulse wave data in the early stages, eliminating invalid or erroneous data caused by factors such as pressure switching, results in input pure pulse wave data with higher quality and reliability. Using the characteristic parameters of these pure pulse waves, such as amplitude, frequency, rise time, and fall time, as input, and employing machine learning algorithms to construct a pulse pattern model, significantly improves the model's quality.

[0072] For example, by employing the Support Vector Machine (SVM) algorithm, a large amount of high-quality, clean pulse wave data provides rich and accurate information for model training. The model can more accurately learn the intrinsic relationships between different pulse characteristics, avoiding interference from invalid or erroneous data, thus establishing a high-performing pulse model. This model can more accurately identify different pulse types, providing reliable support for subsequent pulse analysis and medical diagnosis.

[0073] Although embodiments of the present invention have been described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the present invention, and all such changes and alterations shall not depart from the protection scope of the present invention.

Claims

1. A pulse wave processing method based on a wearable smart pulse oximeter, characterized in that, Includes the following steps: S1. A wearable smart pulse meter that integrates a pulse wave acquisition sensor and a pressure detection element can simultaneously acquire pulse wave signals and pressure signals from the moment buoyancy is applied until the pressure completes the switching between floating, middle and sinking stages. S2. Perform abrupt change detection on the collected pressure data to determine the pressure abrupt change points; S3. Based on the determined pressure change point, the collected pulse wave is divided into effective pulse wave segments corresponding to the buoyancy, medium force, and sinking stable periods; S4. The effective pulse wave segments corresponding to the stable periods of buoyancy, medium force, and sinking are spliced ​​together to obtain pure pulse wave data.

2. The pulse wave processing method based on a wearable smart pulse oximeter as described in claim 1, characterized in that, In step S1, when simultaneously acquiring pulse wave signals and pressure signals, the acquisition time for each of the floating, middle, and sinking stages shall not be less than 10 seconds.

3. The pulse wave processing method based on a wearable smart pulse oximeter as described in claim 1, characterized in that, In step S2, the step of determining abrupt changes in the collected pressure data and identifying pressure abrupt change points includes: The collected pressure data is differentiated to obtain the slope data of pressure change. The starting point where the slope changes significantly from a stable state close to 0 is determined as the pre-pressure segmentation mutation point, and the regression point where the slope returns to a stable state from a significant change is determined as the post-pressure segmentation mutation point.

4. The pulse wave processing method based on a wearable smart pulse oximeter as described in claim 3, characterized in that, In step S3, the step of dividing the acquired pulse wave into effective pulse wave segments corresponding to the buoyancy, medium force, and sinking stable periods based on the determined pressure change points includes: The buoyancy stabilization period segment corresponds to the point from the start of data acquisition to the abrupt transition point before the transition from buoyancy to neutral force. The segment in the stable period of medium force corresponds to the abrupt change point after the transition from buoyancy to medium force to the abrupt change point before the transition from medium force to sinking force; The stable period segment corresponds to the point of abrupt change in the transition from medium force to sinking force until the end of the data collection.

5. The pulse wave processing method based on a wearable smart pulse oximeter as described in claim 1, characterized in that, In step S4, the effective pulse wave segments corresponding to the stable periods of buoyancy, medium force, and sinking are spliced ​​together in the order of buoyancy, medium force, and sinking stages, and the time series and signal amplitude are calibrated during splicing.

6. The pulse wave processing method based on a wearable smart pulse oximeter as described in claim 5, characterized in that, The time series calibration uses a linear interpolation algorithm to insert new data points within the data time interval of adjacent segments to achieve time continuity; the signal amplitude calibration uses a normalization algorithm to map the signal amplitude of each segment to a unified standard range.

7. The pulse wave processing method based on a wearable smart pulse oximeter as described in claim 6, characterized in that, The linear interpolation algorithm includes: The pulse wave data interpolation between two adjacent segments is calculated using the following formula: ； in, This refers to the acquisition time of the last data point in the preceding sub-segment of two adjacent sub-segments. for Pulse wave data collected at all times; This refers to the acquisition time of the first data point in the latter of two adjacent sub-segments. for Pulse wave data collected at all times; for arrive At any point in time between, For calculation Interpolation of pulse wave data corresponding to a given time.

8. The pulse wave processing method based on a wearable smart pulse oximeter as described in claim 6, characterized in that, The normalization algorithm includes: Assuming the maximum amplitude of each segment of the pulse wave is The minimum value is The standard amplitude range is For any pulse wave data point amplitude value within each segment Normalization is performed using the following formula: ； in, for The mapping value.

Images