CN118078234A

CN118078234A - PPG signal processing method, device and storage medium

Info

Publication number: CN118078234A
Application number: CN202410516782.8A
Authority: CN
Inventors: 张晓帆
Original assignee: Honor Device Co Ltd
Current assignee: Honor Device Co Ltd
Priority date: 2024-04-28
Filing date: 2024-04-28
Publication date: 2024-05-28
Anticipated expiration: 2044-04-28
Also published as: CN118078234B

Abstract

The application provides a PPG signal processing method, a device and a storage medium. According to the method, the signal quality evaluation result of the ACC signal is used as a condition whether the PPG signal is in suspension or not, the influence of the motion state on the PPG signal is reduced, peak-valley detection, adjacent valley interception, start-stop point alignment and amplitude normalization are sequentially carried out on the PPG signal to obtain a monocycle signal set, elements in the monocycle signal set are matched with a template signal to obtain a monocycle PPG main signal, the PPG main signal is decomposed and fitted to obtain a main wave and a counterpulsation wave of the PPG signal, so that the operation complexity of PPG signal decomposition is simplified, and the accuracy of PPG signal decomposition is improved.

Description

PPG signal processing method, device and storage medium

Technical Field

The present application relates to the field of terminal devices, and in particular, to a PPG signal processing method, device, and storage medium.

Background

Photoplethysmography (PPG) is a non-invasive detection method for detecting blood volume changes in living tissue by means of electro-optical means. Since the intensity of the reflected light is related to the blood flow velocity, the blood flow velocity is affected by the periodic heart rate. Thus, measurements of blood oxygen saturation (SPO 2), heart rate, blood pressure, etc. can be achieved based on the PPG signal.

However, in the wearable device scenario, the PPG signal that is usually collected is weak and the interference is large, so that the PPG signal of a part of people cannot be directly used for the prediction of the relevant index.

Disclosure of Invention

The embodiment of the application provides a PPG signal processing method, PPG signal processing equipment and a storage medium, which aim to improve the precision of PPG signal decomposition, so that the accuracy of PWV value prediction is improved when the decomposed PPG signal is used for subsequent PWV calculation.

In a first aspect, an embodiment of the present application provides a PPG signal processing method, including: acquiring a photoplethysmography PPG signal and an acceleration ACC signal acquired by a wearable device; preprocessing the PPG signal to obtain a filtered signal, wherein the preprocessing comprises performing time stamp synchronous check on the PPG signal based on the ACC signal, performing packet loss detection based on time stamps of adjacent data packets of the PPG signal, and filtering the PPG signal; performing peak-valley detection on the filtered signal, performing secondary processing on the filtered signal based on a detection result to obtain a single-period PPG signal set, wherein the secondary processing comprises performing single-period signal interception on the filtered signal based on the detection result, and performing start-stop point alignment and amplitude normalization on the intercepted signal set; matching each single-period PPG signal in the single-period PPG signal set with each template signal in a preset template signal set, and setting the template signal with the highest matching success times as a single-period PPG main signal, wherein the preset template signal set comprises at least one template signal; decomposing the single-period PPG main signal based on a double Gaussian function model to obtain a parameter set; fitting the parameter set based on a linear least squares regression iterative algorithm to obtain a first Gaussian wave and a second Gaussian wave, wherein the first Gaussian wave is the main wave of the PPG signal, and the second Gaussian wave is the dicrotic wave of the PPG signal.

Illustratively, the sampling frequency of the PPG signal and the ACC signal may each be set to 100Hz and the packet uploads are performed in 100ms packets.

Wherein filtering the PPG signal may be filtering it by a filter. The filter may be a second order Butterworth (Butterworth) filter of 0.5-4 hz, for example.

The PPG signal and the ACC signal can be prevented from being misplaced by performing time stamp synchronous check on the PPG signal based on the ACC signal, that is, performing time alignment check on the PPG signal and the ACC signal, so that the risk of misjudging the user state is reduced.

The time stamps of the adjacent data packets of the PPG signals are used for carrying out packet loss detection on the PPG signals, so that the PPG signals are prevented from being lost, and the accuracy of the subsequent analysis results is further affected.

Thus, the time stamp synchronization check and the packet loss detection are performed based on the ACC signal and the PPG signal. If the ACC signal and the PPG signal are not synchronous, returning an asynchronous error code of the PPG signal and the ACC signal; and if the time stamp of the adjacent data packet of the PPG signal is larger than the specified threshold value, the PPG signal data packet loss is considered to occur, and the PPG packet loss error code is returned. And under the condition that the PPG signal does not lose packets, carrying out band-pass filtering on the PPG signal to obtain a filtered signal. Performing peak-to-valley detection on the filtered signal, and performing secondary processing on the filtered signal based on a detection result to obtain a single-period PPG signal set; matching the single-period PPG signal with a preset template signal to obtain a single-period PPG main signal; and decomposing and fitting the single-period PPG main signal based on the double Gaussian function model and the least square regression iterative algorithm to obtain a first Gaussian wave and a second Gaussian wave, so that the operation complexity of PPG signal decomposition is simplified, and the accuracy of PPG signal decomposition is improved.

According to a first aspect, the acquiring the photoplethysmography PPG signal and the acceleration ACC signal acquired by the wearable device includes: and acquiring the PPG signal and the ACC signal acquired by the wearable equipment in the same state and the same period.

Therefore, the ACC signal and the PPG signal are set to be in the same state and in the same period for data acquisition, so that the motion state detection through the ACC signal can be realized, and whether the PPG signal has motion artifacts or not can be judged.

According to a first aspect, or any implementation of the first aspect above, the ACC signal comprises an acx signal along an X axis, an acy signal along a Y axis and an acz signal along a Z axis; after the acquiring the photoplethysmography PPG signal and the acceleration ACC signal acquired by the wearable device, the method further includes: performing modular value calculation, differential and absolute processing on the AccX signal, the AccY signal and the AccZ signal in sequence to obtain a differential modular value AccSDiff; calculating the times of the nonstandard triaxial orientation and the gradient mutation times in a preset period based on the AccX signal, the AccY signal and the AccZ signal; calculating the average value, the maximum value and the variance in the preset period based on the differential modulus AccSDiff; and re-acquiring the PPG signal and the ACC signal when the frequency of the tri-axial direction substandard, the frequency of the slope mutation, the average value, the maximum value and the variance meet preset suspension conditions.

The preset period may be set according to actual requirements, and may be set to 1 second in general.

For example, based on the differential mode AccSDiff, the step of calculating the average value, the maximum value and the variance in the preset period may be to slide on the differential mode by using a non-overlapping sliding window, and calculate the average value AccSMean _i per second, the maximum value AccSMaxVal _i per second and the variance value AccSVar _i per second for the data in the window until the calculation is completed to obtain an array of characteristic values of N seconds. Wherein the sliding window size is 1 second.

According to the first aspect, or any implementation manner of the first aspect, the preset suspension condition includes: re-acquiring the PPG signal and the ACC signal under the condition that the frequency of the tri-axial direction substandard is larger than a preset direction substandard threshold value; re-acquiring the PPG signal and the ACC signal under the condition that the slope mutation times are larger than a preset mutation threshold value; acquiring a first number of which the average value is larger than a first preset threshold value, a second number of which the maximum value is larger than a second preset threshold value and a third number of which the variance is larger than a third preset threshold value; re-acquiring the PPG signal and the ACC signal if the first number is greater than a first number threshold and the second number is greater than a second number threshold; and re-acquiring the PPG signal and the ACC signal when the first number is greater than a third number threshold and the third number is greater than a fourth number threshold.

For example, if the accumulated three-axis direction does not reach the standard times DirectCount to be more than 5, judging that the wearing posture of the current wearable device does not meet the wearing requirement, stopping the detection, and prompting the user to wear correctly through the wearable device or the mobile phone.

For example, if the accumulated slope mutation times MaxSlopeCount is greater than 8, it is determined that the current physical state of the user does not satisfy the stationary state, the detection is stopped, and the user is prompted to keep stationary through the wearable device or the mobile phone.

Illustratively, a number C1 of AccSMean _i > 0.01G, a number C2 of AccSMaxVal _i > 0.04G, and a number C3 of AccSVar _i > 0.5G in the upper eigenvalue array are obtained. Under the condition that C1 is more than 3 and C2 is more than 5, or C1 is more than 5 and C3 is more than 4, judging that the current physical state of the user does not meet the static state, stopping the detection, and prompting the user to keep static through the wearable device or the mobile phone.

According to the first aspect, or any implementation manner of the first aspect, the performing monocycle signal interception on the filtered signal based on the detection result to obtain a monocycle signal set S1 includes: and based on two adjacent valley points in the filtered signal, carrying out monocycle signal interception on the filtered signal to obtain the monocycle signal set S1.

Exemplary, peak-valley detection is performed on the filtered signal to obtain a peak-valley corresponding to each heart beat period, and single-period signal interception is performed on the filtered signal by taking two adjacent valley points as starting points to obtain a single-period signal set S1. A schematic diagram of a screenshot thereof is shown in fig. 9 below.

According to the first aspect, or any implementation manner of the first aspect, before said matching each single-period PPG signal in the single-period PPG signal set with each template signal in a preset template signal set, setting a template signal with a highest number of successful matches as a single-period PPG main signal, where the preset template signal set includes at least one template signal, the method further includes: sequentially calculating the skewness index and kurtosis index of each monocycle PPG signal in the monocycle PPG signal set; filtering the monocycle PPG signal set based on the skewness index and the kurtosis index of each monocycle PPG signal to obtain a filtered monocycle PPG signal set; and creating or updating the preset template signal set based on the filtered single-period PPG signal set.

According to a first aspect, or any implementation manner of the first aspect, the creating the preset template signal set based on the filtered single period PPG signal set includes: acquiring the number of templates in the preset template signal set; setting a first monocycle PPG signal in the filtered monocycle PPG signal set as a first template signal and adding the first monocycle PPG signal to the preset template signal set under the condition that the template number is equal to 0; and when the template number is smaller than a preset template number threshold, sequentially calculating the similarity between each single-period PPG signal in the filtered single-period PPG signal set and each template signal in the preset template signal set, setting the corresponding single-period PPG signal as a second template signal under the condition that any similarity is larger than a first similarity threshold, and adding the second template signal into the preset template signal set.

Illustratively, the similarity calculation described above uses a DTW distance to evaluate, with a smaller distance indicating a more similar signal.

The preset template number threshold may be set according to actual requirements, and may be generally set to 3.

The first similarity may be set according to actual requirements, and may be generally set to 0.3.

Specifically, when the similarity is greater than the first similarity threshold, setting the corresponding monocycle PPG signal as a template signal, and adding the template signal into a preset template signal set.

According to a first aspect, or any implementation manner of the first aspect, the creating the preset template signal set based on the filtered single period PPG signal set further includes: setting a corresponding monocycle PPG signal as a potential template signal and adding the potential template signal to the preset template signal set under the condition that the template number is larger than or equal to the preset template number threshold and all the similarity is larger than the first similarity threshold; and under the condition that the successful times of matching the potential template signals are larger than other template signals in any preset template signal set, replacing the corresponding template signals with the potential template signals, and clearing the potential template signals in the preset template signal set.

Specifically, the potential template signals are a special preset template signal, and the number of the potential template signals is generally not more than 1, so that the potential template signals are used for complementary detection under the condition that the single-period PPG signal is in a changeable form.

According to a first aspect, or any implementation manner of the first aspect, the updating the preset template signal set based on the filtered single period PPG signal set includes: sequentially calculating the similarity between each single-period PPG signal in the filtered single-period PPG signal set and each template signal in the preset template signal set; under the condition that the similarity is smaller than the first similarity threshold value, adding 1 to the current matching success times of the template signals; and under the condition that the similarity is smaller than a second similarity threshold, updating the current template signal by using the corresponding monocycle PPG signal until the monocycle PPG signals in the filtered monocycle PPG signal set are completely matched.

For example, the updating of the template signal using the single period PPG signal may be to combine the single period PPG signal with the template signal based on an average method, and update the original template signal using the combined signal as a new template signal.

For example, the above-mentioned averaging method may be a method of point-wise averaging the single-period PPG signal with the template signal with reference to the data correspondence rule of (3) in fig. 11 below.

Specifically, the first similarity threshold is greater than the second similarity threshold. The second similarity threshold may be set to 0.1.

According to the first aspect, or any implementation manner of the first aspect, after updating the current template signal with its corresponding monocycle PPG signal if the similarity is smaller than the second similarity threshold, the method further includes: calculating the similarity between the updated template signal and other template signals in a preset template signal set, combining the two corresponding template signals under the condition that any similarity is smaller than the second similarity threshold value, combining the matching success times of the two template signals, and deleting the updated template signal.

According to the first aspect, or any implementation manner of the first aspect, the decomposing the single period PPG main signal based on the dual gaussian function model to obtain a parameter set includes: based on the maximum value S4Max of the monocycle PPG main signal and a first position S4MaxLoc, a second position S4MaxEndLoc with the amplitude S4Max/5 on the right side of the first position is obtained; carrying out difference on the single-period PPG main signal to obtain a difference signal, and obtaining a zero crossing point position S5ZeroLoc of the difference signal; calculating initial parameters of the single period PPG main signal in the dual gaussian function model based on the maximum value S4Max, the first position S4MaxLoc, the second position S4MaxEndLoc and the zero crossing point position S5ZeroLoc; and decomposing the single-period PPG main signal based on the initial parameters and the double Gaussian function model to obtain a parameter set.

According to the first aspect, or any implementation manner of the first aspect, the fitting the parameter set based on the linear least squares regression iterative algorithm, after obtaining the first gaussian wave and the second gaussian wave, includes: and performing feature calculation on the single-period PPG main signal, the first Gaussian wave and the second Gaussian wave, and training a preset PWV regression model by using the obtained feature set.

Illustratively, the resulting feature set includes the S4 total duration T, the first gaussian wave rise time T1, the fatter T, the reflection index RI, the enhancement index AI, the heart rate interval RR, and the K value (mean (PPG)/max (PPG)). And extending the feature set first gaussian on-line time duty cycle T1/T, hardening exponent si=height/-fatt, normalized hardening exponent SI/RR, BMI exponent Weight/Height ², etc.

Specifically, the preset PWV regression model may be trained based on a machine learning algorithm or a multiple regression algorithm.

For example, the machine learning algorithm may include an SVM, logistic regression, or mainstream decision tree model. Such as random forest algorithms, iterative algorithms (AdaBoost), optimized distributed gradient enhancement libraries (XGBoost), gradient lifting machine learning algorithms (CatBoost), gradient lifting decision tree models (LightGBM), and the like.

In a second aspect, an embodiment of the present application provides an electronic device. The electronic device includes: a memory and a processor, the memory and the processor coupled; the memory stores program instructions that, when executed by the processor, cause the electronic device to perform the instructions of the first aspect or of the method in any possible implementation of the first aspect.

In a third aspect, embodiments of the present application provide a computer readable medium storing a computer program comprising instructions for performing the method of the first aspect or any possible implementation of the first aspect.

In a fourth aspect, embodiments of the present application provide a computer program comprising instructions for performing the method of the first aspect or any possible implementation of the first aspect.

In a fifth aspect, an embodiment of the present application provides a chip, where the chip includes a processing circuit and a transceiver pin. Wherein the transceiver pin and the processing circuit communicate with each other via an internal connection path, the processing circuit performing the method of the first aspect or any one of the possible implementation manners of the first aspect to control the receiving pin to receive signals and to control the transmitting pin to transmit signals.

Drawings

FIG. 1 is a schematic diagram of an exemplary illustrated multipoint method measurement PWV;

Fig. 2 is a schematic diagram illustrating PPG signals of different age groups;

FIG. 3 is a schematic diagram of an exemplary illustrated wearable device;

fig. 4 is a schematic diagram of an exemplary illustrated wearable device PWV measurement scenario;

FIG. 5 is a schematic diagram of an exemplary user interface;

Fig. 6 is a schematic diagram of an exemplary PPG signal processing flow shown;

FIG. 7 is a schematic diagram of an exemplary illustrated ACC signal quality assessment;

Fig. 8 is a schematic diagram of exemplary shown PPG signal peak-to-valley identification;

fig. 9 is a schematic diagram of exemplary shown PPG signal segmentation;

fig. 10 is a schematic diagram of exemplary shown PPG signal secondary processing;

FIG. 11 is a schematic diagram of exemplary DTW method similarity calculation;

fig. 12 is a schematic diagram of PPG signal processing results exemplarily shown;

fig. 13 is a software architecture diagram of an exemplary electronic device;

Fig. 14 is a schematic diagram of a hardware structure of an exemplary illustrated electronic device.

Detailed Description

The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

The term "and/or" is herein merely an association relationship describing an associated object, meaning that there may be three relationships, e.g., a and/or B, may represent: a exists alone, A and B exist together, and B exists alone.

The terms first and second and the like in the description and in the claims of embodiments of the application, are used for distinguishing between different objects and not necessarily for describing a particular sequential order of objects. For example, the first target object and the second target object, etc., are used to distinguish between different target objects, and are not used to describe a particular order of target objects.

In embodiments of the application, words such as "exemplary" or "such as" are used to mean serving as an example, instance, or illustration. Any embodiment or design described herein as "exemplary" or "e.g." in an embodiment of the present application is not to be taken as preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "such as" is intended to present related concepts in a concrete fashion.

In the description of the embodiments of the present application, unless otherwise indicated, the meaning of "a plurality" means two or more. For example, the plurality of processing units refers to two or more processing units; the plurality of systems means two or more systems.

Pulse Wave Velocity (PWV) is an important index for evaluating vascular health, and the risk of arteriosclerosis can be evaluated and classified based on PWV. In the field of wearable health monitoring, PWV detection is one of the hot spots.

Currently, the industry mainly adopts an Electrocardiogram (ECG) +ppg scheme to perform PWV calculation, so as to obtain an arteriosclerosis risk assessment result. Wherein, ECG is a kind of electrophysiological activity of recording heart in time unit through chest, need to paste the electrode on the skin surface of the heart part of human body while detecting; PPG is a noninvasive detection method for detecting a change in blood volume in living tissue by means of photoelectric means, and a corresponding sensor is required to be provided at a detected site during detection.

For example, referring to fig. 1, fig. 1 is a schematic diagram of PWV detection by an ecg+ppg scheme. In fig. 1, signals of positions of wrists, arms and hearts of a human body are acquired respectively to obtain an ECG signal and a PPG signal, and then a PWV index is calculated according to the ECG signal and the PPG signal.

However, ECG measurements require electrodes to be placed on the skin surface of the user's chest and require the user to actively trigger, making this solution relatively inconvenient to implement.

Thus, in some possible implementations, PWV detection of a pure PPG scheme is also provided. The pure PPG scheme is further divided into a multi-point PPG scheme and a single-point PPG scheme.

Illustratively, the multipoint PPG scheme generally includes a sensor disposed at two points of a neck thigh or an arm ankle of a detected human body, so as to perform multipoint PPG signal acquisition. The scheme is also relatively complex to implement because of the need to collect data at two or more detection points that are far apart.

Therefore, the single point PPG scheme is the main stream direction of current research. It can be understood that the single point PPG scheme performs PWV calculation based on the main wave and the dicrotic wave detection in the PPG signal and the related feature analysis, so as to obtain the risk assessment result of arteriosclerosis of the human body.

For example, referring to fig. 2, fig. 2 is a schematic diagram of typical PPG signal contrast analysis of different ages acquired by a wearable device.

As can be seen from fig. 2, the young person has excellent physical functions, the contraction and expansion of blood vessels are obvious at each heartbeat, and the wearable device can obviously collect PPG signals and distinguish the main wave and the dicrotic wave in the PPG signals, so that the PWV index can be accurately calculated. The aged has poor body functions, the PPG signals acquired by the wearable equipment are weak and have large interference, and even no obvious dicrotic wave exists in the PPG signals, so that the PWV index calculation error is large.

In view of this, the embodiment of the application provides a PPG signal processing method, which collects PPG signals of a human body through a wearable device, performs data processing and feature calculation on the PPG signals, and uses the calculated features to predict PWV indexes, thereby improving the calculation accuracy of the PWV indexes.

Currently, in order to accurately monitor PWV, heart rate, blood oxygen concentration, blood pressure, etc. of a human body in real time, a wearable device (taking a smart watch as an example) is generally provided with a PPG sensor on a surface contacting with skin of a user, such as the back surface shown in fig. 3, and then accurately monitors PWV, heart rate, blood oxygen concentration, blood pressure, etc. of the human body by using a PWV algorithm, a heart rate algorithm, a blood oxygen concentration algorithm, a blood pressure algorithm, etc. based on information collected by the PPG sensor.

Wherein the PPG sensor comprises a light emitting diode (LIGHT EMITTING diode, LED) and a receiver. For ease of illustration, the PWV is measured as an example. Referring to fig. 4, an exemplary smart watch including the PPG sensor shown in fig. 3 is worn on an arm by a user, and a blood vessel health detection function is turned on. In one possible implementation, to enhance the user experience, a blood vessel detection picture may be displayed on the user interface of the smart watch, along with the text information "being measured," as shown in fig. 4 (1).

Illustratively, when the user interface of the smart watch is in the style shown in fig. 4 (1), the LEDs within the PPG sensor in the smart watch will constantly project light towards the skin, which is absorbed by the blood through the skin tissue, while the receiver receives the reflected light signal. When the reflected light signal is converted into an electrical signal, the absorption of the light is changed due to the pulsation of blood in the artery, so that an Alternating Current (AC) signal can be obtained, and the PPG signal can be obtained by extracting the AC signal. By analyzing the PPG features, the PWV of the user can be determined. Finally, the measured PWV indicator can be displayed on the user interface of the smart watch, as shown in fig. 4 (2).

In an ideal case, the collected PPG signal can accurately detect PWV indexes at each moment of the human body. However, due to complex scenes such as hardware, temperature, motion and the like, the true signal of the PPG is distorted, so that the accuracy of PWV calculation is greatly restricted. Since the PPG signal is particularly sensitive to motion, i.e. the PPG signal is prone to motion artifacts, the PPG signal is noisy and difficult to process, resulting in inaccurate prediction results. Therefore, in the embodiment of the application, a PPG signal processing method is provided to process a PPG signal, and the processed PPG signal is used for subsequent PWV index prediction, so as to obtain an accurate PWV index.

In a possible implementation manner, in order to further improve the use experience of the user, in the embodiment of the present application, PPG data detected by the smart watch and the calculated PWV index may also be sent to the mobile phone of the user, so as to be referred to by the user.

Illustratively, when the user clicks on the vascular health application on the handset, the handset responds to the operation by displaying a vascular health detection interface, such as interface 10a of fig. 5 (1). The interface 10a includes a pulse wave display area 10a-1 and a detection countdown display area 10a-2. At this time, the smart watch starts the blood vessel detection, and the interface is displayed as the interface shown in fig. 4 (1). The smart watch sends the collected PPG data to the mobile phone in real time, and the mobile phone updates the pulse wave display area 10a-1 and updates the display area 10a-2 according to the detected remaining time based on the PPG data received by the mobile phone in the interface 10a.

After the smart watch completes PPG data acquisition and obtains PWV indexes and sends the PWV indexes to the mobile phone, the mobile phone displays a vascular health detection result interface 10b, as shown in (2) of fig. 5. Included in interface 10b are an arteriosclerosis risk level control 10b-1 and a PWV index control 10b-2.

Illustratively, after receiving the PWV index, the mobile phone updates the control 10b-2 according to the PWV index, and updates the control 10b-1 according to the risk level corresponding to the PWV index.

The embodiment of the application collects the PPG signals in the human body through the PPG sensor, processes the PPG signals based on the PPG signal processing method provided by the embodiment of the application, and uses the processed PPG signals for PWV index prediction, thereby obtaining accurate PWV indexes and improving the calculation precision of the PWV indexes.

In order to better understand the PPG signal processing flow of the PPG signal processing method provided by the embodiment of the present application, the following embodiment still takes the smart watch 100 as an example, and the PPG signal processing flow is described with reference to fig. 6 to 12.

As shown in fig. 6, the PPG signal processing method in the embodiment of the present application may be divided into 4 stages (such as a data acquisition stage, a data preprocessing stage, a data processing stage and a feature calculation stage) and 5 modules (such as a signal acquisition module, a signal preprocessing module, a peak-valley detection module, a monocycle signal extraction module, a monocycle signal decomposition module and a feature extraction module) in the implementation process.

Referring to fig. 6, illustratively, during a data acquisition phase, a human body is data acquired by a data acquisition module in a wearable device, the acquired data including, but not limited to, an Accelerometer (ACC) signal and a PPG signal.

In the process of data acquisition of the wearable equipment, motion artifacts are generated due to the possibility of motion or shaking of a human body. Motion artifacts will result in an inability to accurately determine the monocycle PPG signal during subsequent data use, thus requiring identification of such interfering signals and suspension of data acquisition.

For example, if the wearing state of the wearing device is poor, for example, loose wearing or a dial is turned down to generate a gap, the PPG signal may also be disturbed.

For example, referring to fig. 7, in fig. 7, when the user is in a motion state, the tri-axial ACC signal and the PPG signal appear to fluctuate significantly irregularly; and when the user finishes the motion state and enters the static state, the ACC signal and the PPG signal show regular fluctuation. Therefore, when the user is in a motion state, the PPG signal cannot be used for subsequent PWV index calculation, and detection needs to be stopped; after the user enters a static state, the PPG signal can truly reflect the PWV index of the user and can be detected.

In the embodiment of the application, the ACC signals are synchronously acquired in the PPG signal acquisition process, and the obtained ACC signals are used for detecting the motion state of a human body, so that the detection result is used as a condition whether the PPG signals are stopped or not.

For technical details not described in detail in the data acquisition phase, reference may be made to the existing implementation scheme for acquisition of sensor signals, which are not described here in detail.

With continued reference to fig. 6, exemplary operations in the data preprocessing stage include packet loss detection of the PPG signal by a signal preprocessing module in the wearable device, and bandpass filtering the original PPG signal to obtain a filtered signal S1.

In one possible implementation, the sampling frequency of the ACC signal and the PPG signal are both set to 100 hertz (Hz) and the packet upload is performed in 100 millisecond (ms) packets. After the motion state of the ACC signal passes, the PPG signal needs to be preprocessed, where the preprocessing method includes time stamp verification and bandpass filtering.

Wherein the timestamp check is for: 1. and judging whether the PPG signal and the ACC signal are synchronous or not, and 2, judging whether the PPG data lose packets or not.

Illustratively, if the timestamp difference between the PPG signal data packet and the ACC signal data packet is greater than a preset threshold (which may be empirically set, and its value may be set to 200 ms), the PPG signal is deemed to be out of synchronization with the ACC signal, and the user is prompted to re-detect through the interfaces of (1) in fig. 4 and (1) in fig. 5.

For example, if the difference between the time stamp of the current PPG signal data packet and the previous PPG signal data packet is greater than another preset threshold (the preset threshold may be set empirically, and the value of the preset threshold may be set to 150 ms), the PPG signal is deemed to have a packet loss condition, and the user is prompted through the interface in fig. 4 (1) and fig. 5 (1), and detection is performed again.

And under the condition that the PPG signal has no packet loss, carrying out band-pass filtering by adopting a filter to finally obtain a filtered signal PPGFilt. The filter may be a 0.5-4 Hz second order Butterworth filter.

For details of the band-pass filtering stage in the data preprocessing stage, reference may be made to the filtering implementation scheme of the existing filter, which is not described here.

With continued reference to fig. 6, exemplary operations in the data processing stage include peak-to-valley detection by a peak-to-valley detection module in the wearable device, monocycle signal extraction by a monocycle signal extraction module, and monocycle signal decomposition by a monocycle signal decomposition module.

After the PPG signal is filtered, the peak-to-valley detection is performed on the filtered signal PPGFilt in the embodiment of the present application, so as to obtain a peak-to-valley value corresponding to each heart beat period. And carrying out monocycle signal interception on PPGFilt based on adjacent valley points to obtain a plurality of corresponding monocycle PPG signal sets S1.

For example, referring to fig. 8 and 9, after peak-to-valley detection of PPGFilt, peak-to-valley values for each beat period may be obtained. And carrying out monocycle signal interception on PPGFilt according to two adjacent valley values to obtain a plurality of monocycle PPG signal sets S1. The morphology of the monocycle PPG signal set S1 is shown in fig. 10 (1).

Because of the fluctuation of the PPG amplitude in the detection process of the wearable device, the starting point and dead point leveling process is also required to be performed on each single-period PPG signal to obtain a signal set S2 (see (2) in fig. 10), and the normalization process is also required to obtain a signal set S3 (see (3) in fig. 10).

In the scene detected by the wearable device, even if a user does not have obvious movement or shake, the PPG signal is also easily influenced by factors such as wearing positions, wearing tightness, arrhythmia or human body micro-motion, so that larger difference of the PPG signal under different periods is caused, and even the situation of single-period PPG signal identification errors can occur.

Therefore, the embodiment of the application performs main signal extraction based on the signal set S3 by the monocycle signal extraction module to obtain the monocycle PPG main signal S4.

After the main signal S4 is obtained, the main signal S4 needs to be decomposed to obtain main waves and counterpulsation waves as shown in fig. 2 (1).

Illustratively, after obtaining the main signal S4, the main signal S4 is decomposed using a dual gaussian function model, and then fitted using a least squares method (LM algorithm), to obtain a first gaussian wave S6 and a second gaussian wave S7, where the first gaussian wave is a main wave, and the second gaussian wave is a counterpulsation wave.

With continued reference to fig. 6, the operation in the feature extraction stage is, illustratively, feature computation in a feature extraction module in the wearable device based on the obtained primary signal S4, the first gaussian wave, the second gaussian wave, and the user' S basic information.

Illustratively, performing feature computation according to the main signal S4, the first gaussian wave and the second gaussian wave to obtain a first feature; the first features include, but are not limited to, S4 total time duration T, first gaussian wave rise time T1, duplex wave spacing fatter T, reflection index RI, enhancement index AI, heart rate spacing RR, K values.

Illustratively, the base characteristics are derived from the base information, including, but not limited to, height, age, weight.

Illustratively, an extended combined feature (hereinafter referred to as a second feature) is derived from the first feature and the base feature, the second feature including, but not limited to, the first gaussian wave on-line time duty cycle, the hardening exponent SI, the normalized hardening exponent, the BMI exponent, and the like.

With continued reference to fig. 6, exemplary PPG signals collected by different users are formed into a PPG feature set based on the PPG signal processing method according to the embodiment of the present application, then feature screening is performed on the PPG feature set, and the feature set obtained by screening is trained by using a machine learning algorithm or a multiple linear regression algorithm as a PWV regression model, so as to obtain a PWV regression model for predicting PWV indexes. The PWV regression model can be issued to each intelligent wearable device, so that the intelligent wearable device can detect PWV indexes based on the PWV regression model, and the accuracy of the PWV is improved.

For example, the step of performing feature screening may be to screen a feature set with a correlation coefficient between the reference PWV and the calculated feature greater than 0.1 based on Pearson correlation analysis.

Therefore, the PWV regression model is finally obtained through continuous iterative training. The PWV regression model in the embodiment of the application can reduce the problem that the pulse wave characteristics cannot be calculated due to individual difference or wearing position, and greatly improves the accuracy of PWV indexes.

Based on the description of 4 stages of PPG signal processing shown in fig. 6, an implementation procedure of the PPG signal processing method provided by the embodiment of the present application may include:

s101, acquiring a photoplethysmography PPG signal and an acceleration ACC signal acquired by the wearable device.

The ACC signals are triaxial ACC signals, and specifically comprise an AccX signal along an X axis, an AccY signal along a Y axis and an AccZ signal along a Z axis. For example, referring to fig. 7 (1), curve Accx, curve Accy, and curve Accz are the acquired tri-axial ACC signals.

For example, the PPG signal may be acquired by a PPG sensor in the wearable device. Regarding the location of the PPG sensor, it may be on the side that is in contact with the user's skin, such as the back side of a smart watch as shown in fig. 3.

For example, the ACC signal may be acquired by an ACC sensor integrated in the wearable device.

It should be noted that, the PPG signal and the ACC signal collected by the wearable device are collected by the wearable device in the same state and the same period.

In one possible implementation, since the ACC signal is a signal acquired in the same state and in the same period as the PPG signal, the signal quality of the ACC signal may also be used to evaluate the signal quality of the PPG signal.

The quality evaluation process of the ACC signal may include:

s001, carrying out module value calculation, difference and absolute treatment on the AccX signal, the AccY signal and the AccZ signal in sequence to obtain a difference module value AccSDiff;

Wherein, ACC signal modulus AccS is:

；

wherein, differential modulus AccSDiff is:

；

Wherein, N is the number of ACC signal modulus values.

For example, referring to fig. 7 (2) and (3), curve AccSDiff is the calculated differential mode signal and curve PPG signal is the acquired PPG signal. By comparing the differential mode value signal with the PPG signal, the PPG signal is significantly different in different ACC signal states, so that the wearable equipment can be subjected to motion detection based on the ACC signal, and the PPG signal can be subjected to quality evaluation based on the detection result.

S002, calculating the times of the nonstandard triaxial orientation and the abrupt gradient times in a preset period based on the AccX signal, the AccY signal and the AccZ signal;

the triaxial direction substandard times DirectCount and the slope abrupt change times MaxSlopeCount are used for identifying whether the dial plate faces upwards when the user wears the watch, and the requirement of tight wearing is met, so that the condition of PPG signal quality reduction caused by incorrect wearing posture is reduced.

For example, the number of times of the non-standard of the triaxial orientation in the preset period may be the number of times of the non-standard of the triaxial orientation in each second, wherein the judging condition of the non-standard may be Abs (mean (AccX _j)) <0.5G, abs (mean (AccY _j)) <0.5G, and Abs (mean (AccZ _j)) >0.8G, j is the ACC data index of each axis corresponding to the ith second. When the AccX signal, the AccY signal and the AccZ signal do not meet the conditions, determining that the triaxial ACC signal does not meet the forward wearing conditions in the current time, and adding 1 to the number of times of the substandard triaxial orientation in each second. Where G represents 1 gravitational acceleration, g=9.8 m/s ².

For example, the number of slope abrupt changes in the preset period may be that any one of the maximum or minimum slope values of the tri-axial ACC signal in each second is greater than the preset slope threshold value, and then it is determined that the ACC signal is abrupt in the current time, and the number of slope abrupt changes in each second is increased by 1. Wherein the preset slope threshold may be set to 0.05G.

S003, calculating the average value, the maximum value and the variance in the preset period based on the differential modulus AccSDiff;

Specifically, sliding is performed on the differential model AccSDiff by using a non-overlapping sliding window, and an average value AccSMean _i per second, a maximum value AccSMaxVal _i per second and a variance value AccSVar _i per second are calculated on the data in the window until the calculation is completed to obtain a continuous N-second characteristic value array. Wherein the sliding window size is 1 second.

S004, re-acquiring the PPG signal and the ACC signal when the frequency of the non-standard orientation of the triaxial, the frequency of the slope mutation, the average value, the maximum value and the variance meet a preset suspension condition.

Specifically, the suspension conditions include:

condition 1: if the accumulated times DirectCount of the three-axis directions are not up to the standard, judging that the wearing posture of the current wearable device does not meet the wearing requirement, stopping the detection, and prompting the user to wear correctly through the wearable device or the mobile phone.

And 2, if the accumulated slope mutation times MaxSlopeCount are more than 8, judging that the current physical state of the user does not meet the static state, stopping the detection, and prompting the user to keep static through the wearable equipment or the mobile phone.

And 3, acquiring the number C1 of AccSMean _i to be more than 0.01G, the number C2 of AccSMaxVal _i to be more than 0.04G and the number C3 of AccSVar _i to be more than 0.5G in the upper characteristic value array. Under the condition that C1 is more than 3 and C2 is more than 5, or C1 is more than 5 and C3 is more than 4, judging that the current physical state of the user does not meet the static state, stopping the detection, and prompting the user to keep static through the wearable device or the mobile phone.

According to the embodiment of the application, the quality of the PPG signal is predicted based on the quality evaluation result of the ACC signal, so that the problem that the single-period PPG signal cannot be accurately positioned in the subsequent calculation process is avoided.

S102, preprocessing the PPG signal to obtain a filtered signal, wherein the preprocessing comprises performing time stamp synchronous verification on the PPG signal based on the ACC signal, performing packet loss detection based on time stamps of adjacent data packets of the PPG signal and filtering the PPG signal.

Specifically, the time stamp of the PPG signal and the time stamp of the ACC signal are checked, whether the PPG signal and the ACC data are synchronous or not is judged, and whether the PPG signal loses packets or not is judged according to the time stamp of the necklace data packet of the PPG signal.

Illustratively, if the timestamp difference between the PPG signal and the ACC signal is > 200ms, the PPG signal and the ACC signal are considered unsynchronized, a data anomaly cue is returned, and the PPG signal and the ACC signal are re-acquired.

Illustratively, if the timestamp difference of adjacent PPG signal data packets is greater than 150ms, the PPG data is considered to have packet loss, a data anomaly prompt is returned, and the PPG signal and the ACC signal are collected again.

And when the situation that the PPG signal has no packet loss is judged, performing band-pass filtering by adopting a second-order Butterworth filter with the frequency of 0.5-4 Hz, and finally obtaining a filtering signal PPGFilt.

S103, peak-valley detection is carried out on the filtered signals, secondary processing is carried out on the filtered signals based on detection results to obtain single-period PPG signal sets, the secondary processing comprises single-period signal interception is carried out on the filtered signals based on the detection results, and start-stop point alignment and amplitude normalization are carried out on the intercepted signal sets.

For example, the peak-to-valley detection is performed on the filtered signal PPGFilt, so as to obtain the peak-to-valley value corresponding to each beat period, as shown in fig. 8.

Illustratively, according to the detection of the peak-valley value corresponding to each heart beat period of the filtered signal PPGFilt, monocycle signal interception is performed by taking two adjacent valley points as starting points (an example of which is shown in fig. 9), so as to obtain a plurality of monocycle PPG signal sets S1, an example of which is shown in fig. 10 (1).

Since the PPG amplitude fluctuates during wearing of the wearing device, it is also necessary to perform start-stop alignment processing on each single-period PPG signal S1, respectively, to obtain a signal S2, where the waveform of S2 is shown in (2) in fig. 10.

Exemplary, the pull-up processing method is as follows: taking a monocycle signal S with length N:

。

Setting:

，

and then modifying the signal in the step S1 point by point to obtain a signal S2, wherein each element in the S2 is respectively as follows:

。

illustratively, N may be the number of signal points in the signal S1.

And then carrying out amplitude normalization on the signal S2 subjected to the pull-up processing, normalizing to [0,1] to obtain a single-period PPG signal set S3, wherein the waveform of the S3 is shown in (3) in fig. 10.

Wherein, each element in the signal set S3 is respectively:

。

Wherein, N is the length of the monocycle signal, data _Min is the minimum value in the signal set S3, and data _Max is the maximum value in the signal set S3.

S104, matching each single-period PPG signal in the single-period PPG signal set with each template signal in a preset template signal set, and setting the template signal with the highest matching success times as a single-period PPG main signal, wherein the preset template signal set comprises at least one template signal.

Specifically, the template signal in the preset template signal set may be preset, or may be set according to the single-period PPG signal S3.

In the process of signal acquisition of the wearable equipment, even if a user does not have large movement or shake, the PPG signal is also easily influenced by factors such as wearing positions, wearing tightness, arrhythmia or human body micro-motion, so that large difference of the PPG signal under different periods can occur, and even the situation of single-period PPG signal identification errors can occur. Therefore, the embodiment of the application also provides a set of single-period PPG main signal extraction flow, which comprises the following specific implementation steps:

s401, sequentially calculating the skewness index and the kurtosis index of each single-period PPG signal in the single-period PPG signal set.

The calculation formula of the skewness index Skew is as follows:

。

the kurtosis index Kur is calculated as follows:

。

wherein μ and σ represent the mean and variance of the monocycle PPG signal, N is the number of elements in the monocycle PPG signal set, and PPG _i is the monocycle PPG signal with sequence number i.

S402, filtering the monocycle PPG signal set based on the skewness index and the kurtosis index of each monocycle PPG signal to obtain the filtered monocycle PPG signal set.

And when the skewness index and the kurtosis index simultaneously meet the preset threshold combination, judging that the current single-period PPG signal is not a noise signal, otherwise, setting the current single-period PPG signal as the noise signal and filtering.

Illustratively, the preset threshold combination is 0.28-0.13, skew, and 2.23-0.15.

S403, creating or updating the preset template signal set based on the filtered monocycle PPG signal set.

The step of creating the template signal in the preset template signal set based on the filtered monocycle PPG signal set comprises the following steps:

s4031, obtaining the number of templates in the preset template signal set;

For example, the preset template signal set may include 4 template signals after updating, which are 3 monocycle template signals and 1 potential template signal, respectively.

S4032, setting a first one of the filtered set of monocycle PPG signals as a first template signal and adding it to the set of preset template signals, if the number of templates is equal to 0.

Specifically, under the condition that no template signal exists in the preset template signal set, setting a first element in the filtered monocycle PPG signal set as a first template signal, and adding the first element into the preset template signal set.

And S4033, when the template number is smaller than a preset template number threshold, sequentially calculating the similarity between each single-period PPG signal in the filtered single-period PPG signal set and each template signal in the preset template signal set, setting the corresponding single-period PPG signal as a second template signal under the condition that any similarity is larger than a first similarity threshold, and adding the second template signal into the preset template signal set.

Specifically, if one or more template signals exist in the preset template signal set at this time and the number of the template signals is less than 3, calculating the similarity between each single-period PPG signal in the filtered single-period PPG signal set and each template signal in the preset template signal set, and setting the corresponding single-period PPG signal as a new template signal and adding the new template signal to the preset template signal set under the condition that the similarity is greater than a first preset similarity threshold.

Illustratively, since the peak-valley points of the PPG signal are not completely aligned in different periods and the lengths of the PPG signal are not completely consistent in different periods, a dynamic time warping algorithm (DYNAMIC TIME WARPING, DTW) is used in the embodiment of the present application to calculate the similarity between the single period PPG signal and the template signal.

For an example, to further simplify complexity of similarity calculation, in the embodiment of the present application, an absolute distance method is used to calculate a distance between two points in the process of calculating the DTW distance, and an example is shown in fig. 11, where { B _i } and { T _j } are sample points of two PPG sequences respectively (i.e., sample points corresponding to the monocycle PPG signal and the template signal). Then the two-point distance D (B _i,T_j)=|B_i-T_j l) gives the DTW distance of the two sets of signals.

For example, the first preset similarity threshold may be set according to the requirement, and the size thereof may be set to 0.3.

Further, the step of creating a template signal in the preset template signal set based on the filtered monocycle PPG signal set further includes:

S4034, setting the corresponding monocycle PPG signal as a potential template signal and adding the potential template signal to the preset template signal set under the condition that the template number is larger than or equal to the preset template number threshold and all the similarity is larger than the first similarity threshold.

Specifically, under the condition that the number of templates is greater than or equal to 3 and the similarity is greater than 0.3, setting the corresponding monocycle PPG signal as a potential template signal, and adding the potential template signal into a preset template signal set to participate in similarity calculation.

The filtering module is configured to filter each single-period PPG signal in the single-period PPG signal set, and to filter each single-period PPG signal in the single-period PPG signal set.

In the embodiment of the present application, after the construction of the preset template signal set is completed, the method further includes a step of updating and matching elements in the single period PPG signal set with elements in the preset template signal set, including:

S404, sequentially calculating the similarity between each single-period PPG signal in the filtered single-period PPG signal set and each template signal in the preset template signal set.

And S405, adding 1 to the current successful times of the template signal matching under the condition that the similarity is smaller than the first similarity threshold value.

And S406, under the condition that the similarity is smaller than a second similarity threshold value, updating the current template signal by using the corresponding monocycle PPG signal until all monocycle PPG signals in the filtered monocycle PPG signal set are matched.

Specifically, the step of updating the template signal by using the single-period PPG signal may be to calculate the single-period PPG signal and the template signal by using an average method. For example, referring to the data correspondence rule of (3) in fig. 11, the single period PPG signal is point-wise averaged with the template signal, the calculated signal is set as a new template signal, and the template signal is replaced.

Illustratively, the first similarity threshold is greater than the second similarity threshold.

For example, the second similarity threshold may be set according to the requirement, and the value thereof may be set to 0.1.

Further, after the step of updating the template signal based on the single-period PPG signal, the method further includes:

s407, calculating the similarity between the updated template signal and other template signals in a preset template signal set, combining the two corresponding template signals under the condition that any similarity is smaller than the second similarity threshold value, combining the matching success times of the two template signals, and deleting the updated template signal.

Specifically, the updated template signal is compared with other template signals in the preset template signal set in similarity. When any similarity is smaller than the second similarity threshold, the updated template signal and other template signals corresponding to the similarity are indicated, the two template signals can be identified as the same template signal, the two template signals are combined by using a mean value method, the successful times of matching of the two template signals are also combined, and the updated template signal is deleted.

Further, after the matching of all the elements in the two sets is completed, the template signal with the largest matching frequency is set as the single-period PPG main signal S4 of the current detection.

According to the embodiment of the application, the single period PPG main signal is extracted by combining the template method and the DTW method, so that the complexity of PPG main signal extraction is reduced and the adaptability in similarity judgment of different-length signals is improved compared with the traditional Pearson correlation coefficient method.

S105, decomposing the single-period PPG main signal based on a double Gaussian function model to obtain a parameter set.

Specifically, after obtaining the PPG main signal S4, the embodiment of the present application decomposes the PPG main signal S4 by using a dual gaussian function model to obtain a parameter set, where the dual gaussian function model is defined as follows:

。

Wherein A1 and A2 are the heights of the single Gaussian function curves respectively, [ mu ] 1 and [ mu ] 2 are the peak positions of the single Gaussian function curves respectively, and [ sigma ] 1 and [ sigma ] 2 are the widths of the single Gaussian function curves respectively.

In order to further reduce the parameter searching calculation amount in the subsequent link, an initial parameter value can be set for the single period PPG signal, and the specific flow is as follows:

S501, identifying a maximum value S4Max of a PPG main signal S4 and a first position S4MaxLoc, and acquiring a second position S4MaxEndLoc with the amplitude S4Max/5 on the right side of the first position;

s502, carrying out difference on the single-period PPG main signal to obtain a difference signal, and obtaining a zero crossing point position S5ZeroLoc of the difference signal;

s503, calculating initial parameters of the single-period PPG main signal in the double Gaussian function model based on the maximum value S4Max, the first position S4MaxLoc, the second position S4MaxEndLoc and the zero crossing point position S5 ZeroLoc;

Specifically, the starting value of [ A1, [ mu ] 1, sigma 1] is set to be [ S4Max/2, S4MaxLoc/3, S4 MaxLoc-2X 5zeroloc ]. The start value of [ A2, [ mu ] 1, sigma 2] is set to [0, S4MaxLoc/2,2 (S4 MaxLoc-2S 5 zeroloc) ].

And decomposing the PPG main signal S4 based on the initial parameters and the double Gaussian function model to obtain a parameter set.

S106, fitting the parameter set based on a linear least squares regression iterative algorithm to obtain a first Gaussian wave and a second Gaussian wave, wherein the first Gaussian wave is the main wave of the PPG signal, and the second Gaussian wave is the dicrotic wave of the PPG signal.

Specifically, an optimal parameter set is found among the above parameter sets based on an LM (Levenberg-Marquardt) algorithm. The LM algorithm is a nonlinear least squares fitting algorithm, among other things, that fits the data by minimizing the sum of squares of the residuals. After the fitting is completed, a first Gaussian wave S6 and a second Gaussian wave S7 are obtained, and S4=S6+S7 is within a certain deviation range.

For an example, the fitted first gaussian wave S6 and second gaussian wave S7, see (1) and (2) in fig. 12 for an example. As can be seen from fig. 12, even in the case where there is no significant dicrotic wave in the PPG signal, the embodiment of the present application can still decompose it to obtain accurate main waves (the first gaussian wave S6) and dicrotic waves (the second gaussian wave S7).

In a possible implementation manner, after the first gaussian wave S6 and the second gaussian wave S7 are obtained, the embodiment of the present application further provides a feature extraction method and a PWV regression model training method, including:

And step S107, performing feature calculation on the single-period PPG main signal, the first Gaussian wave and the second Gaussian wave, and training a preset PWV regression model by using the obtained feature set.

Specifically, feature computation is performed according to the PPG main signal S4, the first gaussian wave S6 and the second gaussian wave S7 to obtain a feature set, which includes, but is not limited to, S4 total duration T, first gaussian wave rise time T1, duplex wave interval T, reflection index RI, enhancement index AI, heart rate interval RR, K value (mean (PPG)/max (PPG)).

Meanwhile, the online time duty ratio T1/T of the first Gaussian wave of the extended combined characteristic is calculated according to the acquired basic characteristics of Height, age and Weight of the user, hardening index SI=height/-father T, normalized hardening index SI/RR, BMI index Weight/Height ² and the like.

Combining the calculated characteristic composition characteristics, then carrying out characteristic screening, and training a characteristic set obtained by screening by using a machine learning algorithm or a multiple linear regression algorithm to obtain a PWV regression model. And predicting the PWV value of the user by using the trained PWV regression model.

For example, the feature screening may be to screen out a feature set with a correlation coefficient between the reference PWV and the calculated feature greater than 0.1 based on the Pearson correlation coefficient method.

Therefore, the PWV regression model is finally obtained through continuous iterative training. The PWV regression model in the embodiment of the application can reduce the problem that the pulse wave characteristics cannot be calculated due to individual difference or wearing position, and greatly improves the accuracy of PWV indexes. Compared with the traditional wave-by-wave PPG decomposition, the embodiment of the application obtains the PPG decomposition waveform by using the single-period PPG main signal through the LM algorithm with the initial value, further simplifies the complexity of operation, and simultaneously, the waveform obtained by decomposition is more similar to the real physiological waveform.

In addition, in order to better understand the technical scheme provided by the embodiment of the application, taking the wearable device as an intelligent watch as an example, the functional module and the hardware related to the implementation of the PPG signal processing method provided by the embodiment of the application and the interaction between the functional module and the hardware are explained based on the relation between the software structure and the hardware of the intelligent watch.

Before explaining the software structure of the wearable device, an architecture that the software system of the wearable device can employ is first explained.

Specifically, in practical applications, the software system of the wearable device may employ a layered architecture, an event driven architecture, a microkernel architecture, a microservice architecture, or a cloud architecture.

Furthermore, it is appreciated that software systems currently in use with the mainstream wearable devices include, but are not limited to, windows systems, android systems, and iOS systems. For convenience of explanation, the embodiment of the application takes an Android system with a layered architecture as an example, and illustrates a software structure of a wearable device.

Furthermore, it should be understood that the PPG signal processing method provided later with respect to the embodiment of the present application is equally applicable to other systems in a specific implementation.

Referring to fig. 13, a block diagram of the software structure and the hardware structure of the wearable device according to the embodiment of the present application is shown.

As shown in fig. 13, the layered architecture of the wearable device divides the software into several layers, each with a clear role and division of work. The layers communicate with each other through a software interface. In some implementations, the Android system may be divided into five layers, from top to bottom, an application layer/application layer (Applications) belonging to the application part, a framework layer/application framework layer (Application Framework, FWK) belonging to the core part, a Runtime (run) and a system library, a hardware abstraction layer (Hardware Abstract Layer, HAL), a Linux Kernel (Linux Kernel) layer belonging to the bottom part.

The application layer may include a series of application packages, among other things. As shown in FIG. 13, the application packages may include camera, game, vascular health, setup, etc. applications, which are not explicitly recited herein, and the present application is not limited in this regard.

Among other things, vascular health applications may be specifically provided for applications for turning on vascular health detection functions.

It will be appreciated that in practical applications, the functions implemented by the vascular health application may be integrated into the sports health application that manages various sports information, or may be integrated into the setup application, as the application is not limited in this regard.

Wherein the framework layer may provide an application programming interface (application programming interface, API) and programming framework for application programs of the application layer. In some implementations, these programming interfaces and programming frameworks can be described as functions. As shown in FIG. 13, the framework layers may include functions of a content provider, a window manager, a view system, a resource manager, etc., which are not to be construed as limiting the application.

It should be noted that, the window manager located in the framework layer is used for managing the window program. The window manager can acquire the size of the display screen, judge whether a status bar exists, lock the screen, intercept the screen and the like.

In addition, it should be noted that the view system located in the frame layer includes visual controls, such as a control for displaying text, a control for displaying pictures, and the like. The view system may be used to build applications. The display interface may be composed of one or more views. For example, the display interfaces including the blood vessel health detection pictures/icons shown in (1) and (2) in fig. 4 may include a view displaying letters and a view displaying pictures.

With continued reference to fig. 13, an exemplary Runtime, specifically An Zhuoyun (Android run), may include a core library and virtual machines, primarily responsible for scheduling and management of the Android system.

The core library comprises two parts: one part is a function which needs to be called by java language, and the other part is a core library of android. The application layer and the framework layer run in virtual machines. The virtual machine executes java files of the application layer and the framework layer as binary files. The virtual machine is used for executing the functions of object life cycle management, stack management, thread management, security and exception management, garbage collection and the like.

With continued reference to FIG. 13, an exemplary system library may include a plurality of functional modules. For example: surface manager (surface manager), media Libraries (Media Libraries), three-dimensional (3D) graphics processing Libraries (e.g., openGL ES), two-dimensional (2D) graphics engines (e.g., SGL), etc.

With continued reference to FIG. 13, the exemplary HAL layer is an interface layer between the operating system kernel (kernel layer) and the hardware circuitry, which aims to isolate the FWK from the kernel so that Android does not depend excessively on the kernel, thereby enabling development of the FWK without regard to drivers.

With continued reference to FIG. 13, exemplary HAL layers may include various interfaces therein, such as an audio-visual interface, a GPS interface, a call interface, a WiFi interface, etc., which are not to be construed as limiting the application.

With continued reference to FIG. 13, the kernel layer in the Android system is illustratively the layer between hardware and software. The kernel layer may include various processes/threads, power management, various drivers, such as WiFi drivers, and the like.

As to the software structure of the wearable device, it is to be understood that the layers and the components contained in the layers in the software structure shown in fig. 13 are not limited to the wearable device. In other embodiments of the application, the wearable device may include more or fewer layers than shown, and more or fewer components may be included in each layer, as the application is not limited.

Based on the software structure of the wearable device shown in fig. 13, when a user triggers a vascular health detection operation through a vascular health application/setting application installed in an application layer, the wearable device is happy in response to the operation, the PPG sensor and the ACC sensor of the hardware part synchronously acquire PPG signals and ACC signals and then give the PPG signals and ACC signals to the processor, and the processor performs preprocessing and feature extraction on the sensor signals according to program instructions stored in the internal memory, such as program instructions for preprocessing and feature extraction on the acquired sensor signals, and further performs PWV prediction on feature information of the extracted sensor signals, and inputs the PWV regression model (program) stored in the memory storage, so that an accurate PWV value can be obtained. Finally, detection of vascular health of the user is achieved based on the PWV value.

For specific logic of the PPG signal processing method provided in the embodiment of the present application, refer to fig. 6, and will not be described herein.

In order to better understand the technical solution provided by the embodiment of the present application, still taking the wearable device as an example of the smart watch, referring to fig. 14, a specific structure of the smart watch and a device involved in implementing the embodiment of the present application are specifically described.

Referring to fig. 14, the wearable device 100 may include: processor 110, external memory interface 120, internal memory 121, universal serial bus (universal serial bus, USB) interface 130, charge management module 140, power management module 141, battery 142, antenna 1, antenna 2, mobile communication module 150, wireless communication module 160, audio module 170, speaker 170A, receiver 170B, microphone 170C, headset interface 170D, sensor module 180, keys 190, motor 191, indicator 192, camera 193, display 194, and subscriber identity module (subscriber identification module, SIM) card interface 195, etc.

In particular, in the technical solution provided in the embodiment of the present application, in order to implement PPG signal processing, the sensor module 180 needs to include a PPG sensor 180A and an ACC sensor 180B. Wherein the PPG sensor may be used to collect PPG signals and the ACC sensor may be used to collect ACC signals.

With continued reference to fig. 14, by way of example, since both PPG sensor 180A and ACC sensor 180B are communicatively coupled to processor 110. Therefore, the PPG signals and ACC signals collected by the PPG sensor 180A and the ACC sensor 180B are processed by the processor 110 according to the processing procedure shown in the embodiment shown in fig. 6, so as to obtain a feature set of the single-period PPG signal, and realize accurate measurement of vascular health.

In addition, in practical application, the sensor 180 may further include a pressure sensor, a gyroscope sensor, a barometric sensor, a magnetic sensor, a distance sensor, a proximity sensor, a fingerprint sensor, a temperature sensor, a touch sensor, an ambient light sensor, a bone conduction sensor, etc., which are not listed here, but the present application is not limited thereto.

Further, it should be noted that the processor 110 may include one or more interfaces. The interfaces may include an integrated circuit (inter-INTEGRATED CIRCUIT, I2C) interface, an integrated circuit built-in audio (inter-INTEGRATED CIRCUIT SOUND, I2S) interface, a pulse code modulation (pulse code modulation, PCM) interface, a universal asynchronous receiver transmitter (universal asynchronous receiver/transmitter, UART) interface, a mobile industry processor interface (mobile industry processor interface, MIPI), a general-purpose input/output (GPIO) interface, a subscriber identity module (subscriber identity module, SIM) interface, and/or a universal serial bus (universal serial bus, USB) interface, among others.

With continued reference to fig. 14, the exemplary charge management module 140 is operable to receive a charge input from a charger. The charger can be a wireless charger or a wired charger.

With continued reference to fig. 14, the wireless communication functions of the wearable device 100 may be implemented by the antenna 1, the antenna 2, the mobile communication module 150, the wireless communication module 160, a modem processor, a baseband processor, and so on, as an example.

Specifically, in the technical solution provided by the embodiment of the present application, feature information obtained after PPG signal processing in the wearable device may be sent to a cloud server, and an electronic device with powerful processing capability, such as a PC, trains the PWV regression model according to the feature information, and issues the trained PWV regression model to the wearable device 100. In this way, the wearable device 100 can communicate with the cloud server through the mobile module 150 or the wireless communication module 160, so as to obtain the PWV regression model.

For example, in other possible implementations, the PWV regression model built by the electronic device such as a PC may also be transmitted to the wearable device 100 through the USB interface 130.

It should be understood that the above description is only an example for better understanding of the technical solution of the present embodiment, and is not the only limitation of the present embodiment.

With continued reference to fig. 14, exemplary display 194 is used to display images, videos, and the like. In some implementations, the wearable device 100 may include 1 or N display screens 194, N being a positive integer greater than 1.

With continued reference to fig. 14, exemplary, a camera 193 is used to capture still images or video.

With continued reference to fig. 14, an exemplary external memory interface 120 may be used to connect an external memory card, such as a Micro SD card, to enable expansion of the memory capabilities of the wearable device 100. The external memory card communicates with the processor 110 through an external memory interface 120 to implement data storage functions. For example, files such as music, video, etc. are stored in an external memory card.

With continued reference to fig. 14, by way of example, the internal memory 121 may be used to store computer executable program code that includes instructions. In this way, the processor 110 executes various functional applications of the wearable device 100 and data processing by executing instructions stored in the internal memory 121. The internal memory 121 may include a storage program area and a storage data area. The storage program area may store an application program (such as a sound playing function, an image playing function, etc.) required for at least one function of the operating system, etc. The storage data area may store data created during use of the wearable device 100 (e.g., audio data, phonebook, etc.), and so on. In addition, the internal memory 121 may include a high-speed random access memory, and may further include a nonvolatile memory such as at least one magnetic disk storage device, a flash memory device, a universal flash memory (universal flash storage, UFS), and the like.

In particular, in the technical solution provided in the embodiment of the present application, the PWV regression model may be downloaded into the internal memory 121 in advance.

As to the hardware structure of the wearable device 100, it should be understood that the wearable device 100 shown in fig. 14 is only one example, and in a specific implementation, the wearable device 100 may have more or fewer components than shown in the drawings, may combine two or more components, or may have different component configurations. The various components shown in fig. 14 may be implemented in hardware, software, or a combination of hardware and software, including one or more signal processing and/or application specific integrated circuits.

Furthermore, it will be appreciated that the electronic device, in order to achieve the above-described functions, comprises corresponding hardware and/or software modules that perform the respective functions. The present application can be implemented in hardware or a combination of hardware and computer software, in conjunction with the example algorithm steps described in connection with the embodiments disclosed herein. Whether a function is implemented as hardware or computer software driven hardware depends upon the particular application and design constraints imposed on the solution. Those skilled in the art may implement the described functionality using different approaches for each particular application in conjunction with the embodiments, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

In addition, it should be noted that, in an actual application scenario, the PPG signal processing method provided in each of the foregoing embodiments implemented by the electronic device may also be executed by a chip system included in the electronic device, where the chip system may include a processor. The chip system may be coupled to a memory such that the chip system, when running, invokes a computer program stored in the memory, implementing the steps performed by the electronic device described above. The processor in the chip system can be an application processor or a non-application processor.

In addition, an embodiment of the present application further provides a computer readable storage medium, where computer instructions are stored, which when executed on an electronic device, cause the electronic device to perform the above related method steps to implement the PPG signal processing method in the above embodiment.

In addition, the embodiment of the application further provides a computer program product, when the computer program product runs on the electronic device, the electronic device is caused to execute the related steps, so as to realize the PPG signal processing method in the embodiment.

In addition, embodiments of the present application also provide a chip (which may also be a component or module) that may include one or more processing circuits and one or more transceiver pins; the transceiver pin and the processing circuit communicate with each other through an internal connection path, and the processing circuit executes the related method steps to implement the PPG signal processing method in the above embodiment, so as to control the receiving pin to receive signals and control the transmitting pin to transmit signals.

In addition, as can be seen from the above description, the electronic device, the computer-readable storage medium, the computer program product, or the chip provided by the embodiments of the present application are used to perform the corresponding methods provided above, and therefore, the advantages achieved by the embodiments of the present application can refer to the advantages in the corresponding methods provided above, and are not repeated herein.

The above embodiments are only for illustrating the technical solution of the present application, and not for limiting the same; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the application.

Claims

1. A method of PPG signal processing, the method comprising:

acquiring a photoplethysmography PPG signal and an acceleration ACC signal acquired by a wearable device;

Preprocessing the PPG signal to obtain a filtered signal, wherein the preprocessing comprises performing time stamp synchronous check on the PPG signal based on the ACC signal, performing packet loss detection based on time stamps of adjacent data packets of the PPG signal and filtering the PPG signal;

Performing peak-valley detection on the filtered signal, performing secondary processing on the filtered signal based on a detection result to obtain a single-period PPG signal set, wherein the secondary processing comprises performing single-period signal interception on the filtered signal based on the detection result, and performing start-stop point alignment and amplitude normalization on the intercepted signal set;

matching each single-period PPG signal in the single-period PPG signal set with each template signal in a preset template signal set, and setting the template signal with the highest matching success times as a single-period PPG main signal, wherein the preset template signal set comprises at least one template signal;

decomposing the single-period PPG main signal based on a double Gaussian function model to obtain a parameter set;

fitting the parameter set based on a linear least squares regression iterative algorithm to obtain a first Gaussian wave and a second Gaussian wave, wherein the first Gaussian wave is the main wave of the PPG signal, and the second Gaussian wave is the dicrotic wave of the PPG signal.

2. The method of claim 1, wherein the acquiring the photoplethysmography, PPG, signal and the acceleration, ACC, signal acquired by the wearable device comprises:

and acquiring the PPG signal and the ACC signal acquired by the wearable equipment in the same state and the same period.

3. The method of claim 2, wherein the ACC signal comprises an acx signal along an X axis, an acy signal along a Y axis, and an acz signal along a Z axis;

After the acquiring the photoplethysmography PPG signal and the acceleration ACC signal acquired by the wearable device, the method further includes:

performing modular value calculation, differential and absolute processing on the AccX signal, the AccY signal and the AccZ signal in sequence to obtain a differential modular value AccSDiff;

Calculating the times of the nonstandard triaxial orientation and the gradient mutation times in a preset period based on the AccX signal, the AccY signal and the AccZ signal;

Calculating the average value, the maximum value and the variance in the preset period based on the differential modulus AccSDiff;

And re-acquiring the PPG signal and the ACC signal when the frequency of the tri-axial direction substandard, the frequency of the slope mutation, the average value, the maximum value and the variance meet preset suspension conditions.

4. A method according to claim 3, wherein the preset suspension condition comprises:

Re-acquiring the PPG signal and the ACC signal under the condition that the frequency of the tri-axial direction substandard is larger than a preset direction substandard threshold value;

Re-acquiring the PPG signal and the ACC signal under the condition that the slope mutation times are larger than a preset mutation threshold value;

Acquiring a first number of which the average value is larger than a first preset threshold value, a second number of which the maximum value is larger than a second preset threshold value and a third number of which the variance is larger than a third preset threshold value;

Re-acquiring the PPG signal and the ACC signal if the first number is greater than a first number threshold and the second number is greater than a second number threshold;

And re-acquiring the PPG signal and the ACC signal when the first number is greater than a third number threshold and the third number is greater than a fourth number threshold.

5. The method according to claim 1, wherein the performing monocycle signal interception on the filtered signal based on the detection result to obtain a monocycle signal set S1 includes:

And based on two adjacent valley points in the filtered signal, carrying out monocycle signal interception on the filtered signal to obtain the monocycle signal set S1.

6. The method of claim 1, wherein prior to said matching each of the set of monocycle PPG signals with each of a set of preset template signals, setting the template signal with the highest number of successful matches as the monocycle PPG primary signal, the method further comprises:

Sequentially calculating the skewness index and kurtosis index of each monocycle PPG signal in the monocycle PPG signal set;

Filtering the monocycle PPG signal set based on the skewness index and the kurtosis index of each monocycle PPG signal to obtain a filtered monocycle PPG signal set;

and creating or updating the preset template signal set based on the filtered single-period PPG signal set.

7. The method of claim 6, wherein the creating the set of preset template signals based on the filtered set of monocycle PPG signals comprises:

acquiring the number of templates in the preset template signal set;

setting a first monocycle PPG signal in the filtered monocycle PPG signal set as a first template signal and adding the first monocycle PPG signal to the preset template signal set under the condition that the template number is equal to 0;

and when the template number is smaller than a preset template number threshold, sequentially calculating the similarity between each single-period PPG signal in the filtered single-period PPG signal set and each template signal in the preset template signal set, setting the corresponding single-period PPG signal as a second template signal under the condition that any similarity is larger than a first similarity threshold, and adding the second template signal into the preset template signal set.

8. The method of claim 7, wherein the creating the set of preset template signals based on the filtered set of monocycle PPG signals further comprises:

setting a corresponding monocycle PPG signal as a potential template signal and adding the potential template signal to the preset template signal set under the condition that the template number is larger than or equal to the preset template number threshold and all the similarity is larger than the first similarity threshold;

And under the condition that the successful times of matching the potential template signals are larger than other template signals in any preset template signal set, replacing the corresponding template signals with the potential template signals, and clearing the potential template signals in the preset template signal set.

9. The method of claim 6, wherein the updating the set of preset template signals based on the filtered set of monocycle PPG signals comprises:

Sequentially calculating the similarity between each single-period PPG signal in the filtered single-period PPG signal set and each template signal in the preset template signal set;

under the condition that the similarity is smaller than a first similarity threshold value, adding 1 to the current successful times of matching the template signals;

And under the condition that the similarity is smaller than a second similarity threshold, updating the current template signal by using the corresponding monocycle PPG signal until the monocycle PPG signals in the filtered monocycle PPG signal set are completely matched, wherein the first similarity threshold is larger than the second similarity threshold.

10. The method of claim 9, the updating the current template signal with its corresponding monocycle PPG signal if the similarity is less than a second similarity threshold, further comprising:

Calculating the similarity between the updated template signal and other template signals in a preset template signal set, combining the two corresponding template signals under the condition that any similarity is smaller than the second similarity threshold value, combining the matching success times of the two template signals, and deleting the updated template signal.

11. The method according to claim 1, wherein the decomposing the single period PPG main signal based on the dual gaussian function model to obtain a parameter set comprises:

Based on the maximum value S4Max of the monocycle PPG main signal and a first position S4MaxLoc, a second position S4MaxEndLoc with the amplitude S4Max/5 on the right side of the first position is obtained;

Carrying out difference on the single-period PPG main signal to obtain a difference signal, and obtaining a zero crossing point position S5ZeroLoc of the difference signal;

calculating initial parameters of the single period PPG main signal in the dual gaussian function model based on the maximum value S4Max, the first position S4MaxLoc, the second position S4MaxEndLoc and the zero crossing point position S5 ZeroLoc;

and decomposing the single-period PPG main signal based on the initial parameters and the double Gaussian function model to obtain a parameter set.

12. The method according to claim 1, wherein the fitting the parameter set based on the linear least squares regression iterative algorithm includes, after obtaining the first gaussian wave and the second gaussian wave:

And performing feature calculation on the single-period PPG main signal, the first Gaussian wave and the second Gaussian wave, and training a preset PWV regression model by using the obtained feature set.

13. An electronic device, the electronic device comprising: a memory and a processor, the memory and the processor coupled; the memory stores program instructions that, when executed by the processor, cause the electronic device to perform the PPG signal processing method of any one of claims 1 to 12.

14. A computer readable storage medium comprising a computer program which, when run on an electronic device, causes the electronic device to perform the PPG signal processing method of any one of claims 1 to 12.