Ppg signal processing method and related device

Abstract

The application provides a PPG signal processing method and related equipment, relates to the technical field of intelligent terminals, and aims to improve the quality of a PPG signal. The PPG signal processing method comprises the following steps: obtaining N sampling points of a photoplethysmography (PPG) signal in a sampling time period, each sampling point having a corresponding signal intensity value. N is a positive integer and N >= 2. Iterating the N sampling points, when the difference between the signal intensity value of the i-th sampling point and the signal intensity value of the (i-1)-th sampling point is greater than or equal to a signal intensity threshold, marking the i-th sampling point as an abnormal point; when the difference is less than the signal intensity threshold, marking the i-th sampling point as a normal point, until the iteration is completed. i is a positive integer and i >= 2. Reconstructing the PPG signal in the sampling time period according to the plurality of normal points. The abnormal point corresponds to the position of a glitch signal in the PPG signal. By reconstructing the PPG signal through the normal points, the glitch signal at the abnormal point can be suppressed, thereby improving the quality of the PPG signal.

Description

Technical Field

[0001] This application relates to the field of smart terminal technology, specifically to a PPG signal processing method and related equipment. Background Technology

[0002] Photoplethysmography (PPG) signals reflect the pulsation state of blood vessels and are widely used in measuring physiological parameters such as heart rate, blood oxygen saturation, and blood pressure. PPG signals are measured by PPG sensors, which are widely used in wearable devices. PPG sensors consist of a light-emitting diode (LED) and a photodetector (PD). The LED emits a light signal. The PD collects the attenuated light signal reflected from the skin, tissues, and blood vessels, and converts this attenuated light signal into an alternating current (AC) electrical signal, thus recording the PPG signal reflecting the pulsation state of blood vessels.

[0003] However, when wearable devices are in motion, the PPG signal collected by the PPG sensor is susceptible to noise interference, which may produce motion artifacts and reduce the accuracy of heart rate detection. For example, if the wearable device shakes during exercise, the distance between it and the skin may increase, causing a jump in the PPG signal magnitude, resulting in noticeable spikes and distortion of the PPG signal. Summary of the Invention

[0004] This application provides a PPG signal processing method and related equipment, which aim to improve the quality of PPG signals.

[0005] The first aspect of this application provides a PPG signal processing method, which includes: acquiring N sampling points of a photoplethysmography (PPG) signal within a sampling time period, each sampling point having a corresponding signal intensity value. N is a positive integer and N≥2. Traversing the N sampling points, if the difference between the signal intensity value of the i-th sampling point and the signal intensity value of the (i-1)-th sampling point is greater than or equal to a signal intensity threshold, the i-th sampling point is marked as an abnormal point; if the difference is less than the signal intensity threshold, the i-th sampling point is marked as a normal point, until the traversal ends. i is a positive integer and i≥2. Reconstructing the PPG signal within the sampling time period based on multiple normal points.

[0006] In this embodiment, the signal intensity value of the sampling point reflects the intensity of the attenuated optical signal acquired by the PPG sensor. When the difference between the signal intensity values ​​of two adjacent sampling points is greater than or equal to the signal intensity threshold, it indicates that a magnitude jump has occurred in the PPG signal, thus indicating the generation of a glitch signal. Anomalies correspond to the locations in the PPG signal where glitch signals occur. Reconstructing the PPG signal from normal points can suppress glitch signals at anomalies, thereby improving the quality of the PPG signal.

[0007] In one embodiment, reconstructing the PPG signal within a sampling time period based on multiple normal points includes: traversing N sampling points; when the j-th sampling point is an outlier, determining multiple normal points adjacent to the j-th sampling point in the time domain; fitting the signal intensity value of the j-th sampling point based on the multiple normal points to obtain the fitted signal intensity value of the j-th sampling point; replacing the signal intensity value of the j-th sampling point with the fitted signal intensity value; and repeating this process until the traversal ends. j is a positive integer and j≥2.

[0008] In this embodiment, for the glitch signal at the abnormal point, the signal intensity value of the glitch signal is fitted based on multiple normal points around the abnormal point to obtain a fitted signal intensity value. The fitted signal intensity value is much smaller than the signal intensity value of the glitch signal. The fitted signal intensity value is used to replace the signal intensity value of the glitch signal, thereby reducing the signal intensity value at the abnormal point that experiences a jump in magnitude, thus effectively suppressing the glitch signal.

[0009] In another embodiment, fitting the signal intensity value of the j-th sampling point based on multiple normal points includes: performing linear interpolation on the signal intensity value of the j-th sampling point based on normal points A and B, which are adjacent to the j-th sampling point. The j-th sampling point is located between normal points A and B.

[0010] In another embodiment, the method further includes: counting the number of outliers within the sampling time period; calculating the percentage of outliers based on the number of outliers; and issuing a reminder regarding the tightness of the fit when the percentage of outliers is greater than or equal to a percentage threshold.

[0011] In this embodiment, when the percentage of abnormal points is greater than or equal to a threshold, it indicates that a large number of spikes have been generated in the PPG signal, suggesting a large contact distance between the wearable device and the human skin. The tightness indicator is used to indicate this large contact distance.

[0012] In another embodiment, the method further includes: acquiring N sampling points of the motion signal within a sampling time period. The motion signal includes acceleration signals and / or angular velocity signals. The N sampling points of the motion signal and the N sampling points of the reconstructed PPG signal are encapsulated into N data packets. Feature parameters are extracted from the N data packets. The feature parameters are input into a heart rate detection model to obtain heart rate data.

[0013] A second aspect of this application provides a wearable device including a PPG sensor and a processor coupled to the PPG sensor. The processor is configured to: receive a PPG signal from the PPG sensor during a sampling time period. The PPG signal includes N sampling points, each with a corresponding signal strength value. N is a positive integer and N≥2. The processor iterates through the N sampling points, marking the i-th sampling point as an anomaly if the difference between the signal strength value of the i-th sampling point and the signal strength value of the (i-1)-th sampling point is greater than or equal to a signal strength threshold; and marking the i-th sampling point as a normal point if the difference is less than the signal strength threshold, until the iteration ends. i is a positive integer and i≥2. The processor reconstructs the PPG signal during the sampling time period based on multiple normal points.

[0014] A third aspect of this application provides an electronic device including a wireless communication module and a processor coupled to the wireless communication module. The wireless communication module is used to communicate with a wearable device. The processor is configured to: receive a PPG signal from the wearable device via the wireless communication module during a sampling time period. The PPG signal includes N sampling points, each sampling point having a corresponding signal strength value. N is a positive integer and N≥2. The N sampling points are traversed. When the difference between the signal strength value of the i-th sampling point and the signal strength value of the (i-1)-th sampling point is greater than or equal to a signal strength threshold, the i-th sampling point is marked as an abnormal point; when the difference is less than the signal strength threshold, the i-th sampling point is marked as a normal point, until the traversal ends. i is a positive integer and i≥2. The PPG signal during the sampling time period is reconstructed based on multiple normal points.

[0015] A fourth aspect of this application provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the method provided in the first aspect.

[0016] A fifth aspect of this application provides a computer program product including computer instructions that implement the method provided in the first aspect when a processor executes the computer instructions.

[0017] It is understood that the beneficial effects of the wearable device provided in the second aspect, the electronic device provided in the third aspect, the computer-readable storage medium provided in the fourth aspect, and the computer program product provided in the fifth aspect are substantially the same as the beneficial effects of the method provided in the first aspect, and will not be repeated here. Attached Figure Description

[0018] Figure 1 This is a schematic diagram illustrating the interaction between an electronic device and a wearable device, as provided as an example.

[0019] Figure 2 This is a schematic diagram illustrating a wearable device for measuring heart rate, as provided as an example.

[0020] Figure 3 This is a schematic diagram of the software architecture of a wearable device provided as an example.

[0021] Figure 4 This is a schematic diagram of a PPG signal under a test environment provided as an example.

[0022] Figure 5 This is a schematic diagram of the PPG signal under a test environment provided as another example.

[0023] Figure 6 This is a flowchart of a PPG signal processing method provided as an example.

[0024] Figure 7 This is an illustration of a wearable device providing a feedback on the tightness of the fit.

[0025] Figure 8 This is a flowchart of an example PPG signal reconstruction method.

[0026] Figure 9 This is a schematic diagram of the PPG signal before reconstruction provided as an example.

[0027] Figure 10 This is a schematic diagram of the reconstructed PPG signal provided as an example.

[0028] Figure 11 This is a schematic diagram illustrating heart rate data prior to reconstructing the PPG signal, provided as an example.

[0029] Figure 12 This is a schematic diagram of heart rate data after reconstructing the PPG signal, provided as an example.

[0030] Figure 13 This is a schematic diagram of the hardware structure of a wearable device provided as an example. Detailed Implementation

[0031] It should be noted that in the embodiments of this application, "multiple" refers to two or more. The terms "first," "second," "third," "fourth," etc., in the specification, claims, and drawings of this application are used to distinguish similar objects, not to describe a specific order or sequence. The methods disclosed in the embodiments of this application, or the methods shown in the flowcharts, include one or more steps for implementing the method. Without departing from the scope of the claims, the execution order of multiple steps can be interchanged, and some steps can also be deleted.

[0032] In the embodiments of this application, electronic devices include, but are not limited to, smartphones, tablets, handheld computers, laptops, intelligent robots, mobile internet devices (MID), virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, wireless terminals in self-driving, wireless terminals in remote medical surgery, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals in smart homes, cellular phones, cordless phones, personal digital assistants (PDAs), handheld devices with wireless communication capabilities, in-vehicle devices, terminal devices in 5G networks, or terminal devices in public terrestrial mobile networks (PLMNs).

[0033] Wearable devices include, but are not limited to, smartwatches, smart bracelets, smart glasses, sports headphones, head-mounted devices (HMDs), and terminal devices with accessories such as rings, bracelets, anklets, and necklaces.

[0034] In some embodiments, an electronic device can be paired with a wearable device via a specific application (APP) to communicate with the wearable device and control it to measure physiological parameters such as the user's heart rate, blood oxygen saturation (also known as "blood oxygen concentration"), and blood pressure. For example, a smartphone can control a smartwatch. Figure 1 As shown, launch the Health app on your smartphone, add the smartwatch in the device management interface, and after successful addition, you can manage the smartwatch through the Health app, such as controlling the smartwatch to measure the user's heart rate.

[0035] In other embodiments, wearable devices can provide functions such as heart rate detection, blood oxygen detection, sleep monitoring, and stress detection. For example, a smartwatch... Figure 2 As shown, launching the heart rate monitoring application on the smartwatch triggers the smartwatch to measure the user's heart rate and display the heart rate data on the heart rate interface.

[0036] The following section describes the heart rate detection scenario in detail, taking into account the software structure of wearable devices.

[0037] Wearable device software systems can adopt layered architecture, event-driven architecture, microkernel architecture, microservice architecture, or cloud architecture. Taking the layered architecture of Lite OS as an example, a layered architecture divides the software system into several layers, each with a clear role and division of labor, and the layers communicate with each other through interfaces. Figure 3 As shown, the Lite OS system includes an application layer, a framework layer, an algorithm and system library, a kernel layer, a hardware abstraction layer (HAL), and a hardware layer.

[0038] The application layer comprises a series of application packages. These application packages may include multiple functional applications, such as heart rate monitoring applications, blood oxygen monitoring applications, sleep monitoring applications, and stress monitoring applications.

[0039] The framework layer provides application programming interfaces (APIs) and programming frameworks for the various functional applications in the application layer. The framework layer may include health services and device management services. Health services include heart rate monitoring, blood oxygen monitoring, sleep monitoring, and stress monitoring. Device management services include sensor services, Bluetooth (BT) services, Near Field Communication (NFC) services, location services, and user interface (UI) services.

[0040] The algorithm and system library includes various types of system functions, algorithms, and protocol stacks. Algorithms may include heart rate detection algorithms, blood oxygen detection algorithms, sleep monitoring algorithms, stress detection algorithms, motion detection algorithms, and wear detection algorithms.

[0041] The kernel layer is used to invoke system functions, algorithms, and protocol stacks in algorithms and system libraries. The kernel layer may include packet assembly services and scheduling services. Packet assembly services encapsulate various types of data from the hardware abstraction layer into data packets that meet the input requirements of algorithms. Scheduling services process data packets from the packet assembly service by invoking system functions, algorithms, and protocol stacks according to business needs.

[0042] The hardware abstraction layer includes driver services, interrupt services, inertial measurement unit (IMU) configuration services, and optical path configuration services. Driver services drive the devices in the hardware layer according to business requirements. Interrupt services add timestamps to various types of data from the hardware layer. The IMU configuration service provides IMU configuration information to the hardware layer's accelerometer (ACC) sensors and gyroscopes. The optical path configuration service provides optical path configuration information to the hardware layer's photoplethysmography (PPG) sensors.

[0043] A PPG sensor consists of a light-emitting diode (LED) and a photodetector (PD). The LED emits a light signal. The PD collects the attenuated light signal reflected from the skin, tissues, and blood vessels, and converts it into an alternating current (AC) electrical signal, thus recording the pulse wave signal (also known as the "PPG signal") that reflects the pulsating state of the blood vessels. Wearable devices use the PPG signal to calculate physiological parameters such as heart rate, blood oxygen saturation, and blood pressure.

[0044] An ACC sensor is used to acquire acceleration signals from the wearable device based on IMU configuration information. A gyroscope is used to acquire angular velocity signals from the wearable device based on IMU configuration information. The wearable device determines whether it is in motion or at rest based on the acceleration and / or angular velocity signals.

[0045] In this embodiment, the IMU configuration information includes the sampling frequency and sampling duration. The optical path configuration information includes the sampling frequency and integration duration (also called "integration time" or "exposure time"), where the integration duration refers to the duration for the PPG sensor to control the PD to acquire attenuated optical signals. The sampling frequency of the ACC sensor and gyroscope is equal to that of the PPG sensor, and the sampling duration of the ACC sensor and gyroscope is equal to the integration duration of the PPG sensor.

[0046] The hardware layer includes a sensor module, a BT module, an NFC module, a positioning module, and a display screen. The sensor module includes a PPG sensor, an ACC sensor, a gyroscope, and a capacitive sensor. The capacitive sensor is used to collect impedance data between the wearable device and the wearable skin. The wearable device determines whether it is being worn or not based on the impedance data.

[0047] In a heart rate detection scenario, when the heart rate detection application is launched on a wearable device, the application sends a heart rate detection command to the framework layer. In response, the framework layer launches a heart rate detection service, which sends an algorithm call command to the kernel layer, instructing the invocation of the heart rate detection algorithm. The kernel layer, in response, launches a packet assembly service and a scheduling service, and sends a data request to the hardware abstraction layer. The hardware abstraction layer, in response to the data request, launches a driver service, an interrupt service, an IMU configuration service, and an optical path configuration service. The driver service sends driver commands to the PPG sensor, ACC sensor, and gyroscope in the hardware layer. The IMU configuration service sends IMU configuration information to the ACC sensor and gyroscope. The optical path configuration service sends optical path configuration information to the PPG sensor. In response to the driver command, the PPG sensor, based on the optical path configuration information from the optical path configuration service, controls the LED to emit a light signal and controls the PD to collect the attenuated light signal reflected back from human skin, tissue, and blood vessels, converting the attenuated light signal into a PPG signal and sending the PPG signal to the hardware abstraction layer. The ACC sensor, responding to drive commands, acquires acceleration signals from the wearable device based on IMU configuration information from the IMU configuration service and sends these signals to the hardware abstraction layer. The gyroscope, also responding to drive commands, acquires angular velocity signals from the wearable device based on IMU configuration information from the IMU configuration service and sends these signals to the hardware abstraction layer. The interrupt service adds timestamps to the PPG, acceleration, and angular velocity signals and sends these timestamped PPG, acceleration, and angular velocity signals to the kernel layer. The kernel layer's packet assembly service encapsulates the timestamped PPG, acceleration, and angular velocity signals into data packets that meet the algorithm's input requirements and sends these data packets to the scheduling service; each data packet has a corresponding timestamp. Upon receiving the data packets, the scheduling service calls the algorithm and the system library's heart rate detection algorithm, inputs the data packets into the heart rate detection algorithm, calculates the user's heart rate data, and sends the heart rate data to the framework layer's heart rate detection service. The heart rate detection service sends the heart rate data to the application layer's heart rate detection application. The heart rate detection application displays the heart rate data on the heart rate interface.

[0048] When wearable devices are in motion, the PPG signal collected by the PPG sensor is susceptible to noise interference, which may produce motion artifacts and reduce the accuracy of heart rate detection. For example, if the wearable device shakes during exercise, the distance between it and the skin may increase, causing a jump in the PPG signal magnitude, resulting in noticeable spikes and distortion of the PPG signal.

[0049] For example, such as Figure 4 and Figure 5As shown, to avoid interference from ambient light, the effect of the contact distance between the wearable device and human skin on the PPG signal was tested in a dark room. When the wearable device is in a resting state, the PPG signal collected by the PPG sensor is as follows: Figure 4 As shown, Figure 4 There were no obvious glitches. When the wearable device was intermittently raised to simulate the shaking of the device during movement, gradually increasing the contact distance between the wearable device and the skin, the PPG signal collected by the PPG sensor was as follows: Figure 5 As shown, Figure 5 Several obvious spike signals appeared in the image.

[0050] To improve the accuracy of heart rate detection, this application provides a PPG signal processing method that reconstructs the PPG signal by accurately identifying and repairing glitch signals, thereby effectively reducing glitch signals and improving the quality of the PPG signal.

[0051] The PPG signal processing method of this application embodiment is described in detail below.

[0052] PPG signal processing methods can be applied to wearable devices, such as... Figure 6 As shown, the PPG signal processing method includes the following steps:

[0053] S101. Acquire N sampling points of the PPG signal within the sampling time period.

[0054] Each sampling point has a corresponding signal strength value, where N is a positive integer and N≥2.

[0055] In this embodiment, the signal strength value of the sampling point reflects the intensity of the attenuated optical signal collected by the PPG sensor. The sampling time period can be set as needed, and the number of sampling points depends on the sampling time period and the sampling frequency. For example, when the sampling time period is 1 second (s) and the sampling frequency is 100 Hz, the number of sampling points is equal to the sampling time period multiplied by the sampling frequency, so the number of sampling points is 100.

[0056] S102. Traverse the N sampling points and calculate the difference between the signal strength value of the i-th sampling point and the signal strength value of the (i-1)-th sampling point.

[0057] Where i is a positive integer and i≥2.

[0058] S103. Determine whether the difference is greater than or equal to the signal strength threshold.

[0059] If yes, proceed to step S104; otherwise, proceed to step S105.

[0060] In this embodiment, when the difference between the signal strength values ​​of two adjacent sampling points is greater than or equal to the signal strength threshold, it reflects a magnitude jump in the PPG signal, indicating the generation of a glitch signal. The signal strength threshold can be set as needed, for example, a signal strength threshold of 20000.

[0061] S104. Mark the i-th sampling point as an outlier.

[0062] S105. Mark the i-th sampling point as a normal point.

[0063] In this embodiment, each sampling point is either a normal point or an abnormal point. An abnormal point corresponds to the location in the PPG signal where a glitch signal is generated. The marked sampling points contain attribute information, including a normal flag or an abnormal flag. A normal flag indicates that the sampling point is normal. An abnormal flag indicates that the sampling point is abnormal. For example, the normal flag is "0" and the abnormal flag is "1".

[0064] After executing step S104 or S105, i = i + 1, and return to execute steps S102-S103 until i = N, at which point the traversal ends. After the traversal ends, execute steps S106-S108.

[0065] S106. Count the number of outliers within the sampling period.

[0066] S107. Calculate the percentage of outliers based on the number of outliers.

[0067] The percentage of outliers is equal to the number of outliers divided by the total number of sampling points.

[0068] S108. Determine whether the percentage of outliers is greater than or equal to the percentage threshold.

[0069] If yes, proceed to steps S109-S114; otherwise, proceed to steps S110-S114.

[0070] In this embodiment, when the percentage of abnormal points is greater than or equal to a percentage threshold, it indicates that a large number of spike signals have been generated in the PPG signal, thereby indicating that the contact distance between the wearable device and the human skin is large. The percentage threshold can be set as needed, for example, a percentage threshold of 0.3.

[0071] S109. Issue a reminder regarding the tightness of the fit.

[0072] In this embodiment, the tightness indicator is used to indicate that the wearable device is too far from the skin. For example, such as... Figure 7 As shown, the heart rate monitoring application displays "Please wear it tighter" on the heart rate interface.

[0073] S110. Reconstruct the PPG signal within the sampling time period based on multiple normal points.

[0074] In this embodiment, for the glitch signal at the abnormal point, the signal intensity value of the glitch signal is fitted based on multiple normal points around the abnormal point to obtain a fitted signal intensity value. The fitted signal intensity value is much smaller than the signal intensity value of the glitch signal. The fitted signal intensity value is used to replace the signal intensity value of the glitch signal, thereby reducing the signal intensity value at the abnormal point that experiences a jump in magnitude, thus effectively suppressing the glitch signal.

[0075] It is understood that the fitting process can employ linear regression, polynomial regression, and spline regression, and this embodiment is not limited to any particular method.

[0076] For example, such as Figure 8 As shown, the PPG signal reconstruction method includes the following steps:

[0077] S201. Traverse N sampling points and query the attribute information of the j-th sampling point.

[0078] Where j is a positive integer and j≥2.

[0079] S202. Determine whether the j-th sampling point is an outlier.

[0080] If yes, proceed to step S203; otherwise, j = j + 1, return to steps S201-S202, and continue until j = N, at which point the traversal ends.

[0081] In this embodiment, when the attribute information of the j-th sampling point includes an anomaly flag, it indicates that the point is an anomaly.

[0082] S203. Determine the normal points A and B that are adjacent to the j-th sampling point in the time domain.

[0083] The j-th sampling point is located between normal points A and B.

[0084] S204. Based on normal points A and B, perform linear interpolation on the signal intensity value of the j-th sampling point to obtain the fitted signal intensity value of the j-th sampling point.

[0085] In this embodiment, the fitted value of the signal strength at the j-th sampling point is much smaller than the signal strength value at the j-th sampling point.

[0086] S205. Replace the signal strength value of the j-th sampling point with the fitted signal strength value of the j-th sampling point.

[0087] After executing step S205, j = j + 1, and return to execute steps S201-S202 until the traversal ends.

[0088] For example, such as Figure 9 As shown, when the wearable device is in motion, the PPG signal collected by the PPG sensor exhibits a large number of spike signals. These spike signals include regular and irregular spike signals. Regular spike signals are correlated with the contact distance between the wearable device and the human skin, while irregular spike signals are not correlated with this contact distance. Figure 10 As shown, the reconstructed PPG signal effectively reduces regular glitches, thereby improving the quality of the PPG signal.

[0089] S111. Acquire N sampling points of the motion signal within the sampling time period.

[0090] The motion signal includes acceleration signal and / or angular velocity signal.

[0091] In this embodiment, the motion signal reflects whether the wearable device is in motion or at rest. Since the sampling frequency of the motion signal is the same as that of the PPG signal, the number of sampling points in the motion signal is the same as the number of sampling points in the PPG signal.

[0092] S112. Encapsulate the N sampling points of the motion signal and the N sampling points of the reconstructed PPG signal into N data packets.

[0093] In this embodiment, since the motion signal and PPG signal are sampled synchronously, the sampling points of the motion signal with the same timestamp and the sampling points of the reconstructed PPG signal are encapsulated into a data packet, so that each data packet has a corresponding timestamp.

[0094] S113. Extract the feature parameters of N data packets.

[0095] The characteristic parameters include frequency domain characteristics, waveform characteristics, and correlation characteristics.

[0096] In this embodiment, frequency domain features of the motion signal and the reconstructed PPG signal are extracted through frequency domain transformation. Waveform features, including peak value, period, and amplitude, are extracted from the motion signal and the reconstructed PPG signal through peak detection and morphological analysis. Correlation features reflecting the relationship between the motion signal and the reconstructed PPG signal are extracted through Principal Component Analysis (PCA) and Canonical Correlation Analysis (CCA).

[0097] S114. Input the feature parameters into the heart rate detection model to obtain heart rate data.

[0098] In this embodiment, the heart rate detection model can employ a fusion model, which includes a time-domain analysis model, a frequency-domain analysis model, and a deep learning model. The time-domain analysis model determines the peak positions and number of peaks in the time domain based on feature parameters, with the frequency corresponding to each peak position being the heart rate value. The frequency-domain analysis model determines the main frequency components based on the feature parameters. The deep learning model optimizes the feature parameters. The fusion model optimizes the heart rate data based on the outputs of the time-domain analysis model, the frequency-domain analysis model, and the deep learning model, thereby improving the accuracy of the heart rate data.

[0099] It is understood that deep learning models may include convolutional neural networks (CNNs) and recurrent neural networks (RNNs), but this embodiment is not limited to them.

[0100] For example, such as Figure 11 As shown, when the wearable device is in motion, heart rate data S1 is calculated based on the PPG signal collected by the PPG sensor. The heart rate data S1 differs significantly from the reference heart rate data S0 in certain intervals, reflecting low accuracy of the heart rate data S1. The reference heart rate data S0 can be the heart rate data measured by the Polar heart rate belt, which is equipped with an electrocardiogram (ECG) sensor, which has high accuracy. Figure 12 As shown, after reconstructing the PPG signal, the heart rate data S2 is calculated based on the reconstructed PPG signal. Compared with the heart rate data S1, the heart rate data S2 has a higher similarity to the reference heart rate data S0, reflecting that the heart rate data S2 has higher accuracy.

[0101] The PPG signal processing method has been explained in detail above in conjunction with the software structure of wearable devices. The hardware structure of wearable devices is briefly described below.

[0102] For example, such as Figure 13 As shown, the wearable device includes a processor 110, internal memory 120, charging management module 130, power management module 140, battery 150, sensor module 160, wireless communication module 170, and display screen 180. Sensor module 160 includes a PPG sensor 161, an ACC sensor 162, a gyroscope 163, and a capacitive sensor 164. Wireless communication module 170 includes a BT module 171, an NFC module 172, and a positioning module 173.

[0103] Processor 110 may include one or more processing units, such as an application processor (AP), a graphics processing unit (GPU), an image signal processor (ISP), a controller, a digital signal processor (DSP), a baseband processor, and a neural network processing unit (NPU). These different processing units may be independent devices or integrated into one or more processors. The controller can generate operation control signals based on instruction opcodes and timing signals to control instruction fetching and execution.

[0104] The processor 110 may also include a memory for storing instructions and data. In some embodiments, the memory in the processor 110 is a cache memory. The cache memory can store instructions or data that the processor 110 has just used or that are used repeatedly. If the processor 110 needs to use the instruction or data again, it can retrieve it directly from the cache memory. This avoids repeated accesses, reduces the waiting time of the processor 110, and thus improves system efficiency.

[0105] Internal memory 120 stores executable program code, including instructions. Internal memory 120 may include a program storage area and a data storage area. The program storage area may store the operating system and at least one functional application. The data storage area may store data created during the use of the wearable device. Furthermore, internal memory 120 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, Universal Flash Storage (UFS), etc. Processor 110 executes various functional applications and data processing of the wearable device by running instructions stored in internal memory 120 and / or instructions stored in memory located within the processor.

[0106] The charging management module 130 receives charging input from the wireless charger. The charging management module 130 can also receive wireless charging input via the wireless charging coil of the wearable device. While charging the battery 150, the charging management module 130 can also supply power to the wearable device via the power management module 140.

[0107] The power management module 140 receives input from the battery 150 and / or the charging management module 130, and supplies power to the processor 110, internal memory 120, sensor module 160, wireless communication module 170, and display screen 180. The power management module 140 can also be used to monitor parameters such as battery capacity, battery cycle count, and battery health status (leakage current, impedance). In some embodiments, the power management module 140 may also be located within the processor 110. In other embodiments, the power management module 140 and the charging management module 130 may also be located in the same device.

[0108] Display screen 180 is used to display images. Display screen 180 includes a display panel. The display panel may be an organic light-emitting diode (OLED), an active-matrix organic light-emitting diode (AMOLED), a flexible light-emitting diode (FLED), a miniaturized LED, a microled LED, a quantum dot light-emitting diode (QLED), etc.

[0109] It is understood that the structure illustrated in this embodiment does not constitute a specific limitation on the wearable device. In other embodiments, the wearable device may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements.

[0110] It should be noted that the PPG signal processing method of this application embodiment can also be applied to electronic devices, chips, computer-readable storage media, and computer program products.

[0111] The electronic device includes a processor, a memory, a wireless communication module, and a display screen. The wireless communication module is used to communicate with a wearable device. When the processor executes computer instructions stored in the memory, it implements the PPG signal processing method of this application embodiment.

[0112] The chip includes a processor and an interface circuit, with the processor and interface circuit electrically connected. The interface circuit can read computer instructions stored in the memory and send the computer instructions to the processor. When the processor executes the computer instructions, it implements the PPG signal processing method of this application embodiment.

[0113] A computer-readable storage medium stores computer instructions that, when executed by a processor, implement the PPG signal processing method of the embodiments of this application.

[0114] Computer-readable storage media include volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer-readable storage media include, but are not limited to, random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer.

[0115] The computer program product includes computer instructions that, when executed by a processor, implement the PPG signal processing method of the embodiments of this application.

[0116] The embodiments of this application have been described in detail above with reference to the accompanying drawings. However, this application is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of this application.

Claims

1. A PPG signal processing method, characterized in that, The method includes: N sampling points of the photoplethysmography (PPG) signal are acquired within the sampling time period. Each sampling point has a corresponding signal intensity value, where N is a positive integer and N≥2. Traverse the N sampling points. When the difference between the signal strength value of the i-th sampling point and the signal strength value of the (i-1)-th sampling point is greater than or equal to the signal strength threshold, mark the i-th sampling point as an abnormal point; when the difference is less than the signal strength threshold, mark the i-th sampling point as a normal point, until the traversal ends, where i is a positive integer and i≥2. The PPG signal during the sampling time period is reconstructed based on multiple normal points.

2. The method as described in claim 1, characterized in that, The step of reconstructing the PPG signal within the sampling time period based on multiple normal points includes: Traverse the N sampling points. When the j-th sampling point is the anomalous point, determine multiple normal points that are adjacent to the j-th sampling point in the time domain. Fit the signal intensity value of the j-th sampling point according to the multiple normal points to obtain the signal intensity fitting value of the j-th sampling point. Replace the signal intensity value of the j-th sampling point with the signal intensity fitting value until the traversal ends. j is a positive integer and j≥2.

3. The method as described in claim 2, characterized in that, The process of fitting the signal intensity value of the j-th sampling point based on multiple normal points includes: Linear interpolation is performed on the signal strength value of the j-th sampling point based on normal points A and B, which are adjacent to the j-th sampling point. The j-th sampling point is located between normal points A and B.

4. The method according to any one of claims 1-3, characterized in that, The method further includes: Count the number of outliers within the sampling time period; Calculate the percentage of abnormal points based on the number of abnormal points; When the percentage of abnormal points is greater than or equal to the percentage threshold, a reminder will be issued regarding the tightness of the garment.

5. The method according to any one of claims 1-4, characterized in that, The method further includes: N sampling points of motion signal are acquired within the sampling time period, the motion signal including acceleration signal and / or angular velocity signal; The N sampling points of the motion signal and the N sampling points of the reconstructed PPG signal are encapsulated into N data packets; Extract the feature parameters of the N data packets; The feature parameters are input into the heart rate detection model to obtain heart rate data.

6. A wearable device, characterized in that, It includes a PPG sensor and a processor coupled to the PPG sensor, the processor being configured to: During the sampling time period, a PPG signal is received from the PPG sensor. The PPG signal includes N sampling points, each sampling point has a corresponding signal strength value, where N is a positive integer and N≥2. Traverse the N sampling points. When the difference between the signal strength value of the i-th sampling point and the signal strength value of the (i-1)-th sampling point is greater than or equal to the signal strength threshold, mark the i-th sampling point as an abnormal point; when the difference is less than the signal strength threshold, mark the i-th sampling point as a normal point, until the traversal ends, where i is a positive integer and i≥2. The PPG signal during the sampling time period is reconstructed based on multiple normal points.

7. The wearable device as described in claim 6, characterized in that, The step of reconstructing the PPG signal within the sampling time period based on multiple normal points includes: Traverse the N sampling points. When the j-th sampling point is the anomalous point, determine multiple normal points that are adjacent to the j-th sampling point in the time domain. Fit the signal intensity value of the j-th sampling point according to the multiple normal points to obtain the signal intensity fitting value of the j-th sampling point. Replace the signal intensity value of the j-th sampling point with the signal intensity fitting value until the traversal ends. j is a positive integer and j≥2.

8. The wearable device as described in claim 7, characterized in that, The process of fitting the signal intensity value of the j-th sampling point based on multiple normal points includes: Linear interpolation is performed on the signal strength value of the j-th sampling point based on normal points A and B, which are adjacent to the j-th sampling point. The j-th sampling point is located between normal points A and B.

9. The wearable device as described in any one of claims 6-8, characterized in that, The processor is also configured to: Count the number of outliers within the sampling time period; Calculate the percentage of abnormal points based on the number of abnormal points; When the percentage of abnormal points is greater than or equal to the percentage threshold, a reminder will be issued regarding the tightness of the garment.

10. The wearable device as described in any one of claims 6-9, characterized in that, The wearable device further includes an accelerometer and / or a gyroscope, and the processor is further configured to: Motion signals from the accelerometer and / or gyroscope are received during the sampling time period. The motion signals include N sampling points and include acceleration signals and / or angular velocity signals. The N sampling points of the motion signal and the N sampling points of the reconstructed PPG signal are encapsulated into N data packets; Extract the feature parameters of the N data packets; The feature parameters are input into the heart rate detection model to obtain heart rate data.

11. An electronic device, characterized in that, It includes a wireless communication module and a processor coupled to the wireless communication module, the wireless communication module being used to communicate with a wearable device, and the processor being configured to: During the sampling period, the wireless communication module receives PPG signals from the wearable device. The PPG signals include N sampling points, each sampling point has a corresponding signal strength value, N is a positive integer and N≥2; Traverse the N sampling points. When the difference between the signal strength value of the i-th sampling point and the signal strength value of the (i-1)-th sampling point is greater than or equal to the signal strength threshold, mark the i-th sampling point as an abnormal point; when the difference is less than the signal strength threshold, mark the i-th sampling point as a normal point, until the traversal ends, where i is a positive integer and i≥2. The PPG signal during the sampling time period is reconstructed based on multiple normal points.

12. A computer-readable storage medium, characterized in that, It stores computer instructions that, when executed by a processor, implement the method as described in any one of claims 1-5.

13. A computer program product, characterized in that, It includes computer instructions that, when executed by a processor, implement the method as described in any one of claims 1-5.

Images