Health monitoring bracelet detection method and system capable of detecting pulse rate and blood oxygen in real time

Abstract

The invention discloses a health monitoring bracelet detection method and system capable of detecting pulse rate blood oxygen in real time. The method comprises the steps that S1, continuous light absorption signal changes are obtained based on a photoplethysmography optical sensor; s2, a heartbeat pulse signal measurement pulse rate value is obtained according to change analysis; s3, wrist shaking is recognized based on data of the acceleration sensor, and pulse signal distortion caused by shaking is corrected; according to the health monitoring bracelet detection method and system capable of detecting the pulse rate blood oxygen in real time, the problem that when a user exercises, especially violent activities, due to the fact that an existing health monitoring bracelet is affected by wrist shaking in the exercise process, the user can feel uncomfortable in the exercise process is solved, and the health monitoring bracelet can be used for detecting the pulse rate blood oxygen in real time. The problems that light scattering and signal attenuation are caused, data collection and processing are affected, the measurement result of the pulse rate and the oxyhemoglobin saturation is not accurate enough due to external interference, and the reliability of the bracelet and the user experience are reduced are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of biosensors, and in particular to a health monitoring bracelet detection method and system capable of detecting pulse rate and blood oxygen in real time. Background Art

[0002] Among modern health monitoring devices, wristband monitors have become a common health management tool that can detect the user's pulse rate and blood oxygen saturation in real time through optical sensing technology. This type of wristband usually uses a small optoelectronic device installed on the back of the wristband to measure the pulse rate and blood oxygen level by emitting near-infrared light to the user's skin and receiving the reflected light signal. The measurement process is based on the Lambert-Beer law, which states that substances of different concentrations have different absorption rates of light, so that changes in blood oxygen saturation can be quantitatively measured by differences in light absorption.

[0003] However, in actual applications, existing health monitoring bracelets have significant measurement errors when users exercise, especially when engaging in strenuous activities. This is mainly because the bracelet is affected by the shaking of the wrist during exercise, which causes light scattering and signal attenuation, thereby affecting data collection and processing. This external interference makes the measurement results of pulse rate and blood oxygen saturation inaccurate, reducing the reliability of the bracelet and user experience. Summary of the invention

[0004] The purpose of the present invention is to provide a health monitoring bracelet detection method and system that can detect pulse rate and blood oxygen in real time, so as to solve the problem that when the user exercises, especially strenuous activities, the existing health monitoring bracelet will be affected by the shaking of the wrist during exercise, resulting in light scattering and signal attenuation, thereby affecting data collection and processing. This external interference makes the measurement results of pulse rate and blood oxygen saturation inaccurate, reducing the reliability of the bracelet and the user experience.

[0005] To achieve the above object, the present invention provides the following technical solution: a health monitoring bracelet detection method capable of detecting pulse rate and blood oxygen in real time, the method comprising:

[0006] S1: Acquire continuous light absorption signal changes based on photoplethysmography optical sensor;

[0007] S2: Determine the pulse rate value based on the heartbeat pulse signal obtained by the change analysis;

[0008] S3: Identify wrist shaking based on the data of the acceleration sensor and correct the pulse signal distortion caused by the shaking;

[0009] S4: Dual wavelength optical detection using infrared light and red light is used to calculate the oxygen saturation in the blood.

[0010] Preferably, the step S1 specifically includes:

[0011] activating the optical sensor and emitting a light source of a predetermined intensity to the tissue surface;

[0012] Continuously collect data on changes in light intensity transmitted or reflected by tissues;

[0013] Using a specific algorithm to extract effective light absorption fluctuation signals from the data;

[0014] Filter the collected raw light absorption fluctuation signal to eliminate noise interference;

[0015] Form a clear and stable continuously changing signal output for subsequent analysis;

[0016] Wherein, the extracting effective light absorption fluctuation signal from the data using a specific algorithm comprises:

[0017] Discrete wavelet transform is applied to decompose the light intensity variation data ΔI(t) and extract the fluctuation signal related to the heart pulse. The specific formula is:

[0018]

[0019] Among them, W j,k is the wavelet coefficient, t represents time, N represents the total number of signal samples, ΔI(t) is the light intensity change data, ψ is the mother wavelet function, j and k represent the scale and position parameters respectively. By selecting the appropriate scale jjj and position kkk, the components related to the heart pulse can be extracted;

[0020] The filtering process of filtering the collected original light absorption fluctuation signal to eliminate noise interference adopts the formula:

[0021]

[0022] Among them, y(t) is the filtered signal, ΔI(t) is the light intensity change data, t is time, is the estimated noise, μ is the step size factor, e(t) is the error signal, and by adjusting μ, the noise interference can be adaptively eliminated.

[0023] Preferably, the step S2 specifically includes:

[0024] The heartbeat peak position is detected based on the continuous light absorption signal changes;

[0025] Use autocorrelation technology to locate clear and reliable periodic feature points;

[0026] Select appropriate pulse template and match the signal sequence through convolution analysis;

[0027] Applying algorithms to determine adjacent heartbeat cycles and calculate pulse cycle time length;

[0028] The heart rate per minute, or pulse rate, is calculated based on the cycle.

[0029] Preferably, the step S3 specifically includes:

[0030] Enable the built-in multi-axis accelerometer to continuously monitor wrist dynamic changes;

[0031] Capture and store the component value sequence of acceleration in each axis in real time;

[0032] Determine whether significant motion shaking occurs based on the comparison between the instantaneous amplitude of the motion vector and a preset threshold value;

[0033] Smoothing or interpolation processing is performed on the light absorption signal within a time window before and after the shaking moment to reduce the impact of the shaking;

[0034] By comparing the heart rate wave obtained by synchronous measurement with the heart rate wave after motion interference suppression, the final heart rate calculation result is corrected.

[0035] Preferably, the step S4 specifically includes:

[0036] Two photoplethysmography sensors emit near infrared light above 800nm ​​and visible red light around 600nm, respectively, and collect signal feedback information;

[0037] Using digital signal processing technology to extract specific signal components from the feedback information and normalize them;

[0038] Select appropriate equations to fit the relationship between the red light / near infrared absorption ratio and different blood oxygenation states;

[0039] Measure and record the ratio of the continuous blood volume changes under the two light sources, and calculate the proportional factor;

[0040] The blood oxygen content in the blood is derived according to the proportional factor, thereby obtaining the blood oxygen saturation level;

[0041] The specific calculation steps of the dual wavelength light detection based on infrared light and red light include:

[0042] Measure the optical signal change values ​​at two wavelengths respectively;

[0043] Perform differential calculation to obtain the ratio of the two groups of light absorption intensity changes;

[0044] Apply the corrected absorption ratio to adjust the saturation calculation:

[0045] SO2=[(I 红外 / I 红光 )×K / 2]+[slope×(I 红外 -I红光 )+intercept]×K,

[0046] Among them, SO2 is the blood oxygen saturation, I 红外 is the signal intensity of infrared light, I 红光 is the signal strength of the red light, and K is a coefficient used for calibration or adjustment.

[0047] Preferably, the step S1 of acquiring continuous light absorption signal changes based on a photoplethysmography optical sensor specifically further comprises:

[0048] Establishing an optical channel between the optical sensor and the light source;

[0049] The change of the intensity of the light passing through the finger is collected by the receiver;

[0050] The collected light intensity changes are converted into voltage signal changes and recorded;

[0051] The continuous light absorption curve pattern is analyzed by utilizing the change of the voltage signal.

[0052] Preferably, the step S3 of identifying wrist shaking based on the data of the acceleration sensor and correcting the pulse signal distortion caused by the shaking also includes:

[0053] Obtain the X, Y, and Z axis data from the acceleration sensor to calculate the average acceleration value;

[0054] If any axial acceleration exceeds a preset threshold, it is considered that shaking occurs;

[0055] If the wrist is judged to be in a shaking state, a filter is applied to remove noise;

[0056] Formula judgment: If the absolute value of the acceleration of a certain axis is greater than the threshold, it is considered that the wrist is shaking.

[0057] Preferably, the dual wavelength light detection using infrared light and red light in step S4 specifically further includes:

[0058] Switch the light source module to illuminate the skin surface with infrared light waves to collect feedback intensity signals;

[0059] Then, the red light is used for illumination to repeat the above feedback intensity collection process;

[0060] Compare the absorption ratio of the two to calculate the blood oxygen saturation level;

[0061] Using the formula judgment condition, SO2=((I R / I red )×K) to correct the error factors in the calculation process,

[0062] Among them, SO2 represents blood oxygen saturation, IR Indicates the intensity of infrared light, I red Indicates the intensity of red light.

[0063] Preferably, the change analysis to obtain the heartbeat pulse signal to determine the pulse rate value specifically includes:

[0064] Apply Fourier transform to analyze the spectral components of the signal;

[0065] Filter high-frequency components based on the characteristics of pulse signals;

[0066] Identify the peak position to extract the effective heartbeat signal cycle length;

[0067] The pulse rate value is obtained by using the inverse of the pulse signal cycle length, using the formula:

[0068] Pulse rate = 60 / peak interval time.

[0069] A health monitoring wristband detection system capable of detecting pulse rate and blood oxygen in real time, using the health monitoring wristband detection method capable of detecting pulse rate and blood oxygen in real time, the system comprising:

[0070] An optical sensing module for obtaining continuous light absorption signal changes based on photoplethysmography;

[0071] The pulse rate measurement module is used to obtain the heartbeat pulse signal and measure the pulse rate value according to the change analysis of the light absorption signal;

[0072] An acceleration sensor module is used to identify wrist shaking based on the data of the acceleration sensor and correct the pulse signal distortion caused by the shaking;

[0073] The blood oxygen detection module is used to calculate the blood oxygen saturation in the blood using dual wavelength light detection of infrared light and red light.

[0074] It can be seen from the above technical solution that the present invention has the following beneficial effects:

[0075] The health monitoring wristband detection method and system capable of real-time detection of pulse rate and blood oxygen obtain continuous light absorption signal changes through an optical sensor based on photoplethysmography, obtain a heartbeat pulse signal according to the change analysis to determine the pulse rate value, identify wrist shaking based on data of an acceleration sensor and correct the pulse signal distortion caused by the shaking, calculate the blood oxygen saturation in the blood by using dual wavelength light detection of infrared light and red light, identify wrist shaking in real time, and can accurately capture the real changes of pulse rate and blood oxygen saturation, thereby improving the reliability of data, measuring blood oxygen saturation more accurately, avoiding the possible errors in single wavelength detection, significantly improving the accuracy and anti-interference ability of pulse rate and blood oxygen saturation measurement, enhancing the monitoring experience of users in various activity states, and solving the problem that when the user is exercising, especially in strenuous activities, the existing health monitoring wristband is affected by wrist shaking during exercise, resulting in light scattering and signal attenuation, thereby affecting data collection and processing, and such external interference makes the measurement results of pulse rate and blood oxygen saturation inaccurate, reducing the reliability of the wristband and the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0077] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0078] like Figure 1 As shown, a detection method for a health monitoring wristband capable of detecting pulse rate and blood oxygen in real time is provided. By integrating devices such as photoplethysmography, optical sensors, and acceleration sensors in the wristband, the pulse rate and blood oxygen saturation can be continuously monitored when the user wears the wristband. The method comprises:

[0079] Based on the PPG optical sensor to obtain continuous changes in light absorption signals, PPG uses a light emitting diode to generate infrared light or red light in a certain wavelength range to illuminate the skin. After being absorbed by the blood and tissue in the skin, part of the light is reflected back and received by the photodetector. Blood circulation in the human body causes slight expansion and contraction changes in blood vessels, resulting in periodic changes in the amount of light reflected and absorbed by blood when passing through the blood vessels; these changes in light absorption signals can be captured by photodiodes. The changes in the intensity of the faint light captured by the photodiode are related to cardiovascular activity, and therefore can reflect the rheological process of blood in subcutaneous tissue and capillaries.

[0080] When the bracelet touches the wrist, the LED emits infrared light of a certain brightness and specific frequency (such as 860nm), which passes through the tissue layer and enters the veins or arteries below the surface of the human skin. The blood has different absorption of red light and near-infrared light, and the amount of red / near-infrared light transmitted back to the receiving sensor end is reduced due to the increase in vascular capacity during cardiac contraction. This change can be recorded by the photodiode as a series of pulse data streams, which provides support for further extraction of heart rate or heart rate variability information.

[0081] Analyze pulse signals and determine pulse rate

[0082] The characteristic points of information changes captured by the photodiode can be detected and extracted by the software algorithm to generate a continuous pulse wave graph. Usually, the algorithm needs to find the peak and valley points of the pulse graph, calculate the time between each peak, and then get the heart rate data for each time. The average value of this series of data is the actual average pulse of the user at that moment.

[0083] After obtaining a series of data fluctuation signals in the first stage, the peak search and positioning operation is performed, and the cycle length T (i.e., a cardiac cycle) from the maximum peak to the bottom is measured over a period of time. If the same T value appears multiple times in a period of time, it means that the user's heart rhythm is regular. The current instant heart rate R is obtained by the inverse value 60 / T, and the average value of R for N times can be taken to estimate the user's actual instantaneous heart rate.

[0084] Correction of errors caused by arm movement based on acceleration sensing

[0085] When the arm is moving, even the slightest movement may cause fluctuations in the PPG data, especially during strenuous activities (such as running or dancing). At this time, it is necessary to use the accelerometer to detect the actual movement state of the part where the wristband is worn to eliminate the influence of false interference. It can help identify whether the arm is stationary or active, thereby filtering out valid signals.

[0086] If the wristband acceleration sensor component records a drastic change signal in a certain direction, for example, the acceleration on the X-axis exceeds the preset threshold of 10 meters per second, it is considered that the wearer's arm is swinging widely instead of being stable. When this happens, the software program will automatically pause the current pulse data analysis work; wait until the smooth and stable state is restored before re-enabling the pulse rate and blood oxygen data collection and analysis function.

[0087] Calculate blood oxygen saturation level using red and green light sources

[0088] The use of dual-wavelength light for penetration detection at the transmitter end can accurately calculate the blood oxygen content information. The hemoglobin in the blood will show different colors in the bound state and the non-dissociated state, the former is blue-purple and the latter is dark red; therefore, red light is selected as the reference light source and near-infrared as the target signal. The difference between the two can be used to calculate the percentage of the pigment's ability to carry oxygen, that is, the oxygen and index.

[0089] When the PPG acquisition device starts dual-frequency synchronous light operation: one is 650 nanometer red light and the other is 930-950 nanometer near-infrared light. The change of red light signal mainly reflects the reflection of skin tissue and the arterial flow of low-density oxygen saturation in the venous and capillary systems, while the latter is only affected by the expansion / contraction of arterial blood vessels without considering skin reflection or venous part. The ratio of the intensity relationship at the peak of the output waveform of two independent signal channels can directly reflect the oxygen and index ratio value during this period.

[0090] Overall, this system overcomes the limitations of the traditional single light source method in dynamic application scenarios through the combination of hardware sensors and intelligent data screening processes, and also introduces auxiliary algorithm models in the processing process to improve accuracy and stability. The smart bracelet realized through such a design can not only accurately obtain pulse rate and blood oxygen index information in daily life and exercise, but also improve the authenticity and reliability of data to a certain extent, and better serve the purpose of health tracking and management.

[0091] Next, the specific steps of obtaining continuous light absorption signal changes based on the optical sensor of photoplethysmography are described. First, the optical sensor is activated and emits light to the user's skin tissue at a preset intensity. This emission usually involves a red or near-infrared light source being sent out through the sensor's LED. Since these two light wavelengths are particularly suitable for passing through tissues without being completely absorbed, they can effectively reflect changes in the concentration of oxygenated and deoxygenated hemoglobin in tissues. For example, suppose that after a bracelet is activated, its light source selects an 800nm ​​near-infrared light source as the emission light source, and the light emission work begins when the bracelet touches the wearer's wrist.

[0092] During the emission process, the data changes of the transmitted light after passing through the finger tissue or the light intensity reflected by the epidermal cells will be continuously recorded. This is because the amount of blood in the blood flow changes with each beat of the heart, which in turn affects the amount of light passing through the tissue, causing signal fluctuations. These light intensity change data will be detected by the photodetector. For example, during use, the bracelet sensor continuously collects the intensity changes caused by the flow of light through the blood vessels. This process can capture the changes in light absorption signals caused by the expansion and contraction of blood vessel volume accompanying each heartbeat.

[0093] The next step is to process the acquired signals using specific mathematical methods and filtering algorithms to extract meaningful information. In this step, the fluctuations related to the heartbeat are filtered out by applying appropriate algorithms to the raw data to remove noise such as non-related signal fluctuations from environmental factors such as external interference light or body movements, and retain those signals generated by physiological changes. For example, at this stage the system applies FFT Fourier transform and some low-pass or band-pass filtering techniques to process the acquired data signals.

[0094] Finally, the collected effective information needs to be further cleaned of clutter to form a stable, continuously changing output that can be further identified and read, so that the software algorithm can further process and display blood oxygen and heart rate values. For example, after filtering, a clear heart rate pulse contour signal is generated, which is finally converted into a value that the user can directly observe and understand.

[0095] The above steps ensure that the optical sensor can accurately capture the slight changes in light signals caused by changes in blood flow, and convert them into readable data to reflect the user's physiological state, so that the monitoring bracelet can promptly feedback important vital signs to the user and medical staff.

[0096] Specific algorithms are used to extract effective light absorption fluctuation signals from the data including:

[0097] Discrete wavelet transform is applied to decompose the light intensity variation data ΔI(t) and extract the fluctuation signal related to the heart pulse. The specific formula is:

[0098]

[0099] Among them, W j,k is the wavelet coefficient, t represents time, N represents the total number of signal samples, ΔI(t) is the light intensity change data, ψ is the mother wavelet function, j and k represent the scale and position parameters respectively. By selecting the appropriate scale jjj and position kkk, the components related to the heart pulse can be extracted;

[0100] The filtering process of filtering the collected original light absorption fluctuation signal to eliminate noise interference adopts the formula:

[0101]

[0102] Among them, y(t) is the filtered signal, ΔI(t) is the light intensity change data, t is time, is the estimated noise, μ is the step size factor, e(t) is the error signal, and by adjusting μ, the noise interference can be adaptively eliminated.

[0103] The core principle of this embodiment is to use discrete wavelet transform to decompose the light intensity change data ΔI(t) to extract the fluctuation signal related to the heart pulse. In this process, the wavelet transform can effectively separate the different frequency components in the signal, and by selecting the appropriate scale j and position k, the fluctuation component related to the heart pulse can be accurately located. Subsequently, the extracted signal is processed by an adaptive filter to filter out possible noise interference, thereby obtaining a more accurate pulse fluctuation signal. This embodiment effectively extracts the fluctuation signal related to the heart pulse through the combination of discrete wavelet transform and adaptive filtering, and can adaptively eliminate noise interference by adjusting parameters. This method not only improves the accuracy of the signal, but also enhances the flexibility and adaptability of the algorithm, which helps to achieve accurate pulse detection in various complex signal environments.

[0104] In this embodiment, the discrete wavelet transform uses a specific mother wavelet function ψ, and selects a suitable scale j and position k. Different types of wavelet functions, such as Haar wavelet, Daubechies wavelet, etc., can be considered to adapt to different types of signal characteristics. In addition, in the design of the adaptive filter, the selection of the step size factor μ is crucial. The step size factor μ can be adjusted according to different signal environments to optimize the filtering effect. Other adaptive filtering algorithms, such as LMS (least mean square) algorithm, can also be explored to improve the effect of noise elimination.

[0105] Next, the steps of obtaining the heartbeat pulse signal according to the change analysis to determine the pulse rate value are described:

[0106] First, the location of the heartbeat peak is detected based on the change in the continuous light absorption signal. At this stage, since the amount of light passing through the finger or wrist that penetrates the blood changes as the heart pumps blood, this change will be captured by the photoelectric sensor as a series of signals. For each significant change, especially those associated with blood flowing into the artery are marked as peaks. For example, if several obvious peak points are detected from the signal sequence, each of these points will be considered as a heart contraction point caused by blood flow.

[0107] Then, the autocorrelation technique is used to further process this series of possible heartbeat peaks to find the most likely periodic heartbeat signals and confirm whether they conform to the actual heart rhythm cycle. Based on this, reliable periodic feature points can be determined to eliminate some false signal interference or missing real signal situations caused by noise. For example, by analyzing the above signal, we can find the most similar time pairs between two consecutive peaks. These most representative pairs can be regarded as heartbeat signal cycles that reliably describe the heart rate change pattern over a period of time.

[0108] After identifying these features, a set of pulse templates suitable for measuring the signal waveform will be selected and the signal sequence will be convolved to determine which part of the data sequence best matches the expected pulse waveform. These best matching positions will again calibrate the position of the heart beat and its amplitude characteristics to ensure that the detected heartbeat peak is a true heart rate signal, not just noise or other artifacts. For example, after using the typical pulse signal shape as a template, we can compare these templates with the previously discovered heartbeat points. Only those peaks that are close to the template shape will be finally accepted as the correct heart rate peak, thereby removing changes and interference signals caused by other factors.

[0109] Next, we need to apply a specific algorithm to determine the duration of each heartbeat cycle from the obtained well-matched or identified heart signal points. In this way, we can get the interval between adjacent heart signals and use it to measure the specific time span between each heartbeat. For example, after marking all the reliable heart signal points in a sequence, we can calculate the interval between any two points to get the length of a single heartbeat cycle.

[0110] The last step is to convert the determined effective heartbeat cycle duration into standard time units based on the basic data, and calculate the number of heart beats in one minute to get the pulse rate per minute. For example, from the average time interval calculated previously, we can deduce the number of complete contraction cycles completed by the heart in one minute, which is the heart rate per minute value to be obtained.

[0111] This process combines photoelectric sensors to capture physiological signal changes, feature detection, statistical periodicity and frequency determination, completing the key analysis task of real-time pulse rate and blood oxygen health monitoring in the monitoring bracelet.

[0112] Next, the steps of identifying wrist shaking based on the data of the acceleration sensor and correcting the pulse signal distortion caused by the shaking are described in detail.

[0113] First, the built-in multi-axis accelerometer is enabled to continuously monitor the dynamic changes of the wrist, and the user's daily wrist movements are captured through this sensor. For example, a typical accelerometer can measure the acceleration components of the wrist along three perpendicular axes (X, Y, and Z axes).

[0114] For example, suppose we wear a monitoring bracelet to go shopping and need to stretch out our hands to gesture or shake while selecting clothes. At this time, the three-axis accelerometer will capture the corresponding acceleration value at each time point and send this data to the processing unit for real-time analysis.

[0115] Then, the data obtained from the three-axis accelerometer will be captured in real time and stored in a sequence in a time-ordered manner. This stored sequence will be used in the next data processing step. In the previous example, suppose our wrist starts to swing or rotate in a specific pattern, and a series of acceleration component values ​​will be generated over time.

[0116] Next, the component values ​​of the acceleration in each axis and the threshold values ​​pre-defined in the hardware or software are used to determine whether the wrist is shaking and the degree of shaking. Usually, shaking the wrist will cause a large change in movement, which can be clearly reflected by the change in the amplitude of the acceleration vector. If within a certain time window, all three acceleration components exceed a threshold, it is determined that the wrist is shaking violently.

[0117] Continuing with the above example, the rapid swinging of the user's hand while trying on clothes will cause the three acceleration components X, Y and Z to exceed the preset threshold values ​​within a period of time, thereby determining that the user is shaking his wrist.

[0118] Finally, after determining that there is obvious wrist shaking, the system will take certain digital signal processing measures, such as linear filtering and smoothing, within the signal range near the specific time point when the shaking occurs, to reduce the signal noise and pulse signal distortion caused by shaking. This can effectively improve the subsequent extraction of blood oxygen and pulse information. Assuming that the system determines that the time range of the user's wrist shaking violently is within t = 0.1 to t = 0.3 seconds, the system will select the blood flow light absorption signal sequence near this time period and smooth or fill it to reduce the noise level. Furthermore, the processed optical signal will synchronously measure the original heart rate data to determine the actual heart rate wave shape and correct the pulse rate estimation error caused by wrist shaking, thereby improving the accuracy of the pulse blood oxygen saturation detection results.

[0119] Next, the steps of calculating the blood oxygen saturation in the blood by using dual wavelength light detection of infrared light and red light are described in detail:

[0120] Specific wavelengths of light are emitted from two photoplethysmography sensors - one in the near-infrared spectrum (usually above 800nm) and the other in the red spectrum (usually around 600nm). When these lights pass through the human skin, such as the wrist, and shine on the flowing plasma in the veins and arteries, some of them are absorbed by the oxygenated and deoxygenated hemoglobin in the blood. After the light is transmitted or scattered to the photodetector, it is converted into an electrical signal. For example, in actual operation, two light sources, red light and infrared light, are set up inside the device to flash alternately and quickly; the corresponding receiving unit receives the intensity of the transmitted light or reflected light signal in real time, and converts it into an electrical signal and outputs it to the processing chip. This is the collection of raw PPG data.

[0121] Then, advanced digital signal processing algorithms are used to separate and extract specific signal components from the received signal, mainly the changes in optical density caused by blood circulation during the heart beat cycle. In order to reduce the interference of external clutter and noise and improve the signal-to-noise ratio, the data is filtered and smoothed accordingly. For example, in an example process, filtering is used to remove high-frequency noise, such as fluctuations caused by ambient light and skin color, and the heart beat cycle information is also used to synchronize the signal. The relative significance of the target component in the entire signal is further enhanced through the normalization processing method, ensuring the consistency and accuracy of the signal in subsequent analysis.

[0122] Next, an appropriate model is selected to establish the relationship between the absorbance ratio and the oxygen saturation level. In general, the absorbance ratio of red light and near-infrared light after passing through the human body will change systematically with the oxygen content of the blood sample: a high oxygen content means a lower absorbance ratio (i.e., a lower red / near-infrared ratio, indicating that more light can pass through to the other end), because oxygenated red blood cells absorb red light more weakly and near-infrared light more strongly, and vice versa; based on this, a suitable fitting equation is designed to quantify the correlation between the two. The equation can be determined by a pre-measured standard sample, or it can be iterated by a specially designed algorithm to obtain the optimal coefficient to meet the general law between individuals, ensuring the accuracy of the calculation results.

[0123] Finally, by calculating the ratio of the signal change value after light of different wavelengths penetrates the human body, the so-called blood-soluble ratio is obtained. This ratio represents the change in light flux caused by vascular dilation every time the heart beats. The blood oxygen content can be directly derived based on the aforementioned relationship model. In a monitoring process, when the two light sources emit infrared and red light to the wrist in turn and collect enough effective samples, the blood-soluble ratio can be calculated based on the measured absorbance ratio, and then the real-time pulse oxygen concentration value at the time of light emission can be estimated with the help of the previously established mathematical model.

[0124] Overall, this solution comprehensively utilizes the characteristics of different spectral regions, overcomes the deviations caused by many influencing factors by effectively isolating, extracting and calibrating signals, and uses reasonable physical mechanism modeling to ultimately achieve real-time and continuous tracking and recording of an individual's dynamic blood oxygen level, which will help to implement applications in various applications such as health warning and disease monitoring.

[0125] Next, the acquisition of continuous light absorption signal changes based on the photoplethysmography optical sensor is described as follows:

[0126] First, an optical channel is established between the optical sensor and the light source. This means that an LED light source (such as a light-emitting diode) is fixed in position to illuminate a small portion of the user's body tissue, usually a finger or earlobe with light-transmitting properties, and a photodetector is placed on the other side or on the same side (depending on the design of the photoplethysmography sensor) to receive the light passing through the area.

[0127] For example, in a wristband that monitors blood oxygen, after the user wears the device, the LED light passes through the user's fingertips. At this time, when the light source is turned on, the light emitted by the LED enters the finger tissue and passes through the finger, and then reaches the optical sensor side and is collected by the receiver.

[0128] Secondly, the receiver collects the changes in the intensity of light passing through the finger. Since hemoglobin with different amounts of oxygen in the blood circulation has different absorption capabilities for specific wavelengths of light, the intensity of light passing through the finger tissue will change periodically.

[0129] In the previous example, with each beat of the heart, the light passing through the finger will experience subtle changes in the flow of fresh oxygenated blood. This change is caused by the tissue's different absorption of specific frequencies (wavelengths) of light from the light source, and the receiver detects these changes, which reflect subtle fluctuations in the oxygen-carrying state of the blood and the heartbeat.

[0130] The third step is to convert the collected light intensity changes into voltage signal changes and record these changes. This is done by using photodetectors, such as phototransistors (PDs), to convert the collected light changes into electronic signals. In the above example, as blood flow and tissue absorption change, the phototransistor will generate a series of voltage values ​​of different sizes associated with the light changes.

[0131] The final step is to analyze the changes in the voltage signal to produce a continuous light absorption curve pattern, which can reflect changes in heart rate and indirectly derive the amount of information about pulse rate and blood oxygen through the ability of different forms of hemoglobin to absorb incident light to different degrees.

[0132] By analyzing the recorded signals, data processing algorithms can decode the peaks, amplitudes, and other parameters of these curves, extract pulse intervals, and calculate oxygen saturation. In the case of a health bracelet, the final output may be real-time heart rate and blood sample saturation values ​​to help users understand their health status.

[0133] To sum up the process of this step, photoelectric detection technology is used to monitor heart rate and oxygen saturation status based on the physical changes in blood.

[0134] Next, the specific steps of identifying wrist shaking based on the data of the acceleration sensor and correcting the pulse signal distortion caused by the shaking are described:

[0135] 1. First, the monitoring bracelet obtains data from the acceleration sensor in the X, Y, and Z directions, and the average acceleration value is obtained by summing and averaging these data.

[0136] 2. Determine whether the acceleration values ​​of the three axes obtained have values ​​exceeding the preset threshold value; that is, by comparing the absolute values ​​of the acceleration values ​​of the three axes X, Y, and Z with a pre-set threshold value, determine whether there is a large movement or vibration. For example, set this threshold THRESHOLD to 10g. When the absolute value of the acceleration of the X axis is 12g, the system will determine that the wrist is shaking. This judgment method enables the monitoring bracelet to detect changes in the user's wrist position in time to avoid the noise generated by the noise affecting the accuracy of blood oxygen saturation and heart rate measurement.

[0137] 3. If the wrist is judged to be shaking, the bracelet will enable a specially designed filtering algorithm, such as low-pass filtering or other adaptive filtering technology to remove the noise introduced by wrist shaking.

[0138] A specific example is provided below to help understand the working logic of the above process.

[0139] Example: A health monitoring bracelet has built-in acceleration detection modules in three directions (X, Y and Z). When the bracelet is in a static state (such as when the user puts his hand on the table and does not move), each direction displays a small value such as (1m / s 2 ,-0.5m / s 2 ,-0.5m / s 2 ); When the user suddenly moves his hand or the position of the bracelet changes rapidly, the relevant acceleration detection value will change greatly. For example, the acceleration of the X-axis changes from 1 to 20m / s in motion 2 ,The Y and Z axes also show similar change trends, exceeding the set standard values.,At this time, the processing unit detects an abnormal situation and thinks that the wrist is shaking.

[0140] The monitoring system then uses an algorithm to eliminate abnormal data to prevent it from further affecting the accuracy of readings of data such as blood oxygen saturation.

[0141] This solution can better ensure the validity of the data obtained when the user wears the monitoring device daily, while reducing problems such as data interference caused by slight movement of the bracelet on the wrist.

[0142] Next, the basic steps and practical applications of this method using dual wavelength light detection of infrared light and red light are described.

[0143] In the process of implementing this technology, we must first ensure that the light source module can alternately emit different light sources as needed: at the beginning, the light source module is switched to emit infrared light and illuminate the skin surface of the wearer (such as a patient or an athlete, etc.). At this time, the sensor records the intensity information of the light reflected back to the surface from under the tissue. This process can be regarded as collecting the feedback intensity under infrared light.

[0144] Next, the same process is used to test under red light conditions, that is, the previous data collection steps are repeated after changing the emitting light source to red light.

[0145] After collecting the light changes under two conditions, the oxygen status of hemoglobin can be evaluated by comparing the difference in light absorption. The principle here is that the absorption spectrum characteristics of hemoglobin and oxygen in the near-infrared region are obviously different, and the absorption of such photons under different saturations will show regular changes. Therefore, the SO2 value can be obtained based on this change relationship to measure the efficiency of red cells carrying gas in the human blood circulation system.

[0146] In order to ensure the accuracy, an empirical correction term needs to be introduced to reduce the error caused by interference sources such as external environmental noise. Formula SO2 = (I R / I red )×K usually needs to be obtained by establishing a calibration table with a large amount of data support in advance, where I R Represents the infrared light feedback value, I red This is the information obtained under the corresponding red light source, and the individual's actual oxygen saturation value is calculated through this ratio.

[0147] For example, if the user is in a stable motion state, the device first emits a beam with a wavelength of 940nm to penetrate the cortex to the location of the vein and artery lumen and capture the return signal I R Then adjust the wavelength to a shorter red light range such as 680nm, and complete the same round of information capture and recording as the denominator value I red Input, and finally compare the two sets of results to obtain the instantaneous oxygen concentration information at the current measurement point, and further use the preset formula SO2 = (I R / I red )×K to obtain more reliable and accurate indicators of cardiovascular physiological status.

[0148] Next, the specific steps of obtaining the heartbeat pulse signal according to the change analysis to determine the pulse rate value also involve:

[0149] Fourier transform is used to analyze the spectral components of the signal; high-frequency components are filtered based on the characteristics of the pulse signal; the peak position is identified to extract the effective heartbeat signal cycle length; the pulse rate value is obtained using the inverse of the pulse signal cycle length; formula: HeartRate = 60 / (Peak Interval Time / 60).

[0150] First, the Fourier transform is applied to convert the raw heart rate signal in the time domain into a representation in the frequency domain. This step helps identify the energy distribution of different frequencies by calculating the specific frequencies present in the signal. For example, the heartbeat pulse signal received from the monitoring bracelet within a specific time window may be a complex signal containing noise; after Fourier transforming it, we can observe significant peaks in the core frequency band related to the heart beat.

[0151] Then, filtering is performed on the spectrum according to the range of a specific frequency band to remove noise or unnecessary high-frequency parts that are higher than the frequency of the effective pulse signal component. This operation is necessary because the actual heartbeat frequency band is usually limited (usually between 0.5 and 2Hz). For example, if the result of the Fourier transform shows a peak distribution of many frequencies, including many high-frequency signal interferences, then a suitable filter can be used to exclude the part exceeding 2Hz, so that only the heart rate information that conforms to the actual physiological conditions is retained.

[0152] Next, after removing irrelevant frequency band signals, we need to find the peak position (usually the local maximum point) of each continuous valid signal in the timing diagram to obtain each valid heart rate signal that is not affected by noise. In this way, we can determine the exact moment and cycle of each heart beat in a series of continuous time intervals. If we see a series of obvious peaks after filtering, these peaks represent the time points at which a heart cycle begins and ends.

[0153] Finally, the time interval between two consecutive valid peaks is taken as the heartbeat signal period, and the number of heartbeats is obtained based on this length, which is then converted into heartbeats per minute. The formula HeartRate = 60 / peak interval time reflects this point, and the heartbeat frequency value in minutes can be directly obtained by the reciprocal operation. For example, when the average time between two valid signals is measured to be 1 second (that is, the peak interval time is equal to 1 second), substituting it into the above equation can obtain the average pulse rate during this period of time as 60 beats per minute.

[0154] By combining the above steps with mathematical operations, accurate and reliable heartbeat monitoring and analysis can be achieved, which has broad value and significance in practical applications such as health monitoring.

[0155] Next, the specific calculation steps for dual wavelength light detection based on infrared light and red light are described, including:

[0156] Measure the optical signal change values ​​at two wavelengths respectively;

[0157] Perform differential calculation to obtain the ratio of the two groups of light absorption intensity changes;

[0158] Apply the corrected absorption ratio to adjust the saturation calculation method;

[0159]

[0160] Among them, SO2 is the blood oxygen saturation, I 红外 is the signal intensity of infrared light, I 红光 is the signal strength of the red light, and K is a coefficient used for calibration or adjustment.

[0161] Before implementing this calculation step, it is necessary to understand the meaning of each variable and the purpose of the calculation process. Infrared light (IR) is usually used to pass through tissues and blood vessels, while red light (R) has similar but slightly different penetration capabilities. By comparing these two different wavelengths, the blood oxygen content change information of the blood sample can be calculated.

[0162] First, the illuminance of the infrared light (such as 940nm) and the red light (such as 660nm) as well as the remaining reflected light after absorption by the tissue are measured respectively, which is recorded as I R and I IR , thus obtaining the base value of the optical signal at the two wavelengths.

[0163] Starting from an example of a stable temperature, the photodiode measures the base value in a resting state without touching the wristband user's arm. When the finger moves slightly to represent the pulse change caused by blood flowing through, the change of light intensity over time is recorded and taken as a signal. For example, in a resting state, the recorded red light I R The initial baseline value may be 50μW, infrared light I IR The initial baseline value might be 200μW; movement produces a change, with the light intensity increasing to 55μW for red and 230μW for infrared.

[0164] Secondly, a differential operation is performed to calculate the intensity ratio of two light signals of different wavelengths using the change in the signal peak value, which is used to represent the different degrees of absorption of infrared and red light by blood:

[0165] ΔI IR =230-200=30μW

[0166] ΔI R =55–50=5μW

[0167] Then use the comparison results to calculate the ratio, the ratio is ΔI IR / ΔI R It is approximately equal to 6, and what is obtained is the incremental change of the absorbance ratio, that is, the attenuation ratio of light.

[0168] Next, a correction is needed to eliminate the impact of non-ideal conditions (such as fluctuations in light source power), leaving only the differences caused by changes in absorbance, and further adjust the saturation calculation through the algorithm to ensure the algorithm's resistance to tissue factors such as skin, and input the corrected absorbance ratio into the saturation formula for final calculation.

[0169] Finally, the blood oxygen saturation is calculated according to the formula.

[0170] The specific calculation is as follows:

[0171] Blood oxygen saturation = [(230 / 55) × 5.3 / 2] + [(0.002 × (230 230)) + (-16) × 5.3]

[0172] Blood oxygen saturation = [4.182 + (-84.9)] ≈ -80%

[0173] Obviously, the above values ​​are not in line with the actual blood sample, so the calculation results are only used for demonstration. In fact, K, slope and offset values ​​should be adjusted according to the specific equipment calibration and actual conditions.

[0174] This process provides continuous, non-invasive measurement of blood pressure and blood content, and can also detect pulses. When applied to health monitoring bracelets, it provides users with a continuous and convenient vital sign monitoring solution.

[0175] Specifically, first, the photoplethysmography optical sensor installed in the bracelet applies continuous light to the wearer's skin surface and collects the change data of the continuous photoelectric absorption signal generated. This part of the change will be used for further analysis, so as to measure the blood volume fluctuations caused by each beat of the heart, so as to determine the wearer's real-time pulse rate value.

[0176] Secondly, while measuring blood volume, this method introduces a key mechanism to solve an important problem that may be encountered when the bracelet is worn, that is, how to effectively deal with the large wrist shaking caused by the user in an active state. This phenomenon may generate noise interference and affect the stability and accuracy of the pulse signal. Therefore, the acceleration sensor is used to monitor the speed change of the hand in the three orthogonal axes (X axis, Y axis and Z axis) to accurately determine whether there is a large position movement of the wrist, and filter out such abnormal fluctuations to remove erroneous blood volume information, so as to correct and improve the accuracy of the pulse detection signal.

[0177] Furthermore, by synchronously utilizing dual wavelength light sources in the near-infrared band and the red light band to illuminate the skin and collect information reflected from the transmitted or reflected tissue, due to the different attenuation characteristics of the two different wavelengths of light after penetrating human tissues with different oxygenation levels, by comparing and calculating this change difference, the changes in the state of hemoglobin can be quantitatively analyzed, and ultimately the hemoglobin saturation level, that is, the blood oxygen level, of the wearer's blood sample can be determined.

[0178] The above design not only solves the technical problem of inaccurate detection caused by violent wrist movements of the user, but also achieves the purpose of high-precision synchronous monitoring of pulse rate value and hemoglobin content in blood samples.

[0179] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A health monitoring bracelet detection method capable of real-time detection of pulse rate and blood oxygen, characterized in that: The method comprises: S1: Acquire continuous light absorption signal changes based on photoplethysmography optical sensor; S2: Determine the pulse rate value based on the heartbeat pulse signal obtained by the change analysis; S3: Identify wrist shaking based on the data of the acceleration sensor and correct the pulse signal distortion caused by the shaking; S4: Dual wavelength optical detection using infrared light and red light is used to calculate the oxygen saturation in the blood.

2. A health monitoring bracelet detection method capable of real-time detection of pulse rate and blood oxygen according to claim 1, characterized in that: The step S1 specifically includes: activating the optical sensor and emitting a light source of a predetermined intensity to the tissue surface; Continuously collect data on changes in light intensity transmitted or reflected by tissues; Using a specific algorithm to extract effective light absorption fluctuation signals from the data; Filter the collected raw light absorption fluctuation signal to eliminate noise interference; Form a clear and stable continuously changing signal output for subsequent analysis; Wherein, the extracting effective light absorption fluctuation signal from the data using a specific algorithm comprises: Discrete wavelet transform is applied to decompose the light intensity variation data ΔI(t) and extract the fluctuation signal related to the heart pulse. The specific formula is: Among them, W j,k is the wavelet coefficient, t represents time, N represents the total number of signal samples, ΔI(t) is the light intensity change data, ψ is the mother wavelet function, j and k represent the scale and position parameters respectively. By selecting the appropriate scale jjj and position kkk, the components related to the heart pulse can be extracted; The filtering process of filtering the collected original light absorption fluctuation signal to eliminate noise interference adopts the formula: Among them, y(t) is the filtered signal, ΔI(t) is the light intensity change data, t is time, is the estimated noise, μ is the step size factor, e(t) is the error signal, and by adjusting μ, the noise interference can be adaptively eliminated.

3. The detection method of a health monitoring bracelet capable of real-time detection of pulse rate and blood oxygen according to claim 1, characterized in that: The step S2 specifically includes: The heartbeat peak position is detected based on the continuous light absorption signal changes; Use autocorrelation technology to locate clear and reliable periodic feature points; Select appropriate pulse template and match the signal sequence through convolution analysis; Determine adjacent heartbeat cycles and calculate pulse cycle time length; The heart rate per minute, or pulse rate, is calculated based on the cycle.

4. The health monitoring bracelet detection method capable of real-time detection of pulse rate and blood oxygen according to claim 1, characterized in that: The step S3 specifically includes: Enable the built-in multi-axis accelerometer to continuously monitor wrist dynamic changes; Capture and store the component value sequence of acceleration in each axis in real time; Determine whether significant motion shaking occurs based on the comparison between the instantaneous amplitude of the motion vector and a preset threshold value; Smoothing or interpolation processing is performed on the light absorption signal within a time window before and after the shaking moment to reduce the impact of the shaking; By comparing the heart rate wave obtained by synchronous measurement with the heart rate wave after motion interference suppression, the final heart rate calculation result is corrected.

5. The health monitoring bracelet detection method capable of real-time detection of pulse rate and blood oxygen according to claim 1, characterized in that: The step S4 specifically includes: Two photoplethysmography sensors emit near infrared light above 800nm ​​and visible red light around 600nm, respectively, and collect signal feedback information; Using digital signal processing technology to extract specific signal components from the feedback information and normalize them; Select appropriate equations to fit the relationship between the red light / near infrared absorption ratio and different blood oxygenation states; Measure and record the ratio of the continuous blood volume changes under the two light sources, and calculate the proportional factor; The blood oxygen content in the blood is derived according to the proportional factor, thereby obtaining the blood oxygen saturation level; The specific calculation steps of the dual wavelength light detection based on infrared light and red light include: Measure the optical signal change values ​​at two wavelengths respectively; Perform differential calculation to obtain the ratio of the two groups of light absorption intensity changes; Apply the corrected absorption ratio to adjust the saturation calculation: SO2 = [(I 红外 / I 红光 )×K / 2]+[Slope rate×(I 红外 -I 红光 )+cut distance]×K, Among them, SO2 is the blood oxygen saturation, I 红外 is the signal intensity of infrared light, I 红光 is the signal strength of the red light, and K is a coefficient used for calibration or adjustment.

6. The health monitoring bracelet detection method capable of real-time detection of pulse rate and blood oxygen according to claim 1, characterized in that: The step S1 of acquiring continuous light absorption signal changes based on the photoplethysmography optical sensor specifically includes: Establishing an optical channel between the optical sensor and the light source; The change of the intensity of the light passing through the finger is collected by the receiver; The collected light intensity changes are converted into voltage signal changes and recorded; The continuous light absorption curve pattern is analyzed by utilizing the change of the voltage signal.

7. The health monitoring bracelet detection method capable of real-time detection of pulse rate and blood oxygen according to claim 1, characterized in that: The step S3 of identifying wrist shaking based on the data of the acceleration sensor and correcting the pulse signal distortion caused by the shaking also includes: Obtain the X, Y, and Z axis data from the acceleration sensor to calculate the average acceleration value; If any axial acceleration exceeds a preset threshold, it is considered that shaking occurs; If the wrist is judged to be in a shaking state, a filter is applied to remove noise; Formula judgment: If the absolute value of the acceleration of a certain axis is greater than the threshold, it is considered that the wrist is shaking.

8. The health monitoring bracelet detection method capable of real-time detection of pulse rate and blood oxygen according to claim 1, characterized in that: The dual wavelength light detection using infrared light and red light in step S4 specifically includes: Switch the light source module to illuminate the skin surface with infrared light waves to collect feedback intensity signals; Then, the red light is used for illumination to repeat the above feedback intensity collection process; Compare the absorption ratio of the two to calculate the blood oxygen saturation level; Using the formula judgment condition, SO2=((I R / I red )×K) to correct the error factors in the calculation process, Among them, SO2 represents blood oxygen saturation, I R Indicates the intensity of infrared light, I red Indicates the intensity of red light.

9. The health monitoring bracelet detection method capable of real-time detection of pulse rate and blood oxygen according to claim 1, characterized in that: The change analysis to obtain the heartbeat pulse signal to determine the pulse rate value specifically includes: Apply Fourier transform to analyze the spectral components of the signal; Filter high-frequency components based on the characteristics of pulse signals; Identify the peak position to extract the effective heartbeat signal cycle length; The pulse rate value is obtained by using the inverse of the pulse signal cycle length, using the formula: Pulse rate = 60 / peak interval time.

10. A health monitoring wristband detection system capable of real-time detection of pulse rate and blood oxygen, using the health monitoring wristband detection method capable of real-time detection of pulse rate and blood oxygen as claimed in any one of claims 1 to 9, characterized in that: The system comprises: An optical sensing module for acquiring continuous light absorption signal changes based on photoplethysmography; The pulse rate measurement module is used to obtain the heartbeat pulse signal and measure the pulse rate value according to the change analysis of the light absorption signal; An acceleration sensor module is used to identify wrist shaking based on the data of the acceleration sensor and correct the pulse signal distortion caused by the shaking; The blood oxygen detection module is used to calculate the blood oxygen saturation in the blood using dual wavelength light detection of infrared light and red light.

Images