Electrocardio and blood oxygen monitoring method and system based on wearable device

Abstract

The invention discloses an electrocardio and blood oxygen monitoring method and system based on wearable equipment, and relates to the technical field of electrocardio monitoring. The AD8232 chip is used for electrocardiosignal acquisition, and the MQ-LAU-002 multi-wavelength emitter and the MQ237LF photoelectric converter are used for blood oxygen signal acquisition, so that the signal acquisition precision is improved, the environmental interference and noise influence are reduced, accurate electrocardiosignal data acquisition is ensured, and the accuracy of electrocardiosignal acquisition is improved. The problem that an existing wearable device is insufficient in signal acquisition precision and stability can be solved, and more reliable health monitoring data are provided. Collection and analysis of electrocardio and blood oxygen signals are integrated, the limitation that traditional health monitoring equipment is single in function is broken through, two different types of physiological data are subjected to joint analysis, more comprehensive health assessment can be provided, association between electrocardio and blood oxygen data can be recognized, and early warning can be conducted in time.

Description

Technical Field

[0001] This invention relates to the field of electrocardiogram (ECG) monitoring technology, and in particular to a method and system for monitoring ECG and blood oxygenation based on wearable devices. Background Technology

[0002] Existing ECG and pulse oximetry monitoring devices are widely used in clinical and home health management, but they generally suffer from several technical defects and shortcomings, affecting monitoring effectiveness and user experience. Traditional ECG devices and pulse oximeters have problems with signal acquisition accuracy. Due to factors such as poor electrode contact, noise interference, and environmental changes, the ECG signals of existing devices are often affected, leading to unstable ECG data and even incorrect assessments of heart health. Traditional ECG and pulse oximetry monitoring devices are usually large and heavy, making them unsuitable for daily wear and prolonged use. Existing ECG monitoring devices generally require connecting wires and can only be used in hospitals or clinics, failing to meet patients' continuous monitoring needs in daily life. Pulse oximeters are mostly handheld and cannot be linked with other health data for real-time analysis, limiting the implementation of personalized health management.

[0003] While existing wearable devices possess wireless transmission capabilities such as Bluetooth or Wi-Fi, they still suffer from transmission latency and instability when handling large amounts of data. The transmission of real-time ECG and blood oxygen data can be affected by transmission speed and network fluctuations, preventing users from obtaining health data in a timely manner and impacting the accuracy of health alerts.

[0004] Battery life is a significant issue for current wearable devices, especially when long-term, continuous data collection and transmission are required, as batteries often cannot support sustained operation. Frequent charging is not only inconvenient but also reduces device availability and reliability. Existing devices typically offer only a few hours of continuous use, which is insufficient for the needs of 24 / 7, long-term monitoring.

[0005] Existing monitoring devices typically only have data acquisition capabilities, lacking intelligent analysis and automated processing. Most devices rely on users to interpret the data themselves or display data results through mobile applications, failing to perform in-depth data analysis or health prediction. Even some devices with basic health data processing functions have low depth and accuracy in intelligent analysis, unable to achieve personalized health management, and failing to provide efficient solutions in early warning, disease prediction, and health intervention.

[0006] Furthermore, existing ECG monitoring devices and pulse oximeters are typically standalone devices, unable to simultaneously monitor and analyze ECG signals and pulse oxygenation data in a linked manner. This results in isolated health data, lacking effective comprehensive assessment and limiting the system's application in personalized health management. Summary of the Invention

[0007] The purpose of this invention is to provide a method and system for monitoring electrocardiogram and blood oxygen based on wearable devices, so as to improve the above-mentioned technical problems.

[0008] To achieve the above-mentioned objectives, the embodiments of the present invention provide the following technical solutions:

[0009] A method for monitoring electrocardiogram and blood oxygen based on wearable devices, comprising:

[0010] Electrocardiogram (ECG) signals are acquired using an ECG acquisition device, and blood oxygen saturation is acquired and uploaded using a blood oxygen saturation signal acquisition device.

[0011] Hilbert-Huang transform and wavelet transform were used to extract features from electrocardiogram (ECG) signals, obtain ECG-instantaneous frequency features, and generate a two-dimensional time-frequency distribution map.

[0012] An initial health assessment model was constructed and trained using the ETO algorithm to obtain the health assessment model; the two-dimensional time-frequency distribution spectrum was processed through the health assessment model to obtain the user's electrocardiogram health assessment results;

[0013] Based on the user's electrocardiogram health assessment results, blood oxygen saturation, and personal health indicators, health recommendations are generated through a large language model.

[0014] An electrocardiogram and blood oxygen monitoring system based on a wearable device, comprising:

[0015] An electrocardiogram (ECG) acquisition device is used to acquire raw ECG signals and perform contact failure detection and noise reduction to obtain contact detection results and ECG signals.

[0016] The blood oxygen saturation acquisition module is used to acquire the initial blood oxygen saturation and perform electrode contact detection to obtain the electrode contact detection result and blood oxygen saturation.

[0017] Communication components are used for data transmission;

[0018] The two-dimensional time-frequency distribution map generation module is used to extract features from electrocardiogram signals using Hilbert-Huang transform and wavelet transform, obtain electrocardiogram-instantaneous frequency features, and generate a two-dimensional time-frequency distribution map.

[0019] The health assessment model building and training module is used to build an initial health assessment model and train it using the ETO algorithm to obtain the health assessment model.

[0020] The ECG health assessment result generation module is used to process the two-dimensional time-frequency distribution spectrum through the health assessment model to obtain the user's ECG health assessment results;

[0021] The health advice generation module is used to generate health advice based on the user's electrocardiogram health assessment results, blood oxygen saturation, and personal health indicators, using a large language model.

[0022] The client is used to remind users to adjust electrodes based on electrode contact detection results and contact detection results; it also receives, displays, and invokes health assessment analysis results and health suggestions.

[0023] The beneficial effects of this invention are as follows:

[0024] This invention uses the AD8232 chip to acquire electrocardiogram (ECG) signals and the MQ-LAU-002 multi-wavelength transmitter and MQ237LF photoelectric converter to acquire blood oxygen signals. This improves the accuracy of signal acquisition, reduces environmental interference and noise, and ensures accurate ECG data acquisition. It can solve the problems of insufficient signal acquisition accuracy and stability of existing wearable devices and provide more reliable health monitoring data.

[0025] This invention uses Bluetooth Low Energy as a communication component to achieve real-time transmission of ECG and blood oxygen data with extremely low power consumption, solving the problems of slow transmission speed and high latency of traditional devices. Furthermore, due to the low power consumption characteristics of Bluetooth Low Energy, the device's battery life is extended, meeting the needs of long-term wear.

[0026] This invention integrates the acquisition and analysis of electrocardiogram (ECG) and blood oxygenation signals, breaking the limitations of the single function of traditional health monitoring devices. By jointly analyzing two different types of physiological data, it can provide a more comprehensive health assessment, identify the correlation between ECG and blood oxygenation data, and provide timely warnings of potential health problems. Attached Figure Description

[0027] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is a flowchart of the method in an embodiment of the present invention;

[0029] Figure 2 This is an internal circuit diagram of the AD8232 chip in an embodiment of the present invention;

[0030] Figure 3 This is a circuit diagram of the AD8232 module in an embodiment of the present invention;

[0031] Figure 4 This is a circuit diagram of the blood oxygen saturation signal acquisition device in an embodiment of the present invention;

[0032] Figure 5 This is a diagram of the CNN-Transformer-BiLSTM model structure in an embodiment of the present invention;

[0033] Figure 6 This is a system structure diagram in an embodiment of the present invention;

[0034] Figure 7 This is a waveform diagram of an electrocardiogram (ECG) signal in an embodiment of the present invention;

[0035] Figure 8 This is a blood oxygen saturation display diagram in an embodiment of the present invention. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0037] Please see Figure 1 This embodiment provides a method for monitoring electrocardiogram and blood oxygen based on wearable devices. Figure 1 The execution entity of the method shown can be a software and / or hardware device. The execution entity of this application can include, but is not limited to, at least one of the following: user equipment, network equipment, etc. User equipment can include, but is not limited to, computers, smartphones, personal digital assistants (PDAs), and the aforementioned electronic devices. Network equipment can include, but is not limited to, a single network server, a server group consisting of multiple network servers, or a cloud based on cloud computing consisting of a large number of computers or network servers. Cloud computing is a type of distributed computing, consisting of a super virtual computer composed of a group of loosely coupled computers. This embodiment does not impose any limitations on this.

[0038] A method for monitoring electrocardiogram and blood oxygen based on wearable devices, comprising:

[0039] S1. Collect electrocardiogram (ECG) signals through an ECG acquisition device and collect and upload blood oxygen saturation signals through a blood oxygen saturation signal acquisition device.

[0040] S1 includes:

[0041] S1-1. Construct an electrocardiogram (ECG) acquisition device and a blood oxygen saturation signal acquisition device;

[0042] S1-2. Acquire raw ECG signals through an ECG acquisition device and perform contact failure detection to obtain contact detection results; the ECG acquisition device includes custom adhesive electrodes, wearable fabric electrodes, and an AD8232 module; the AD8232 module includes an AD8232 chip, a 0.5Hz dual-pole high-pass filter, and a 40Hz dual-pole low-pass filter.

[0043] If the AD8232 module detects an original ECG signal amplitude of <0.1mV or a baseline drift of >0.5mV within three consecutive cardiac cycles (i.e., 2 seconds), the contact detection result is determined to be poor contact.

[0044] S1-3. Based on the contact detection results, the original ECG signal is denoised using the ECG acquisition device to obtain the ECG signal. When the contact detection result indicates poor contact, the AD8232 chip sends the corresponding signal and transmits it to the client via Bluetooth, displaying a red warning message "Electrode loose, please adjust position," further reminding the user to confirm whether the customized adhesive electrode or wearable fabric electrode is making good contact. When the contact detection result is normal, the original ECG signal is denoised using a 0.5Hz dual-pole high-pass filter and a 40Hz dual-pole low-pass filter.

[0045] Customized adhesive electrodes use flexible materials suitable for long-term wear as electrode material and employ full-coverage conductive hydrogel to provide a larger conductive area, ensuring stable contact between the electrode and the skin, reducing signal loss and interference, and improving the quality of the original electrocardiogram signal.

[0046] The wearable fabric electrodes are made of silver fiber. Silver fiber is soft and elastic, which allows it to conform better to the skin and increase wearing comfort. It also has high chemical stability, corrosion resistance, abrasion resistance and antibacterial properties, which extend the life of the wearable fabric electrodes and reduce the risk of infection.

[0047] Custom-fit adhesive electrodes and wearable fabric electrodes are typically placed on the chest, close to the heart, to more effectively acquire raw electrocardiogram (ECG) signals. Custom-fit adhesive electrodes include a left chest electrode (LA), a right chest electrode (RA), and an RL electrode.

[0048] This invention reduces the size of the device and increases wearing comfort by employing flexible materials and a lightweight design, making it suitable for prolonged wear and solving the problems of bulky and inconvenient traditional devices. This embodiment uses the AD8232 chip (model AD8232ACPZ-RL), which internally includes:

[0049] The instrumentation amplifier IA is used to amplify weak raw ECG signals while suppressing common-mode noise. It features high input impedance and high common-mode rejection ratio.

[0050] Operational amplifier A1 is used for additional gain and low-pass filtering. The gain and filtering characteristics are set by an external resistor and capacitor network to process the amplified original ECG signal so that the frequency range and amplitude of its output signal are suitable for subsequent analog-to-digital conversion and analysis.

[0051] The right leg drive amplifier A2 is used to feed a weak current back to the human body to counteract the effects of common-mode voltage. The feedback network enables A2 to adjust the output current, balance the potential difference between the electrodes, and reduce noise.

[0052] Reference voltage amplifier A3 is used to generate a stable reference voltage for other modules through an internal reference voltage source and operational amplifier, avoiding the impact of power supply fluctuations on circuit performance; lead detachment monitoring capacitor is used to detect poor contact or detachment of electrodes with skin by monitoring voltage changes at the input terminal, and to provide feedback.

[0053] A 0.5Hz dual-pole high-pass filter is used to filter out low-frequency signals other than baseline drift in the original electrocardiogram signal by setting the cutoff frequency;

[0054] A 40Hz dual-pole low-pass filter is used to filter out high-frequency noise in the original ECG signal, so as to retain the low-frequency ECG signal components, making the signal waveform smoother and reducing sharp waveform changes.

[0055] The lead detachment monitoring capacitors include capacitors C1 and C2, which are used to detect whether the electrode is making poor contact.

[0056] like Figure 2As shown, internally, the first output, non-inverting input, and inverting input of the instrumentation amplifier IA serve as pins 1 to 3 of the AD8232 chip, respectively. The fifth output of the instrumentation amplifier IA is connected to one end of a 150kΩ resistor, the other end of which is connected to the inverting input of the right-leg drive amplifier A2 and serves as pin 4 of the AD8232 chip. The non-inverting input of the right-leg drive amplifier A2 serves as pin 5 of the AD8232 chip. One end of a 10kΩ resistor serves as pin 6 of the AD8232 chip, and the other end... One end of switch S2 is connected to one end of the AD8232 chip; the non-inverting input of operational amplifier A1 serves as pin 7 of the AD8232 chip; the other end of switch S2 is connected to the non-inverting input of right leg drive amplifier A2, the fourth output of instrumentation amplifier IA, and the output of reference voltage amplifier A3, and serves as pin 8 of the AD8232 chip; the inverting input and output of operational amplifier A1 serve as pins 9 and 10 of the AD8232 chip, respectively; capacitors C1 and C2 serve as pins 11 and 12 of the AD8232 chip, respectively; fault pin FR and input pin... Shutdown pin FR, ground pin GND, +V s The pins are respectively used as pins 13 to 17 of the AD8232 chip; the input terminal of the reference voltage amplifier A3 is used as pin 18 of the AD8232 chip; the third output terminal of the instrumentation amplifier IA is connected to one end of another 10KΩ resistor and is used as pin 19 of the AD8232 chip; the second output terminal of the instrumentation amplifier IA is connected to one end of the switch S1 and is used as pin 20 of the AD8232 chip; the other end of the other 10KΩ resistor is connected to the other end of the switch S1.

[0057] Compared to single-pole high-pass filters, double-pole high-pass filters exhibit faster attenuation near the cutoff frequency, resulting in better suppression of low-frequency noise. When designing a double-pole high-pass filter, the quality factor must be considered. A higher quality factor indicates a sharper filter response, while a lower quality factor indicates a flatter response. The quality factor is directly related to the resistors and capacitors in the double-pole high-pass filter circuit, and can therefore be controlled by adjusting the resistors to achieve narrowband bandpass filtering or maximum bandpass flatness.

[0058] The transfer function H(s), frequency domain gain ||H(jω)||, and cutoff frequency f of a two-pole high-pass filter. c,h The corresponding formula is:

[0059]

[0060]

[0061] Where s represents the complex frequency variable, R 16 R17 Both represent resistance, C 17 C 20 All represent capacitance, j represents the imaginary unit, ω represents angular frequency, and π represents pi.

[0062] In this embodiment, R 16 =R 17 ≥100kΩ and C 17 =C 20 When the resistance R 18 The smaller the value of R, the larger the quality factor Q, and the larger the corresponding peak value of the filter; however, when R... 18 If R is too small, it will affect the stability of the filter. Therefore, R... 18 Set the value to R 16 Or R 17 The filter's bandpass is flattest when it is 0.14 times that of the original filter. (Set R...) 16 =R 17 =10MΩ, making the roll-off of the dual-pole high-pass filter 20dB / decibels. Since the initial roll-off was 40dB / decibels, reducing it to 20dB / decibels will not significantly affect the out-of-band low-frequency signal suppression performance of the dual-pole high-pass filter. Therefore, setting the cutoff frequency to 0.5Hz can reduce the phase distortion of the signal and improve the signal fidelity.

[0063] A dual-pole low-pass filter effectively removes high-frequency noise while preserving low-frequency ECG signal components, resulting in a smoother signal waveform, reduced sharp waveform changes, and a relatively flat response over a wide frequency range, avoiding distortion of the useful signal. Compared to a single-pole low-pass filter, a dual-pole low-pass filter provides better signal quality. In this embodiment, a Sallen-Key filter topology is used to design the low-pass filter. This structure is an active filter using operational amplifiers, resistors, and capacitors, and can be configured as a low-pass, high-pass, band-pass, or band-stop filter. The low-pass Sallen-Key filter is simple and stable, making it ideal for filtering high-frequency noise. Its cutoff frequency fc c,l The formulas for Gain and Quality Factor Q are:

[0064]

[0065]

[0066] Among them, R 26 R 27 R 28 R 29 Both represent resistance, C 20 C 21All represent capacitors. In this embodiment, adjusting the gain of the dual-pole low-pass filter will affect its quality factor Q. The quality factor Q is usually 0.5 or 0.7. To achieve maximum flatness and a sharp cutoff frequency, Q is set to 0.7. Gain < 3 because when Gain > 3, the circuit of the dual-pole low-pass filter may become unstable. To achieve the highest possible gain, C can be modified. 20 C 21 The capacitance value of C 21 The capacitance value is set to at least C. 20 Four times the capacitance value extends the RC filter time constant, thereby achieving higher gain and further improving circuit stability. A 40Hz bipolar low-pass filter simultaneously drives electrode RL to improve common-mode rejection performance, and the gain of operational amplifier A1 is set to 11, thus achieving a total gain of 1100 for the ECG acquisition device.

[0067] like Figure 3As shown, the AD8232 module includes a chip U10 with model number AD8232ACPZ-RL; pin 1 of U10 is connected to one end of capacitor C17; pin 2 of U10 is connected to one end of resistor R24; pin 3 of U10 is connected to one end of resistor R23; pin 4 of U10 is connected to one end of capacitor C19; pin 5 of U10 is connected to the other end of capacitor C19 and one end of resistor R21; pin 6 of U10 is connected to one end of resistor R20, one end of resistor R18 and one end of capacitor C18; pin 7 of U10 is connected to one end of capacitor C20 and one end of resistor R26; pin 8 of U10 is connected to the other end of capacitor C20. One end of resistor R28 is connected as the REFOUT terminal; pin 9 of U10 is connected to the other end of resistor R28 and one end of resistor R29; pin 10 of U10 is connected to the other end of resistor R29 and one end of capacitor C21, and serves as the VP terminal; pins 11-13 and pin 15 of U10 serve as IO26, IO25, SDN, and VDD3V3 terminals, respectively; pin 14 of U10 is grounded; pin 16 of U10 is connected to one end of capacitor C23 and grounded; pin 17 of U10 is connected to the other end of capacitor C23 and serves as the VDD3V3 terminal; pin 18 of U10 is connected to one end of resistor R31 and one end of resistor R30, respectively. One end of U10 is connected to one end of capacitor C22; pin 19 of U10 is connected to one end of resistor R17; pin 20 of U10 is connected to the other end of capacitor C17 and one end of resistor R16; pin 21 of U10 is grounded; the other end of resistor R24 ​​is connected to one end of resistor R22 and serves as the LA terminal; the other end of resistor R23 is connected to one end of resistor R25 and serves as the RA terminal; the other end of resistor R21 serves as the RL terminal; the other end of resistor R20 is connected to the other ends of resistor R22, resistor R25, and one end of resistor R19; the other end of resistor R19 serves as the VDD3V3 terminal; the other end of resistor R18 is connected to the other end of resistor R16 and one end of capacitor C22; pin 19 of U10 is connected to one end of resistor R17; pin 20 of U10 is connected to the other end of capacitor C17 and one end of resistor R16; pin 21 of U10 is grounded; the other end of resistor R24 ​​is connected to one end of resistor R22 and serves as the LA terminal; the other end of resistor R23 is connected to one end of resistor R25 and serves as the RA terminal; the other end of resistor R21 serves as the RL terminal; the other end of resistor R20 is connected to the other end of resistor R16 and one end of resistor R17; the other end of resistor R18 is connected to the other end of resistor R16 and one end of capacitor C22; the other end of U10 is connected to one end of capacitor C22 and one end of resistor R17; the other end of U10 is connected to one end of capacitor C17 and one end of resistor R18 ... The other end of resistor R17 and the other end of capacitor C18 serve as the REFOUT terminal; the other end of resistor R26 is connected to one end of resistor R27 and the other end of capacitor C21 respectively; the other end of resistor R27 serves as the IAOUT terminal; the other end of resistor R31 serves as the VDD3V3 terminal; the other end of resistor R30 is connected to the other end of capacitor C22 and grounded; the LA, RA, and RL terminals are all externally connected to the AUDIO-JACKSMD2 chip JP1; the VP terminal is externally connected to the positive terminal of photodiode D5, and the negative terminal of photodiode D5 is connected to the grounding resistor R9; the SDN terminal is externally connected to one end of resistor R10, and the other end of resistor R10 is connected to the VDD3V3 terminal.

[0068] S1-4. The blood oxygen saturation signal acquisition device emits red and infrared light and measures the changes in reflected / transmitted light intensity.

[0069] The blood oxygen saturation signal acquisition device employs the photoelectric solvent pulse wave method, which includes:

[0070] The MQ-LAU-002 multi-wavelength transmitter is used to emit 660nm red light and 940nm infrared light to measure blood oxygen saturation.

[0071] The MQ237LF photoelectric converter is used to receive the intensity changes of reflected / transmitted light after red / infrared light penetrates tissue and convert it into an electrical signal.

[0072] The TPW3221-VR analog switch chip is used to receive control commands; based on the control commands, it switches the electrical signals corresponding to red light / infrared light and sends them to the 2078TWE0b microcontroller.

[0073] The 2078TWE0b microcontroller is used to send control commands; it receives and analyzes the electrical signals corresponding to red / infrared light to obtain blood oxygen saturation.

[0074] like Figure 4As shown, the blood oxygen saturation signal acquisition device includes chip U1, chip U2, chip U3 (model 2078TWE0bQFN32), and chips J4-J8; pin 1 of U1 serves as the VSS terminal; pin 2 of U1 is connected to one end of inductor L1; the other end of inductor L1 is connected to one end of capacitor C4 and pin 1 of J6, serving as the VCC terminal; the other end of capacitor C4 serves as the VSS terminal; pin 3 of U1 is connected to one end of capacitor C3 and pin 2 of J5, serving as the VDD terminal; the other end of capacitor C3 serves as the VSS terminal; pins 1, 3-7 of J5 serve as the VSS, SCL, SDA, RES, DC, and CS terminals, respectively; pin 2 of J6 is connected to pin 3 of J6, serving as the VSS terminal; pins 4-5 of J6 serve as the RXD and TXD terminals, respectively. Pin 1 of U2 is connected to one end of resistor R6, one end of capacitor C2, pins 2 and 10 of U2, and serves as the VDD terminal; the other end of capacitor C2 serves as the VSS terminal, U2-IN2 terminal, and RD terminal; pins 3-4, pin 6, and pins 8-9 of U2 serve as the IR terminal, U2-IN1 terminal, and VSS terminal, respectively; pin 5 of U2 is connected to one end of resistor R1; the other end of resistor R1 is connected to one end of resistor R2 and one end of resistor R3, respectively; the other end of resistor R3 serves as the VSS terminal; pin 7 of U2 is connected to the other end of resistor R2. Pins 1-3, 6-9, 14, 22, 25-30, and 32 of U3 are designated as SCL, DIO, DAT, U2-IN2, U2-IN1, VDD, VSS, PD, VSS, IICSCL, IICSDA, RES, DC, SCL1, CS, and SDA terminals, respectively. Pin 15 of U3 is connected to one end of resistor R5, with the other end of R5 serving as the RXD terminal. Pin 6 of U3 is connected to one end of resistor R4, with the other end of R4 serving as the TXD terminal. Pin 24 of U3 is connected to one end of capacitor C1 and serves as the VDD terminal; the other end of capacitor C1 serves as the VSS terminal. Pins 1-6 of J4 are used as DAT, RD, IR, PD, VSS, and VDD terminals, respectively; pins 1-4 of J7 are used as VDD, IISSCL, IICSDA, and VSS terminals, respectively; pins 1-4 of J8 are used as VSS, SCL, DIO, and VDD terminals, respectively.

[0075] The photoelectric solvent pulse wave method is a non-invasive detection method that avoids the risks of trauma and infection, and enables continuous, real-time health monitoring, effectively reducing the economic and time costs for patients.

[0076] S1-5. Based on the change in light intensity, obtain the blood oxygen electrode contact detection results and calculate the blood oxygen saturation using the Lambert-Beer law.

[0077] Initial blood oxygen saturation was calculated based on changes in light intensity using the Lambert-Beer law; the standard deviation σ of the signal amplitude and the noise variance of the light intensity changes were also calculated. According to the formula: Calculate the Quality of Life (SQI); based on the initial blood oxygen saturation, calculate the corresponding fluctuation value; if the SQI is less than 15dB or the fluctuation value is greater than 3% / min, the blood oxygen electrode contact detection result is determined to be abnormal, the corresponding signal is transmitted to the client, and a prompt "Please adjust the finger clip position" pops up; otherwise, the blood oxygen electrode contact detection result is determined to be normal, and the initial blood oxygen saturation is used as the blood oxygen saturation value. 10 (·) represents the logarithmic function with base 10.

[0078] According to the Lambert-Beer law, when there is no pulsation in the arterial wall, the reflected light has only a DC component I. DC The corresponding formula is:

[0079]

[0080] When blood vessels pulsate due to the heartbeat, the optical path changes, and the reflected light becomes a combination of direct current and alternating current (AC) components. DC +I AC The corresponding formula is:

[0081]

[0082] Where I0 represents the incident light intensity, F and e represent the tissue absorption coefficient and constant, respectively, and L represents the optical path length. ε Hb These represent the absorbance coefficients of oxyhemoglobin and deoxyhemoglobin, respectively. c Hb ΔL represents the concentrations of oxyhemoglobin and deoxyhemoglobin, respectively, and ΔL represents the change in optical path length.

[0083] By integrating the formulas for DC plus AC component and DC component, we obtain:

[0084]

[0085] Divide both sides of the integrated formula by I. DC Taking the logarithm, we get:

[0086]

[0087] Because of I DC >>I AC ,thereby If it can be considered as infinitesimal, then:

[0088]

[0089] Therefore, we get:

[0090] When irradiated with light of different wavelengths, we can obtain:

[0091]

[0092] Subtracting the formulas corresponding to two different wavelengths of light, we obtain the proportionality constant R:

[0093]

[0094] According to the formula:

[0095]

[0096] The blood oxygen saturation SpO2 is obtained; where ln(·) represents the logarithmic function, and λ1 and λ2 represent two different wavelengths of light.

[0097] This invention monitors the contact status of electrodes in real time, ensuring accurate signal acquisition and avoiding signal loss or inaccuracy caused by poor contact, thereby improving the quality of monitoring data and enabling continuous health monitoring in daily life.

[0098] S1-6. Transmit ECG signals and blood oxygen saturation to the client via Bluetooth Low Energy (BLE) and upload them to the cloud. In the client, the ECG signals are visualized in the form of waveforms. Users and doctors can intuitively understand the changes in ECG signals through waveforms, allowing users to check their ECG health status at any time.

[0099] In this embodiment, the client is a WeChat mini-program that receives transmitted electrocardiogram (ECG) signals and blood oxygen saturation, and displays the user's health information in real time. ECG signals record cardiac electrical activity, providing information such as heart rate, rhythm, P wave, QRS complex, T wave, ST segment, and QT interval, used to identify arrhythmias, myocardial ischemia, and other problems. Blood oxygen saturation, along with pulse rate, reflects blood oxygen supply levels and circulatory function, used to identify hypoxemia and respiratory dysfunction. The combined analysis of these two metrics provides a more comprehensive assessment of health status. For example, arrhythmias accompanied by decreased blood oxygen levels may indicate cardiopulmonary problems, and long-term trend analysis helps identify chronic diseases and potential risks, providing crucial support for disease diagnosis and health management. Furthermore, users can view their ECG waveforms, blood oxygen levels, and other data on the WeChat mini-program, ensuring convenient and real-time health monitoring.

[0100] S2. Use Hilbert-Huang transform and wavelet transform to extract features from electrocardiogram (ECG) signals, obtain ECG-instantaneous frequency features, and generate a two-dimensional time-frequency distribution map.

[0101] S2 includes:

[0102] S2-1. The ECG signal is processed using the wavelet thresholding denoising algorithm. The sym4 wavelet is selected to decompose the ECG signal. The coefficients obtained by the decomposition are processed by soft thresholding, and then the signal is reconstructed to remove noise and obtain the ECG reconstructed signal.

[0103] S2-2. R-wave localization and ECG segmentation are performed on the ECG reconstructed signal to obtain the segmented ECG signal. The Pan-Tompkins algorithm is used to locate the R-wave of the ECG reconstructed signal. Centered on the R-wave, 179 points are selected forward and 180 points are selected backward. Including the R-wave peak, a total of 360 data points are selected for ECG segmentation to ensure that a complete heartbeat cycle is captured.

[0104] S2-3. Standardize, augment, and undersample the ECG segmented signal to obtain the ECG sampled signal.

[0105] The ECG segmentation signal is linearly transformed using the Z-score standardization method to achieve a distribution with a mean of 0 and a standard deviation of 1, thus eliminating individual differences. A data augmentation method with noise addition is then used to augment the standardized ECG segmentation signal, resulting in an enhanced ECG signal. Finally, the enhanced ECG signal is subjected to a preliminary adjustment to balance the dataset and prevent overfitting of the model to the majority class, yielding the ECG sampling signal.

[0106] S2-4. Extract the instantaneous frequency characteristics of the ECG sampling signal using the Hilbert-Huang transform;

[0107] S2-5. Use wavelet transform to generate a two-dimensional time-frequency distribution map of instantaneous frequency characteristics.

[0108] S3. Construct an initial health assessment model and train it using the ETO algorithm (triangular exponential optimization algorithm) to obtain the health assessment model; process the two-dimensional time-frequency distribution spectrum through the health assessment model to obtain the user's electrocardiogram health assessment results;

[0109] The initial health assessment model is a CNN-Transformer-BiLSTM model, which includes a cascaded CNN model, Transformer model, BiLSTM model, and output layer; for example... Figure 5As shown, the CNN model includes a first convolutional module, a second convolutional module, and a third convolutional module in series. Each third convolutional module includes a corresponding convolutional layer, a normalization layer, and a max-pooling layer. The first convolutional module's convolutional layer includes 128 filters, a kernel size of 20, and a stride of 3. The second convolutional module's convolutional layer includes 64 filters, a kernel size of 7, and a stride of 1. The third convolutional module's convolutional layer includes 64 filters, a kernel size of 10, and a stride of 1. The Transformer model includes a multi-head attention layer and a max-pooling layer. The multi-head attention layer uses 128 dimensions and 4 heads. The BiLSTM model includes a first LSTM layer, a second LSTM layer, a flattening layer, a dropout layer, a first fully connected layer, and a second fully connected layer in series. The first fully connected layer includes 100 neurons and uses the ReLU function. The second fully connected layer includes 50 neurons and uses the ReLU function. The output layer includes 5 neurons, and the activation function is the softmax function.

[0110] S3 includes:

[0111] S3-1. Construct an initial health assessment model; obtain historical two-dimensional time-frequency distribution maps;

[0112] S3-2. Input the historical two-dimensional time-frequency distribution map into the CNN model to obtain the spatial features of the electrocardiogram;

[0113] The historical two-dimensional time-frequency distribution map is input into the first convolution module to obtain the first ECG spatial convolution feature; the first ECG spatial convolution feature is input into the second convolution module to obtain the second ECG spatial convolution feature; the second ECG spatial convolution feature is input into the third convolution module to obtain the ECG spatial feature.

[0114] CNN models are used to extract spatial features of electrocardiogram (ECG) signals. They utilize convolutional layers to extract local features, such as specific ECG waveform features, reducing data dimensionality and model computation while retaining key features.

[0115] S3-3. Input the ECG spatial features into the Transformer model to obtain the key ECG features;

[0116] By inputting the ECG spatial features into a multi-head attention layer and extracting information from different representation subspaces through parallel computation of multiple attention heads, the model's ability to capture various features and relationships in the data can be enhanced, resulting in key ECG sub-features from different heads. Multiple key ECG sub-features are merged to obtain initial key ECG features. The initial key ECG features are then input into a max pooling layer to obtain key ECG features.

[0117] By using the Transformer model to focus on the features of ECG signals from different perspectives, we can better handle long-distance dependencies in ECG signals and improve the model's ability to recognize complex ECG patterns.

[0118] S3-4. Input the key ECG features into the BiLSTM model to obtain the spatiotemporal features of the ECG;

[0119] The key ECG features are input into the first LSTM layer to obtain the first ECG spatiotemporal features;

[0120] The first ECG spatiotemporal features are input into the second LSTM layer to obtain the second ECG spatiotemporal features;

[0121] The second ECG spatiotemporal feature is input into the flattening layer to obtain a one-dimensional ECG spatiotemporal feature vector.

[0122] The one-dimensional spatiotemporal feature vector of ECG is input into the Dropout layer to obtain the spatiotemporal regularized feature vector of ECG, which prevents overfitting of the BiLSTM model.

[0123] The ECG spatiotemporal regularized feature vector is input into the first fully connected layer and the second fully connected layer to obtain the ECG spatiotemporal features.

[0124] The BiLSTM model processes both forward and reverse ECG signal sequences simultaneously, effectively capturing the temporal characteristics of ECG signals. Through the mechanisms of input gate, forget gate, and output gate, it controls the flow of information, solving the gradient vanishing and gradient explosion problems existing in traditional RNNs.

[0125] S3-5. Input the spatiotemporal features of the electrocardiogram into the output layer to obtain the electrocardiogram health assessment results of the training users;

[0126] S3-6. Based on the ECG health assessment results of the training users, the weight parameters of the initial health assessment model are adjusted using the ETO algorithm to obtain the health assessment model. The ETO algorithm optimizes the training process of the health assessment model through a unique triangular exponent calculation method, effectively adjusting the weight parameters, accelerating convergence, finding the optimal solution more efficiently, avoiding getting trapped in local optima, and improving model performance.

[0127] S3-7. Process the two-dimensional time-frequency distribution spectrum using a health assessment model to obtain the user's ECG health assessment results. The user's ECG health assessment results include ECG abnormality detection results and ECG classification results. ECG abnormality detection results include normal ECG and abnormal ECG; ECG classification results are the types of arrhythmias, including tachyarrhythmias and bradycardia.

[0128] S4. Based on the user's ECG health assessment results, blood oxygen saturation, and personal health indicators, generate health recommendations through a large language model. Personal health indicators include structured and unstructured health indicators; structured health indicators include quantifiable physiological and biochemical indicators with clear data formats, such as age, gender, height, weight, blood pressure, heart rate, blood glucose, and blood lipids; unstructured health indicators include medical records, treatment records, and doctor's diagnosis records.

[0129] S4 includes:

[0130] S4-1. Use the TensorFlow machine learning model to process the user's ECG health assessment results, blood oxygen saturation and personal health indicators, and extract multimodal key fusion features.

[0131] In TensorFlow machine learning models, for time-series signals in the input data, time-series features are extracted using LSTM or CNN-LSTM models, such as "frequency of premature ventricular contractions in ECG" and "duration of blood oxygen saturation below 90%", and converted into structured labels, namely "arrhythmia risk feature: moderate". For structured indicators in the input data, anomalies are detected using regression or rule models, such as "fasting blood glucose 7.8 mmol / L, exceeding the normal range by 1.3 mmol / L", generating "abnormal indicator features", such as "hyperglycemia feature: significant". For text data in the input data, keywords are extracted using TensorFlow's text classification model, such as "history of hypertension" and "penicillin allergy", and converted into "medical history feature labels".

[0132] By fusing temporal features, structured labels, abnormal indicator features, keywords, and medical history feature labels, multimodal key fusion features are obtained.

[0133] S4-2. Input the multimodal key fusion features into the LLM large language model to dynamically generate personalized health recommendations, including health trend predictions and lifestyle adjustment plans. Transform these recommendations into natural language prompts using the LLM large language model's built-in feature adaptation module. Based on these natural language prompts, predict potential health risks (health trend predictions) and lifestyle adjustment plans using the LLM large language model. Return the predicted potential health risks (health trend predictions) and lifestyle adjustment plans to the client for easy viewing by users and access by doctors.

[0134] like Figure 6 As shown, a wearable device-based electrocardiogram and blood oxygen monitoring system includes:

[0135] An electrocardiogram (ECG) acquisition device is used to acquire raw ECG signals and perform contact failure detection and noise reduction to obtain contact detection results and ECG signals.

[0136] The blood oxygen saturation acquisition module is used to acquire the initial blood oxygen saturation and perform electrode contact detection to obtain the electrode contact detection result and blood oxygen saturation.

[0137] A communication component for data transmission; in this embodiment, the communication component is Bluetooth Low Energy.

[0138] The two-dimensional time-frequency distribution map generation module is used to extract features from electrocardiogram signals using Hilbert-Huang transform and wavelet transform, obtain electrocardiogram-instantaneous frequency features, and generate a two-dimensional time-frequency distribution map.

[0139] The health assessment model construction and training module is used to build an initial health assessment model and train it using the ETO algorithm (triangular exponential optimization algorithm) to obtain the health assessment model.

[0140] The ECG health assessment result generation module is used to process the two-dimensional time-frequency distribution spectrum through the health assessment model to obtain the user's ECG health assessment results;

[0141] The health advice generation module is used to generate health advice based on the user's electrocardiogram health assessment results, blood oxygen saturation, and personal health indicators, using a large language model.

[0142] The client (mobile terminal) is used to remind users to adjust electrodes based on electrode contact detection results and to receive, display, and invoke health assessment analysis results and health suggestions.

[0143] The ECG acquisition device includes custom-made adhesive electrodes, wearable fabric electrodes, and an AD8232 module; the AD8232 module includes an AD8232 chip, a 0.5Hz dual-pole high-pass filter, and a 40Hz dual-pole low-pass filter; wherein:

[0144] Customized adhesive electrodes and wearable fabric electrodes are both used to collect raw electrocardiogram signals;

[0145] The AD8232 chip is used to detect whether the raw ECG signal has poor contact. It obtains the contact detection result and sends an alert command to the client through the communication component. Based on the contact detection result, it amplifies the raw ECG signal and sends a noise reduction command to a 0.5Hz dual-pole high-pass filter and a 40Hz dual-pole low-pass filter.

[0146] A 0.5Hz dual-pole high-pass filter is used to perform high-frequency filtering and noise reduction on the amplified original ECG signal to obtain the initial ECG signal.

[0147] A 40Hz dual-pole low-pass filter is used to perform low-frequency filtering and noise reduction on the initial ECG signal to obtain the ECG signal.

[0148] The blood oxygen saturation acquisition module includes:

[0149] The MQ-LAU-002 multi-wavelength transmitter is used to emit 660nm red light and 940nm infrared light to measure blood oxygen saturation.

[0150] The MQ237LF photoelectric converter is used to receive the intensity changes of reflected / transmitted light after red / infrared light penetrates tissue and convert it into an electrical signal.

[0151] The TPW3221-VR analog switch chip is used to receive control commands; based on the control commands, it switches the electrical signals corresponding to red light / infrared light and sends them to the 2078TWE0b microcontroller.

[0152] The 2078TWE0b microcontroller is used to send control commands; it receives and analyzes the electrical signals corresponding to red / infrared light to obtain blood oxygen saturation.

[0153] The ECG and blood oxygen monitoring system also includes a feedback module, which receives feedback information extracted after the user follows the health advice; updates the user's health record based on the feedback information and uploads it to the web platform for doctors to access and make judgments; and sends the feedback information to the health assessment model building and training module to facilitate the training process of the health assessment model.

[0154] In the experiment of this embodiment, an electrocardiogram (ECG) acquisition device was used to collect ECG signals, such as... Figure 7 As shown, the acquisition accuracy is improved compared to traditional ECG equipment, and the ECG waveform is clear and stable.

[0155] Blood oxygen saturation signals were collected using a blood oxygen saturation signal acquisition device and a traditional pulse oximeter, respectively. The device display interfaces are as follows: Figure 8 As shown, the measurement results are basically consistent, with an error within 1%.

[0156] The verification ECG signal was acquired and its features were extracted to obtain the corresponding verification-two-dimensional time-frequency distribution map. The verification-two-dimensional time-frequency distribution map was processed using the CNN-Transformer-BiLSTM model, and the corresponding model index was calculated, as shown in Table 1.

[0157] Table 1

[0158] Accuracy Recall rate F1 score normal heartbeat 0.98 0.99 0.99 Left bundle branch block pulsation 1.00 0.99 1.00 Right bundle branch block pulsation 1.00 1.00 1.00 Premature atrial contractions 0.99 0.99 0.99 Premature ventricular contractions 0.99 1.00 1.00 accuracy 0.99

[0159] As shown in Table 1, the CNN-Transformer-BiLSTM model achieved an accuracy of 99.50%, and its arrhythmia prediction performance was significantly better than that of traditional algorithms.

[0160] It should be noted that the specific methods by which each module performs operations in the system described in the above embodiments have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0161] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0162] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for monitoring electrocardiogram and blood oxygenation based on wearable devices, characterized in that, include: Electrocardiogram (ECG) signals are acquired using an ECG acquisition device, and blood oxygen saturation is acquired and uploaded using a blood oxygen saturation signal acquisition device. Hilbert-Huang transform and wavelet transform were used to extract features from electrocardiogram (ECG) signals, obtain ECG-instantaneous frequency features, and generate a two-dimensional time-frequency distribution map. An initial health assessment model was constructed and trained using the ETO algorithm to obtain the health assessment model; the two-dimensional time-frequency distribution spectrum was processed through the health assessment model to obtain the user's electrocardiogram health assessment results; Based on the user's electrocardiogram health assessment results, blood oxygen saturation, and personal health indicators, health recommendations are generated through a large language model.

2. The method for monitoring electrocardiogram and blood oxygen based on a wearable device according to claim 1, characterized in that, The process of acquiring electrocardiogram (ECG) signals via an ECG acquisition device and acquiring and uploading blood oxygen saturation signals via a blood oxygen saturation signal acquisition device includes: Construct an electrocardiogram (ECG) acquisition device and a blood oxygen saturation signal acquisition device; Raw electrocardiogram (ECG) signals are acquired using an ECG acquisition device, and poor contact is detected to obtain contact detection results. Based on the contact detection results, the original ECG signal is denoised using an ECG acquisition device to obtain the ECG signal; The blood oxygen saturation signal acquisition device emits red and infrared light and measures the changes in reflected / transmitted light intensity. Based on changes in light intensity, the results of blood oxygen electrode contact detection are obtained, and blood oxygen saturation is calculated using Lambert-Beer's law.

3. The method for monitoring electrocardiogram and blood oxygen based on a wearable device according to claim 1, characterized in that, The electrocardiogram (ECG) acquisition device includes: Customized adhesive electrodes and wearable fabric electrodes are both used to collect raw electrocardiogram signals; The AD8232 chip is used to detect whether the raw ECG signal has poor contact. It obtains the contact detection result and sends an alert command to the client through the communication component. Based on the contact detection result, it amplifies the raw ECG signal and sends a noise reduction command to a 0.5Hz dual-pole high-pass filter and a 40Hz dual-pole low-pass filter. A 0.5Hz dual-pole high-pass filter is used to perform high-frequency filtering and noise reduction on the amplified original ECG signal to obtain the initial ECG signal. A 40Hz dual-pole low-pass filter is used to perform low-frequency filtering and noise reduction on the initial ECG signal to obtain the ECG signal.

4. The method for monitoring electrocardiogram and blood oxygen based on a wearable device according to claim 1, characterized in that, The blood oxygen saturation acquisition module includes: The MQ-LAU-002 multi-wavelength emitter is used to emit 660nm red light and 940nm infrared light to measure blood oxygen saturation. The MQ237LF photoelectric converter is used to receive the intensity changes of reflected / transmitted light after red / infrared light penetrates tissue and convert it into an electrical signal; The TPW3221-VR analog switch chip is used to receive control commands; based on the control commands, it switches the electrical signals corresponding to red light / infrared light and sends them to the 2078TWE0b microcontroller. The 2078TWE0b microcontroller is used to send control commands; it receives and analyzes the electrical signals corresponding to red / infrared light to obtain blood oxygen saturation.

5. The method for monitoring electrocardiogram and blood oxygen based on a wearable device according to claim 1, characterized in that, The initial health assessment model includes a cascaded CNN model, a Transformer model, a BiLSTM model, and an output layer. The CNN model includes a cascaded first convolutional module, a second convolutional module, and a third convolutional module. Each third convolutional module includes a corresponding convolutional layer, a normalization layer, and a max-pooling layer. The first convolutional module's convolutional layer includes 128 filters, a kernel size of 20, and a stride of 3. The second convolutional module's convolutional layer includes 64 filters, a kernel size of 7, and a stride of 1. The third convolutional module's convolutional layer includes 64 filters... The kernel size is 10, and the stride is 1. The Transformer model includes a multi-head attention layer and a max pooling layer. The multi-head attention layer uses 128 dimensions and 4 heads. The BiLSTM model includes a first LSTM layer, a second LSTM layer, a flattening layer, a dropout layer, a first fully connected layer, and a second fully connected layer, which are connected in sequence. The first fully connected layer has 100 neurons and uses the ReLU function. The second fully connected layer has 50 neurons and uses the ReLU function. The output layer has 5 neurons and uses the softmax function as the activation function.

6. The method for monitoring electrocardiogram and blood oxygen based on a wearable device according to claim 5, characterized in that, The training process of the initial health assessment model includes: Construct an initial health assessment model; obtain historical two-dimensional time-frequency distribution maps; The historical two-dimensional time-frequency distribution map is input into the CNN model to obtain the spatial features of the electrocardiogram; The spatial features of the electrocardiogram (ECG) are input into the Transformer model to obtain the key ECG features; Key ECG features are input into a BiLSTM model to obtain spatiotemporal ECG features; The spatiotemporal features of electrocardiogram (ECG) are input into the output layer to obtain the ECG health assessment results of the training users; Based on the ECG health assessment results of the training users, the weight parameters of the initial health assessment model are adjusted using the ETO algorithm to obtain the health assessment model.

7. A wearable device-based electrocardiogram (ECG) and blood oxygen monitoring system, used to implement the wearable device-based ECG and blood oxygen monitoring method according to any one of claims 1 to 6, characterized in that, include: An electrocardiogram (ECG) acquisition device is used to acquire raw ECG signals and perform contact failure detection and noise reduction to obtain contact detection results and ECG signals. The blood oxygen saturation acquisition module is used to acquire the initial blood oxygen saturation and perform electrode contact detection to obtain the electrode contact detection result and blood oxygen saturation. Communication components are used for data transmission; The two-dimensional time-frequency distribution map generation module is used to extract features from electrocardiogram signals using Hilbert-Huang transform and wavelet transform, obtain electrocardiogram-instantaneous frequency features, and generate a two-dimensional time-frequency distribution map. The health assessment model building and training module is used to build an initial health assessment model and train it using the ETO algorithm to obtain the health assessment model. The ECG health assessment result generation module is used to process the two-dimensional time-frequency distribution spectrum through the health assessment model to obtain the user's ECG health assessment results; The health advice generation module is used to generate health advice based on the user's electrocardiogram health assessment results, blood oxygen saturation, and personal health indicators, using a large language model. The client application is used to remind users to adjust the electrodes based on the electrode contact detection results and other information. Receive, display, and access health assessment and analysis results and health recommendations.

8. The ECG and blood oxygen monitoring system based on a wearable device according to claim 7, characterized in that, The ECG acquisition device includes custom adhesive electrodes, wearable fabric electrodes, and an AD8232 module; The AD8232 module includes a chip U10 with model number AD8232ACPZ-RL; pin 1 of U10 is connected to one end of capacitor C17; pin 2 of U10 is connected to one end of resistor R24; pin 3 of U10 is connected to one end of resistor R23; pin 4 of U10 is connected to one end of capacitor C19; pin 5 of U10 is connected to the other end of capacitor C19 and one end of resistor R21; pin 6 of U10 is connected to one end of resistor R20, one end of resistor R18, and one end of capacitor C18; pin 7 of U10 is connected to one end of capacitor C20 and one end of resistor R26; pin 8 of U10 is connected to the other end of capacitor C20 and one end of resistor R28, and serves as the REFOUT terminal; pin 9 of U10 is connected to resistor R24. The other end of U10 is connected to one end of resistor R29; pin 10 of U10 is connected to the other end of resistor R29 and one end of capacitor C21, and serves as the VP terminal; pins 11-13 and pin 15 of U10 serve as IO26, IO25, SDN, and VDD3V3 terminals, respectively; pin 14 of U10 is grounded; pin 16 of U10 is connected to one end of capacitor C23 and grounded; pin 17 of U10 is connected to the other end of capacitor C23 and serves as the VDD3V3 terminal; pin 18 of U10 is connected to one end of resistor R31, one end of resistor R30, and one end of capacitor C22, respectively; pin 19 of U10 is connected to one end of resistor R17; pin 20 of U10 is connected to the other end of capacitor C17 and one end of resistor R16, respectively; pin 21 of U10 is grounded. The other end of resistor R24 ​​is connected to one end of resistor R22 and serves as the LA terminal; the other end of resistor R23 is connected to one end of resistor R25 and serves as the RA terminal; the other end of resistor R21 serves as the RL terminal; the other end of resistor R20 is connected to the other ends of resistors R22, R25, and R19 respectively; the other end of resistor R19 serves as the VDD3V3 terminal; the other end of resistor R18 is connected to the other ends of resistors R16 and R17 respectively; the other end of capacitor C18 serves as the REFOUT terminal; the other end of resistor R26 is connected to one end of resistor R27 and the other end of capacitor C21 respectively; the other end of resistor R27 serves as the IAOUT terminal; the other end of resistor R31 serves as the VDD3V3 terminal; the other end of resistor R30 is connected to the other end of capacitor C22 and grounded. The LA, RA, and RL terminals are all connected to an external AUDIO-JACKSMD2 chip JP1; the VP terminal is connected to the positive terminal of an external photodiode D5, and the negative terminal of the photodiode D5 is connected to a grounding resistor R9; the SDN terminal is connected to one end of an external resistor R10, and the other end of the resistor R10 is connected to the VDD3V3 terminal.

9. A wearable device-based electrocardiogram and blood oxygen monitoring system according to claim 7, characterized in that, The blood oxygen saturation acquisition module includes chip U1, chip U2, chip U3 (model 2078TWE0bQFN32), and chips J4-J8; Pin 1 of U1 is used as the VSS terminal; pin 2 of U1 is connected to one end of inductor L1; the other end of inductor L1, one end of capacitor C4, and pin 1 of J6 are used as the VCC terminal; the other end of capacitor C4 is used as the VSS terminal. Pin 3 of U1 is connected to one end of capacitor C3 and pin 2 of J5 respectively, and serves as the VDD terminal; the other end of capacitor C3 serves as the VSS terminal. Pins 1 and 3-7 of J5 are used as VSS, SCL, SDA, RES, DC and CS terminals respectively; pin 2 of J6 is connected to pin 3 of J6 and serves as VSS terminal. Pins 4 and 5 of J6 are used as the RXD and TXD terminals, respectively. Pin 1 of U2 is connected to one end of resistor R6, one end of capacitor C2, pins 2 and 10 of U2, and serves as the VDD terminal; the other end of capacitor C2 serves as the VSS terminal, U2-IN2 terminal, and RD terminal; pins 3-4, 6, and 8-9 of U2 serve as the IR terminal, U2-IN1 terminal, and VSS terminal, respectively; pin 5 of U2 is connected to one end of resistor R1; the other end of resistor R1 is connected to one end of resistor R2 and one end of resistor R3, respectively; the other end of resistor R3 serves as the VSS terminal; pin 7 of U2 is connected to the other end of resistor R2. Pins 1-3, 6-9, 14, 22, 25-30, and 32 of U3 are designated as SCL, DIO, DAT, U2-IN2, U2-IN1, VDD, VSS, PD, VSS, IICSCL, IICSDA, RES, DC, SCL1, CS, and SDA terminals, respectively. Pin 15 of U3 is connected to one end of resistor R5, with the other end serving as the RXD terminal. Pin 6 of U3 is connected to one end of resistor R4, with the other end serving as the TXD terminal. Pin 24 of U3 is connected to one end of capacitor C1 and serves as the VDD terminal; the other end of capacitor C1 serves as the VSS terminal. Pins 1-6 of J4 are designated as DAT, RD, IR, PD, VSS, and VDD terminals, respectively. Pins 1-4 of J7 are used as VDD, IISSCL, IICSDA, and VSS terminals, respectively. Pins 1-4 of J8 are used as VSS, SCL, DIO, and VDD terminals, respectively.

Images