Capacitive wearable noninvasive sensor and blood vessel blood flow change detection method

Abstract

The invention provides a capacitive wearable noninvasive sensor and a blood vessel blood flow change detection method, and belongs to the technical field of blood vessel blood flow change monitoring. Two semi-arc-shaped copper foil electrodes are respectively attached to two sides of the wrist of a human body, and the two semi-arc-shaped copper foil electrodes are oppositely arranged and surround the wrist of the human body by a circle. And the self-adaptive variable-frequency excitation module is used for outputting alternating excitation signals of which the frequency can be self-adaptively adjusted in a set range to the two semicircular copper foil electrodes. A coupling capacitor is formed between the two semicircular copper foil electrodes and the wrist of the human body, the coupling capacitor generates a capacitance signal under the action of an alternating excitation signal, and the capacitance signal forms a capacitance change signal along with the blood pulsation change of the artery of the wrist of the human body. And the programmable gain amplification module is used for automatically adjusting the gain according to the strength of the capacitance change signal. The problem that in the prior art, due to the fact that a traditional sensor is low in detection precision, the blood flow change detection accuracy of the blood vessel is not high is solved.

Description

Technical Field

[0001] This invention belongs to the field of vascular blood flow change monitoring technology, specifically relating to a capacitive wearable non-invasive sensor and a method for detecting vascular blood flow changes. Background Technology

[0002] With the increasing aging of the population and changes in lifestyle, changes in vascular blood flow have become a significant threat to human health, characterized by high morbidity, high health burden, and high physiological risks. Early identification, continuous monitoring, and scientific intervention of key physiological states such as vascular elasticity decline, hemodynamic imbalance, and abnormal pulse wave transmission are crucial for maintaining circulatory homeostasis and reducing health risks.

[0003] Currently, traditional methods for detecting changes in vascular blood flow mainly rely on specialized medical equipment such as electrocardiograms, blood pressure monitors, and vascular ultrasound. These methods have limitations such as limited testing scenarios, inability to achieve long-term continuous monitoring, complex operation, and the need for professional assistance, making it difficult to meet the needs of daily prevention for the general population and home monitoring for high-risk groups.

[0004] Capacitive sensing technology has the advantages of being non-invasive, having a fast response speed, simple structure, and being integrable. Most capacitive sensors are used in the industrial field, but their application in the medical and health field is relatively limited. Even capacitive sensors that are widely used in physiological monitoring equipment still have many shortcomings: they are bulky, inconvenient to wear, and unsuitable for daily health monitoring.

[0005] Meanwhile, the sensors currently available for smart wearables are mainly photoelectric sensors based on photoplethysmography (PPG), which can only monitor a single physiological indicator (such as heart rate). They are point-like detectors with limited detection range and accuracy, and are greatly affected by skin color, sweat, and exercise status. They also lack high-precision signal processing capabilities and cannot accurately detect changes in vascular blood flow.

[0006] Therefore, developing a capacitive sensor and detection system that is stable in signal, accurate in detection, convenient to wear, and capable of detecting changes in blood flow throughout the entire process is of great practical significance and application value, and also aligns with the current trend of intelligent medical devices developing towards non-invasiveness, convenience, and intelligence. Summary of the Invention

[0007] The purpose of this invention is to provide a capacitive wearable non-invasive sensor and a method for detecting changes in vascular blood flow, which solves the problem that the detection accuracy of traditional sensors in the prior art is low, resulting in low accuracy in detecting changes in vascular blood flow.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a capacitive wearable non-invasive sensor, comprising two semi-circular copper foil electrodes, an adaptive frequency conversion excitation module, a programmable gain amplification module, and a signal acquisition interface. The two semi-circular copper foil electrodes are respectively attached to both sides of the human wrist, and the two semi-circular copper foil electrodes are arranged opposite each other and surround the human wrist. The adaptive frequency conversion excitation module is used to output an alternating excitation signal with an adaptively adjustable frequency within a set range to the two semi-circular copper foil electrodes. The two semi-circular copper foil electrodes form a coupling capacitor with the human wrist. The coupling capacitor generates a capacitance signal under the action of the alternating excitation signal. The capacitance signal changes with the change of blood pulse in the artery of the human wrist to form a capacitance change signal. The programmable gain amplifier module is used to automatically adjust the gain according to the strength of the capacitance change signal; The signal acquisition interface is used to transmit the gain-adjusted capacitance change signal to the signal acquisition chip for conversion processing. The programmable gain amplifier module is disposed between the signal acquisition interface and the signal acquisition chip.

[0009] A further improvement of the present invention is that the surfaces of the two semi-circular copper foil electrodes are covered with an insulating layer.

[0010] A further improvement of the present invention is that the insulating layer is a silicone insulating layer.

[0011] A further improvement of the present invention is that the setting range is 50kHz~150kHz.

[0012] A further improvement of the present invention is that the gain adjustment range is 1 to 8 times.

[0013] A further improvement of this invention is that the automatic gain adjustment based on the intensity of the capacitance change signal specifically involves: Increase the gain when the pulse signal is weak, resulting in a low intensity of the capacitance change signal; When a strong pulse signal results in a large intensity of capacitance change signal, reduce the gain.

[0014] A further improvement of the present invention is that the capacitive wearable non-invasive sensor has a flexible structure.

[0015] Secondly, the present invention provides a method for detecting changes in vascular blood flow, using the capacitive wearable non-invasive sensor described above, comprising the following steps: The alternating excitation signal and the capacitance change signal generated by the pulsation of blood in the human wrist artery are coupled to obtain the coupled signal. After adjusting the strength of the coupled signal, the coupled signal is converted into a digital signal; The digital signal is transmitted to the terminal device, and the digital signal is feature extracted in the terminal device to obtain heart rate features, pulse wave amplitude features, pulse wave morphology features and hemodynamic features; By fusing heart rate characteristics, pulse wave amplitude characteristics, pulse wave morphology characteristics, and hemodynamic characteristics, multi-dimensional features are obtained. By inputting multi-dimensional features into a trained physiological index analysis model, the results of vascular blood flow change detection are obtained.

[0016] A further improvement of this invention is that, after obtaining the detection results of changes in vascular blood flow, an early warning strategy and an intervention strategy are set based on the detection results of changes in vascular blood flow.

[0017] A further improvement of this invention is that, after setting the early warning strategy and the intervention strategy, health management is carried out on the health status corresponding to the blood pulsation of the wrist artery based on the early warning strategy and the intervention strategy.

[0018] Compared with the prior art, the present invention has the following beneficial effects: The capacitive wearable non-invasive sensor proposed in this invention features an adaptive frequency conversion excitation module. This module dynamically adjusts the excitation frequency based on the characteristics of different wrist tissues, adapting to various populations (the elderly, young people, and obese individuals), avoiding power frequency interference, and improving the accuracy of electric field penetration and the ability to capture weak signals, thereby enhancing the accuracy of detecting changes in vascular blood flow. Furthermore, a programmable gain amplification module automatically adjusts the intensity of the capacitance change signal, preventing signal saturation or loss of weak signals, thus also improving the accuracy of detecting changes in vascular blood flow.

[0019] Furthermore, the present invention discloses that the surfaces of the two semi-circular arc-shaped copper foil electrodes are covered with an insulating layer. On the one hand, this can achieve insulation and isolation between the copper foil electrodes and the skin, preventing polarization of the copper foil electrodes and skin irritation. On the other hand, it can fill the tiny gaps between the copper foil electrodes and the skin, reducing the influence of the air medium on capacitance detection and improving the stability of capacitance change signals. Attached Figure Description

[0020] Figure 1 This is a side view of the overall structure of the capacitive wearable non-invasive sensor of the present invention. Figure 2 This is a top view of the overall structure of the capacitive wearable non-invasive sensor of the present invention. Figure 3 This is a schematic diagram of the electrode arrangement in the capacitive wearable non-invasive sensor of the present invention; Figure 4This is a flowchart of the method for detecting changes in vascular blood flow according to the present invention; Figure 5 This is a schematic diagram of the vascular blood flow change detection system in Embodiment 4 of the present invention; Figure 6 This is a schematic diagram of the internal units of the deep learning signal processing module in Embodiment 4 of the present invention; Figure 7 This is a flowchart of the CNN-LSTM deep learning algorithm in Embodiment 4 of the present invention; In the diagram: 1. Copper foil electrode; 2. Copper foil electrode lead-out; 3. Insulating layer; 4. Human wrist wrapped with copper foil electrode; 5. Data intelligent processing module. Detailed Implementation

[0021] To further understand the content of this invention, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention.

[0022] Example 1: This embodiment discloses a capacitive wearable non-invasive sensor. The capacitive wearable non-invasive sensor of the present invention includes two semi-circular copper foil electrodes, an adaptive frequency conversion excitation module, a programmable gain amplification module, and a signal acquisition interface.

[0023] Two semi-circular copper foil electrodes are attached to both sides of the human wrist, with the two semi-circular copper foil electrodes facing each other and encircling the human wrist (in... Figure 3 (represented by the number 4 in the text).

[0024] The adaptive frequency conversion excitation module is used to output an alternating excitation signal with an adaptively adjustable frequency within a set range to two semi-circular copper foil electrodes.

[0025] Two semi-circular copper foil electrodes form a coupling capacitor between themselves and the human wrist. The coupling capacitor generates a capacitance signal under the action of an alternating excitation signal. The capacitance signal changes with the pulse of blood in the artery of the human wrist, forming a capacitance change signal.

[0026] The programmable gain amplifier module is used to automatically adjust the gain based on the signal strength of capacitance changes.

[0027] The signal acquisition interface is used to transmit the gain-adjusted capacitance change signal to the signal acquisition chip for conversion and processing; The programmable gain amplifier module is located between the signal acquisition interface and the signal acquisition chip.

[0028] Example 2: This embodiment discloses a capacitive wearable non-invasive sensor. A side view of the overall structure of the capacitive wearable non-invasive sensor is shown below. Figure 1As shown, a top view of the overall structure of the capacitive wearable non-invasive sensor is as follows: Figure 2 As shown in the diagram, the electrode arrangement in a capacitive wearable non-invasive sensor is as follows: Figure 3 As shown.

[0029] The capacitive wearable non-invasive sensor of the present invention includes two semi-circular copper foil electrodes. Figure 1 and Figure 2 (Represented by reference numeral 1, and reference numeral 2, indicating copper foil electrode leads), adaptive frequency conversion excitation module, and programmable gain amplification module (the adaptive frequency conversion excitation module and the programmable gain amplification module are in...) Figure 3 (represented by label 5) and signal acquisition interface.

[0030] Two semi-circular copper foil electrodes are attached to both sides of the human wrist, with the two semi-circular copper foil electrodes facing each other and encircling the human wrist.

[0031] The adaptive frequency conversion excitation module is used to output an alternating excitation signal with an adaptively adjustable frequency within a set range (50kHz~150kHz in this embodiment) to the two semi-circular copper foil electrodes. The adaptive frequency conversion excitation module is described in detail below: The adaptive frequency conversion excitation module (specifically composed of a signal generator and a frequency adjustment chip, model AD9850, capable of precise frequency tuning within the range of 50kHz to 150kHz with an adjustment accuracy of 1Hz) abandons the existing fixed-frequency excitation method. Its core function is to automatically adjust the excitation frequency within the 50kHz to 150kHz range based on the user's wrist tissue characteristics (analyzing fat thickness and skin moisture through initial capacitance signals). Specifically, for users with thicker fat, the frequency is increased (100kHz to 150kHz) to ensure the electric field penetrates to the arterial blood vessel layer; for users with thinner / dryer skin, the frequency is decreased (50kHz to 100kHz) to reduce signal attenuation. Simultaneously, it avoids 50Hz power frequency and harmonic interference throughout, ensuring the stability of the capacitance change signal. The adaptive frequency conversion excitation module uses a small-amplitude alternating voltage excitation with an average DC value of 0. The voltage amplitude can be adaptively adjusted within the range of 3V to 5V, adjusting according to skin contact conditions to further improve wearing comfort. Simultaneously, it enables the capacitive wearable non-invasive sensor to generate a stable charging and discharging current, facilitating the acquisition of capacitance change signals.

[0032] Two semi-circular copper foil electrodes form a coupling capacitor between themselves and the human wrist. This coupling capacitor generates a capacitance signal under the influence of an alternating excitation signal. This capacitance signal changes with the pulse of blood in the wrist artery, creating a capacitance variation signal. The dimensions of the two semi-circular copper foil electrodes are described below: The two semi-circular arc-shaped copper foil electrodes are made of highly conductive copper foil material, which balances conductivity and flexibility. In this embodiment, the width of the two semi-circular arc-shaped copper foil electrodes is 18mm (along the arm direction), the arc length is 35mm, and the thickness is 0.2mm. The width, arc length, and thickness of the copper foil electrodes can be adjusted according to the size of the human wrist.

[0033] The two semi-circular copper foil electrodes can be made of metals such as gold, silver, aluminum, and silver nanowires (AgNWs), conductive non-metals such as ITO (indium tin oxide), carbon nanotubes (CNTs), and graphene, or composite materials of graphene / copper foil, conductive polymer materials, and organic and inorganic conductive materials.

[0034] The number of copper foil electrodes in this embodiment can be adjusted according to actual needs. For example, it can be set to four semi-circular copper foil electrodes (two sets of copper foil electrodes). Multiple semi-circular copper foil electrodes can form an array electrode to realize the synchronous detection of multiple sets of capacitance signals.

[0035] This embodiment can change the shape of the capacitive wearable non-invasive sensor and apply it to other parts of the human body, such as the waist, legs, and neck, or to animal health monitoring.

[0036] The programmable gain amplifier module (specifically using a PGA280 programmable gain amplifier) ​​is used to automatically adjust the gain based on the intensity of the capacitance change signal (in this embodiment, the gain adjustment range is 1 to 8 times), ensuring consistent acquisition accuracy of capacitance change signals from different populations and solving the problem of poor adaptability of existing sensors to different populations. Specifically, the automatic gain adjustment based on the intensity of the capacitance change signal is as follows: When the pulse signal is weak (such as in the elderly and people with poor vascular elasticity) resulting in a low intensity of capacitance change signal, increase the gain; When a strong pulse signal results in a strong capacitance change signal, reduce the gain to avoid saturation distortion of the capacitance change signal.

[0037] The signal acquisition interface is used to transmit the gain-adjusted capacitance change signal to the signal acquisition chip for conversion and processing; The programmable gain amplifier module is positioned between the signal acquisition interface and the signal acquisition chip (in this embodiment, the signal acquisition chip is the FDC2214 chip manufactured by TI). Other signal acquisition chip types can be selected based on actual needs.

[0038] The surfaces of the two semi-circular copper foil electrodes are covered with an insulating layer. Figure 1 and Figure 2(represented by reference numeral 3 in the text) On the one hand, it can achieve insulation between the copper foil electrode and the skin, preventing copper foil electrode polarization and skin irritation. On the other hand, it can fill the tiny gap between the copper foil electrode and the skin, reducing the influence of the air medium on capacitance detection and improving the stability of capacitance change signals.

[0039] In this embodiment, a silicone insulating layer is used. In other embodiments, insulating layers of other materials can also be used, as long as the insulating effect can be achieved.

[0040] The capacitive wearable non-invasive sensor in this embodiment has a flexible structure that conforms to the curvature of the human wrist, making it comfortable to wear. The size of the two semi-circular copper foil electrodes can be adjusted according to different human wrist sizes to avoid affecting the detection accuracy due to wearing misalignment, changes in tightness, and the influence of wearing tightness and slight movements of the human wrist on the capacitance change signal, making it suitable for long-term continuous monitoring.

[0041] The assembly process of the capacitive wearable non-invasive sensor is described below: The prepared electrodes (copper foil electrodes), adaptive frequency conversion excitation module, programmable gain amplification module and signal acquisition interface are assembled and fixed with a flexible substrate to ensure that the capacitive wearable non-invasive sensor fits the curvature of the human wrist as a whole, which can adapt to different adult wrist sizes and form a uniform through alternating electric field.

[0042] The working principle of capacitive wearable non-invasive sensors is explained below: Two opposing electrodes (two semi-circular copper foil electrodes) and wrist tissue (skin, muscle, blood vessels, and bone) form an arc-shaped equivalent capacitor, with the wrist tissue acting as the capacitor's dielectric. The capacitance value is calculated using the following formula:

[0043] in, The vacuum permittivity, The equivalent dielectric constant of wrist tissue is given by S, where S is the effective facing area of ​​the electrodes, and d is the distance between the electrodes; the dielectric constant of arterial blood is given by (…). As the heart beats, the arteries expand and become congested with blood during cardiac contraction, increasing blood volume and the equivalent dielectric constant. As the heart relaxes, the capacitance C increases; during diastole, arteries constrict, blood volume decreases, and the equivalent dielectric constant increases. As the capacitance value C decreases, the dynamic change in capacitance value is detected. This allows us to reverse-engineer changes in arterial blood pulsation and blood flow volume, and then extract cardiovascular-related physiological indicators such as heart rate, pulse wave, and vascular elasticity.

[0044] The equivalent dielectric constant change caused by the periodic filling / draining of arterial blood is continuously measured, and the real-time capacitance change ΔC(t) is calculated. The formula for calculating ΔC(t) is as follows: ΔC(t)=C(t) C0

[0045] Where C(t) is the real-time capacitance, C0 is the static baseline capacitance (determined by unchanging tissues such as bone, muscle, and fat), and the real-time capacitance change ΔC(t) is generated only by blood pulsation.

[0046] Testing revealed that the sensor's capacitance in this embodiment is approximately 220 pF at rest and approximately 385 pF during cardiac congestion. The capacitance change caused by the pulse... It has a signal-to-noise ratio of approximately 165pF and a relative change rate of about 75%, and can accurately acquire signals using the FDC2214 chip manufactured by TI.

[0047] The capacitive wearable non-invasive sensor in this embodiment can be used independently or integrated into smart terminals such as smartwatches.

[0048] Example 3: This embodiment discloses a method for detecting changes in vascular blood flow, and the flowchart of the detection method is as follows: Figure 4 As shown, using the capacitive wearable non-invasive sensor described above includes the following steps: S1. Couple the alternating excitation signal with the capacitance change signal generated by the pulsation of blood in the human wrist artery to obtain the coupled signal; S2. After adjusting the strength of the coupled signal, the coupled signal is converted into a digital signal; S3. Transmit the digital signal to the terminal device, and extract features from the digital signal in the terminal device to obtain heart rate features, pulse wave amplitude features, pulse wave morphology features and hemodynamic features; S4. The heart rate characteristics, pulse wave amplitude characteristics, pulse wave morphology characteristics and hemodynamic characteristics are fused to obtain multi-dimensional characteristics; S5. Input the multi-dimensional features into the trained physiological index analysis model to obtain the detection results of changes in vascular blood flow.

[0049] After obtaining the results of vascular blood flow change detection, early warning and intervention strategies are set based on these results.

[0050] After setting early warning and intervention strategies, health management is carried out based on the health status corresponding to the blood pulsation of the wrist artery.

[0051] Example 4: The method for detecting changes in vascular blood flow in this invention is specifically implemented through a vascular blood flow change detection system (hereinafter referred to as the system), such as... Figure 5 As shown, the vascular blood flow change detection system includes a sensor acquisition module, a data transmission module, a deep learning signal processing module, an analysis module, an early warning and intervention module, and a health management module. The following provides a detailed description of the power supply module, sensor acquisition module, data transmission module, deep learning signal processing module, analysis module, early warning and intervention module, and health management module: A. Power Module In this embodiment, the vascular blood flow change detection system is powered by a power module (battery).

[0052] B. Sensor acquisition module As the core of the system, it consists of the capacitive wearable non-invasive sensor described above and the FDC2214 signal acquisition chip manufactured by TI. After the system is powered on, it generates an alternating excitation signal (achieved through the adaptive frequency conversion excitation module of the capacitive wearable non-invasive sensor). The alternating excitation signal is coupled with the capacitance change signal generated by the pulsation of blood in the wrist artery. The coupled signal is adjusted by the programmable gain amplification module, and then the analog signal (coupled signal) is converted into a digital signal by the FDC2214 signal acquisition chip manufactured by TI and transmitted to the data transmission module.

[0053] C. Data transmission module Employing a Bluetooth Low Energy (BLE) transmission module (model nRF52840), the system transmits the collected digital signals to terminal devices (such as smartphones, smartwatches, and home monitoring terminals) in real time. The transmission distance can reach up to 10 meters, and the low power consumption (standby power consumption ≤10μA) makes it suitable for long-term wearable use. It also supports local data storage and network transmission, with cloud backup processing (using Alibaba Cloud servers, supporting encrypted data storage) to prevent data loss, facilitate subsequent traceability and analysis, and provide reference for clinical diagnosis.

[0054] The data transmission module in this embodiment can be replaced with other transmission technologies such as mobile networks, China Unicom networks, China Telecom networks (3G, 4G and 5G), WiFi and NFC.

[0055] D Deep Learning Signal Processing Module As the core signal processing unit of the system, the CNN-LSTM deep learning algorithm is introduced (other deep learning algorithms, such as Transformer, can also be used to further improve the accuracy of clutter filtering and feature extraction). Simultaneously, the baseline calibration method is optimized to achieve accurate clutter filtering and precise feature extraction. The deep learning signal processing module is mainly implemented through the following units, such as... Figure 6 As shown, the following is an introduction to each unit: Clutter filtering unit: It filters out nonlinear interference such as motion interference, sweat interference, and power frequency noise through a CNN (Convolutional Neural Network) network. Compared with traditional adaptive filtering algorithms, the clutter filtering capability is improved by 60%, while retaining pure dynamic capacitance change signals. Signal timing analysis unit: Analyzes the time series characteristics of capacitance signals through LSTM (Long Short-Term Memory) network, captures subtle changes in pulse waves (such as changes in pulse wave amplitude and rise time caused by decreased vascular elasticity), and avoids the limitations of a single filtering algorithm; Dynamic baseline calibration unit: It adopts a dynamic baseline calibration algorithm, calibrating the baseline every 100ms. It has a faster response and 30% higher calibration accuracy than the existing moving average calibration algorithm, effectively eliminating baseline drift caused by exercise, sweat and changes in wearing tightness. Multi-feature fusion extraction unit: Based on existing heart rate and pulse wave amplitude features, new pulse wave morphology and hemodynamic features are extracted. Heart rate features include heart rate, heart rate variability (HRV), and pulse wave rhythm regularity; pulse wave amplitude features include pulse wave amplitude, perfusion index (PI), and respiratory modulated pulse wave amplitude (RPA); pulse wave morphology features include pulse wave conduction velocity, pulse wave rise time, pulse wave fall time, diastolic decay rate, pulse wave reflex index (RI), pre-ejection time (PEP), left ventricular ejection time (LVET), arteriosclerosis index (AI), and diastolic blood flow recovery rate; hemodynamic features include peripheral resistance index (PRI). Through multi-feature fusion analysis, the accuracy of vascular blood flow change detection is improved, and misjudgment based on a single feature is reduced.

[0056] The flowchart of the CNN-LSTM hybrid deep learning algorithm is as follows: Figure 7 As shown, the specific steps of implementing the CNN-LSTM hybrid deep learning algorithm are as follows: S1. Data Preprocessing The converted digital signal is normalized and segmented to form a fixed-length input sequence; S2. CNN Feature Extraction The preprocessed signal is input into the CNN network, which contains 3 convolutional layers and 2 pooling layers. The convolutional kernel size is 3×1, and the activation function is the ReLU function, which is used to extract the spatial features of the signal and filter out nonlinear interference such as motion interference, sweat interference, and power frequency noise. S3. Feature Map Flattening Flatten the multi-dimensional feature map output by the CNN into a one-dimensional feature vector; S4. LSTM Timing Analysis The flattened feature vector is input into the LSTM network, which contains two LSTM layers and one fully connected layer with 64 hidden neurons. It is used to capture the temporal dependence of the pulse wave signal and output temporal features. S5. Multi-feature extraction Based on LSTM output, core physiological characteristic parameters such as heart rate, pulse wave amplitude, pulse wave conduction velocity, pulse wave rise time, fall time, and diastolic decay rate are extracted through algorithms such as peak detection and waveform analysis. S6 Dynamic Baseline Calibration A sliding window mean filtering algorithm is adopted, with a window size of 100ms. The baseline of the capacitance signal is calibrated every 100ms to eliminate baseline drift caused by exercise, sweat, and changes in wearing tightness. S7. Feature Fusion and Output The extracted features are fused at the feature level (combined into a feature vector) or at the decision level and then transmitted to the analysis module.

[0057] E. Analysis Module As the core analysis unit of the system, it constructs a physiological indicator analysis model based on machine learning algorithms and clinical data on vascular blood flow changes (including physiological indicator data from healthy individuals, hypertensive patients, and patients with arteriosclerosis). Combined with multi-dimensional feature parameters extracted by the deep learning signal processing module, it enables the detection of vascular blood flow changes. The specific functions implemented by the analysis module include: Health status assessment combines indicators such as heart rate, pulse wave, and vascular elasticity to evaluate the user's current cardiovascular health status and distinguish between normal and abnormal states (such as heart rate that is too fast / too slow, or decreased vascular elasticity); vascular blood flow change detection uses trend analysis of long-term monitoring data to detect changes in vascular blood flow and issue early warnings; personalized analysis combines information such as the user's age, gender, and medical history to provide personalized health assessment reports, avoiding misjudgments based on a single indicator.

[0058] F. Early Warning and Intervention Module When the analysis module detects abnormal physiological indicators (such as heart rate exceeding the normal range or significantly decreased vascular elasticity) or changes in vascular blood flow, it issues tiered alerts via terminal devices (sound, vibration, pop-up window) and simultaneously pushes targeted intervention suggestions. Tiered alerts include Level 1 (minor abnormality), Level 2 (significant abnormality), and Level 3 (serious abnormality), each corresponding to different intervention suggestions. Level 3 alerts can link with family members or community medical institutions for rapid intervention, reducing the risk of sudden cardiovascular events.

[0059] The following is a detailed explanation of the tiered early warning system: A tiered early warning mechanism is set up in the terminal device: Level 1 warning (minor abnormality): When physiological indicators deviate slightly from the normal range, the terminal device pops up a window and pushes lifestyle adjustment suggestions (such as a reasonable diet, moderate exercise, and regular work and rest); Level 2 warning (significant abnormality): When physiological indicators deviate significantly from the normal range, the terminal device emits a vibration and pop-up warning, reminding the user to retest in time and consult a doctor if necessary; Level 3 warning (serious abnormality): When physiological indicators deviate significantly from the normal range or significant changes in vascular blood flow are detected, the terminal device emits an audible, vibration, and pop-up warning, pushes an emergency warning message to the family's mobile phone, and provides access to nearby medical institutions for appointment booking, enabling rapid intervention.

[0060] G. Health Management Module It provides users with long-term health management services, including historical data query of physiological indicators (supporting daily, weekly and monthly queries), health report generation (generating one personalized health report per month), personalized health plan creation (pushing diet and exercise plans based on the user's health status), and medical appointment connection (linking with local community hospitals and tertiary hospitals, supporting online appointments), etc.

[0061] Users can view their own cardiovascular health trends through terminal devices, and doctors can understand users' long-term monitoring status through cloud data, providing a reference for clinical diagnosis and prevention, realizing a closed-loop service of "monitoring-analysis-intervention-management", which meets the development needs of full-cycle intelligent health monitoring.

[0062] The vascular blood flow change detection system in this embodiment is also equipped with a control module, which allows users to perform human-computer interaction and operation control.

[0063] The capacitive wearable non-invasive sensor proposed in this invention is a skin-touch flexible capacitive dielectric sensor. It can detect changes in vascular blood flow by detecting the capacitance change signal ΔC caused by the periodic change in the dielectric constant of arterial blood. The working principle of the skin-touch flexible capacitive dielectric sensor is explained below: The electrodes and wrist tissue form a variable capacitor (capacitive wearable non-invasive sensor), which is filled / emptied by arterial blood during the heart cycle, resulting in an equivalent dielectric constant. The capacitance changes periodically, generating a detectable capacitance change signal ΔC, thereby enabling the detection of changes in vascular blood flow.

[0064] Compared with the prior art, the present invention has the following beneficial effects: A. Sensor accuracy significantly improved This invention presents a capacitive wearable non-invasive sensor with an adaptive frequency conversion excitation module. This module dynamically adjusts the excitation frequency according to the different characteristics of human wrist tissue, adapting to different groups (elderly, young people, and obese individuals, etc.), avoiding power frequency interference, and improving the accuracy of electric field penetration and the ability to capture weak signals. The sensor also features a programmable gain amplification module, which automatically adjusts the intensity of capacitance change signals to avoid signal saturation or loss of weak signals. Combined with CNN-LSTM hybrid deep learning signal processing technology, it accurately filters out various noises, fuses multiple features to extract core indicators, achieving a heart rate detection error of ≤1 beat / minute, a vascular elasticity detection error of ≤3%, and consistency with professional medical equipment detection results exceeding 98%. The capacitance change detection accuracy reaches 0.01pF, and the signal-to-noise ratio under motion interference is improved by 60%, far exceeding existing similar sensors.

[0065] B. Strong anti-interference ability and excellent stability The adaptive frequency conversion excitation module of the capacitive wearable non-invasive sensor of this invention can effectively avoid power frequency and environmental electromagnetic interference. The programmable gain amplification module can be adjusted to ensure the acquisition stability under different signal strengths. The deep learning algorithm accurately filters out nonlinear interference such as motion and sweat. Dynamic baseline calibration eliminates baseline drift in real time, making the signal loss rate ≤0.5% and the baseline drift ≤0.05pF for continuous 24-hour monitoring. The accuracy rate of abnormal indicator warning reaches more than 95%, which can effectively capture abnormal cardiovascular signals and avoid missed or false judgments.

[0066] C. Easy to wear, non-invasive, and highly comfortable The capacitive wearable non-invasive sensor proposed in this invention adopts a flexible design that conforms to the contour of the human wrist, can adapt to different wrist sizes, is snug and comfortable to wear, has no puncture or stinging sensation, has no electrolysis or polarization reaction in the alternating electric field, and adaptive voltage adjustment further improves wearing comfort. It is suitable for long-term continuous wear and is compatible with various scenarios such as home and office, solving the problems of inconvenience and poor comfort of traditional devices.

[0067] D. Comprehensive functionality, enabling closed-loop management. The proposed method for detecting changes in vascular blood flow is implemented through a vascular blood flow change detection system. This system integrates high-precision data acquisition, deep learning signal processing, intelligent analysis, early warning intervention, and health management functions. It can not only monitor cardiovascular-related physiological indicators in real time and with high precision, but also realize abnormal early warning, detection of changes in vascular blood flow, and personalized health management. It breaks through the limitations of existing systems in terms of single function and insufficient accuracy, and realizes a closed-loop service of "monitoring-analysis-intervention-management" throughout the entire process. It can effectively help in the early detection and prevention of changes in vascular blood flow.

[0068] E. Costs are controllable and easy to promote. The capacitive wearable non-invasive sensor proposed in this invention has a simple structure, using conventional copper foil, medical-grade silicone, and other materials. Its adaptive frequency conversion excitation module and programmable gain amplification module are small in size and cost-effective. The deep learning algorithm can be integrated into smart hardware or a cloud server, eliminating the need for extensive additional hardware and facilitating mass production. With low hardware and software costs and no need for specialized operation, it can be widely applied in home monitoring, community screening, and chronic disease management, making it suitable for large-scale deployment. It is particularly suitable for daily monitoring of the elderly and high-risk groups for changes in vascular blood flow, possessing high practical value and market potential, and significantly improving the feasibility of monitoring changes in vascular blood flow.

[0069] Example 5: To verify the effectiveness of the vascular blood flow change detection method of the present invention, this embodiment recruited 100 volunteers (including 20 healthy individuals, 20 patients with hypertension, and 10 patients with arteriosclerosis) for a one-month test. The test results are as follows: Detection accuracy: Heart rate detection error ≤ 1 beat / minute, vascular elasticity detection error ≤ 3%, with consistency with professional medical equipment (electrocardiograph, vascular ultrasound) results exceeding 98%; pulse wave conduction velocity detection error ≤ 2%, enabling accurate assessment of vascular elasticity; Stability: Continuous 24-hour monitoring with a signal loss rate of ≤0.5% and a baseline drift of ≤0.05pF. It has strong resistance to interference from wrist micro-movements, dry or wet skin, and environmental electromagnetic interference. It can stably collect signals during daily activities (such as walking, working, resting, and light exercise). Comfort: Volunteers wore the masks for extended periods (8 hours a day) without experiencing any stinging, allergies, or other discomfort, with a satisfaction rate of 92%. Early warning accuracy: The accuracy rate of abnormal indicator early warning is over 95%, which can effectively capture abnormal cardiovascular signals (such as sudden increase in heart rate, atrial fibrillation, etc.) without any missed or false judgments. Adaptability: It can be adapted to people with different wrist sizes and different fat thicknesses, with a 100% compatibility rate and good consistency in detection accuracy among different groups of people.

[0070] Test results show that the capacitive wearable non-invasive sensor and vascular blood flow change detection method proposed in this invention can stably and accurately achieve real-time monitoring of vascular blood flow changes, early warning of abnormal vascular blood flow changes, and long-term management through adaptive frequency conversion excitation, programmable gain control, and deep learning signal processing technology. It meets the needs of home monitoring, community screening, chronic disease management and other scenarios, and has extremely high practicality and feasibility.

[0071] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A capacitive wearable non-invasive sensor, characterized in that, It includes two semi-circular copper foil electrodes, an adaptive frequency conversion excitation module, a programmable gain amplification module, and a signal acquisition interface; The two semi-circular copper foil electrodes are respectively attached to both sides of the human wrist, and the two semi-circular copper foil electrodes are arranged opposite each other and surround the human wrist. The adaptive frequency conversion excitation module is used to output an alternating excitation signal with an adaptively adjustable frequency within a set range to the two semi-circular copper foil electrodes. The two semi-circular copper foil electrodes form a coupling capacitor with the human wrist. The coupling capacitor generates a capacitance signal under the action of the alternating excitation signal. The capacitance signal changes with the change of blood pulse in the artery of the human wrist to form a capacitance change signal. The programmable gain amplifier module is used to automatically adjust the gain according to the strength of the capacitance change signal; The signal acquisition interface is used to transmit the gain-adjusted capacitance change signal to the signal acquisition chip for conversion processing. The programmable gain amplifier module is disposed between the signal acquisition interface and the signal acquisition chip.

2. The capacitive wearable non-invasive sensor according to claim 1, characterized in that, The surfaces of the two semi-circular copper foil electrodes are covered with an insulating layer.

3. The capacitive wearable non-invasive sensor according to claim 2, characterized in that, The insulating layer is a silicone insulating layer.

4. The capacitive wearable non-invasive sensor according to claim 1, characterized in that, The setting range is 50kHz to 150kHz.

5. The capacitive wearable non-invasive sensor according to claim 1, characterized in that, The gain adjustment range is 1 to 8 times.

6. The capacitive wearable non-invasive sensor according to claim 1, characterized in that, The automatic gain adjustment based on the strength of the capacitance change signal specifically includes: Increase the gain when the pulse signal is weak, resulting in a low intensity of the capacitance change signal; When a strong pulse signal results in a large intensity of capacitance change signal, reduce the gain.

7. The capacitive wearable non-invasive sensor according to claim 1, characterized in that, The capacitive wearable non-invasive sensor has a flexible structure.

8. A method for detecting changes in vascular blood flow, using the capacitive wearable non-invasive sensor described in claims 1-7, characterized in that, Includes the following steps: The alternating excitation signal and the capacitance change signal generated by the pulsation of blood in the human wrist artery are coupled to obtain the coupled signal. After adjusting the strength of the coupled signal, the coupled signal is converted into a digital signal; The digital signal is transmitted to the terminal device, and the digital signal is feature extracted in the terminal device to obtain heart rate features, pulse wave amplitude features, pulse wave morphology features and hemodynamic features; By fusing heart rate characteristics, pulse wave amplitude characteristics, pulse wave morphology characteristics, and hemodynamic characteristics, multi-dimensional features are obtained. By inputting multi-dimensional features into a trained physiological index analysis model, the results of vascular blood flow change detection are obtained.

9. The method for detecting changes in vascular blood flow according to claim 8, characterized in that, After obtaining the results of vascular blood flow change detection, early warning and intervention strategies are set based on these results.

10. The method for detecting changes in vascular blood flow according to claim 9, characterized in that, After setting early warning and intervention strategies, health management is carried out based on the health status corresponding to the blood pulsation of the wrist artery.

Images