Blood pressure monitoring device and method based on pulse sensor array
By combining a three-layer flexible structure and a deep neural network algorithm, the problems of comfort and accurate positioning of pulse sensors during long-term wear have been solved, realizing a highly integrated, small-volume blood pressure monitoring device and improving the portability and accuracy of blood pressure measurement.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SUN YAT SEN UNIV
- Filing Date
- 2025-12-22
- Publication Date
- 2026-04-24
AI Technical Summary
Existing pulse sensors are uncomfortable for users when worn tightly for extended periods and are difficult to pinpoint the exact location of the pulse, affecting the accuracy of blood pressure monitoring.
It adopts a three-layer flexible structure design, including conductive flexible material, flexible interdigital array electrodes, power supply bias module and acquisition circuit module, combined with leather and carboxylated carbon nanotube composite material, to realize a highly integrated small-volume device, and combines deep neural network algorithm for blood pressure prediction.
It improves the portability, real-time performance, and measurement accuracy of blood pressure monitoring, ensures stable pulse signal transmission and sensor permeability, reduces signal interference and distortion, and achieves non-invasive, highly accurate blood pressure measurement.
Smart Images
Figure CN121910344A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of blood pressure monitoring technology, and more particularly to a device and method for blood pressure monitoring based on a pulse sensor array. Background Technology
[0002] Human pulse wave signals are closely related to blood pressure. The multiple peaks, troughs, and progressive, reflected, and replay waves contained within the pulse wave correspond to the physiological processes of heartbeat and blood reflection within blood vessels. Therefore, pulse wave signals have been widely used as an indispensable technical means for disease diagnosis and treatment. With the rapid development of flexible electronics and machine learning, non-invasive continuous blood pressure monitoring based on pulse wave signals has become a research hotspot. Artificial intelligence technology can be used to conduct in-depth analysis of physiological signals collected by wearable electronic devices, thereby enhancing the application capabilities of flexible electronic devices in the field of disease diagnosis and treatment. Currently, the research and development of wearable flexible pulse sensors has achieved many results, forming various technical routes such as photoplethysmography (PPG), piezoresistive, capacitive, piezoelectric, triboelectric, and bioimpedance sensing. The mode of acquiring high-quality pulse signals through flexible pulse sensors and combining them with artificial intelligence for data analysis provides a new approach to the design of non-invasive blood pressure monitoring systems. Among various flexible pulse sensors, piezoresistive pulse sensors are widely used in pulse signal detection scenarios due to their advantages such as high sensitivity and ease of data acquisition.
[0003] However, existing pulse sensors still have many problems that need to be solved: on the one hand, many pulse sensors are designed without considering the user's comfort under long-term close-fitting conditions. The non-breathable sensor structure can easily cause skin discomfort or even allergic reactions after long-term wear. On the other hand, the pulse is a weak physiological signal of the human body, and it is difficult for users to accurately locate the pulse position. Moreover, the pulse morphology is easily affected by external pressure and other acquisition conditions. Different degrees of tightness of the sensor will change the morphology of the pulse signal. If these factors are ignored, it will have an adverse effect on the accuracy of subsequent blood pressure monitoring.
[0004] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention
[0005] The main objective of this application is to propose a device and method for blood pressure monitoring based on a pulse sensor array, which can improve pressure sensing performance and breathability, and achieve non-invasive and highly accurate blood pressure measurement.
[0006] To achieve the above objectives, one aspect of this application proposes a blood pressure monitoring device based on a pulse sensor array, the device comprising: A first flexible layer, comprising a conductive flexible material and the pulse sensor array, wherein the pulse sensor array is disposed in the conductive flexible material; The second flexible layer includes a flexible interdigital array electrode, a power supply bias module, a signal output interface array, and a data acquisition circuit module. The power supply bias module, the signal output interface array, and the data acquisition circuit module are respectively disposed on the flexible interdigital array electrode. One side of the second flexible layer is disposed above the first flexible layer. The signal output interface array is correspondingly connected to the pulse sensor array. A third flexible layer, comprising a single flexible material, is connected to the other side of the second flexible layer; The host computer is connected to the acquisition circuit module for communication.
[0007] In some embodiments, the conductive flexible material includes a leather-carboxylated carbon nanotube composite material; The single flexible material includes leather.
[0008] In some embodiments, the leather-carboxylated carbon nanotube composite material is obtained through the following steps: Carboxylated carbon nanotube powder, N-methylpyrrolidone, and a dispersant were added to deionized water to obtain a mixed solution. The mixed solution was ultrasonically dispersed in an ice bath to obtain a carboxylated carbon nanotube dispersion. The carboxylated carbon nanotube dispersion was diluted by drawing it with a syringe to obtain a filtrate. The leather and the filtrate are vacuum filtered through a shaping template to obtain the filtered leather. The filtered leather is dried to obtain the leather-carboxylated carbon nanotube composite material.
[0009] In some embodiments, one side of the second flexible layer is connected to the first flexible layer by sewing; The other side of the second flexible layer is connected to the third flexible layer by sewing.
[0010] In some embodiments, the acquisition circuit module includes an analog signal processing circuit and a main control power supply circuit; The analog signal processing circuit is equipped with a sensor input array and a power supply bias output. The analog signal processing circuit and the main control power supply circuit are fixedly connected by upper and lower pin headers.
[0011] In some embodiments, the acquisition circuit module is controlled by an STM32F103RET6 microprocessor and outputs a 0.1V voltage through a 12-bit digital-to-analog converter to bias the sensor. The acquisition circuit module also includes Bluetooth Low Energy and UART interfaces, through which data is transmitted.
[0012] In some embodiments, the data processing procedure of the acquisition circuit module includes the following steps: The pulse sensor array obtains the pulse current signal by biasing the output voltage of the digital-to-analog converter; The pulse current signal is amplified and converted by transimpedance to obtain a voltage signal; The voltage signal is processed by the first branch and the second branch respectively; In this process, the voltage signal of the first branch is amplified by an amplifier and then converted from analog to digital to obtain preliminary pulse pressure data, which is used for extracting wear pressure. The voltage signal of the second branch is adjusted by a bandpass filter, an amplifier, and an adder to obtain a voltage signal within a suitable voltage range. The voltage signal within the suitable voltage range is then processed by an analog-to-digital converter to obtain pulse pressure analysis data, which is used for pulse signal analysis.
[0013] In some embodiments, the data processing procedure of the acquisition circuit module further includes the following steps: The microprocessor performs an arithmetic average filter on the acquired preliminary pulse pressure data and the pulse pressure analysis data to obtain blood pressure monitoring data. The blood pressure monitoring data is transmitted to the low-power Bluetooth via an asynchronous transceiver; The low-power Bluetooth transmits the blood pressure monitoring data to the host computer.
[0014] To achieve the above objectives, another aspect of this application proposes a method for blood pressure monitoring based on a pulse sensor array, the method comprising the following steps: It is confirmed that the blood pressure monitoring device based on the pulse sensor array described above has been started and the blood pressure monitoring data has been transmitted to the host computer; The host computer preprocesses the acquired blood pressure monitoring data; The host computer extracts features from the preprocessed blood pressure monitoring data using a feature extraction algorithm to obtain blood pressure monitoring features; The host computer uses a blood pressure prediction model to predict the blood pressure monitoring characteristics and obtain the predicted blood pressure value. The host computer displays the predicted blood pressure value in real time through the user interface.
[0015] In some embodiments, the blood pressure prediction model is constructed using a deep neural network algorithm model; The deep neural network algorithm model consists of an input layer, three fully connected hidden layers, and an output layer, wherein the output layer includes systolic blood pressure output and diastolic blood pressure output. Each of the hidden layers includes a Dropout layer, the activation function of the hidden layer includes a modified linear unit, and the loss function of the hidden layer includes mean squared error.
[0016] The embodiments of this application include at least the following beneficial effects: This application provides a device and method for blood pressure monitoring based on a pulse sensor array. This scheme, through the integrated design of a three-layer flexible structure, embeds the pulse sensor array within the conductive flexible material of the first flexible layer, while integrating the flexible interdigital array electrodes, power supply bias module, signal output interface array, and acquisition circuit module into the second flexible layer. Combined with the encapsulation of the third flexible layer, it achieves a highly integrated and compact device form, significantly improving the flexibility and portability of use. The corresponding connection between the signal output interface array and the pulse sensor array ensures the stable transmission of pulse signals. Combined with the preprocessing function of the acquisition circuit module, it can effectively reduce interference and distortion during signal transmission. With the communication interaction with the host computer, it can achieve non-invasive, real-time, and highly accurate blood pressure measurement. Compared with existing blood pressure measuring instruments, the overall design of the three-layer flexible structure has significant advantages in portability, real-time performance, comfort, and measurement accuracy. Attached Figure Description
[0017] Figure 1 This is a schematic diagram showing the disassembled structure of the blood pressure monitoring device based on a pulse sensor array provided in the embodiments of this application; Figure 2 This is a schematic diagram of the assembly structure of a blood pressure monitoring device based on a pulse sensor array; Figure 3 This is a schematic diagram of the preparation process of composite materials of leather and carboxylated carbon nanotubes; Figure 4 This is a schematic diagram of the sensitivity curve of sensing unit A; Figure 5 This is a schematic diagram of the sensitivity curve of sensing unit B; Figure 6 This is a schematic diagram of the sensitivity curve of sensing unit C; Figure 7 This is a schematic diagram of the sensitivity curve of sensing unit D; Figure 8 This is a diagram illustrating response and recovery time. Figure 9 This is a diagram illustrating the limit of detection. Figure 10 This is a schematic diagram of the crosstalk-free response of the array; Figure 11This is a schematic diagram comparing the water vapor transmission rate of a blood pressure monitoring device based on a pulse sensor array and that of plastic wrap. Figure 12 This is a schematic diagram of the data acquisition circuit. Figure 13 This is a structural schematic diagram of the signal acquisition circuit system block diagram; Figure 14 This is the first physical diagram of the signal acquisition circuit; Figure 15 This is the second physical diagram of the signal acquisition circuit; Figure 16 This is a schematic diagram of the minimum system circuit of a microprocessor. Figure 17 This is a schematic diagram of the power module circuit; Figure 18 This is a schematic diagram of the low-power Bluetooth and UART interface circuit principles; Figure 19 This is a schematic diagram of the pulse signal processing circuit. Figure 20 This is a flowchart of a blood pressure monitoring method based on a pulse sensor array; Figure 21 This is a schematic diagram of the data flow of a blood pressure monitoring device based on a pulse sensor array; Figure 22 This is a logical schematic diagram of a blood pressure monitoring method based on a pulse sensor array; Figure 23 This is a scatter plot comparing the predicted and measured systolic and diastolic blood pressure. Figure 24 This is a schematic diagram of the user interface. Detailed Implementation
[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.
[0019] It is understood that the terms “first,” “second,” etc., used in this application may be used herein to describe various concepts, but unless otherwise stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the words “if,” “when,” or “in response to a determination” as used herein may be interpreted as “when…” or “when…” or “in response to a determination.”
[0020] As used in this application, the terms "at least one", "multiple", "each", "any", etc., "at least one" includes one, two or more, "multiple" includes two or more, "each" refers to each of the corresponding multiples, and "any" refers to any one of the multiples.
[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.
[0022] Before providing a detailed description of the embodiments of this application, some of the nouns and terms involved in the embodiments of this application will be explained first. The nouns and terms involved in the embodiments of this application are subject to the following interpretations.
[0023] Carboxylated carbon nanotubes: Carbon nanotubes are chemically modified to introduce carboxyl groups (-COOH), which enhances their hydrophilicity and dispersibility, and improves their compatibility in composite materials.
[0024] N-Methylpyrrolidone: A polar organic solvent commonly used to dissolve and disperse materials such as carbon nanotubes.
[0025] Dispersants: can reduce the interparticle forces in a liquid, so that carboxylated carbon nanotubes and other materials can be uniformly dispersed, avoiding agglomeration and ensuring material performance.
[0026] Vacuum filtration: This method uses negative pressure to accelerate the passage of liquid through a filter membrane, achieving solid-liquid separation. It is commonly used in the molding and composition control of composite materials.
[0027] MCU: Microcontroller Unit, responsible for signal processing, data transmission and module control.
[0028] UART interface: Universal Asynchronous Receiver / Transmitter, used to enable wired data communication between the acquisition circuit and the host computer.
[0029] BLE module: Bluetooth Low Energy module, used for short-range wireless data transmission between the acquisition circuit and the host computer, adapted for wearable devices.
[0030] GUI: Graphical User Interface, which displays blood pressure data in a visual form and enables interaction on the host computer.
[0031] Data processing algorithm: The collected pulse signals are processed by noise reduction and feature extraction to provide high-quality data for blood pressure prediction.
[0032] Among numerous cardiovascular disease risk factors, hypertension is considered one of the most dangerous and common, leading to stroke, myocardial infarction, heart failure, and other diseases. Therefore, monitoring blood pressure and its changes is crucial for understanding disease progression and guiding clinical treatment. Currently, clinical practice primarily relies on cuff-type blood pressure monitors to measure resting blood pressure. These instruments can only record isolated systolic and diastolic blood pressure, making continuous dynamic blood pressure measurement difficult. Furthermore, continuous blood pressure monitoring in clinical practice involves implanting invasive pressure sensors in the arteries. Such methods are invasive, causing discomfort to patients and posing a risk of infection, making them unsuitable for routine monitoring. Therefore, non-invasive continuous blood pressure monitoring methods have become a research hotspot.
[0033] Numerous studies have shown a close correlation between human pulse wave signals and blood pressure, and it has been widely used as an indispensable technical means for disease diagnosis and treatment. The pulse wave has multiple peaks and troughs, including advancing waves, reflected waves, and replay waves, corresponding to the heartbeat and blood reflection in blood vessels. With the rapid development of flexible electronics and machine learning, continuous blood pressure monitoring using pulse wave signals has become a research hotspot in non-invasive continuous blood pressure monitoring. Artificial intelligence can be used to achieve in-depth analysis of physiological signals collected by wearable electronic devices, thereby improving the capabilities of flexible electronic devices in disease diagnosis and treatment. Among existing flexible pulse sensors, piezoresistive pulse sensors are widely used for pulse signal detection due to their high sensitivity and ease of data acquisition. However, many reported pulse sensors neglect user comfort under prolonged close-fitting conditions; non-breathable pulse sensors may cause skin discomfort or even allergies with prolonged wear. Furthermore, as a weak physiological signal, the pulse is difficult for users to pinpoint precisely, and the pulse morphology is also easily affected by acquisition conditions such as external pressure. Consequently, the tightness of the pulse sensor worn can affect the morphology of the pulse signal. If these factors are not considered, it will affect the accuracy of subsequent blood pressure monitoring.
[0034] In view of this, such as Figure 1 and Figure 2As shown in the figure, this application provides a blood pressure monitoring device based on a pulse sensor array. The solution consists of a three-layer flexible structure and a host computer. Each layer and component works together to realize the functions of pulse signal acquisition, processing, and blood pressure monitoring. The first layer is the first flexible layer 100, the main body of which is a conductive flexible material. The pulse sensor array is embedded in the conductive flexible material. The properties of the conductive flexible material ensure the fit between the sensor and human skin and the stability of signal transmission. The second layer is the second flexible layer 200, which integrates a flexible interdigital array electrode, a power supply bias module, a signal output interface array, and a data acquisition circuit module. These components are directly mounted on the flexible interdigital array electrode. One side of the second flexible layer is attached to the top of the first flexible layer. The signal output interface array and the pulse sensor array form a one-to-one connection. The power supply bias module provides a stable operating voltage / current to the pulse sensor array and the data acquisition circuit module. The signal output interface array transmits the weak pulse electrical signals acquired by the pulse sensor array to the data acquisition circuit module, which then performs preprocessing such as amplification, filtering, and analog-to-digital conversion on the input signal. The third layer is the third flexible layer 300, made of a single flexible material. It connects to the side of the second flexible layer opposite to the first flexible layer and provides protection for the power supply bias module, signal output interface array, and data acquisition circuit module within the second flexible layer. It also enhances the overall wearability and flexibility of the device. In addition, the device is also equipped with a host computer, which establishes a communication connection with the acquisition circuit module to receive the preprocessed digital signals transmitted by the acquisition circuit module, and can realize the calculation, analysis and visualization of blood pressure data through built-in algorithms.
[0035] In some embodiments, the conductive flexible material includes a composite material of leather and carboxylated carbon nanotubes. The second flexible layer is a flexible interdigital array electrode designed for the pulse sensor, formed by sewing the leather and carboxylated carbon nanotube composite material of the first flexible layer into a laminated structure. The third flexible layer is ordinary leather, combined with the conductive flexible material to sandwich the second flexible layer in the middle of the device, also tightly connected by sewing. As a flexible resistive pressure sensor, each unit in the pulse sensor array can conformally fit the skin, thereby acquiring high-quality pulse signals. Furthermore, the array design increases the sensing area and ensures no crosstalk between sensing units, allowing users to more easily locate the pulse position, thus increasing the success rate of pulse acquisition. Since the sensor substrate material is leather, the sensor array also inherits the natural breathability of leather, ensuring user comfort during long-term wear.
[0036] like Figure 3As shown, carboxylated carbon nanotube powder, N-methylpyrrolidone, and a dispersant were added to deionized water to obtain a mixed solution. The mixed solution was ultrasonically dispersed in an ice bath to obtain a carboxylated carbon nanotube dispersion. Leather of appropriate size was cut as filter paper for vacuum filtration. An appropriate amount of the carboxylated carbon nanotube dispersion was drawn with a syringe and further diluted. A template was added between the leather and the filtrate, and the leather was vacuum filtered to obtain filtered leather. A pulse sensor array was placed on the filtered leather, and the filtered leather was placed in an oven to dry. Subsequently, the dried filtered leather was sewn together with electrodes and another layer of leather using a sewing machine to create a blood pressure monitoring device based on a pulse sensor array.
[0037] In some embodiments, such as Figure 4 , Figure 5 , Figure 6 , Figure 7 , Figure 8 , Figure 9 , Figure 10 and Figure 11 As shown, the sensing unit possesses high-sensitivity pressure sensing performance, fast response and recovery time, no crosstalk between sensing units, and good breathability. Notably, the leather-based pulse sensor array has the capability to sense both dynamic and static pressure, thus decoupling the wear pressure and pulse signals for subsequent calibration, thereby improving the accuracy of blood pressure calculation.
[0038] Specifically, Figures 4 to 7 The table presents the sensitivity curves of four sensing units, with the slope of the sensitivity fitting for sensing unit A being -0.0705 kPa. - ¹Measured sensitivity: -0.18 kPa - ¹, The fitting slope of sensor unit B is -0.0554 kPa - ¹, Measured -0.18 kPa - ¹, The fitting slope of sensing unit C is -0.117 kPa - ¹, The fitting slope of sensing unit D is -0.269 kPa - ¹, The degree of fit between the experimental data points and the fitted lines in each curve reflects the linearity of the response of different sensing units to pressure changes. Figure 8 The response-recovery curves show that during the loading / unloading process of the device at a pressure of 4 kPa, The value can fluctuate rapidly with changes in pressure, reflecting a short response and recovery time; Figure 9 The signal curve demonstrates the device's ability to detect low pressures of 2 Pa. The significant fluctuations in the signal demonstrate that the device has a low detection limit. Figure 10 In the three-dimensional response diagram, the independent signal changes of different sensing units in the time dimension, such as only a specific unit showing changes... The peak value indicates that the pulse sensor array has no crosstalk problem; Figure 11 The comparison chart of water vapor transmission rates shows that the water vapor transmission rate of the leather-based sensor array is significantly higher than that of the plastic wrap, demonstrating its superior breathability and contributing to improved wearing comfort. More specifically, during the pressure response and response recovery time tests, a 0.3V DC voltage was applied to the sensor. A self-made moving platform was used to achieve different degrees of vertical displacement of the pressure gauge, thereby applying different levels of pressure to the sensor. An LCR meter was used to record the resistance changes in real time. In the water vapor transmission rate test, three beakers were filled with the same mass of water, and then sealed with leather, filtered leather, and plastic wrap, respectively. The change in mass of the three beakers over time was measured using a weighing balance to determine the breathability of the three materials.
[0039] In some embodiments, such as Figure 12 As shown, the acquisition circuit module includes an analog signal processing circuit 212 and a main control power supply circuit 211. The analog signal processing circuit has a sensor input array 214 and a power supply bias output 213. The analog signal processing circuit and the main control power supply circuit are fixedly connected via upper and lower pin headers. Specifically, the acquisition circuit mainly comprises two PCBs: the main control power supply circuit and the analog signal processing circuit, which are connected via upper and lower pin headers. A blood pressure monitoring device based on a pulse sensor array acquires the pulse signal from the radial artery of the human wrist and connects it to the lower PCB. This signal undergoes transimpedance amplification, bandpass filtering, and amplification by the analog signal processing circuit. Subsequently, it is converted from analog to digital by the analog-to-digital converter (ADC) in the main control power supply circuit. The signal is then transmitted to the blood pressure calculation terminal, i.e., the host computer, via Bluetooth Low Energy.
[0040] The pulse signal is acquired by a data acquisition circuit controlled by an STM32F103RET6 microprocessor. The circuit diagram of the STM32F103RET6 microprocessor is shown below. Figure 16 As shown, a 0.1V voltage is output from a 12-bit digital-to-analog converter (DAC) for sensor bias. The circuit block diagram is as follows. Figure 13 As shown, the physical diagram of the circuit is as follows. Figure 14 and Figure 15 shown. Specifically, Figure 13The signal processing flow of the circuit is presented. The output signals of sensors 1 to 4 are first connected to the corresponding transimpedance amplifier modules (labeled "3") to convert the weak current signals into voltage signals. Then, they are processed in two ways. One way is filtered and the gain is adjusted by the bandpass filter and amplification module (labeled "3"), and the other way is directly connected to the amplification module (labeled "3"). Both types of processed signals are input to the ADC (analog-to-digital converter). The ADC is connected to the MCU (labeled "5"). The MCU also integrates the DAC (labeled "1") and the UART interface (labeled "4"), and realizes data transmission through Bluetooth Low Energy (labeled "6"). The power module (labeled "2") supplies power to the entire circuit system. Figure 14 and Figure 15 This corresponds to the core module in the block diagram: Figure 14 On the circuit board, multiple chips marked "3" correspond to signal conditioning modules such as transimpedance amplification and filtering amplification, while the interface below corresponds to the DAC (marked "1"). Figure 15 In the diagram, the chip marked "8" is the Bluetooth Low Energy module (corresponding to "6" in the block diagram), the chip marked "5" is the MCU, the interface marked "4" corresponds to UART, and the power module (marked "2") provides operating power to the circuit. Overall, it realizes the functions of sensor signal acquisition, conditioning, conversion and transmission.
[0041] like Figure 13 , Figure 14 , Figure 15 , Figure 16 , Figure 17 , Figure 18 and Figure 19 As shown, the pulse signal acquisition process is as follows: The sensor outputs a pulse current signal under the bias of the DAC output voltage. This current signal is then converted into a voltage signal through transimpedance amplification. This voltage signal is then processed in two ways: one path is amplified and input to a 12-bit analog-to-digital converter (ADC) for analog-to-digital conversion, used for wearable pressure extraction. The other path is bandpass filtered from 0.5-15Hz, then amplified and adjusted to a suitable voltage range by an adder, before being input to the ADC for pulse signal analysis. The microprocessor performs an arithmetic average filter on the recorded data, and then sends it to a Bluetooth Low Energy (BLE) chip via an asynchronous transceiver. The BLE module wirelessly transmits the data to the blood pressure calculation terminal via a specific protocol. Simultaneously with the pulse signal acquisition, a cuff blood pressure monitor is used to collect the volunteer's blood pressure values at that moment, collecting a total of 240 sets of pulse and blood pressure data.
[0042] Figure 20 This is an optional flowchart of a blood pressure monitoring method based on a pulse sensor array provided in the embodiments of this application. Figure 20 The method may include, but is not limited to, steps S110 to S150.
[0043] Step S110: Confirm that the blood pressure monitoring device based on the pulse sensor array has been started and the blood pressure monitoring data has been transmitted to the host computer; Step S120: The host computer preprocesses the acquired blood pressure monitoring data; Step S130: The host computer extracts features from the preprocessed blood pressure monitoring data using a feature extraction algorithm to obtain blood pressure monitoring features; Step S140: The host computer uses the blood pressure prediction model to predict the blood pressure monitoring characteristics and obtain the predicted blood pressure value. In step S150, the host computer displays the predicted blood pressure value in real time through the user interface.
[0044] Steps S110 to S150 as illustrated in this embodiment first confirm that the blood pressure monitoring device based on the pulse sensor array has been successfully started. The device collects blood pressure-related data through its built-in pulse sensor array, processes this data through the acquisition circuit, and transmits it to the host computer via Bluetooth Low Energy or UART interface. After receiving the blood pressure monitoring data, the host computer first preprocesses the raw data to remove noise, correct outliers, etc., and improve data quality. Then, the host computer calls a feature extraction algorithm to extract features from the preprocessed blood pressure monitoring data, and filters and obtains key blood pressure monitoring features that can reflect the blood pressure change pattern. Next, the host computer inputs the extracted blood pressure monitoring features into a preset blood pressure prediction model, and obtains an accurate blood pressure prediction value through model calculation. Finally, the host computer displays the calculated blood pressure prediction value in real time on the user interface, allowing users to intuitively view the blood pressure monitoring results.
[0045] Specifically, such as Figure 21 , Figure 22 , Figure 23 and Figure 24 As shown, blood pressure monitoring data, i.e., pulse signals, are processed through feature extraction and then input into a deep neural network (DNN) model. The DNN model consists of an input layer, three fully connected hidden layers, and an output layer. The preferred number of neurons in the three hidden layers are 512, 256, and 128, respectively. The outputs of the DNN model are diastolic and systolic blood pressure, respectively. 240 sets of pulse data were collected from three volunteers using a blood pressure monitoring device based on a pulse sensor array, and corresponding blood pressure values were collected using a cuff blood pressure monitor. After feature extraction and training, the final results are as follows: Figure 23 As shown. The mean error and standard deviation of systolic blood pressure are 0.200. 4.245 mmHg, with a mean error and standard deviation of 0.121 for diastolic blood pressure. 3.162 mmHg. This meets the US Association for the Advancement of Medical Devices (AAMD) requirements for blood pressure monitoring devices (5±8 mmHg), demonstrating the application potential of blood pressure monitoring devices based on pulse sensor arrays. Figure 24 A schematic diagram of the user interface of the entire device during non-invasive blood pressure monitoring is shown.
[0046] After collecting pulse and blood pressure data, feature extraction was performed on the collected pulse data. The `find_peaks` function in Python 3.10.11 was used to find the eigenvalues of the pulse signal and extract time-domain features. The final dataset consisted of 240 sets containing 20 features and two blood pressure labels. These were then divided into training and testing sets in an 8:2 ratio. To accelerate model convergence, the features were subjected to min-max normalization.
[0047] A deep neural network algorithm for blood pressure calculation was implemented using the PyTorch framework in Python 3.10.11. The DNN algorithm model consists of an input layer, three fully connected hidden layers, and an output layer. The input layer has 20 neurons, the hidden layers have 512, 256, and 128 neurons respectively, and the output layer contains two outputs: systolic blood pressure and diastolic blood pressure. Each hidden layer is followed by a Dropout layer with a probability of 0.2 to improve generalization ability and reduce overfitting. The Corrected Linear Unit (ReLU) is set as the activation function of the hidden layers, and the Mean Squared Error (MSE) is the loss function. The optimizer is the Adam algorithm with a learning rate of 0.001. The network was trained for 300 iterations using the training set. Finally, the test set was input into the trained model to evaluate its accuracy. Once the expected accuracy is achieved, it will be used for subsequent deployment on a host computer.
[0048] Subsequently, using a trained deep neural network algorithm model, i.e., a blood pressure prediction model, combined with data processing algorithms and feature extraction algorithms, a real-time blood pressure monitoring host computer was built. For example... Figure 21As shown in the figure, the complete hardware-to-algorithm chain of the blood pressure monitoring device, from signal acquisition to blood pressure calculation, is as follows: On the left is the sensor-skin interface, where the pulse sensor is attached to the skin to obtain raw pulse information by detecting physical signals corresponding to arterial pulsation (such as pressure and resistance changes); in the middle is the signal acquisition circuit, where the signal output by the sensor is first converted into a voltage signal by a transimpedance amplifier, then filtered by a bandpass filter to remove noise, amplified by an amplifier, and finally converted into a digital signal by an ADC (analog-to-digital converter); on the right is the blood pressure calculation terminal, where the digital signal first extracts key features such as pressure from the pulse data through a feature extraction stage, and then inputs into the blood pressure calculation model. This prediction model contains a neural network with a Dropout layer, which passes through hidden layers of 512, 256, and 128 neurons in sequence, and finally outputs the predicted results of systolic and diastolic blood pressure, thus fully realizing the entire process of physiological signal acquisition, hardware signal conditioning, algorithm feature extraction, and blood pressure numerical calculation.
[0049] like Figure 22 As shown in the diagram, a modular, layered architecture is used to implement data processing and blood pressure prediction functions: First, the initialization module completes serial port configuration, neural network model loading, and GUI initialization; then, the data acquisition module reads the signals transmitted from the external pulse sensor and acquisition circuit in real time from the serial port, parses them into raw data sequences, and caches them; after entering the data processing module, it first updates the latest data through a sliding window, and then completes data preprocessing through wavelet denoising (wden function), baseline extraction / removal (wavelet decomposition and reconstruction), and signal standardization (mean / standard deviation normalization); next, the feature extraction module sequentially completes the extraction of feature points (peak / valley), time features (cycle / rise time, etc.), amplitude features (threshold counting), and area ratio / heart rate features (systolic / diastolic), while optimizing the feature sequence through outlier removal (standard deviation method) and moving average processing; then, the blood pressure prediction module normalizes the features to a standard range, inputs them into a pre-trained PyTorch neural network model to obtain systolic / diastolic predicted values; finally, the result display module updates blood pressure values and system status information (serial port information / processing status) in real time in the GUI, and realizes real-time scheduling and updating of the process through the system main loop.
[0050] like Figure 24 As shown, the host computer is developed based on Python 3.10.11. The overall process includes serial transmission of circuit signals to the host computer, data preprocessing, feature extraction, pre-trained model prediction of blood pressure, and real-time display of results via GUI. Based on this host computer and a leather-based pulse sensor array, non-invasive continuous blood pressure monitoring is achieved. The pulse signal is transmitted from the sensor to the circuit for processing, and then communicates with the host computer in real time via serial port through Bluetooth Low Energy. The data is sequentially processed and input into the pre-trained model to obtain the blood pressure value, which is then displayed in real time.
[0051] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.
[0052] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.
[0053] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.
[0054] It is understood that the content of the above method embodiments is applicable to the embodiments of this program product. The specific functions implemented by the embodiments of this program product are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.
[0055] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.
[0056] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.
[0057] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.
[0058] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.
[0059] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or appropriate combinations thereof.
[0060] Furthermore, the terms “comprising” and “having”, and any variations thereof, are intended to cover non-exclusive inclusion, such that a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such process, method, product, or apparatus.
[0061] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.
[0062] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.
[0063] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0064] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.
Claims
1. A device for blood pressure monitoring based on a pulse sensor array, characterized in that, The device includes: A first flexible layer, comprising a conductive flexible material and the pulse sensor array, wherein the pulse sensor array is disposed in the conductive flexible material; The second flexible layer includes a flexible interdigital array electrode, a power supply bias module, a signal output interface array, and a data acquisition circuit module. The power supply bias module, the signal output interface array, and the data acquisition circuit module are respectively disposed on the flexible interdigital array electrode. One side of the second flexible layer is disposed above the first flexible layer. The signal output interface array is correspondingly connected to the pulse sensor array. A third flexible layer, comprising a single flexible material, is connected to the other side of the second flexible layer; The host computer is connected to the acquisition circuit module for communication.
2. The blood pressure monitoring device based on a pulse sensor array according to claim 1, characterized in that, The conductive flexible material includes a composite material of leather and carboxylated carbon nanotubes; The single flexible material includes leather.
3. The blood pressure monitoring device based on a pulse sensor array according to claim 2, characterized in that, The leather-carboxylated carbon nanotube composite material is obtained through the following steps: Carboxylated carbon nanotube powder, N-methylpyrrolidone, and a dispersant were added to deionized water to obtain a mixed solution. The mixed solution was ultrasonically dispersed in an ice bath to obtain a carboxylated carbon nanotube dispersion. The carboxylated carbon nanotube dispersion was diluted by drawing it with a syringe to obtain a filtrate. The leather and the filtrate are vacuum filtered through a shaping template to obtain the filtered leather. The filtered leather is dried to obtain the leather-carboxylated carbon nanotube composite material.
4. The blood pressure monitoring device based on a pulse sensor array according to claim 2, characterized in that, One side of the second flexible layer is connected to the first flexible layer by sewing; The other side of the second flexible layer is connected to the third flexible layer by sewing.
5. The blood pressure monitoring device based on a pulse sensor array according to claim 1, characterized in that, The acquisition circuit module includes an analog signal processing circuit and a main control power supply circuit; The analog signal processing circuit is equipped with a sensor input array and a power supply bias output. The analog signal processing circuit and the main control power supply circuit are fixedly connected by upper and lower pin headers.
6. The blood pressure monitoring device based on a pulse sensor array according to claim 1, characterized in that, The acquisition circuit module is controlled by an STM32F103RET6 microprocessor and outputs a 0.1V voltage through a 12-bit digital-to-analog converter to bias the sensor. The acquisition circuit module also includes Bluetooth Low Energy and UART interfaces, through which data is transmitted.
7. The blood pressure monitoring device based on a pulse sensor array according to claim 6, characterized in that, The data processing procedure of the acquisition circuit module includes the following steps: The pulse sensor array obtains the pulse current signal by biasing the output voltage of the digital-to-analog converter; The pulse current signal is amplified and converted by transimpedance to obtain a voltage signal; The voltage signal is processed by the first branch and the second branch respectively; In this process, the voltage signal of the first branch is amplified by an amplifier and then converted from analog to digital to obtain preliminary pulse pressure data, which is used for extracting wear pressure. The voltage signal of the second branch is adjusted by a bandpass filter, an amplifier, and an adder to obtain a voltage signal within a suitable voltage range. The voltage signal within the suitable voltage range is then processed by an analog-to-digital converter to obtain pulse pressure analysis data, which is used for pulse signal analysis.
8. The blood pressure monitoring device based on a pulse sensor array according to claim 7, characterized in that, The data processing procedure of the acquisition circuit module also includes the following steps: The microprocessor performs an arithmetic average filter on the acquired preliminary pulse pressure data and the pulse pressure analysis data to obtain blood pressure monitoring data. The blood pressure monitoring data is transmitted to the low-power Bluetooth via an asynchronous transceiver; The low-power Bluetooth transmits the blood pressure monitoring data to the host computer.
9. A method for blood pressure monitoring based on a pulse sensor array, characterized in that, The method includes the following steps: It is determined that the blood pressure monitoring device based on the pulse sensor array according to any one of claims 1-8 has been started and the blood pressure monitoring data has been transmitted to the host computer; The host computer preprocesses the acquired blood pressure monitoring data; The host computer extracts features from the preprocessed blood pressure monitoring data using a feature extraction algorithm to obtain blood pressure monitoring features; The host computer uses a blood pressure prediction model to predict the blood pressure monitoring characteristics and obtain the predicted blood pressure value. The host computer displays the predicted blood pressure value in real time through the user interface.
10. The method according to claim 9, characterized in that, The blood pressure prediction model is constructed using a deep neural network algorithm. The deep neural network algorithm model consists of an input layer, three fully connected hidden layers, and an output layer, wherein the output layer includes systolic blood pressure output and diastolic blood pressure output. Each of the hidden layers includes a Dropout layer, the activation function of the hidden layer includes a modified linear unit, and the loss function of the hidden layer includes mean squared error.