Research device for arteriosclerosis detection

Abstract

The invention provides a research device for arteriosclerosis detection, and the device consists of the following parts: a multi-modal data synchronous collection module which achieves the omnibearing data capture of arteriosclerosis through integrating a flexible ultrasonic array, a PPG module, a laser speckle imaging module and a multi-modal sensor, the blood vessel structure thickness, plaque form and microcirculation dynamic parameters are synchronously obtained in the wearable device, the limitation of single function of traditional equipment is broken through, comprehensive diagnosis from macroscopic to microscopic is achieved, continuous wearing is supported through the ultrathin flexible design, the self-adaptive ultrasonic array and high-precision time sequence synchronization are combined, blood vessel elastic changes can be captured in an all-weather mode, and the diagnosis accuracy is improved. And real-time data support is provided for disease course evolution.

Description

Technical Field

[0001] The present invention relates to the technical field of arteriosclerosis detection, and in particular to a research device for arteriosclerosis detection. Background Art

[0002] Arteriosclerosis detection devices are medical devices that comprehensively assess the degree of vascular elasticity degeneration and the risk of cardiovascular disease by collecting vascular structure, hemodynamics and microcirculation parameters (such as ultrasound imaging to measure the intima-media thickness IMT, pulse wave velocity PWV to analyze large artery stiffness, photoplethysmography PPG to obtain pulse wave morphological characteristics, and laser speckle imaging to quantify microvascular pulsation status). However, existing technologies have significant defects: at the device structure level, traditional detection relies on split devices (such as ultrasound machines, PWV detectors), which are complicated to operate and cannot be integrated and portable. Professionals need to hold a rigid ultrasound probe to locate blood vessels, and the sensor has difficulty adapting to the skin curvature (an incident angle deviation of >10° results in an IMT measurement error of ±0.05mm). In addition, multimodal data is collected asynchronously due to the discrete operation of the devices (timestamp alignment error >50ms). At the detection method level, existing technologies rely on a single parameter (such as only IMT or PWV), resulting in a high rate of missed detection of early sclerosis (a missed detection rate of microcirculatory disorders >40%). It only supports a single short-term test (such as a 5-minute measurement in an outpatient clinic) and cannot capture diurnal fluctuations and long-term progression trends. At the same time, due to the lack of the ability to quantify microvascular pulsation entropy (MPE) and blood flow oscillation frequency, it cannot identify early signals of endothelial dysfunction (such as microcirculatory abnormalities in prediabetes). Summary of the Invention

[0003] Technical problems solved

[0004] In response to the shortcomings of the existing technology, the present invention provides a research device for detecting arteriosclerosis, which solves the problems of existing research devices for detecting arteriosclerosis, such as traditional detection relying on split equipment, complex operation and inability to integrate and portability, and the high rate of missed detection of early sclerosis caused by the existing technology relying on a single parameter.

[0005] Technical Solution

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a research device for detecting arteriosclerosis, the device comprising the following parts:

[0007] Multimodal data synchronous acquisition module: The multimodal data synchronous acquisition module integrates a flexible ultrasound array, a PPG module, laser speckle imaging, and a multimodal sensor to achieve all-round data capture of arteriosclerosis;

[0008] Edge-cloud collaborative data processing module: This module uses a collaborative architecture of edge computing and cloud AI to achieve efficient data processing and intelligent analysis. The edge-end real-time pre-processing in this module reduces data transmission bandwidth requirements and improves response speed. The cloud-end in this module uses a deep learning model to fuse multimodal data to generate comprehensive indicators such as arterial stiffness index and plaque vulnerability score.

[0009] Dynamic feedback and health management module: The dynamic feedback and health management module converts the analysis results into real-time warnings and personalized health intervention plans, forming a closed-loop management.

[0010] Preferably, the multimodal data synchronous acquisition module includes a flexible ultrasonic array unit, a PPG and blood oxygen transmission unit, a laser speckle microblood flow unit and a multimodal synchronous controller. The flexible ultrasonic array unit is an 8×8 array composed of micro piezoelectric sensors, embedded in the inner contact surface of the bracelet, with an operating frequency of 10-20MHz, a resolution of 0.1mm, and a built-in shape memory alloy bracket. The array angle is automatically adjusted according to the curvature of the skin surface to ensure that the ultrasonic beam enters the blood vessel wall vertically.

[0011] Preferably, the PPG and blood oxygen transmission unit uses a dual-wavelength LED light source and a photodiode with a sampling rate of 100Hz and is integrated on the outside of the bracelet.

[0012] Preferably, the laser speckle micro-blood flow unit adopts a 785nm laser diode and a CMOS image sensor, and calculates the microvascular pulsation entropy value through a speckle contrast algorithm. The multimodal synchronization controller is based on an FPGA hardware synchronization circuit, which uniformly manages the clock signals of the ultrasound, PPG, and laser modules to ensure that the data timestamp alignment accuracy is <1ms.

[0013] Preferably, the acquisition steps of the multimodal data synchronous acquisition module are:

[0014] After the flexible ultrasound array is activated, it automatically identifies the centerline position of the blood vessel by scanning the maximum echo signal intensity. When the ultrasound pulse signal is emitted, the PPG and laser speckle modules are synchronously triggered to start acquisition, avoiding timing deviation and continuously acquiring the following data: ultrasound image: blood vessel cross-sectional IMT, plaque three-dimensional contour; PPG waveform: pulse wave rise time, reflection wave enhancement index (AIx); laser speckle sequence: microvascular pulsation entropy (MPE) and blood flow oscillation frequency.

[0015] Preferably, the edge-cloud collaborative data processing module includes a cloud AI analysis platform and an edge computing unit;

[0016] The remote AI analysis unit is integrated into the low-power AI chip in the bracelet, supporting real-time signal filtering and feature extraction;

[0017] The edge computing unit is based on the deep learning model library of the GPU cluster, which includes three sub-models: vascular elasticity assessment, plaque risk grading, and microcirculation disorder detection.

[0018] Preferably, the data processing steps of the edge-cloud collaborative data processing module include edge preprocessing, feature compression transmission, and cloud multi-task AI modeling;

[0019] The edge preprocessing utilizes ultrasonic noise reduction: an adaptive wavelet threshold algorithm is used to eliminate tissue clutter; PPG motion artifact suppression: dynamic compensation is performed in combination with accelerometer data (built-in 6-axis IMU); laser speckle analysis: the spatiotemporal variation of speckle contrast is calculated to generate a microvascular pulsation entropy curve;

[0020] The feature compression transmission encrypts and uploads the pre-processed feature data (not the original signal) to the cloud, reducing bandwidth requirements by 90%;

[0021] The cloud-based multi-task AI modeling establishes a vascular elasticity assessment model: input: IMT, PWV, AIx → output: arterial stiffness index; a plaque risk grading model: input: plaque volume, calcification ratio, surface roughness → output: plaque vulnerability score (PVS); a microcirculation disorder detection model: input: MPE, blood flow oscillation frequency → output: microcirculation health index.

[0022] Preferably, the dynamic feedback and health management module includes a mobile APP interaction module and a wearable device actuator;

[0023] The mobile APP interactive module: visual report generation engine, early warning decision tree, and personalized intervention plan library;

[0024] The wearable device actuator: The bracelet has a built-in vibration motor and LED indicator light for real-time early warning prompts.

[0025] Preferably, the dynamic feedback step of the dynamic feedback and health management module is:

[0026] Instant risk assessment unit: If ASI>70 or PVS>80%, the app triggers a red alert and vibrates, and pushes information about nearby medical institutions;

[0027] Long-term trend unit: Cloud AI compares historical data and calculates the rate of hardening progress;

[0028] Personalized intervention unit: Matches intervention plan library according to patient data.

[0029] Beneficial effects

[0030] The present invention provides a research device for detecting arteriosclerosis. It has the following beneficial effects:

[0031] 1. This invention simultaneously acquires vascular structure thickness, plaque morphology, and microcirculatory dynamic parameters in a wearable device, breaking through the single-function limitations of traditional devices and achieving comprehensive diagnosis from the macro to the micro level. Its ultra-thin and flexible design supports continuous wear. Combined with an adaptive ultrasound array and high-precision timing synchronization, it can capture changes in vascular elasticity around the clock, providing real-time data support for disease progression. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is an overall flow chart of a research device for detecting arteriosclerosis proposed by the present invention;

[0033] Figure 2 This is a flow chart of a multimodal data synchronous acquisition module of a research device for arteriosclerosis detection proposed by the present invention;

[0034] Figure 3 This is a flow chart of the edge-cloud collaborative data processing module of a research device for arteriosclerosis detection proposed by the present invention;

[0035] Figure 4 This is a flow chart of the dynamic feedback and health management module of the research device for arteriosclerosis detection proposed by the present invention. DETAILED DESCRIPTION

[0036] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0037] Example:

[0038] like Figure 1-4 As shown, an embodiment of the present invention provides a research device for detecting arteriosclerosis, which is composed of the following parts:

[0039] Multimodal data synchronous acquisition module: The multimodal data synchronous acquisition module integrates a flexible ultrasound array, a PPG module, laser speckle imaging, and a multimodal sensor to achieve comprehensive data capture of arteriosclerosis;

[0040] Edge-cloud collaborative data processing module: The edge-cloud collaborative data processing module achieves efficient data processing and intelligent analysis through the collaborative architecture of edge computing and cloud AI. The edge-end real-time pre-processing in the edge-cloud collaborative data processing module reduces the data transmission bandwidth requirement and improves the response speed. The cloud-end in the edge-cloud collaborative data processing module uses deep learning models to fuse multimodal data to generate comprehensive indicators such as arterial stiffness index and plaque vulnerability score.

[0041] Dynamic feedback and health management module: The dynamic feedback and health management module converts analysis results into real-time warnings and personalized health intervention plans, forming a closed-loop management.

[0042] The multimodal data synchronization acquisition module includes a flexible ultrasound array unit, a PPG and blood oxygen sensor unit, a laser speckle micro-blood flow unit, and a multimodal synchronization controller. The flexible ultrasound array unit consists of an 8×8 array composed of micro piezoelectric sensors, embedded in the inner contact surface of the bracelet, with an operating frequency of 10-20MHz and a resolution of 0.1mm. It has a built-in shape memory alloy bracket and automatically adjusts the array angle according to the curvature of the skin surface to ensure that the ultrasound beam is perpendicular to the blood vessel wall. Micro piezoelectric sensors (model PZT-5H, size 5×5×0.5mm, frequency 10-20MHz) form an 8×8 array embedded in the inner layer of the wearable device. The shape memory alloy (SMA) bracket dynamically adjusts the array curvature according to the skin surface pressure feedback to ensure that the ultrasound beam is perpendicular to the blood vessel wall (angle deviation <2°), blood vessel cross-sectional images (resolution 0.1mm), and real-time calculation of intima-media thickness (IMT) and plaque three-dimensional volume (accuracy ±0.05mm) 3 );

[0043] The PPG and blood oxygen sensor units use dual-wavelength LED light sources and photodiodes with a sampling rate of 100Hz and are integrated on the outside of the bracelet. The laser speckle micro-blood flow unit uses a 785nm laser diode and a CMOS image sensor to calculate the microvascular pulsation entropy value through a speckle contrast algorithm. The multi-modal synchronization controller is based on an FPGA-based hardware synchronization circuit, which uniformly manages the clock signals of the ultrasound, PPG, and laser modules to ensure data timestamp alignment accuracy of <1ms. The hardware composition of the PPG and laser speckle modules is as follows:

[0044] (1) Dual-wavelength PPG module (660nm / 940nm LED, sampling rate 100Hz), integrated photodiode (response time 10ns);

[0045] (2) 785nm laser speckle imaging unit (CMOS sensor frame rate 30fps, speckle contrast algorithm);

[0046] Data output:

[0047] (1) Pulse wave velocity (PWV, accuracy ±0.2 m / s) and reflection enhancement index (AIx) were extracted from the PPG waveform.

[0048] (2) The laser speckle sequence generates microvascular pulsation entropy (MPE, range 0-1) to quantify the degree of microcirculatory oscillation disorder;

[0049] Multi-mode synchronization controller implementation:

[0050] (1) Hardware-level clock synchronization of the FPGA chip (model Xilinx Zynq-7000) aligns the ultrasound emission pulse, PPG sampling, and laser speckle frame acquisition, with a timing error of <1ms.

[0051] The acquisition steps of the multimodal data synchronous acquisition module are as follows: after the flexible ultrasound array is activated, the blood vessel centerline position is automatically identified by scanning the maximum echo signal intensity. When the ultrasound pulse signal is emitted, the PPG and laser speckle modules are synchronously triggered to start acquisition to avoid timing deviation and continuously acquire the following data: ultrasound image: blood vessel cross-sectional IMT, plaque three-dimensional contour; PPG waveform: pulse wave rise time, reflection wave enhancement index (AIx); laser speckle sequence: microvascular pulsation entropy (MPE) and blood flow oscillation frequency.

[0052] The edge-cloud collaborative data processing module includes a cloud-based AI analysis platform and an edge computing unit. The remote AI analysis unit is a low-power AI chip integrated into the wristband, supporting real-time signal filtering and feature extraction. The edge computing unit is based on a deep learning model library based on a GPU cluster, which includes three sub-models: vascular elasticity assessment, plaque risk grading, and microcirculatory disorder detection. The data processing steps of the edge-cloud collaborative data processing module include edge preprocessing, feature compression and transmission, and cloud-based multi-task AI modeling.

[0053] Edge preprocessing utilizes ultrasonic noise reduction: an adaptive wavelet threshold algorithm is used to eliminate tissue clutter; PPG motion artifact suppression: dynamic compensation is performed by combining accelerometer data (built-in 6-axis IMU); laser speckle analysis: the spatiotemporal variation of speckle contrast is calculated to generate a microvascular pulsation entropy curve;

[0054] Feature compression transmission encrypts and uploads pre-processed feature data (not raw signals) to the cloud, reducing bandwidth requirements by 90%;

[0055] Cloud-based multi-task AI modeling establishes a vascular elasticity assessment model: input: IMT, PWV, AIx → output: arterial stiffness index; plaque risk grading model: input: plaque volume, calcification ratio, surface roughness → output: plaque vulnerability score (PVS); microcirculation disorder detection model: input: MPE, blood flow oscillation frequency → output: microcirculation health index.

[0056] The dynamic feedback and health management module includes a mobile APP interaction module and a wearable device actuator;

[0057] Mobile app interactive module: Visual report generation engine, early warning decision tree, personalized intervention plan library, visual interface displays ASI, PVS, and MHI three-dimensional radar charts, overlaid with the baseline range of healthy people of the same age (gray background contrast), plaque location heat map (red highlights calcified areas) and microcirculation entropy value curve (24-hour trend). Early warning mechanism: If ASI>70 or PVS>80, a vibration alarm (frequency 100Hz) is triggered and the app pop-up pushes emergency medical advice (including map navigation of nearby cardiovascular specialists);

[0058] Wearable device actuator: The bracelet has a built-in vibration motor and LED indicator light for real-time warning prompts.

[0059] The dynamic feedback steps of the dynamic feedback and health management module are:

[0060] (1) Instant risk assessment unit: If ASI>70 or PVS>80%, the APP triggers a red alert and vibrates, and pushes information about nearby medical institutions;

[0061] (2) Long-term trend unit: Cloud AI compares historical data and calculates the rate of hardening progress;

[0062] (3) Personalized intervention unit: Match the intervention plan library based on patient data.

[0063] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A research device for detecting arteriosclerosis, characterized in that: The device consists of the following parts: Multimodal data synchronous acquisition module: The multimodal data synchronous acquisition module integrates a flexible ultrasound array, a PPG module, laser speckle imaging, and a multimodal sensor to achieve all-round data capture of arteriosclerosis; Edge-cloud collaborative data processing module: This module uses a collaborative architecture of edge computing and cloud AI to achieve efficient data processing and intelligent analysis. The edge-end real-time pre-processing in this module reduces data transmission bandwidth requirements and improves response speed. The cloud-end in this module uses a deep learning model to fuse multimodal data to generate comprehensive indicators such as arterial stiffness index and plaque vulnerability score. Dynamic feedback and health management module: The dynamic feedback and health management module converts the analysis results into real-time warnings and personalized health intervention plans, forming a closed-loop management.

2. The research device for detecting arteriosclerosis according to claim 1, characterized in that: The multimodal data synchronization acquisition module includes a flexible ultrasound array unit, a PPG and blood oxygen sensor unit, a laser speckle micro-blood flow unit, and a multimodal synchronization controller. The flexible ultrasound array unit is an 8×8 array composed of micro piezoelectric sensors, embedded in the inner contact surface of the bracelet, with an operating frequency of 10-20MHz and a resolution of 0.1mm. It has a built-in shape memory alloy bracket and automatically adjusts the array angle according to the curvature of the skin surface to ensure that the ultrasound beam enters the blood vessel wall vertically.

3. The research device for detecting arteriosclerosis according to claim 2, characterized in that: The PPG and blood oxygen transmission unit uses a dual-wavelength LED light source and a photodiode, with a sampling rate of 100Hz, and is integrated on the outside of the bracelet.

4. The research device for detecting arteriosclerosis according to claim 2, characterized in that: The laser speckle micro-blood flow unit uses a 785nm laser diode and a CMOS image sensor to calculate the microvascular pulsation entropy value through a speckle contrast algorithm. The multimodal synchronization controller is based on an FPGA-based hardware synchronization circuit to uniformly manage the clock signals of the ultrasound, PPG, and laser modules, ensuring data timestamp alignment accuracy of <1ms.

5. The research device for detecting arteriosclerosis according to claim 1, characterized in that: The acquisition steps of the multimodal data synchronous acquisition module are: After the flexible ultrasound array is activated, it automatically identifies the centerline position of the blood vessel by scanning the maximum echo signal intensity. When the ultrasound pulse signal is emitted, the PPG and laser speckle modules are synchronously triggered to start acquisition, avoiding timing deviation and continuously acquiring the following data: ultrasound image: blood vessel cross-sectional IMT, plaque three-dimensional contour; PPG waveform: pulse wave rise time, reflection wave enhancement index (AIx); laser speckle sequence: microvascular pulsation entropy (MPE) and blood flow oscillation frequency.

6. The research device for detecting arteriosclerosis according to claim 1, characterized in that: The edge-cloud collaborative data processing module includes a cloud AI analysis platform and an edge computing unit; The remote AI analysis unit is integrated into the low-power AI chip in the bracelet, supporting real-time signal filtering and feature extraction; The edge computing unit is based on the deep learning model library of the GPU cluster, which includes three sub-models: vascular elasticity assessment, plaque risk grading, and microcirculation disorder detection.

7. The research device for detecting arteriosclerosis according to claim 1, characterized in that: The data processing steps of the edge-cloud collaborative data processing module include edge preprocessing, feature compression transmission, and cloud multi-task AI modeling; The edge preprocessing utilizes ultrasonic noise reduction: an adaptive wavelet threshold algorithm is used to eliminate tissue clutter; PPG motion artifact suppression: dynamic compensation is performed in combination with accelerometer data (built-in 6-axis IMU); laser speckle analysis: the spatiotemporal variation of speckle contrast is calculated to generate a microvascular pulsation entropy curve; The feature compression transmission encrypts and uploads the pre-processed feature data (not the original signal) to the cloud, reducing bandwidth requirements by 90%; The cloud-based multi-task AI modeling establishes a vascular elasticity assessment model: input: IMT, PWV, AIx → output: arterial stiffness index; a plaque risk grading model: input: plaque volume, calcification ratio, surface roughness → output: plaque vulnerability score (PVS); a microcirculation disorder detection model: input: MPE, blood flow oscillation frequency → output: microcirculation health index.

8. The research device for detecting arteriosclerosis according to claim 1, characterized in that: The dynamic feedback and health management module includes a mobile APP interaction module and a wearable device actuator; The mobile APP interactive module: visual report generation engine, early warning decision tree, and personalized intervention plan library; The wearable device actuator: The bracelet has a built-in vibration motor and LED indicator light for real-time early warning prompts.

9. The research device for detecting arteriosclerosis according to claim 1, characterized in that: The dynamic feedback steps of the dynamic feedback and health management module are: Instant risk assessment unit: If ASI>70 or PVS>80%, the app triggers a red alert and vibrates, and pushes information about nearby medical institutions; Long-term trend unit: Cloud AI compares historical data and calculates the rate of hardening progress; Personalized intervention unit: Matches intervention plan library according to patient data.

Images