Non-contact millimeter wave radar human precise measurement and health assessment system for medical examination

Abstract

The present application relates to the technical field of digital medical instruments and biomedical engineering, and particularly relates to a non-contact millimeter wave radar human precise measurement and health assessment system for medical examination, which comprises a medical-grade synchronous control gateway, a distributed multiple-input multiple-output (MIMO) radar array and a medical data processing workstation, the distributed multiple-input multiple-output (MIMO) radar array comprises a plurality of distributed millimeter wave radar front ends, is used for full-coverage scanning of a standard anatomic station posture of a subject to obtain human original point cloud data. The present application adopts a full-electronic non-contact millimeter wave radar scanning mode, effectively eliminates the reading error of traditional manual measurement and the infection risk of contact type, avoids the privacy leakage problem caused by optical structured light imaging, greatly improves the clinical compliance of the subject, and is different from a traditional single radar or a multi-radar system in disorderly stacking.

Description

Technical Field

[0001] This invention relates to the interdisciplinary field of digital medical devices and biomedical engineering, and in particular to a non-contact millimeter-wave radar human body precision measurement and health assessment system for medical examinations. Background Technology

[0002] In modern clinical medical examinations and health assessment systems, human body shape characteristics and circumference data (such as height, waist circumference, hip circumference, and waist-to-hip ratio) are important medical anthropometry indicators for measuring health status and monitoring signs of metabolic and other chronic diseases. However, current medical examination scenarios mainly rely on the following traditional methods to obtain these indicators, all of which have revealed significant limitations in practical applications: First, traditional anthropometric measurements mainly rely on manual contact measurements by medical staff using soft measuring tapes. This method is not only labor-intensive, but also prone to significant reading errors in high-concurrency scenarios such as health check-up centers due to differences in the subjective techniques of different measurers. More seriously, contact measurements can easily become a medium for the spread of pathogens, bringing significant clinical risks of cross-infection.

[0003] Secondly, with the development of machine vision technology, some medical examination institutions have begun to introduce optical imaging systems based on optical cameras or 3D structured light for non-contact measurement. However, since three-dimensional reconstruction of the human body requires obtaining a realistic body surface contour, examinees often need to wear tight clothing or expose parts of their bodies. This raises significant privacy and ethical issues in medical examination scenarios, leading to psychological resistance from examinees and making it difficult to popularize in rapid health screenings for large populations. Finally, millimeter-wave radar, due to its ability to penetrate clothing and protect privacy, has been gradually introduced into the field of biomedical engineering. Existing single radar systems often have limited field of view and cannot cover the full body outline under a standard anatomical standing posture. If multiple radars are simply combined for distributed measurement, due to the lack of an underlying hardware-level synchronization mechanism, each radar will generate severe co-frequency interference when working at the same time, resulting in the acquisition of point cloud data containing a lot of noise, which cannot meet the needs of medical-grade precision measurement at all. In summary, there is an urgent need in the clinical and health screening fields for a system that can effectively eliminate reading errors and the risk of contact infection, address privacy and ethical issues, possess strong anti-interference capabilities, and perform non-contact precision measurement and shape modeling of large populations. Summary of the Invention

[0004] The purpose of this invention is to address the clinical pain points of traditional manual contact measurement in medical examination centers when acquiring human body circumference characteristics, such as the risk of cross-infection and subjective reading errors, as well as the privacy and ethical constraints faced by optical imaging systems. The invention proposes a non-contact millimeter-wave radar human body precision measurement and health assessment system for medical examinations.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A non-contact millimeter-wave radar precision measurement and health assessment system for medical examinations includes a medical-grade synchronous control gateway, a distributed multiple-input multiple-output (MIMO) radar array, and a medical data processing workstation. The distributed multiple-input multiple-output (MIMO) radar array contains multiple distributed millimeter-wave radar front-ends for performing full-coverage scanning of subjects in a standard anatomical standing posture to acquire raw point cloud data of the human body.

[0006] Preferably, the medical-grade synchronous control gateway is physically connected to the radar array, and the gateway has a built-in hard real-time time division multiplexing (TDM) protocol to schedule the radar array in order to avoid electromagnetic interference between multiple radars.

[0007] Preferably, the gateway is configured to preprocess the raw point cloud data to generate valid point cloud data.

[0008] Preferably, the medical data processing workstation is communicatively connected to the medical-grade synchronous control gateway to receive valid point cloud data. The workstation generates a 3D model of the subject through imaging and reconstruction algorithms, and automatically calculates key health indicators such as height, waist circumference, hip circumference, and waist-to-hip ratio (WHR) using an adaptive slicing algorithm.

[0009] Preferably, the medical-grade synchronous control gateway is constructed from an Arduino embedded unit; The scheduling mechanism of the hard real-time time division multiplexing (TDM) protocol is as follows: the Arduino connects to the SYNC_IN pin of each radar through general purpose input / output (GPIO) physical connections, strictly controls the transmission window of each radar, and eliminates co-channel interference between multiple radars at the physical level.

[0010] Preferably, the system adopts a heterogeneous computing task hierarchical architecture; The embedded terminal is responsible for data cleaning, coordinate system ID alignment, and packet loss detection. The embedded terminal implements a point cloud validity pre-screening algorithm to remove obvious background noise before data upload, thereby reducing transmission bandwidth pressure.

[0011] Preferably, the medical data processing workstation serves as the host computer, responsible for performing high-performance back projection imaging and Alpha Shape surface reconstruction.

[0012] Preferably, the adaptive slicing algorithm combines slicing logic verified by MATLAB simulation to realize the calculation from sparse point cloud to precise gravitational data.

[0013] Preferably, the adaptive slicing algorithm employs different adaptive fitting strategies when performing physical point cloud human body numerical calculations: For waist circumference calculation, a waist-induction fitting strategy is used to extract accurate waist circumference data; For hip circumference calculation, a hip-outward expansion fitting strategy is used to extract accurate hip circumference data.

[0014] Preferably, the distributed multiple-input multiple-output (MIMO) radar array is specifically a multi-source radar cooperative scanning layout, including at least three radar nodes that surround the subject at a specific angle.

[0015] Compared with the prior art, the present invention has the following advantages: 1. This invention effectively eliminates the reading errors and contact infection risks of traditional manual measurement by adopting a fully electronic, non-contact millimeter-wave radar scanning mode. At the same time, it avoids the privacy leakage problems caused by optical structured light imaging, greatly improving the clinical compliance of examinees. Unlike traditional single radar or disordered stacked multi-radar systems, this system's innovative hardware-level time-division multiplexing (TDM) triggering mechanism ensures the purity of signals during multi-source radar collaborative scanning from the physical base, providing high-quality original sparse point clouds for subsequent high-precision imaging.

[0016] 2. This invention utilizes a unique heterogeneous computing task hierarchical mode and an edge-side (embedded) point cloud validity pre-screening algorithm to significantly reduce useless background noise, greatly alleviating communication bandwidth pressure during data upload and ensuring the overall system's hard real-time performance and high concurrency processing capabilities. Combined with a customized adaptive slice fitting algorithm (such as waist retraction and hip expansion), the system can accurately extract anatomical circumferences from sparse point clouds. The measured values ​​of this system are highly consistent with the true values, with waist circumference error controlled at +5.0% and hip circumference error as low as -0.9%, fully meeting the precise clinical requirements for rapid health screening and chronic disease sign monitoring in large-scale populations. Attached Figure Description

[0017] Figure 1 This is a flowchart of the heterogeneous computation process of the non-contact millimeter-wave radar human body precision measurement and health assessment system for medical examination proposed in this invention. Figure 2 This is a 3D topology diagram of the non-contact millimeter-wave radar human body precision measurement and health assessment system for medical examinations proposed in this invention. Figure 3This is a hardware connection logic diagram of the non-contact millimeter-wave radar human body precision measurement and health assessment system for medical examination proposed in this invention. Figure 4 This is a 3D model of the human body generated by the non-contact millimeter-wave radar human body precision measurement and health assessment system for medical physical examination proposed in this invention. Figure 5 This is a schematic diagram of multi-source radar collaborative scanning for the non-contact millimeter-wave radar human body precision measurement and health assessment system for medical examinations proposed in this invention. Figure 6 This is a physical point cloud human body numerical calculation diagram of the non-contact millimeter-wave radar human body precision measurement and health assessment system for medical physical examination proposed in this invention. Detailed Implementation

[0018] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0019] Reference Figures 1-6 A non-contact millimeter-wave radar precision measurement and health assessment system for medical checkups, including a medical-grade synchronous control gateway, a distributed multiple-input multiple-output (MIMO) radar array, and a medical data processing workstation; A distributed multiple-input multiple-output (MIMO) radar array, comprising multiple distributed millimeter-wave radar front-ends, is used to perform full-coverage scanning of subjects in a standard anatomical standing posture, acquiring raw human point cloud data. Serving as the system's front-end perception and execution layer, it contains multiple distributed millimeter-wave radar front-ends. Its main function is to perform non-contact full-coverage scanning of subjects in a standard anatomical standing posture, acquiring three-dimensional spatial reflection signals from the human body surface, thereby generating raw sparse point cloud data. This module effectively avoids the privacy and ethical issues associated with optical cameras and completely eliminates the infection risk associated with manual contact measurements.

[0020] The medical-grade synchronous control gateway is physically connected to the radar array. The gateway has a built-in hard real-time time division multiplexing (TDM) protocol to schedule the radar array to avoid electromagnetic interference between multiple radars. As the core control and edge computing hub of the system, it plays a key role in connecting the upper and lower layers. On the one hand, the embedded control gateway schedules the radar array through the hard real-time time division multiplexing (TDM) protocol to perform full-coverage scanning of the subject in the standard anatomical standing posture while avoiding electromagnetic interference. On the other hand, as the front end of the heterogeneous computing architecture, the embedded end (Arduino) is responsible for data cleaning, coordinate system ID alignment and packet loss detection, and performs pre-screening before data is uploaded to reduce the pressure on transmission bandwidth.

[0021] The gateway is configured to preprocess the raw point cloud data to generate valid point cloud data. This setting filters out irrelevant background noise such as walls, ground, and equipment in the environment, and only retains the valid point cloud data belonging to the subject's body for output. This reduces the load on subsequent transmission and calculation, and also avoids interference from irrelevant noise on reconstruction accuracy.

[0022] The medical data processing workstation, communicating with a medical-grade synchronous control gateway, receives valid point cloud data. The workstation generates a 3D model of the subject using imaging and reconstruction algorithms, and automatically calculates key health indicators such as height, waist circumference, hip circumference, and waist-to-hip ratio (WHR) using an adaptive slicing algorithm. Serving as the system's backend computing power and decision analysis layer, it is responsible for receiving the pre-screened valid point cloud data. The host computer handles high-performance back projection imaging and AlphaShape surface reconstruction to generate the subject's 3D model. Finally, based on medical anthropometry standards, the host computer automatically calculates key health indicators such as height, waist circumference, hip circumference, and waist-to-hip ratio (WHR) using an adaptive slicing algorithm.

[0023] The medical-grade synchronous control gateway is built using an Arduino embedded unit; The scheduling mechanism of the hard real-time time division multiplexing (TDM) protocol is as follows: Arduino connects to the SYNC_IN pin of each radar through general purpose input / output (GPIO) physical connections, strictly controlling the transmission window of each radar, eliminating co-channel interference between multiple radars at the physical level. With this setting, it is ensured that only one radar node is in the transmission working state at the same time, while the other nodes remain silent. This avoids signal crosstalk caused by multiple radars working at the same frequency from the physical source, and can output raw point cloud data with higher signal-to-noise ratio and less noise, providing a reliable data foundation for subsequent medical-grade precision 3D reconstruction.

[0024] The system adopts a heterogeneous computing task hierarchical architecture; The embedded system is responsible for data cleaning, coordinate system ID alignment, and packet loss detection. It implements a point cloud validity pre-screening algorithm to remove obvious background noise before data upload, thereby reducing transmission bandwidth pressure. By doing so, a large number of invalid background noise that would otherwise need to be processed on the host computer are filtered out in advance, and only valid human point cloud data is uploaded, reducing the amount of data transmission and alleviating communication bandwidth pressure.

[0025] The medical data processing workstation, acting as the host computer, is responsible for performing high-computational-powered back projection imaging and AlphaShape surface reconstruction. This setup leverages the superior computing power of the host computer, concentrating resources on processing imaging and reconstruction tasks that demand high computational capabilities. This ensures the overall operational efficiency of the system and is well-suited to the high-concurrency screening needs of the health checkup center.

[0026] The adaptive slicing algorithm combines slicing logic verified by MATLAB simulation to realize the calculation of precise circumference data from sparse point clouds. By setting it up in this way, the calculation strategy is optimized in combination with the characteristics of human anatomy, avoiding the systematic errors of general algorithms in human circumference calculation, and ensuring that the calculation accuracy of key health indicators meets the requirements of clinical measurement.

[0027] The adaptive slicing algorithm employs different adaptive fitting strategies when performing physical point cloud human body numerical calculations: For waist circumference calculation, a waist-induction fitting strategy is used to extract accurate waist circumference data; For hip circumference calculation, a hip-outward expansion fitting strategy is adopted to extract accurate hip circumference data. By setting it up in this way, the different physiological contour characteristics of waist circumference and hip circumference are matched, the systematic bias caused by the general fitting strategy is corrected, and the measurement results are made to better fit the standards of clinical anatomical measurement. This further improves the calculation accuracy of key health indicators and meets the requirements of medical-grade measurement.

[0028] The distributed multiple-input multiple-output (MIMO) radar array is specifically a multi-source radar collaborative scanning layout, which includes at least three radar nodes that surround the subject at a specific angle. With this setup, blind-spot-free coverage scanning of the entire human body contour under a standard anatomical standing posture can be achieved, avoiding the loss of contour data caused by insufficient field of view of a single radar, ensuring complete coverage of the original point cloud data, and providing complete basic data support for subsequent high-precision 3D reconstruction.

[0029] Example 1: Multi-radar anti-jamming imaging mechanism based on distributed MCU scheduling Combination Figure 1 and Figure 2 As shown, this system includes an Arduino-based central control unit and multiple distributed millimeter-wave radar front-ends. To overcome the co-channel interference problem during multi-source radar cooperative scanning, this embodiment proposes a hardware-level time-division multiplexing (TDM) triggering mechanism.

[0030] Specifically, the Arduino medical-grade synchronous control gateway is directly connected to the synchronization input (SYNC_IN) pin of each distributed radar front-end via general purpose input / output (GPIO) physical connections. During system operation, the gateway's internal high-precision timer strictly controls the transmission window of each radar.

[0031] For example, at time 1, the gateway sends a high-level trigger pulse only to the SYNC_IN pin of radar front-end 1, while the other radars are in sleep or receive-only mode; at time 2, radar front-end 2 is triggered, and so on. This hard real-time scheduling based on physical connections completely eliminates co-channel interference between multiple radars at the physical level, ensuring the purity of the raw sparse point cloud subsequently acquired.

[0032] Example 2: Point Cloud Pre-screening and 3D Reconstruction under a Heterogeneous Computing Architecture After acquiring the raw radar echo, this system innovatively adopts a heterogeneous computing task hierarchical mode, processing the computing tasks in a hierarchical manner. (Reference) Figure 3 Logical flow and Figure 4 Reconstruction results: First, the embedded Arduino, acting as an edge computing node, is responsible not only for low-level communication scheduling but also for data cleaning, coordinate system ID alignment, and packet loss detection. Most importantly, the embedded Arduino implements a lightweight "point cloud validity pre-screening algorithm." Because the medical examination environment (such as walls and medical carts) generates a large amount of static reflection, this algorithm uses a set distance threshold and Doppler threshold to remove obvious background noise in real time before data upload. This design significantly reduces the bandwidth pressure on data transmission to the medical data processing workstation.

[0033] Subsequently, the pre-screened high-purity valid point cloud data is transmitted to the host computer via a high-speed communication interface (such as USB or LAN). Leveraging its powerful computing capabilities, the host computer is responsible for high-performance back projection imaging and the AlphaShape surface reconstruction algorithm, ultimately generating data such as... Figure 4 The 3D sparse point cloud model covering the entire subject is shown.

[0034] Example 3: Numerical Calculation of Human Body from Physical Point Cloud Based on Adaptive Slicing Algorithm Based on the aforementioned hardware architecture, after completing the 3D shape modeling, the system uses a slicing algorithm verified by MATLAB simulation to automatically calculate the girth data from sparse point clouds.

[0035] Specific combination Figure 5 and Figure 6 As shown, due to the non-uniform distribution of physical point clouds on the human body surface, the system employs a highly customized adaptive fitting algorithm. The host computer first adaptively layers the data based on the subject's overall Z-axis height to accurately locate the feature slice planes of the waist and hips. When performing two-dimensional plane fitting, the algorithm introduces an asymmetric convergence strategy: For waist circumference calculation, a waist-induction fitting strategy is adopted (e.g.) Figure 6 (As shown by the red line in the image), to eliminate external outliers caused by drooping arms or clothing wrinkles.

[0036] For hip circumference calculation, a hip-outward expansion fitting strategy is adopted (e.g. Figure 6 (As shown by the blue line in the image), to ensure complete envelopment of the point cloud profile at the maximum bulge of the buttocks.

[0037] System validation and its value in downstream clinical applications; Through the aforementioned hardware and software collaborative system, subjects can automatically generate high-precision health measurement indicators while maintaining a standard anatomical standing posture. Data verification from the final acceptance report shows that, in a non-contact state, the system's waist circumference measurement error is controlled within +5.0%, and the hip circumference measurement error is extremely low at -0.9%. Based on this high-precision circumference calculation, data such as the waist-to-hip ratio (WHR) have extremely high clinical reliability.

[0038] The precise waist-to-hip ratio (WHR) and body shape data of subjects acquired seamlessly and concurrently by this system can be directly used as high-quality structured medical feature vectors and seamlessly input into deep learning models such as TabNet or FT-Transformer. This provides crucial and highly accurate underlying physiological data support for clinical medical prediction of chronic syndromes such as metabolic-associated fatty liver disease (MAFLD) and for in-depth research on the correlation between lower limb muscle mass and metabolic characteristics, greatly expanding the industrial applicability and clinical research value of this digital medical device.

[0039] The functional principle of this invention can be explained through the following operational methods: The system adopts a multi-radar anti-interference mechanism based on distributed MCU scheduling in its hardware architecture: the Arduino central control unit connects to the synchronization input (SYNC_IN) pin of each distributed millimeter-wave radar front end through general purpose input / output (GPIO) physical connection to build a hardware-level time division multiplexing (TDM) triggering mechanism. By strictly scheduling the transmission window of each radar, the co-channel interference during multi-radar cooperative scanning is eliminated from the physical level. The system adopts a heterogeneous computing hierarchical architecture for data processing: the embedded terminal (Arduino) serves as the front-end node, responsible for executing a lightweight point cloud validity pre-screening algorithm, performing data cleaning, coordinate system ID alignment, and packet loss detection, and removing obvious background noise before the data is uploaded to the host computer; the medical data processing workstation serves as the host computer, receiving the pre-screened valid point cloud and executing high-computing-power back projection imaging and AlphaShape surface reconstruction algorithms to generate a 3D model of the examinee.

[0040] In terms of shape modeling and measurement, the system incorporates a non-contact automatic human body circumference measurement algorithm: based on the reconstructed 3D human body model, an adaptive slicing algorithm is used to perform targeted fitting on the extracted physical point cloud. Specifically, a waist-inward fitting strategy is used for the waist circumference feature plane, and a hip-outward fitting strategy is used for the hip circumference feature plane, thereby automatically calculating the subject's precise waist and hip circumference, and solving key health indicators such as waist-hip ratio (WHR).

[0041] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A non-contact millimeter-wave radar precision human body measurement and health assessment system for medical examinations, comprising a medical-grade synchronous control gateway, a distributed multiple-input multiple-output (MIMO) radar array, and a medical data processing workstation, characterized in that: The distributed multiple-input multiple-output (MIMO) radar array contains multiple distributed millimeter-wave radar front-ends, used to perform full-coverage scanning of subjects in a standard anatomical standing posture to acquire raw human point cloud data.

2. The non-contact millimeter-wave radar human body precision measurement and health assessment system for medical physical examinations according to claim 1, characterized in that, The medical-grade synchronization control gateway is physically connected to the radar array. The gateway has a built-in hard real-time time division multiplexing (TDM) protocol to schedule the radar array in order to avoid electromagnetic interference between multiple radars.

3. The non-contact millimeter-wave radar human body precision measurement and health assessment system for medical physical examinations according to claim 2, characterized in that, The gateway is configured to preprocess the raw point cloud data to generate valid point cloud data.

4. The non-contact millimeter-wave radar human body precision measurement and health assessment system for medical physical examinations according to claim 3, characterized in that, The medical data processing workstation is communicatively connected to the medical-grade synchronous control gateway and is used to receive effective point cloud data. The workstation generates a 3D model of the subject through imaging and reconstruction algorithms and automatically calculates key health indicators such as height, waist circumference, hip circumference and waist-to-hip ratio (WHR) using an adaptive slicing algorithm.

5. The non-contact millimeter-wave radar human body precision measurement and health assessment system for medical physical examinations according to claim 1, characterized in that, The medical-grade synchronous control gateway is constructed from an Arduino embedded unit; The scheduling mechanism of the hard real-time time division multiplexing (TDM) protocol is as follows: the Arduino connects to the SYNC_IN pin of each radar through general purpose input / output (GPIO) physical connections, strictly controls the transmission window of each radar, and eliminates co-channel interference between multiple radars at the physical level.

6. The non-contact millimeter-wave radar human body precision measurement and health assessment system for medical physical examinations according to claim 1, characterized in that, The system adopts a heterogeneous computing task hierarchical architecture; The embedded terminal is responsible for data cleaning, coordinate system ID alignment, and packet loss detection. The embedded terminal implements a point cloud validity pre-screening algorithm to remove obvious background noise before data upload, thereby reducing transmission bandwidth pressure.

7. The non-contact millimeter-wave radar human body precision measurement and health assessment system for medical physical examinations according to claim 1, characterized in that, The medical data processing workstation, acting as the host computer, is responsible for performing high-performance back projection imaging and Alpha Shape surface reconstruction.

8. The non-contact millimeter-wave radar human body precision measurement and health assessment system for medical physical examinations according to claim 1, characterized in that, The adaptive slicing algorithm combines slicing logic verified by MATLAB simulation to realize the calculation of sparse point cloud to precise gravitational data.

9. The non-contact millimeter-wave radar human body precision measurement and health assessment system for medical physical examinations according to claim 1, characterized in that, The adaptive slicing algorithm employs different adaptive fitting strategies when performing physical point cloud human body numerical calculations: For waist circumference calculation, a waist-induction fitting strategy is used to extract accurate waist circumference data; For hip circumference calculation, a hip-outward expansion fitting strategy is used to extract accurate hip circumference data.

10. The non-contact millimeter-wave radar human body precision measurement and health assessment system for medical physical examinations according to claim 1, characterized in that, The distributed multiple-input multiple-output (MIMO) radar array is specifically a multi-source radar cooperative scanning layout, including at least three radar nodes that surround the subject at a specific angle.

Images