Health data monitoring system and method based on Internet of Things

Abstract

The invention provides a health data monitoring system and method based on the Internet of Things, and relates to the technical field of health monitoring, the health data monitoring system based on the Internet of Things comprises a data acquisition unit, a data storage management unit and a health reminding push unit; the data acquisition unit is used for acquiring hardware data monitored by the sensor and image data shot by the camera, converting the image data into eye disease analysis indexes and sending the eye disease analysis indexes to the data storage management unit; and the data storage management unit is used for respectively storing the user data, the equipment data and the health monitoring data according to a preset framework. According to the monitoring system and method provided by the invention, complex health data can be processed more accurately, a more accurate health analysis result is provided for a user, and the monitoring system and method have profound significance for scenes such as disease prediction and health risk assessment.

Description

Technical Field

[0001] The present invention relates to the technical field of health monitoring, and particularly to a health data monitoring system and method based on the Internet of Things. Background Art

[0002] Currently, health monitoring technologies mainly include technologies such as blood pressure, blood oxygen, heart rate, and fundus imaging. These technologies provide basic data support for health management.

[0003] However, the limitations of existing health monitoring technologies are that each technology usually focuses on a single health indicator, and the devices are large in size and high in cost, making it impossible to achieve all-round and multi-index integrated detection. For example, although some non-invasive technologies for blood glucose detection are under research, due to the high technical difficulty, the application is still not mature and cannot meet the daily convenience requirements. In addition, existing solutions generally lack personalized monitoring for the elderly and chronic disease patients, and cannot provide timely and effective health warning and intervention means. The lack of customization function, unable to provide specific detection plans or suggestions according to the health status of each user, greatly affects the effect and user experience of health management. Summary of the Invention

[0004] In view of the above-mentioned disadvantages of the prior art, the purpose of the present invention is to provide a health data monitoring system and method based on the Internet of Things, which is used to solve the limitations of existing health monitoring technologies that each technology usually focuses on a single health indicator, and the devices are large in size and high in cost, making it impossible to achieve all-round and multi-index integrated detection. For example, although some non-invasive technologies for blood glucose detection are under research, due to the high technical difficulty, the application is still not mature and cannot meet the daily convenience requirements. In addition, existing solutions generally lack personalized monitoring for the elderly and chronic disease patients, and cannot provide timely and effective health warning and intervention means. The lack of customization function, unable to provide specific detection plans or suggestions according to the health status of each user, greatly affects the effect and user experience of health management.

[0005] To achieve the above and other related objectives, the present invention provides a health data monitoring system based on the Internet of Things, including: a data acquisition unit, a data storage and management unit, and a health reminder push unit; the data acquisition unit is used to acquire the hardware data monitored by sensors and the image data captured by a camera, convert the image data into eye disease analysis indicators, and send them to the data storage and management unit; the data storage and management unit is used to store user data, device data, and health monitoring data separately according to a predetermined architecture, and send the health monitoring data to the health reminder push unit; wherein, the health monitoring data includes hardware data and eye disease analysis indicators; the health reminder push unit is used to send health reminders and warning messages to users according to the health monitoring data and preset health management strategies.

[0006] In an embodiment of the present invention, the data acquisition unit includes: a sensing signal module, which is used to acquire the hardware data monitored by different types of sensors through encapsulated different sensor interfaces, wherein the encapsulated different sensor interfaces are obtained by dynamically creating different types of sensing signal modules; an image acquisition module, which is used to acquire the image data captured by the camera; an image processing module, which is used to extract features from the image data to obtain a fundus image; and an image recognition module, which is used to analyze the possibility of eye diseases for the fundus image to obtain eye disease analysis indicators, and send them to the data storage and management unit.

[0007] In an embodiment of the present invention, the image processing module includes: a preprocessing module, which is used to preprocess the image data through the Canny edge detection algorithm and the image grayscale processing method to obtain high-quality image data; a model recognition module, which is used to classify and recognize the high-quality image data through a deep convolutional neural network model trained by the TensorFlow framework to obtain different types of classified image data; a feature extraction module, which is used to calculate the histogram of oriented gradients of local regions for different types of classified image data through the HOG algorithm to obtain HOG feature samples of the shape and structure information corresponding to different types of classified image data; and an object detection module, which is used to search for and segment the best segmentation hyperplane for the HOG feature samples of the shape and structure information corresponding to different types of classified image data through a support vector machine classifier to obtain a fundus image.

[0008] In one embodiment of the present invention, the preprocessing module includes: an edge detection module, which is used to smooth and remove noise from an image by using a Gaussian filtering formula through the Canny edge detection algorithm to obtain a filtered image, and calculate the gradient of the filtered image through a Sobel operator to obtain the edge of the filtered image; wherein, the magnitude of the gradient is determined by calculating the changes of each pixel in the horizontal and vertical directions; and an exposure correction module, which is used to adjust the color temperature of the filtered image through an automatic white balance algorithm and optimize the brightness and contrast of the filtered image under corresponding lighting conditions through an exposure correction algorithm to obtain high-quality image data.

[0009] In one embodiment of the present invention, the model recognition module includes a model construction module for constructing a deep convolutional neural network model; the model construction module includes: a dataset construction module, which is used to expand the training dataset according to data augmentation techniques, wherein the data augmentation techniques include at least one of an image rotation method, a scaling method, and a cropping method; a model training module, which is used to adopt a transfer learning method to obtain features from a pre-trained image classification network for the training dataset, so as to train and obtain an initial model through the TensorFlow framework; and an optimization module, which is used to quantize and prune the initial model to obtain a deep convolutional neural network model.

[0010] In one embodiment of the present invention, the image recognition module includes: an image analysis module, which is used to analyze the possibility of eye diseases for fundus images through OpenCV image processing algorithms to obtain eye disease analysis indicators; and a data transmission module, which is used to send the eye disease analysis indicators to the data storage and management unit.

[0011] In one embodiment of the present invention, the health reminder push unit includes: a data analysis module, which is used to analyze health monitoring data to obtain the real-time change situation of the health monitoring data; and a reminder sending module, which is used to generate health reminders and warning messages according to the real-time change data and the user's health goals by using a health trend prediction model trained and constructed through historical health data, so as to send health reminders and warning messages to the user.

[0012] In an embodiment of the present invention, the health data monitoring system further includes: a user interface display module, which separates the interface, application logic, and data through the MVC design pattern, where the interface part is responsible for user interaction display, the controller part is responsible for capturing user operations and performing corresponding processing, and the model part is responsible for data storage and management; a communication interface management module, which establishes real-time data interaction between the data acquisition unit, the data storage management unit, the health reminder push unit, and the user interface display module through multiple communication protocols, where the communication protocols include at least one of the MQTT protocol and the HTTP protocol; a power supply module, which provides stable power support for the health data monitoring system; a motor drive module, which is equipped with a dual-motor drive chip, and each chip is correspondingly connected to two DC motors to drive and control the speed and steering of each motor through the PWM signal of each chip; and a voice recognition module, which is used for audio data acquisition, processing, and output control, and the voice recognition module includes a microphone circuit and an audio amplification circuit for audio signal processing, and an input / output interface for microphone and speaker access.

[0013] In an embodiment of the present invention, the sensing signal module includes: a wireless communication module, through which the sensing signal module is communicatively connected to the sensor; a crystal oscillator circuit, which provides a stable system clock signal for the sensing signal module; a Type-C interface module, which establishes USB communication functions; a power management circuit, which provides stable power for the sensing signal module; a microcontroller, which includes multiple pins for serial communication of data, and the pins include GPIO pins, serial communication interface pins, and USB pins; a decoupling capacitor circuit, which provides a stable voltage for the sensing signal module; a reset button circuit, which is used for starting control of the health data monitoring system; and a BOOT0 button module, which is used for starting mode control of the health data monitoring system.

[0014] To achieve the above and other related objectives, the present invention also provides a method for monitoring health data in the Internet of Things, including the following steps: obtaining hardware data monitored by the sensor and image data captured by the camera through the data acquisition unit, converting the image data into eye disease analysis indicators, and sending them to the data storage management unit; storing user data, device data, and health monitoring data separately by the data storage management unit according to a predetermined architecture, and sending the health monitoring data to the health reminder push unit, where the health monitoring data includes hardware data and eye disease analysis indicators; and sending health reminders and warning messages to the user by the health reminder push unit according to the health monitoring data and a preset health management strategy.

[0015] As described above, a health data monitoring system and method based on the Internet of Things according to the present invention have the following beneficial effects: By integrating multiple health monitoring technologies, the limitation of single-index detection is broken, and multiple health data such as blood glucose, blood pressure, heart rate, and fundus images can be monitored simultaneously. The synchronous monitoring of multiple health data can not only provide more comprehensive reference data for users, but also generate daily reports, reducing the resistance for doctors to conduct comprehensive diagnosis and health risk assessment. The present invention adopts a miniaturized and low-cost design of the device, making it suitable for grass-roots medical environments, promoting the development of non-invasive blood glucose detection technology, and combining intelligent interaction functions to improve the user experience and health management efficiency. Grass-roots medical institutions usually face problems of resource and equipment shortages, and high-end multifunctional medical devices are often difficult to popularize due to their high cost and complexity. The design of the present invention enables the device to have a smaller volume and lower manufacturing cost, and can be widely applied in grass-roots medical institutions, community and family health management. By reducing the device cost, the burden on grass-roots medical institutions can be greatly reduced, and at the same time, the accessibility and quality of grass-roots medical services can be improved. Especially in resource-poor areas, the health monitoring coverage rate can be greatly increased. By utilizing intelligent interaction functions, through user interface design and intelligent health reminder push through machine learning, the user experience is improved. Users can view health data in real time through terminal devices such as mobile phones and tablets, receive personalized health advice, and achieve more comprehensive health management. The intelligent interaction function not only makes the user more convenient and efficient in the health monitoring process, but also improves the user's sense of participation and enthusiasm, and promotes the long-term adherence to health management. Through innovation and optimization at the technical, housing, and algorithm levels, it has unique advantages in the synchronous detection of multiple health monitoring indicators, miniaturization of the device, low-cost design, low-power consumption design, and intelligent interaction. It not only promotes the development of intelligent health monitoring devices, but also provides a more practical technical solution for grass-roots medical services and family health management. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a structural block diagram of a health data monitoring system based on the Internet of Things provided by an embodiment of the present invention.

[0017] Figure 2 It shows a schematic diagram of the data transmission process provided by an embodiment of the present invention.

[0018] Figure 3 It shows a schematic connection circuit diagram of a heart rate and blood oxygen sensor and a Cortex-M single-chip microcomputer provided by an embodiment of the present invention

[0019] Figure 4 It shows a schematic diagram of a crystal oscillator circuit provided by an embodiment of the present invention.

[0020] Figure 5Shown is a schematic diagram of the BOOTO button circuit provided by an embodiment of the present invention.

[0021] Figure 6 Shown is a schematic diagram of the Type-C interface circuit provided by an embodiment of the present invention.

[0022] Figure 7 Shown is a schematic diagram of the wireless communication module circuit provided by an embodiment of the present invention.

[0023] Figure 8 Shown is a schematic diagram of the programming & USART interface circuit provided by an embodiment of the present invention.

[0024] Figure 9 Shown is a schematic diagram of the STM32F103C8T6 circuit provided by an embodiment of the present invention.

[0025] Figure 10 Shown is a schematic diagram of the power supply circuit provided by an embodiment of the present invention.

[0026] Figure 11 Shown is a schematic diagram of the reset button circuit provided by an embodiment of the present invention.

[0027] Figure 12 Shown is a schematic diagram of the decoupling capacitor circuit provided by an embodiment of the present invention.

[0028] Figure 13 Shown is a schematic diagram of the boost-buck circuit of the power supply module provided by an embodiment of the present invention.

[0029] Figure 14 Shown is a schematic diagram of the four-way motor drive circuit provided by an embodiment of the present invention.

[0030] Figure 15 Shown is a schematic diagram of the 5V output circuit provided by an embodiment of the present invention.

[0031] Figure 16 Shown is a schematic diagram of the power interface circuit provided by an embodiment of the present invention.

[0032] Figure 17 Shown is a schematic diagram of the motor encoder and control signal pin circuit provided by an embodiment of the present invention.

[0033] Figure 18 Shown is a schematic diagram of the motor interface circuit provided by an embodiment of the present invention.

[0034] Figures 19 to 27 Shown is a schematic diagram of the voice recognition module circuit provided by an embodiment of the present invention.

[0035] Figure 28 Shown is a schematic diagram of the flow of the health data monitoring method provided by an embodiment of the present invention.

[0036] Reference Signs:

[0037] Data acquisition unit 111; data storage and management unit 112; health reminder push unit 113. Detailed Implementation Manner

[0038] The following specific examples illustrate the implementation manners of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific implementation manners. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0039] It should be noted that the diagrams provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner. Therefore, only the components related to the present invention are shown in the diagrams, rather than being drawn according to the number, shape, and size of the components in actual implementation. The types, quantities, and proportions of the components in actual implementation can be arbitrarily changed, and the component layout type may also be more complex.

[0040] In the following description, a large number of details are explored to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0041] Please refer to Figure 1 , the present invention provides a health data monitoring system based on the Internet of Things, including: a data acquisition unit 111, a data storage and management unit 112, and a health reminder push unit 113; the data acquisition unit 111 is used to acquire the hardware data monitored by the sensor and the image data captured by the camera, convert the image data into eye disease analysis indicators, and send them to the data storage and management unit 112; the data storage and management unit 112 is used to store user data, device data, and health monitoring data separately according to a predetermined architecture, and send the health monitoring data to the health reminder push unit 113; wherein, the health monitoring data includes hardware data and eye disease analysis indicators; the health reminder push unit 113 is used to send health reminders and warning messages to the user according to the health monitoring data and a preset health management strategy.

[0042] It is not difficult to find from the above that in the process of health data monitoring based on the Internet of Things, the data acquisition unit 111 can collect the hardware data corresponding to the sensors and the image data corresponding to the camera, and can also send the image data to the eye disease analysis index and store it in the data storage and management unit 112. When the data storage and management unit 112 stores the health monitoring data including the hardware data and the eye disease analysis index, it will also send the health monitoring data to the health reminder push unit 113. Thus, the health reminder push unit 113 can send health reminders and warning messages to the user according to the health monitoring data and the preset health management strategy. In the above way, integrating multiple health monitoring technologies can break through the limitations of single-index detection, and can simultaneously monitor multiple health data such as blood sugar, blood pressure, heart rate, and fundus images. The synchronous monitoring of multiple health data can not only provide more comprehensive reference data for users, but also generate daily reports, reducing the resistance for doctors to conduct comprehensive diagnosis and health risk assessment.

[0043] In an embodiment, the data acquisition unit 111 may further include: a sensing signal module for obtaining hardware data monitored by different types of sensors through different encapsulated sensor interfaces, where the different encapsulated sensor interfaces are obtained by dynamically creating different types of sensing signal modules; an image acquisition module for obtaining image data captured by a camera; an image processing module for extracting features from the image data to obtain a fundus image; and an image recognition module for analyzing the possibility of eye diseases in the fundus image to obtain an eye disease analysis index and sending it to the data storage and management unit 112.

[0044] In this embodiment, the sensing signal module corresponding to the data acquisition unit 111 may adopt a low-power Cortex-M architecture single-chip microcomputer to solve the deficiencies of the Cortex-A architecture single-chip microcomputer in terms of ADC and DAC accuracy. Through precise sensor acquisition and real-time data transmission, the data acquisition accuracy and real-time performance of the system are ensured. The image acquisition module can collect the image data captured by the camera, and the image processing module can extract the fundus image from the image data, and the image recognition module can analyze the possibility of eye diseases in the fundus image, so as to obtain an eye disease analysis index; thus, the possibility of eye diseases can be effectively analyzed.

[0045] The sensors may include a heart rate and blood oxygen sensor, a respiratory system sensor, and a blood glucose detection sensor. Among them, Figure 3 shows the connection method between the heart rate and blood oxygen sensor and the Cortex-M single-chip microcomputer. The Stm32 single-chip microcomputer (that is, the Cortex-M single-chip microcomputer) does not have a built-in wireless communication system, so the esp8266-01s module is used for the wireless communication system.

[0046] Among them, the image recognition module can adopt a single-board computer based on the ARM architecture with a computing power of 5-8 TOPS and running on the Linux operating system, equipped with a high-definition camera of the IMX series or OV series. By combining the images captured by the camera with the OpenCV image processing algorithm, it can analyze the possibility of eye diseases and upload the images and corresponding analysis indicators to the server. The design of this image recognition module is as shown in Figure 2 shown, which shows the connection relationship between the camera and the processing unit, as well as the way to achieve real-time analysis and upload of images through data transmission. It conducts real-time data interaction with the server in the cloud through high-speed data transmission to provide reliable disease diagnosis support.

[0047] In one embodiment, the image processing module includes: a preprocessing module for preprocessing image data through the Canny edge detection algorithm and image grayscale processing method to obtain high-quality image data; a model recognition module for classifying and recognizing the high-quality image data through a deep convolutional neural network model trained by the TensorFlow framework to obtain different types of classified image data; a feature extraction module for calculating the histogram of oriented gradients in the local area of different types of classified image data through the HOG algorithm to obtain HOG feature samples of the shape and structure information corresponding to different types of classified image data; and an object detection module for finding and segmenting the best segmentation hyperplane for the HOG feature samples of the shape and structure information corresponding to different types of classified image data through a support vector machine classifier to obtain fundus images.

[0048] During the process of image processing by the image processing module, the main function of the image data processing module is to process the raw image data collected from the camera on the Cortex-A architecture single-board computer to generate data that can be used for subsequent decision-making analysis. Through this module, a series of processing steps can be implemented on the raw image data, including image preprocessing, feature extraction, and pattern recognition, etc., and finally it is transformed into information with practical value.

[0049] In the image processing module, through the preprocessing module, the Canny edge detection algorithm and image grayscale processing can be used for the preliminary processing of images. Through the model recognition module, a deep convolutional neural network model trained using the TensorFlow framework can be used to classify and recognize high-quality image data to obtain different types of classified image data. Through the feature extraction module, the histogram of oriented gradients (HOG) algorithm can be used to calculate the histogram of oriented gradients in the local area of different types of classified image data, obtaining HOG feature samples of the shape and structure information corresponding to different types of classified image data; then through the target detection module, a support vector machine classifier is used to search for and segment the optimal separation hyperplane for the HOG feature samples of the shape and structure information corresponding to different types of classified image data to obtain fundus images.

[0050] Specifically, in the feature extraction stage of the feature extraction module, the HOG algorithm is introduced to calculate the histogram of oriented gradients in the local area of different types of classified image data. Through this algorithm, the histogram of oriented gradients in the local area of the image can be calculated to describe the features of the image. Specifically, for each small area, the HOG algorithm calculates the gradient of each pixel and constructs an orientation histogram according to the gradient direction to describe the edge features of the area. The features of the entire image are obtained by stitching the histograms of each small area into a large feature vector. This process can effectively capture the shape and structure information in the image and is particularly suitable for target detection tasks. Then, in the target detection stage of the target detection module, a support vector machine (SVM) is used as the classifier. The goal of this support vector machine (SVM) is to find an optimal separation hyperplane to separate samples of different classes, ensuring the accuracy of classification. Through training, the support vector machine (SVM) learns how to classify images based on HOG features and can achieve efficient target detection in many practical applications.

[0051] Moreover, in order to improve the efficiency of the deep learning model, especially its operation on embedded devices, the model can be quantized. The core idea of quantization is to convert floating-point numbers into integer form, thereby reducing the computational resource consumption. In this way, the model is not only more computationally efficient but also can run faster on embedded platforms with limited memory. The quantized model can perform inference more quickly, support real-time processing, and improve the overall system response speed.

[0052] To make full use of hardware resources, it is also possible to optimize the memory management of a single-board computer with the Cortex-A architecture. Through reasonable memory pool design and data flow control, frequent memory allocation and release are avoided, the memory fragmentation problem is reduced, and the efficient and stable operation of the system is ensured. At the same time, the image processing operations of OpenCV can work in cooperation with the NPU in the Cortex-A board. By means of asynchronous programming, the computing efficiency is optimized. During the inference process, the preprocessing and feature extraction operations of the image can be carried out in parallel, effectively reducing the overall processing latency and improving the real-time performance of image data processing.

[0053] In one embodiment, the preprocessing module includes: an edge detection module for smoothing and removing noise from an image using a Gaussian filtering formula through the Canny edge detection algorithm to obtain a filtered image, and calculating the gradient of the filtered image through a Sobel operator to obtain the edges of the filtered image; wherein the magnitude of the gradient is determined by calculating the changes of each pixel in the horizontal and vertical directions; and an exposure correction module for adjusting the color temperature of the filtered image through an automatic white balance algorithm and optimizing the brightness and contrast of the filtered image under corresponding lighting conditions through an exposure correction algorithm to obtain high-quality image data.

[0054] When the preprocessing module is processing, it uses the Canny edge detection algorithm through the edge detection module for processing. The first step of the Canny edge detection algorithm is to use a Gaussian filtering formula to smooth the image and remove noise. Then, the Sobel operator is used to calculate the gradient of the image to identify the edges. The magnitude of the gradient is determined by calculating the changes of each pixel in the horizontal and vertical directions, and the greater the intensity of the change, the more obvious the edge. Next, the image is thinned by non-maximum suppression to ensure that only the most significant edges are retained. Finally, double-threshold processing is used to connect the edges, making the edge detection more accurate and coherent. Through the automatic white balance algorithm of the exposure correction module, the color temperature of the filtered image is adjusted, and through the exposure correction algorithm, the brightness and contrast of the filtered image are optimized under the corresponding lighting conditions to obtain high-quality image data. That is, for image processing, light change is an important factor affecting image quality. To solve the problems of uneven lighting or too dark / too strong ambient light, an automatic white balance (AWB) and an exposure correction algorithm can be added in the image preprocessing stage. Automatic white balance can adjust the color temperature of the image and eliminate the influence of different lighting environments on the image color, while exposure correction can optimize the brightness and contrast of the image under different lighting conditions, making the image clearer. Especially in indoor or low-light environments, this processing can significantly improve the image quality.

[0055] In one embodiment, the model recognition module includes a model construction module for constructing a deep convolutional neural network model; the model construction module includes: a dataset construction module for augmenting the training dataset according to data augmentation techniques, where the data augmentation techniques include at least one of an image rotation method, a scaling method, and a cropping method; a model training module for using transfer learning to obtain features from a pre-trained image classification network for the training dataset to train an initial model through the TensorFlow framework; and an optimization module for quantifying and pruning the initial model to obtain a deep convolutional neural network model.

[0056] In this embodiment, in terms of model construction, by selecting the TensorFlow framework, a deep convolutional neural network model tailored to the specific requirements of the project is designed and trained. This model is specifically designed to identify specific objects or classify image content from images captured by a camera. And to ensure the accuracy and efficiency of the model, the dataset construction module uses data augmentation techniques, including image rotation, scaling, cropping, etc., to augment the training dataset and improve the generalization ability of the model. During the training process, the model training module also uses the transfer learning method to obtain features from a pre-trained image classification network, and then fine-tunes based on the application scenario through the optimization module to accelerate the training process and improve the model accuracy.

[0057] In one embodiment, the image recognition module includes: an image analysis module for analyzing the possibility of eye diseases in fundus images through OpenCV image processing algorithms to obtain eye disease analysis indicators; and a data transmission module for sending the eye disease analysis indicators to the data storage management unit 112.

[0058] Through the image analysis module, it is possible to analyze the possibility of eye diseases by combining the images captured by the camera with the OpenCV image processing algorithms, that is, to obtain eye disease analysis indicators, and send the eye disease analysis indicators to the data storage management unit 112 through the data transmission module. Among them, OpenCV is an open-source computer vision library widely used for image processing and analysis, providing rich algorithms and tools suitable for operations such as image preprocessing and feature extraction. Through OpenCV, the original image can be preprocessed such as denoising, grayscale conversion, and edge detection to lay a foundation for subsequent analysis. In the feature extraction stage, key information in the image is extracted by extracting key points, texture features, etc., and this information will be used for tasks such as pattern recognition and object detection.

[0059] In one embodiment, the health reminder push unit 113 includes: a data analysis module for analyzing health monitoring data to obtain the real-time change situation of the health monitoring data; and a reminder sending module for generating health reminders and warning messages according to the real-time change data and the user's health goals by using a health trend prediction model constructed through training with historical health data, so as to send health reminders and warning messages to the user.

[0060] The main function of the health reminder push unit 113 is to send health reminders or warning messages to the user in real time according to the user's health data and the preset health management strategy, so as to help the user maintain a healthy lifestyle. Through the personalized health reminder and intelligent health management of the present invention, by using reasonable algorithm design and architecture design, it is ensured that the user can receive relevant health information in a timely and accurate manner.

[0061] Specifically, in order to further improve the accuracy and timeliness of the reminder, machine learning algorithms can be used to optimize the reminder mechanism. By analyzing a large amount of historical health data, the system can establish a health trend prediction model and dynamically adjust the health reminder content. Specifically, regression analysis and classification models can be used to model the user's health data. Regression analysis can help quickly identify the relationships and trend changes between the user's health indicators, especially when monitoring key indicators such as body temperature, blood pressure, and heart rate. Through regression analysis, potential health risks can be predicted in advance, such as predicting that the user's heart rate is too high or blood pressure fluctuates, and relevant health reminders can be pushed in a timely manner. The classification model is used to identify whether the user has certain health risks, such as diabetes or hypertension, and push different health reminder contents according to different risk types.

[0062] In terms of health risk prediction, time series analysis algorithms can be used to process the user's health data. By analyzing the change trend of health data over time, the system can identify the fluctuations and abnormalities of the user's health status. For example, through the ARIMA model, the system can model the user's blood glucose level, heart rate and other data and predict the future change trend. If it is predicted that a certain health indicator may exceed the safe range, the system will push a health warning in advance to help the user take measures in a timely manner.

[0063] For the intelligent improvement of this module, the natural language processing (NLP) technology is introduced to enhance the personalization of reminders and the user experience. The system uses NLP technology to automatically generate reminder content and adjusts the tone and expression of reminders according to the user's health status and past interaction records. For example, the reminder tone will vary for users of different age groups. For users in good health, the system will adopt a relatively gentle tone; while for users with potential health risks, the reminder content will be more urgent and direct. In addition, NLP technology also makes the reminder content more natural, avoiding mechanical and cold statements, and improving the user's acceptance.

[0064] In terms of the security of push reminders, encryption technology can be adopted to ensure the security of users' health data during transmission. At the same time, the content of push notifications will also be encrypted to prevent unauthorized third parties from accessing users' health data. The system also supports users to set their own reminder receiving frequencies and types to ensure that users will not be bothered by frequent reminders.

[0065] In some embodiments, the health data monitoring system further includes: a user interface display module, which separates the interface, application logic, and data through the MVC design pattern, where the interface part is responsible for user interaction display, the controller part is responsible for capturing users' operations and performing corresponding processing, and the model part is responsible for data storage and management; a communication interface management module, which establishes real-time data interaction between the data acquisition unit 111, the data storage management unit 112, the health reminder push unit 113, and the user interface display module through multiple communication protocols, where the communication protocols include at least one of the MQTT protocol and the HTTP protocol; a power supply module, which provides stable power support for the health data monitoring system; a motor drive module, which is equipped with a dual-motor drive chip, and each chip is correspondingly connected to two DC motors to drive and control the speed and steering of each motor through the PWM signal of each chip; and a voice recognition module, which is used for audio data acquisition, processing, and output control. The voice recognition module includes a microphone circuit for audio signal processing and an audio amplification circuit, and an input / output interface for microphone and speaker access.

[0066] The user interface display module can use the Qt framework to implement the GUI design. By leveraging the powerful cross-platform features of Qt, this module can run seamlessly on different operating systems. Moreover, Qt provides rich controls and layout management tools, making the interface design more concise and efficient, especially showing good performance when dealing with complex interfaces and graphics. In terms of software architecture, the C / S (client / server) architecture mode can be selected. By using this architecture mode, the front-end user interface and the back-end data processing can be separated, ensuring the high scalability and maintainability of the system. In the specific design, the MVC design pattern can be used. This pattern separates the interface, application logic, and data, achieving a highly cohesive and loosely coupled structure. The interface part (view) is responsible for user interaction display, the controller part is responsible for capturing user operations and performing corresponding processing, and the model part is responsible for data storage and management. Through this pattern, the interface and the background logic can be developed, modified, and extended independently, improving the development efficiency and the maintainability of the system. At the same time, by using the signal and slot mechanism of Qt, the coupling degree between the interface and the business logic can be further reduced, achieving more flexible event handling. The signal and slot mechanism can efficiently handle user input, interface updates, and responses to background calculations, enhancing the user interaction experience.

[0067] The communication interface management module is mainly used to implement data communication between various modules within the system and the connection to external devices, ensuring reliable data transmission and synchronization. By reasonably designing the communication protocol and interface, data interaction with sensors, single-board computers, mobile devices, and cloud servers can be achieved, guaranteeing the efficient operation of the entire system. In the implementation of the communication interface, the MQTT protocol can be adopted as the core communication protocol. MQTT is a lightweight message transmission protocol, especially suitable for communication between resource-constrained devices. It uses the publish / subscribe model, which can effectively reduce network bandwidth consumption and communication latency, and is very suitable for application in the Internet of Things environment. Through the MQTT protocol, real-time data exchange between various modules can be achieved, such as the reporting of sensor data, the pushing of health reminders, and the feedback of user behavior. In addition to the MQTT protocol, interaction with other systems can also be achieved through the HTTP API interface and the HTTP protocol. For example, the mobile application interacts with the server through the HTTP API to obtain information such as user health data and historical records. The API interface follows the stateless design principle, and each request carries complete context information to ensure the independence of requests from each other, enhancing the maintainability and scalability of the system. In terms of hardware communication, data exchange with external sensors is carried out through the serial port. The data collected by the sensor is sent to the single-board computer through the serial port and uploaded to the cloud server after data parsing. A data packet verification mechanism can be adopted during the communication process to ensure the integrity and accuracy of the data and prevent abnormal health data caused by transmission errors. The communication interface management module can use multiple communication protocols to achieve efficient and stable communication between various modules. At the same time, it also provides a secure and reliable interface for external devices and cloud services, ensuring the interconnection and interoperability of the entire system. Through this module, the system can realize functions such as real-time health data collection, push notifications, and remote control, further improving the user's health management experience.

[0068] In one embodiment, the power supply module provides stable power support for the entire system. The input voltage is 12VDC, and the rated current is 3A, which can meet the power requirements of each module. According to the requirements of different modules, the output voltages are 3.3V, 5V, and 12V respectively. The buck-boost circuit is as Figure 13As shown, the power supply module uses the MP2315 chip as the core DC-DC buck converter. The input voltage is 12V, which is filtered by the inductor L1 (4.7μH) and external capacitors C6 and C5 to output a stable 5V voltage. The input terminal (IN pin) of the MP2315 is decoupled by capacitors C1 (22μF) and C2 (0.1μF) to effectively reduce high-frequency noise and power supply ripple; the diode D1 provides input protection to prevent damage to the circuit when the power supply is reverse-connected. The feedback control of the circuit is adjusted by the resistor voltage division network of R5 (20kΩ), R6 (40.2kΩ), and R7 (7.5kΩ) to ensure that the output voltage is accurately stabilized at 5V. R3 (9.09kΩ) and R2, R1 (both 100kΩ) provide appropriate voltage distribution for the EN (enable pin) to control the state when the circuit starts. The inductor L1 and the output capacitor C6 (22μF) further filter out the pulsation of the output current to ensure the stability of the voltage on the load side, which is suitable for powering sensitive digital circuits. The SW pin of the MP2315 is connected to the capacitor C4 (0.1μF) through the resistor R4 (10Ω) to achieve current adjustment and stable operation during the buck process. At the same time, the BST pin provides the gate drive voltage for the internal high-side MOSFET through the capacitor C3 (0.1μF) to improve the conversion efficiency. The overall circuit structure is reasonable, with low power loss, stable output voltage, and input protection and noise reduction design, which is very suitable for the power supply requirements of embedded systems and motor control systems.

[0069] In one embodiment, the motor drive module uses two TB6612 motor drive chips. Each chip can drive two DC motors, so a total of four motors can be driven. The speed and steering of each motor are controlled by PWM signals. Figures 14 to 18It is a schematic diagram of the motor drive module circuit. In the design of the module, the TB6612FNG chip receives PWM signals from the main control microcontroller through PWMA, PWMB, PWMC, and PWMD to achieve precise adjustment of the motor speed; the direction control pins AIN1 / AIN2, BIN1 / BIN2, CIN1 / CIN2, and DIN1 / DIN2 are used to control the forward rotation, reverse rotation, or stop of the motor. The power supply part includes two voltages of 12V and 5V, which are filtered by C9, C10, C11, and C12 to ensure the stability of the power input and prevent voltage fluctuations from interfering with the circuit; at the same time, D1 provides reverse connection protection to prevent damage to the circuit caused by reverse connection of the power supply. In the module design, the four groups of motor output interfaces are connected to the motor through the AO1 / AO2, BO1 / BO2, CO1 / CO2, and DO1 / DO2 ports, corresponding to the output signals of Motor_A, Motor_B, Motor_C, and Motor_D respectively, to achieve independent drive of the motor. In addition, the P1 and P2 pin headers are used to connect control signals, providing the input of PWM signals and direction control signals respectively, ensuring the controllability of the motor speed and direction. The module also designs a protection circuit, including the functions of current limiting by R1, R2, R3, and R4 resistors and diode protection, to prevent the back electromotive force of the motor from impacting the chip and improve the reliability of the entire drive module. The circuit is also equipped with an LED status indicator, which limits the current through R5 to indicate the power supply status of the power supply, facilitating fault troubleshooting and circuit detection. Overall, this module realizes the precise drive of four motors through dual TB6612FNG chips, supports PWM speed regulation and direction control, and has multiple protection measures, suitable for multi-motor control application scenarios such as robots and electric vehicles.

[0070] In some embodiments, the circuit of the voice recognition module is as Figures 19 to 27As shown, the core of this module lies in the microcontroller U1 (US516FB), which is mainly responsible for the acquisition, processing, and output control of audio data. Multiple GPIO pins of U1 can be used to connect external devices, serving as both general input / output ports and supporting serial communication interfaces such as UART (TXD / RXD) and I2C (SCL / SDA). The OSC_IN and OSC_OUT pins of U1 are connected to the 12MHz crystal oscillator Y1 in the lower right corner, together with two 18pF capacitors C16 and C17, forming a stable oscillation circuit to provide an accurate clock signal for the microcontroller and ensure the stable operation of the system. The audio signal processing part includes a microphone circuit and an audio amplification circuit. The microphone input is connected through the MIC+ and MIC- pins. Two 1kΩ current-limiting resistors R1 and R2 protect the input signal, and the decoupling capacitors C11 and C12 (0.1uF) filter out high-frequency noise to ensure a pure audio signal. After being processed by the audio amplification chip U2, the signal is amplified, and the gain is set by the external resistors R5 (10kΩ) and R6 (22kΩ). Capacitors C18 (0.1uF) and C19 (0.22uF) are used for signal coupling to filter out the DC component and ensure the distortion-free transmission of the audio signal to the next stage. The amplified audio signal is output through SPEK+ and SPEK- to drive the speaker. The input / output interface is realized through connectors J1 and J2, which are used for connecting the microphone and the speaker respectively. The status indication is achieved by connecting the red LED D5 in series with R9 (4.7kΩ), which can display the normal power supply or the system operation status. In addition, ESD protection devices D2, D3, and D4 are added to the interface design to effectively prevent external static electricity from damaging the microphone, speaker, and system circuit, improving the stability and durability of the circuit. The power management part provides stable power supply for the entire system. The module uses 3.3V and 5V power supplies. Capacitors C1 - C6, C14, and C15 at the input end play a decoupling role to filter out power supply noise and ensure pure power supply. The reverse connection protection diode D1 (INS5819W) provides protection to prevent damage to the circuit when the power supply is reversely connected. In addition, multiple test points TP1 to TP8 are reserved in the design for debuggers to detect the voltage and signal status of each part, improving the efficiency of maintenance and troubleshooting. The clock circuit, audio input / output, ESD protection, and power management modules coordinate with each other to support the core functions of the voice recognition module. The microcontroller collects the input signal from the microphone, processes it through the amplification circuit, and then drives the speaker to output, while realizing the status indication and communication functions. Components such as capacitors, resistors, and diodes perform their respective functions to provide functions such as filtering, protection, and current limiting, enabling the system to operate stably and efficiently in a complex environment.

[0071] In some embodiments, the sensing signal module includes: a wireless communication module through which the sensing signal module is communicatively connected to the sensor; a crystal oscillator circuit for providing a stable system clock signal to the sensing signal module; a Type-C interface module for establishing USB communication functions; a power management circuit for providing a stable power supply to the sensing signal module; a microcontroller including multiple pins for serial communication of data, the pins including GPIO pins, serial communication interface pins, and USB pins; a decoupling capacitor circuit for providing a stable voltage to the sensing signal module; a reset button circuit for controlling the startup of the health data monitoring system; and a BOOT0 button module for controlling the startup mode of the health data monitoring system.

[0072] In some embodiments, when the heart rate and blood oxygen sensor is connected to the sensing signal module (Cortex-M single-chip microcomputer), the Stm32 single-chip microcomputer (i.e., the Cortex-M single-chip microcomputer) does not have a built-in wireless communication system. Therefore, the esp8266-01s module is used for the wireless communication system. For the stm32 minimum system board, its specific circuit includes a wireless communication module, a crystal oscillator circuit, a power management circuit, a Type-C interface module, a power management circuit, a microcontroller, a decoupling capacitor circuit, a reset button circuit, and a BOOT0 button module.

[0073] In some embodiments, as Figure 4 shown, in the crystal oscillator circuit, two crystal oscillators can be seen. One of them is X1, and X1 is an 8 MHz main clock crystal oscillator that provides a stable system clock signal to the STM32F103C8T6 single-chip microcomputer, which is crucial for the normal operation of the single-chip microcomputer. The other is X3, and X3 is a 32.768 kHz crystal oscillator. This crystal oscillator is mainly used for the RTC (real-time clock) module inside the STM32 to implement a low-power time function and is suitable for timing applications. The capacitors C1, C2 (12.5 pF) and C3, C4 (20 pF) connected in parallel with these two crystal oscillators play a role in matching and stabilizing the oscillation frequency of the crystal oscillators, ensuring the accuracy and reliability of the clock signal.

[0074] In some embodiments, as Figure 5 shown, the BOOT0 button module includes R4 (10 kΩ resistor) and SW2. The BOOT0 pin of the STM32 determines the startup mode of the single-chip microcomputer: when the BOOT0 pin is at a high level during power-on, the single-chip microcomputer starts from the system memory; otherwise, it starts from the main Flash. By pressing the SW2 button, the BOOT0 level can be changed, thus achieving different startup modes.

[0075] In some embodiments, as Figure 6As shown, the Type-C interface module circuit is used for USB communication functions. USB1 is connected to the Type-C socket, and the D+ and D- signal lines are connected to the PA12 and PA11 pins of the STM32 single-chip microcomputer. These two pins respectively correspond to the USB DP (Data Positive) and DM (Data Negative) signal lines of the STM32 to achieve USB communication functions. In addition, there are some matching resistors, R39 (1kΩ) and R40 (1.5kΩ), for the pull-up and pull-down matching of the USB data lines to ensure stable signal transmission.

[0076] In some embodiments, such as Figure 7 As shown, the wireless communication module: The P1 interface is connected to an external wireless module to provide the circuit with wireless data transmission functions. In this module, the IO0 and EN pins control the enabling function of the wireless module, while the TXD and RXD are connected to the USART pins of the STM32 to complete data sending and receiving operations.

[0077] In some embodiments, such as Figure 8 As shown, the sensing signal module also includes a programming & USART interface circuit for writing program codes or data to various types of chips.

[0078] In some embodiments, such as Figure 9 As shown, the microcontroller, with the model number STM32F103C8T6. It has multiple pin functions, including GPIO, serial communication interfaces, and USB functions. Its pins PA9 and PA10 are used for the communication of USART1, and these pins are connected to the USART interface module below to connect to external devices through CN2 to achieve serial communication of data. The resistors R8 and R9 beside this module play a role in current limiting and protection.

[0079] In some embodiments, such as Figure 10 As shown, the power management circuit uses a voltage regulator chip U3, with the model number AMS1117-3.3. AMS1117-3.3 is a 3.3V low-dropout linear voltage regulator. Its input voltage VIN is connected to 5V, and it outputs a 3.3V voltage through internal circuit regulation to provide a stable 3.3V power supply for the STM32 and other modules. Capacitors C5, C6, and C7 (all 100nF) are also connected in series in the circuit, and these capacitors are used for decoupling to filter out power supply noise and ensure the purity of the power supply output.

[0080] In some embodiments, such as Figure 12 As shown, in the decoupling capacitor circuit, it mainly includes three 100nF capacitors C5, C6, and C7. These capacitors are close to the STM32 power pins to provide a stable voltage for the single-chip microcomputer, remove high-frequency interference, and play a filtering role.

[0081] In some embodiments, such asFigure 11 As shown, the reset button circuit includes R2 (a 10kΩ pull-up resistor) and SW1 (a push-button switch). In the normal working state, R2 pulls up the NRST pin to the 3.3V level. When the SW1 button is pressed, the NRST pin is pulled low, triggering a reset of the STM32 and restarting the system.

[0082] The sensor network composed of a wireless communication module, a crystal oscillator circuit, a power management circuit, a Type-C interface module, a power management circuit, a microcontroller, a decoupling capacitor circuit, a reset button circuit, a BOOT0 button module, and a programming & USART interface circuit realizes the acquisition and processing operations of external natural signals.

[0083] Please refer to FIG. 28. The present invention also provides a method for monitoring health data of the Internet of Things, including the following steps:

[0084] Step S10: Obtain the hardware data monitored by the sensor and the image data captured by the camera through the data acquisition unit 111, convert the image data into eye disease analysis indicators, and send them to the data storage and management unit 112;

[0085] Step S20: Store the user data, device data, and health monitoring data separately according to a predetermined architecture through the data storage and management unit 112, and send the health monitoring data to the health reminder push unit 113, where the health monitoring data includes hardware data and eye disease analysis indicators;

[0086] Step S30: Send health reminders and warning messages to the user through the health reminder push unit 113 according to the health monitoring data and the preset health management strategy.

[0087] In summary, a health data monitoring system and method based on the Internet of Things disclosed by the present invention integrates multiple health monitoring technologies, breaks through the limitations of single-index detection, and can simultaneously monitor multiple health data such as blood glucose, blood pressure, heart rate, and fundus images. The synchronous monitoring of multiple health data can not only provide more comprehensive reference data for users, but also generate daily reports, reducing the resistance for doctors to conduct comprehensive diagnosis and health risk assessment. The present invention adopts a miniaturized and low-cost design of the device, making it suitable for the grass-roots medical environment, promoting the development of non-invasive blood glucose detection technology, and combining intelligent interaction functions to improve the user experience and health management efficiency. Grass-roots medical institutions usually face problems of resource and equipment shortages, and high-end multifunctional medical devices are often difficult to popularize due to their high cost and complexity. The design of the present invention enables the device to have a smaller volume and lower manufacturing cost, and can be widely used in grass-roots medical institutions, community and home health management. By reducing the device cost, the burden on grass-roots medical institutions can be greatly reduced, while the accessibility and quality of grass-roots medical services can be improved. Especially in resource-poor areas, the health monitoring coverage rate can be greatly increased. By using intelligent interaction functions, through user interface design and intelligent health reminder push through machine learning, the user experience is improved. Users can view health data in real time through terminal devices such as mobile phones and tablets, receive personalized health advice, and achieve more comprehensive health management. The intelligent interaction function not only makes the health monitoring process more convenient and efficient for users, but also improves the user's sense of participation and enthusiasm, and promotes the long-term adherence to health management. Through innovation and optimization at the technical, housing, and algorithm levels, the present invention has unique advantages in the synchronous detection of multiple health monitoring indicators, miniaturization of the device, low-cost design, low-power consumption design, and intelligent interaction, not only promoting the development of intelligent health monitoring devices, but also providing a more practical technical solution for grass-roots medical services and home health management. Therefore, the present invention effectively overcomes various disadvantages in the prior art and has high industrial utilization value.

[0088] The above embodiments are only illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Any person familiar with this technology can modify or change the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or changes made by those with ordinary knowledge in the technical field without departing from the spirit and technical idea disclosed by the present invention should still be covered by the claims of the present invention.

Claims

1. A health data monitoring system based on the Internet of Things, characterized in that: include: Data acquisition unit, data storage management unit and health reminder push unit; The data acquisition unit is used to acquire hardware data monitored by the sensor and image data captured by the camera, and convert the image data into eye disease analysis indicators, and send them to the data storage management unit; The data storage management unit is used to store user data, device data, and health monitoring data respectively according to a predetermined architecture, and send the health monitoring data to the health reminder push unit; wherein the health monitoring data includes the hardware data and the eye disease analysis index; The health reminder push unit is used to send health reminders and warning information to the user according to the health monitoring data and the preset health management strategy.

2. The health data monitoring system based on the Internet of Things according to claim 1, characterized in that: The data acquisition unit comprises: A sensor signal module is used to obtain hardware data monitored by different types of sensors through different encapsulated sensor interfaces, wherein the encapsulated different sensor interfaces are obtained by dynamically creating different types of sensor signal modules; An image acquisition module is used to obtain image data captured by a camera; An image processing module, used for extracting features from the image data to obtain fundus images; and The image recognition module is used to perform a possibility analysis of eye diseases on the fundus image to obtain the eye disease analysis index and send it to the data storage management unit.

3. The health data monitoring system based on the Internet of Things according to claim 2, characterized in that: The image processing module comprises: A preprocessing module, used to preprocess the image data by using a Canny edge detection algorithm and an image grayscale processing method to obtain high-quality image data; A model recognition module is used to classify and recognize the high-quality image data using a deep convolutional neural network model trained by the TensorFlow framework to obtain different types of classified image data; A feature extraction module, used to calculate the gradient direction histogram of the local area of ​​different types of classified image data by using the HOG algorithm to obtain HOG feature samples of shape and structure information corresponding to different types of classified image data; and The target detection module is used to search and segment the best segmentation hyperplane for HOG feature samples of shape and structure information corresponding to different types of classified image data through a vector machine classifier to obtain the fundus image.

4. The health data monitoring system based on the Internet of Things according to claim 3 is characterized in that: The preprocessing module comprises: An edge detection module, used to smooth and remove noise from an image using a Gaussian filter formula using a Canny edge detection algorithm to obtain a filtered image, and to calculate the gradient of the filtered image using a Sobel operator to obtain the edge of the filtered image; wherein the magnitude of the gradient is determined by calculating the change of each pixel in the horizontal and vertical directions; and The exposure correction module is used to adjust the color temperature of the filtered image through an automatic white balance algorithm, and optimize the brightness and contrast of the filtered image under corresponding lighting conditions through an exposure correction algorithm to obtain the high-quality image data.

5. The health data monitoring system based on the Internet of Things according to claim 3 is characterized in that: The model identification module includes a model construction module for constructing the deep convolutional neural network model; The model building module includes: A data set construction module, used to expand the training data set according to a data enhancement technique, wherein the data enhancement technique includes at least one of an image rotation method, a scaling method, and a cropping method; A model training module, used to acquire features of the training data set from a pre-trained image classification network using a transfer learning method, so as to obtain an initial model through training using a TensorFlow framework; and The optimization module is used to quantize and prune the initial model to obtain the deep convolutional neural network model.

6. The health data monitoring system based on the Internet of Things according to claim 2, characterized in that: The image recognition module comprises: An image analysis module, used to perform an eye disease possibility analysis on the fundus image using an OpenCV image processing algorithm to obtain the eye disease analysis index; and A data transmission module is used to send the eye disease analysis index to the data storage management unit.

7. The health data monitoring system based on the Internet of Things according to claim 1, characterized in that: The health reminder push unit includes: A data analysis module, used to analyze the health monitoring data to obtain real-time changes in the health monitoring data; and The reminder sending module is used to use the health trend prediction model constructed through historical health data training to generate the health reminder and the warning information according to the real-time change data and the user's health goals, so as to send the health reminder and the warning information to the user.

8. The health data monitoring system based on the Internet of Things according to claim 1, characterized in that: The health data monitoring system further includes: The user interface display module is used to separate the interface, application logic and data through the MVC design pattern, so that the interface part is responsible for user interaction display, the controller part is responsible for capturing user operations and performing corresponding processing, and the model part is responsible for data storage and management; A communication interface management module, used to establish real-time data interaction between the data acquisition unit, the data storage management unit, the health reminder push unit and the user interface display module through multiple communication protocols, wherein the communication protocol includes at least one of the MQTT protocol and the HTTP protocol; A power supply module, used to provide stable power support for the health data monitoring system; The motor drive module is equipped with dual motor drive chips, each chip is connected to two DC motors to control the speed and direction of each motor through the PWM signal drive of each chip; and The speech recognition module is used for collecting, processing and outputting audio data. The speech recognition module includes a microphone circuit and an audio amplifier circuit for processing audio signals, and an input and output interface for accessing a microphone and a speaker.

9. The health data monitoring system based on the Internet of Things according to claim 2, characterized in that: The sensor signal module comprises: A wireless communication module, wherein the sensing signal module is connected to the sensor through the wireless communication module; A crystal oscillator circuit, used to provide a stable system clock signal for the sensor signal module; Type-C interface module, used to establish USB communication function; A power management circuit is used to provide a stable power supply for the sensor signal module; A microcontroller comprising a plurality of pins for serial communication of data, wherein the pins include GPIO pins, serial communication interface pins and USB pins; A decoupling capacitor circuit, used to provide a stable voltage for the sensor signal module; A reset button circuit, used for starting control of the health data monitoring system; and The BOOT0 button module is used to control the startup mode of the health data monitoring system.

10. A health data monitoring method for the Internet of Things, characterized in that: The steps include: The data acquisition unit acquires hardware data monitored by the sensor and image data captured by the camera, converts the image data into eye disease analysis indicators, and sends the indicators to the data storage management unit; The data storage management unit stores user data, device data, and health monitoring data separately according to a predetermined architecture, and sends the health monitoring data to a health reminder push unit, wherein the health monitoring data includes the hardware data and the eye disease analysis index; The health reminder push unit sends health reminders and warning information to the user based on the health monitoring data and the preset health management strategy.

Images