An intelligent health monitoring device

Abstract

The application discloses a kind of intelligent health monitoring equipment;Including control module, the control module is electrically connected with processing module, the processing module is electrically connected with detection module and communication module, the communication module is electrically connected with cloud server and distress module;The control module is electrically connected calculation model, the calculation model includes diagnosis model and prognosis model;The distress module includes one-key distress, address report, remote video, data upload and emotion call, the present application obtains the various data information of family person by detection module, is analyzed and predicted by diagnosis model and prognosis model in calculation model, and when family person appears abnormal, it is rescued by distress module, and it is convenient to be rescued in time.

Description

Technical Field

[0001] This invention belongs to the field of home monitoring technology, specifically relating to an intelligent health monitoring device. Background Technology

[0002] Living alone is a reality many elderly people must face. The biggest concerns for society and family are safety and health. Therefore, continuous remote monitoring of the elderly using modern sensors and IoT technology is essential. Home medical devices are medical devices designed for detection, treatment, health maintenance, and rehabilitation, primarily suitable for home settings. In the era of big data, the scale of medical data has exceeded the processing capacity of traditional data processing technologies. To better unlock the value of medical data, artificial intelligence (AI) technology is needed to improve the efficiency and accuracy of data processing. The development of AI technology has also provided more possibilities for medical data analysis; however, various home monitoring systems on the market still have various problems.

[0003] For example, the multifunctional health monitoring smart device disclosed in authorization announcement number CN105997025A includes a wristband, a circuit controller, and a signal receiving module. The signal receiving module is set in the circuit controller. The wristband is equipped with sensors connected to the CPU, a blood pressure module, a blood glucose module, a heart rate module, a body temperature module, a time module, a signal transceiver module, a power supply module, and a display screen. The blood pressure module, blood glucose module, heart rate module, and body temperature module collect human body posture data, including body temperature, blood pressure, blood glucose, and heart rate values, through the sensors and transmit them to the CPU. The CPU combines the current time and body posture data and outputs signals through the signal transceiver module. The circuit controller uses the signals received by the signal receiving module and the current time to intelligently control the circuit, achieving the purpose of safe electricity use, intelligent and user-friendly electricity use, and energy saving, making life more convenient, especially for the elderly. However, it does not solve the problems of existing home monitoring, such as the inconvenience of collecting and detecting multiple data, the inconvenience of data analysis and processing, resulting in data redundancy, only obtaining simple digital information, and the inconvenience of calling for help. Therefore, we propose a smart health monitoring device. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent health monitoring device to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent health monitoring device, comprising a control module, a processing module electrically connected to the control module, a detection module and a communication module electrically connected to the processing module, the detection module being used to process and transmit data information collected by the detection module, and to communicate, transmit and download control commands and data from the control module, the detection module being used to monitor and warn of the health of people living at home, especially to monitor elderly people living at home, the communication module being used to upload and download data information, the communication module being electrically connected to a cloud server and an emergency call module, the cloud server being used to upload and store data information, or to perform calculations on data information, the emergency call module being used to call for help after the detection module detects an abnormality, or to contact the family members of people living at home, especially the elderly living at home, so that guardians can monitor in a timely manner, obtain information about family members and establish communication connections; The control module is electrically connected to the calculation model, which is used to calculate and analyze the data information collected by the detection module. The calculation model includes a diagnostic model and a prognostic model. The diagnostic model is used to analyze the physical condition of the family members based on the data information collected by the detection module, and then analyze the health of the family members by combining various data information to facilitate the diagnosis and treatment of pathological diseases. The prognostic model is used to estimate the risk of a certain outcome of the family members in a certain period of time in the future, so as to facilitate the assessment of the physical condition of the family members, identify high-risk pathological diseases of the family members, and predict the physical condition of the family members in advance by monitoring high-risk pathological diseases in real time. The prognostic model uses regression prediction, which employs MAE and RMSE. The MAE (Mean Absolute Error) is used, which better reflects the actual error between the predicted and actual values. Its calculation formula is as follows: , The RMSE, or root mean square error, is used to measure the deviation between observed and true values. It is particularly sensitive to exceptionally large or small errors in a set of predictions. Its calculation formula is as follows: ; Furthermore, in the two formulas above, m represents the amount of data. The average of the data. For the data information to be calculated; The emergency call module includes one-click SOS, address reporting, remote video, data upload, and family / guardian calls. The one-click SOS is used to call for help when a family member experiences a health problem. It includes calling a specific hospital, guardian, and property management company. When calling a specific hospital, the address reporting and data upload functions simultaneously upload the community address, floor information, and portal information. The system also transmits the family member's health monitoring data to the hospital via data upload, allowing the hospital to promptly understand the family member's condition and facilitate emergency treatment. The family / guardian call function allows the one-click SOS to call the guardian. The remote video allows guardians to conduct video calls with family members during normal times or when data is abnormal, facilitating communication and understanding of the situation.

[0006] Preferably, the data upload is also used to periodically transmit data information to the family call center, so that the guardian can be aware of the health status of the family members. In the event that the data information collected by the detection module shows certain abnormalities, but does not reach the level of a sudden pathological event, the data upload will also transmit the data information to the family call center.

[0007] Preferably, the computational model mainly includes random forest, k-nearest neighbors, support vector machine and artificial neural network algorithms; The steps for establishing the computational model are as follows: S1. Data Collection and Data Processing: The data used to build the computational model includes re-collected data or data from existing experimental, cohort studies, registry, or managed datasets. The data is then cleaned and tested, and classified according to male / female, age, and serum sodium levels. The data is divided into two categories: model calculation data and model validation data, which are used for building the model and validating the model, respectively. S2. Model building and variable selection: After the computational model is developed, variables need to be included in the computational model, and stepwise regression is used to select variables. S3. Model Performance: After the computational model is built, its performance needs to be evaluated. The accuracy of the computational model has two components: discrimination and calibration. S4. Model Validation: The computational model is validated. Validation methods include cross-validation and bootstrapping. S5. Model Usage and Updates: The detection module transmits the collected data to the computational model for calculation and analysis, and the computational model needs to be continuously evolved and dynamically updated.

[0008] Preferably, the calculation formula of the artificial neural network algorithm is as follows: It is determined by the input neuron information and the corresponding weight matrix W. X is a vector, and W is a weight matrix. X is transformed into another vector through W. The information transmitted from the hidden layer neurons of the previous time step to the neurons of the current time step has the following values: The weight matrix U determines the weight. , It is a vector, and U is a weight matrix, through which U is used to... Convert it into another vector; Mainly the neurons at the current moment The inputs from the two parts mentioned above are integrated to activate and generate the output of the hidden layer neurons at the current time step. The integration process is simply vector addition: + ; Assuming the activation function is f, the value of the neuron at the current time step is generated after activation: ; Information transmission mainly involves transferring the values ​​of the hidden layer neurons at the current time step to the output neurons at the current time step: ; The formula for moving forward is as follows: ; The entire neural network contains three weight matrices. The first matrix is ​​W, with dimensions (N, K), where N represents the vector dimension of the hidden layer neuron values ​​and K represents the vector dimension of the input neuron. The second weight matrix is ​​V, with dimensions (L, N), where P represents the vector dimension of the output layer neuron. The third is the weight matrix U, with dimensions (N, N).

[0009] Preferably, the detection module includes a blood pressure detector, a blood glucose detector, a respiration detector, a weight analyzer, a video monitor, an electrocardiogram (ECG) detector, a blood oxygen detector, a temperature detector, and a sleep detector. The blood pressure detector and blood oxygen detector are used to detect the blood pressure level and blood oxygen concentration of the family member. The blood glucose detector is used to detect the blood glucose concentration of the family member. The respiration detector and sleep detector are used to detect the breathing and sleep quality of the family member during sleep. The weight analyzer and the temperature detector are used to detect the weight and temperature of the family member. The weight analyzer also calculates the BMI based on the collected data. The ECG detector is used to collect the electrocardiogram of the family member. The video monitor is used to detect the family member for safety and prevent falls.

[0010] Preferably, the respiration detector uses millimeter-wave radar for detection, and the calculation formula for millimeter-wave radar is as follows: ; in, The amplitude of the heartbeat signal. The frequency of the heartbeat signal. Let z be the phase and z be the noise. The VMD algorithm decomposes a signal into a set of IMF components. The optimal number of IMFs is determined by the cross-correlation coefficient between the residual and the superimposed IMF signal. A smaller cross-correlation coefficient indicates a better value. Let the residual after VMD decomposition be mn, and the reconstructed signal from all IMFs be fn. Then the normalized cross-correlation function is: , in, Let be the cross-correlation function between the two, N be the sequence length, and k be the time difference. The autocorrelation function of the residual. The autocorrelation function of the modal reconstruction signal is given by the formula below, where the correlation between signals is maximized when k=1. The number of optimal decomposition modes is determined by the following formula: .

[0011] Preferably, the detection module and the processing module are connected via a communication cable and a communication interface, or via a wireless network, to facilitate the transmission of data information. The processing module includes signal receiving, signal conversion, signal filtering, and signal gain. The signal receiving module is electrically connected to the detection module to acquire and receive data information. The signal conversion module is electrically connected to the signal receiving module to perform analog-to-digital conversion on the data information. The signal filtering module is electrically connected to the signal conversion module to perform filtering on the data information. The signal gain module is electrically connected to the signal filtering module to amplify the data information.

[0012] Preferably, the processing module further includes data cleaning and data detection. Data cleaning refers to processing missing values, erroneous values, and duplicate values. Data cleaning improves data quality through data filtering and data repair. First, statistical analysis methods are used to identify erroneous values ​​before erroneous data is removed to achieve the purpose of data cleaning. The data detection module detects potential errors in inconsistent data based on the consistency between related data and repairs them to complete the cleaning of data from multiple data sources.

[0013] Preferably, the control module is electrically connected to an auxiliary module, which includes an indicator module, a timer, a storage module, adjustment buttons, a voice reminder, an alarm module, a display, a reset circuit, and an outing aid. The adjustment buttons and the display are used to adjust and control the system, and to set thresholds and family member numbers. The indicator module displays the device's operating status using LEDs. The timer is used for timing processing, facilitating the voice reminder to remind family members to perform health checks and take medication within specific time periods. The alarm module is used to sound an alarm when the detection module and the calculation model detect abnormal data. The reset circuit is used to reset and restart the device when it malfunctions. The outing aid is worn by family members when they go out, facilitating the detection of falls and changes in their physical condition, and sounding an alarm to call for help from passersby in case of abnormalities. The storage module is used to store data information.

[0014] Preferably, the control module is electrically connected to a voltage regulating module, which is electrically connected to a power supply module. The power supply module is also electrically connected to the detection module. The power supply module uses a household power grid. The voltage regulating module includes a step-down circuit, a rectifier circuit, a filter circuit, and a voltage regulator circuit. The step-down circuit reduces the high voltage of the power supply module to a low voltage. The rectifier circuit converts AC voltage to DC voltage. The filter circuit removes AC voltage from the DC voltage. The voltage regulator circuit outputs a stable voltage to power the control module.

[0015] Compared with the prior art, the beneficial effects of the present invention are: When in use, this invention uses a detection module to monitor family members, facilitating the acquisition of various data and information. This data is then transmitted to a computational model, which includes a diagnostic model and a prognostic model. The diagnostic model analyzes the family members' pathology to determine if they are ill, while the prognostic model analyzes existing normal data or data within a certain range to predict and prevent potential pathological changes in the family members' future lives, thus eliminating the possibility of pathological events. Furthermore, when a family member exhibits abnormalities, an emergency call module can be used to call for help, ensuring timely medical attention. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the system structure of the present invention; Figure 2 This is a schematic diagram of the detection module of the present invention; Figure 3 This is a schematic diagram of the distress call module of the present invention; Figure 4 This is a schematic diagram of the auxiliary module of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figures 1-4 This invention provides a technical solution: an intelligent health monitoring device, including a control module, a processing module electrically connected to the control module, a detection module and a communication module electrically connected to the processing module, the detection module being used to process and transmit data information collected by the detection module, and to communicate, transmit and download control commands and data from the control module, the detection module being used to monitor and warn of the health of people living at home, especially to monitor elderly people living at home, the communication module being used to upload and download data information, the communication module being electrically connected to a cloud server and an emergency call module, the cloud server being used to upload and store data information, or to perform calculations on data information, the emergency call module being used to call for help after the detection module detects an abnormality, or to contact the family members of the people living at home, especially the elderly living at home, so that the guardian can monitor in a timely manner, obtain information about the family members and establish communication connections; The control module is electrically connected to the calculation model, which is used to calculate and analyze the data information collected by the detection module. The calculation model includes a diagnostic model and a prognostic model. The diagnostic model is used to analyze the physical condition of the family members based on the data information collected by the detection module, and then analyze the health of the family members by combining various data information to facilitate the diagnosis and treatment of pathological diseases. The prognostic model is used to estimate the risk of a certain outcome of the family members in a certain period of time in the future, so as to facilitate the assessment of the physical condition of the family members, identify high-risk pathological diseases of the family members, and predict the physical condition of the family members in advance by monitoring high-risk pathological diseases in real time. The prognostic model uses regression prediction, which employs MAE and RMSE. The MAE (Mean Absolute Error) is used, which better reflects the actual error between the predicted and actual values. Its calculation formula is as follows: , The RMSE, or root mean square error, is used to measure the deviation between observed and true values. It is particularly sensitive to exceptionally large or small errors in a set of predictions. Its calculation formula is as follows: ; Furthermore, in the two formulas above, m represents the amount of data. The average of the data. For the data information to be calculated; The emergency call module includes one-click SOS, address reporting, remote video, data upload, and family / guardian calls. The one-click SOS is used to call for help when a family member experiences a health problem. It includes calling a specific hospital, guardian, and property management company. When calling a specific hospital, the address reporting and data upload functions simultaneously upload the community address, floor information, and portal information. The system also transmits the family member's health monitoring data to the hospital via data upload, allowing the hospital to promptly understand the family member's condition and facilitate emergency treatment. The family / guardian call function allows the one-click SOS to call the guardian. The remote video allows guardians to conduct video calls with family members during normal times or when data is abnormal, facilitating communication and understanding of the situation.

[0019] To enable the owner of the family member number to obtain real-time information on the health status of family members and to facilitate the calling of guardians within the range of data anomalies, in this embodiment, preferably, the data upload is also used to periodically transmit data information to the family member number during normal times, so that the guardian can be aware of the health status of family members. In addition, when the data information collected by the detection module shows certain anomalies, but does not reach the level of a sudden pathological event, the data upload will also transmit the data information to the family member number.

[0020] In order to perform calculations on the collected data to facilitate the calculation of whether family members have pathological symptoms and to build a calculation model, in this embodiment, preferably, the calculation model mainly includes random forest, k-nearest neighbors, support vector machine and artificial neural network algorithms. The steps for establishing the computational model are as follows: S1. Data Collection and Data Processing: The data used to build the computational model includes re-collected data or data from existing experimental, cohort studies, registry, or managed datasets. The data is then cleaned and tested, and classified according to male / female, age, and serum sodium levels. The data is divided into two categories: model calculation data and model validation data, which are used for building the model and validating the model, respectively. S2. Model building and variable selection: After the computational model is developed, variables need to be included in the computational model, and stepwise regression is used to select variables. S3. Model Performance: After the computational model is built, its performance needs to be evaluated. The accuracy of the computational model has two components: discrimination and calibration. S4. Model Validation: The computational model is validated. Validation methods include cross-validation and bootstrapping. S5. Model Usage and Updates: The detection module transmits the collected data to the computational model for calculation and analysis, and the computational model needs to be continuously evolved and dynamically updated.

[0021] In order for the computational model to perform computational analysis and processing on the data, in this embodiment, preferably, the calculation formula of the artificial neural network algorithm is as follows: It is determined by the input neuron information and the corresponding weight matrix W. X is a vector, and W is a weight matrix. X is transformed into another vector through W. The information transmitted from the hidden layer neurons of the previous time step to the neurons of the current time step has the following values: The weight matrix U determines the weight. , It is a vector, and U is a weight matrix, through which U is used to... Convert it into another vector; Mainly the neurons at the current moment The inputs from the two parts mentioned above are integrated to activate and generate the output of the hidden layer neurons at the current time step. The integration process is simply vector addition: + ; Assuming the activation function is f, the value of the neuron at the current time step is generated after activation: ; Information transmission mainly involves transferring the values ​​of the hidden layer neurons at the current time step to the output neurons at the current time step: ; The formula for moving forward is as follows: ; The entire neural network contains three weight matrices. The first matrix is ​​W, with dimensions (N, K), where N represents the vector dimension of the hidden layer neuron values ​​and K represents the vector dimension of the input neuron. The second weight matrix is ​​V, with dimensions (L, N), where P represents the vector dimension of the output layer neuron. The third is the weight matrix U, with dimensions (N, N).

[0022] To monitor various data of family members, in this embodiment, preferably, the detection module includes a blood pressure detector, a blood glucose detector, a respiration detector, a weight analyzer, a video monitor, an electrocardiogram (ECG) detector, a blood oxygen detector, a temperature detector, and a sleep detector. The blood pressure detector and blood oxygen detector are used to detect the blood pressure level and blood oxygen concentration of the family members. The blood glucose detector is used to detect the blood glucose concentration of the family members. The respiration detector and sleep detector are used to detect the breathing and sleep quality of the family members during sleep. The weight analyzer and the temperature detector are used to detect the weight and temperature of the family members. The weight analyzer also calculates the BMI based on the collected data. The ECG detector is used to collect the electrocardiogram of the family members. The video monitor is used to monitor the safety of the family members and prevent falls.

[0023] To achieve remote, non-contact monitoring of family members' conditions and improve the usability and comfort of the device, in this embodiment, preferably, the respiration detector uses millimeter-wave radar for detection, and the calculation formula for millimeter-wave radar is as follows: ; in, The amplitude of the heartbeat signal. The frequency of the heartbeat signal. Let z be the phase and z be the noise. The VMD algorithm decomposes a signal into a set of IMF components. The optimal number of IMFs is determined by the cross-correlation coefficient between the residual and the superimposed IMF signal. A smaller cross-correlation coefficient indicates a better value. Let the residual after VMD decomposition be mn, and the reconstructed signal from all IMFs be fn. Then the normalized cross-correlation function is: , in, Let be the cross-correlation function between the two, N be the sequence length, and k be the time difference. The autocorrelation function of the residual. The autocorrelation function of the modal reconstruction signal is given by the formula below, where the correlation between signals is maximized when k=1. The number of optimal decomposition modes is determined by the following formula: .

[0024] To facilitate data processing and transmission, in this embodiment, preferably, the detection module and the processing module are connected via a communication cable and interface, or via a wireless network, to facilitate data transmission. The processing module includes signal receiving, signal conversion, signal filtering, and signal gain. The signal receiving module is electrically connected to the detection module to acquire and receive data. The signal conversion module is electrically connected to the signal receiving module to perform analog-to-digital conversion on the data. The signal filtering module is electrically connected to the signal conversion module to perform filtering on the data. The signal gain module is electrically connected to the signal filtering module to amplify the data.

[0025] To process dirty data and improve data quality, in this embodiment, the processing module preferably includes data cleaning and data detection. Data cleaning refers to removing missing, erroneous, and duplicate values. Data cleaning improves data quality through data filtering and repair. First, statistical analysis is used to identify erroneous values ​​before they are removed, thus achieving the purpose of data cleaning. The data detection module detects potential errors in inconsistent data based on the consistency between related data and repairs them to complete the cleaning of data from multiple data sources.

[0026] To improve ease of use and status display, and enhance the overall effectiveness of the device, in this embodiment, preferably, the control module is electrically connected to an auxiliary module. This auxiliary module includes an indicator module, a timer, a storage module, adjustment buttons, a voice reminder, an alarm module, a display, a reset circuit, and an outing aid. The adjustment buttons and the display are used to adjust and control the system, and to set thresholds and family member numbers. The indicator module displays the device's operating status using LEDs. The timer handles timing, allowing the voice reminder to alert family members to perform health checks and take medication within specific time periods. The alarm module triggers an alarm when the detection module and calculation model detect abnormal data. The reset circuit resets and restarts the device in case of malfunction. The outing aid is worn by family members when they go out, allowing for detection of falls and changes in their physical condition, and triggering an alarm to alert passersby in case of abnormalities. The storage module stores the data.

[0027] To power the device and maintain its stable operation, in this embodiment, preferably, the control module is electrically connected to a voltage regulating module, which is electrically connected to the power supply module. The power supply module is also electrically connected to the detection module. The power supply module uses a household power grid. The voltage regulating module includes a step-down circuit, a rectifier circuit, a filter circuit, and a voltage regulator circuit. The step-down circuit reduces the high voltage of the power supply module to a low voltage. The rectifier circuit converts AC voltage to DC voltage. The filter circuit removes AC voltage from the DC voltage. The voltage regulator circuit outputs a stable voltage to power the control module.

[0028] The working principle and usage process of this invention are as follows: During use, a power supply module and a voltage regulation module ensure stable power supply to the system, facilitating control and adjustment of the equipment by the control module. A detection module collects data, including a blood pressure detector, blood glucose detector, respiration detector, weight analyzer, video monitor, electrocardiogram detector, blood oxygen detector, temperature detector, and sleep detector. These collect various data from the residents. The respiration and sleep detectors are located in the bedroom for monitoring the residents' sleep processes. Several video monitors are distributed throughout the room to monitor the residents' daily activities. The main purpose of this invention is for detection... When a family member falls, the detection module transmits the data to the processing module. After processing, the control module transmits the data to the calculation model. The calculation model uses the diagnostic and prognostic models to perform calculations and analysis. The diagnostic model analyzes the family member's physical condition based on the data collected by the detection module. By combining various data points, it analyzes the family member's health to facilitate pathological diagnosis and treatment. The prognostic model estimates the risk of a certain outcome for the family member at a future time, facilitating the assessment of the family member's physical condition and enabling the identification of high-risk pathologies. By monitoring high-risk pathologies in real time, the family member's physical condition can be predicted in advance. Furthermore, all detected data is uploaded to a cloud server for storage via a communication module. In case of data anomalies, an emergency call module is used, which includes one-click SOS, address reporting, remote video, data upload, and family / guardian calls. The one-click SOS is used to call for help when a family member experiences a health problem. It includes calling a specific hospital, guardian, or property management. When calling a specific hospital, the one-click SOS simultaneously uploads the community address, floor information, and portal information via address reporting and data upload. It also uploads the family member's health monitoring data. Data is transmitted to the hospital so that the hospital can promptly obtain information about the family member's condition and facilitate emergency treatment. The family call function allows for one-click emergency calls to guardians. Remote video allows guardians to have remote video calls with family members during normal times or when data is abnormal, so as to understand the situation. Data upload is also used to periodically transmit data information to the family call function so that guardians can be aware of the family member's physical condition. In addition, when the data information collected by the detection module is abnormal, but does not reach the level of a sudden pathological event, the data upload function will also transmit the data information to the family call function.

[0029] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent health monitoring device, comprising a control module, characterized in that: The control module is electrically connected to a processing module, which in turn is electrically connected to a detection module and a communication module. The detection module is used to process and transmit the data collected by the detection module, as well as to communicate, transmit, and download control commands and data from the control module. The detection module is used to monitor and provide early warnings for the health of people living at home, especially for elderly people living at home. The communication module is used to upload and download data. The communication module is electrically connected to a cloud server and an emergency call module. The cloud server is used to upload and store data or to perform calculations on the data. The emergency call module is used to call for help after the detection module detects an abnormality, or to contact the family members of the people living at home, especially the elderly, so that the caregiver can monitor in a timely manner, obtain information about the family members, and establish communication connections. The control module is electrically connected to the calculation model, which is used to calculate and analyze the data information collected by the detection module. The calculation model includes a diagnostic model and a prognostic model. The diagnostic model is used to analyze the physical condition of the family members based on the data information collected by the detection module, and then analyze the health of the family members by combining various data information to facilitate the diagnosis and treatment of pathological diseases. The prognostic model is used to estimate the risk of a certain outcome of the family members in a certain period of time in the future, so as to facilitate the assessment of the physical condition of the family members, identify high-risk pathological diseases of the family members, and predict the physical condition of the family members in advance by monitoring high-risk pathological diseases in real time. The prognostic model uses regression prediction, which employs MAE and RMSE. The MAE (Mean Absolute Error) is used, which better reflects the actual error between the predicted and actual values. Its calculation formula is as follows: ， The RMSE, or root mean square error, is used to measure the deviation between observed and true values. It is particularly sensitive to exceptionally large or small errors in a set of predictions. Its calculation formula is as follows: ； Furthermore, in the two formulas above, m represents the amount of data. The average of the data. For the data information to be calculated; The emergency call module includes one-click SOS, address reporting, remote video, data upload, and family / guardian calls. The one-click SOS is used to call for help when a family member experiences a health problem. It includes calling a specific hospital, guardian, and property management company. When calling a specific hospital, the address reporting and data upload functions simultaneously upload the community address, floor information, and portal information. The system also transmits the family member's health monitoring data to the hospital via data upload, allowing the hospital to promptly understand the family member's condition and facilitate emergency treatment. The family / guardian call function allows the one-click SOS to call the guardian. The remote video allows guardians to conduct video calls with family members during normal times or when data is abnormal, facilitating communication and understanding of the situation.

2. The intelligent health monitoring device according to claim 1, characterized in that: In normal times, the data upload is also used to periodically transmit data information to the family call system, so that guardians can be aware of the health status of family members. In the event that the data information collected by the detection module shows certain abnormalities, but does not reach the level of a sudden pathological event, the data upload will also transmit the data information to the family call system.

3. The intelligent health monitoring device according to claim 1, characterized in that: The computational model mainly includes random forest, k-nearest neighbors, support vector machine and artificial neural network algorithms; The steps for establishing the computational model are as follows: S1. Data Collection and Data Processing: The data used to build the computational model includes re-collected data or data from existing experimental, cohort studies, registry, or managed datasets. The data is then cleaned and tested, and classified according to male / female, age, and serum sodium levels. The data is divided into two categories: model calculation data and model validation data, which are used for building the model and validating the model, respectively. S2. Model building and variable selection: After the computational model is developed, variables need to be included in the computational model, and stepwise regression is used to select variables. S3. Model Performance: After the computational model is built, its performance needs to be evaluated. The accuracy of the computational model has two components: discrimination and calibration. S4. Model Validation: The computational model is validated. Validation methods include cross-validation and bootstrapping. S5. Model Usage and Updates: The detection module transmits the collected data to the computational model for calculation and analysis, and the computational model needs to be continuously evolved and dynamically updated.

4. The intelligent health monitoring device according to claim 3, characterized in that: The calculation formula of the artificial neural network algorithm is as follows: It is determined by the input neuron information and the corresponding weight matrix W. X is a vector, and W is a weight matrix. X is transformed into another vector through W. The information transmitted from the hidden layer neurons of the previous time step to the neurons of the current time step has the following values: The weight matrix U determines the weight. , It is a vector, and U is a weight matrix, through which U is used to... Convert it into another vector; Mainly the neurons at the current moment The inputs from the two parts mentioned above are integrated to activate and generate the output of the hidden layer neurons at the current time step. The integration process is simply vector addition: + ; Assuming the activation function is f, the value of the neuron at the current time step is generated after activation: ; Information transmission mainly involves transferring the values ​​of the hidden layer neurons at the current time step to the output neurons at the current time step: ； The formula for moving forward is as follows: ； The entire neural network contains three weight matrices. The first matrix is ​​W, with dimensions (N, K), where N represents the vector dimension of the hidden layer neuron values ​​and K represents the vector dimension of the input neuron. The second weight matrix is ​​V, with dimensions (L, N), where P represents the vector dimension of the output layer neuron. The third is the weight matrix U, with dimensions (N, N).

5. The intelligent health monitoring device according to claim 1, characterized in that: The detection module includes a blood pressure detector, a blood glucose detector, a respiration detector, a weight analyzer, a video monitor, an electrocardiogram (ECG) detector, a blood oxygen detector, a temperature detector, and a sleep detector. The blood pressure and blood oxygen detectors are used to detect the blood pressure and blood oxygen concentration of the residents. The blood glucose detector is used to detect the blood glucose concentration of the residents. The respiration and sleep detectors are used to detect the breathing and sleep quality of the residents during sleep. The weight analyzer and the temperature detector are used to detect the weight and temperature of the residents, and the weight analyzer also calculates the BMI based on the collected data. The ECG detector is used to collect the electrocardiogram of the residents. The video monitor is used to detect the residents for safety and prevent falls.

6. The intelligent health monitoring device according to claim 5, characterized in that: The respiration detector uses millimeter-wave radar for detection, and the calculation formula for millimeter-wave radar is as follows: ； in, The amplitude of the heartbeat signal. The frequency of the heartbeat signal. Let z be the phase and z be the noise. The VMD algorithm decomposes a signal into a set of IMF components. The optimal number of IMFs is determined by the cross-correlation coefficient between the residual and the superimposed IMF signal. A smaller cross-correlation coefficient indicates a better value. Let the residual after VMD decomposition be mn, and the reconstructed signal from all IMFs be fn. Then the normalized cross-correlation function is: ， in, Let be the cross-correlation function between the two, N be the sequence length, and k be the time difference. The autocorrelation function of the residual. The autocorrelation function of the modal reconstruction signal is given by the formula below, where the correlation between signals is maximized when k=1. The number of optimal decomposition modes is determined by the following formula: 。 7. The intelligent health monitoring device according to claim 1, characterized in that: The detection module and the processing module are connected via communication cables and communication interfaces, or via a wireless network, to facilitate the transmission of data. The processing module includes signal receiving, signal conversion, signal filtering, and signal gain. The signal receiving module is electrically connected to the detection module to acquire and receive data. The signal conversion module is electrically connected to the signal receiving module to perform analog-to-digital conversion on the data. The signal filtering module is electrically connected to the signal conversion module to perform filtering on the data. The signal gain module is electrically connected to the signal filtering module to amplify the data.

8. The intelligent health monitoring device according to claim 7, characterized in that: The processing module also includes data cleaning and data detection. Data cleaning refers to processing missing values, erroneous values, and duplicate values. Data cleaning improves data quality through data filtering and data repair. First, statistical analysis methods are used to identify erroneous values ​​before they are removed to achieve the purpose of data cleaning. The data detection module detects potential errors in inconsistent data based on the consistency between related data and repairs them to complete the cleaning of data from multiple data sources.

9. The intelligent health monitoring device according to claim 1, characterized in that: The control module is electrically connected to an auxiliary module, which includes an indicator module, a timer, a storage module, adjustment buttons, a voice reminder, an alarm module, a display, a reset circuit, and an outing aid. The adjustment buttons and the display are used to adjust and control the system, and to set thresholds and family member numbers. The indicator module displays the device's operating status using LEDs. The timer is used for timing processing, facilitating the voice reminder to remind family members to perform health checks and take medication within specific time periods. The alarm module is used to sound an alarm when the detection module and the calculation model detect abnormal data. The reset circuit is used to reset and restart the device when it malfunctions. The outing aid is worn by family members when they go out, facilitating the detection of falls and changes in their physical condition, and sounding an alarm to call for help from passersby in case of abnormalities. The storage module is used to store data information.

10. The intelligent health monitoring device according to claim 1, characterized in that: The control module is electrically connected to a voltage regulating module, which is electrically connected to a power supply module. The power supply module is also electrically connected to the detection module. The power supply module uses a household power grid. The voltage regulating module includes a step-down circuit, a rectifier circuit, a filter circuit, and a voltage regulator circuit. The step-down circuit reduces the high voltage of the power supply module to a low voltage. The rectifier circuit converts AC voltage to DC voltage. The filter circuit removes AC voltage from the DC voltage. The voltage regulator circuit provides a stable output voltage to power the control module.

Images