Intelligent health management system and method based on Internet of Things

Abstract

The invention discloses a smart health management system and method based on Internet of Things, and particularly relates to the technical field of privacy protection. Physiological data, health data and identity information of old people are encrypted by applying a homomorphic encryption technology; after encryption is completed, the physiological data and health data of the old people are integrated and then are separated from identity information to be stored in the intelligent equipment terminal, the data transmission and storage process is always kept in an encrypted state, and the risk of data leakage is effectively reduced; according to the invention, data training is carried out on the health comprehensive early warning model at the nodes, and then the health comprehensive early warning model training results of the nodes are aggregated to obtain the global health comprehensive early warning model, so that the data privacy can be ensured, and machine learning modeling can be carried out by using dispersed data; the privacy risk alarm is sent to the system management center when it is detected that the system has the privacy risk, then the privacy protection scheme is adjusted, and the duration of the privacy risk can be shortened.

Description

Technical Field

[0001] The present invention relates to the field of privacy protection technology, and more specifically, to a smart health management system and method based on the Internet of Things. Background Art

[0002] With the acceleration of the global aging process, the health management of the elderly population has become one of the major challenges facing the medical systems of various countries. Chronic diseases (such as diabetes, hypertension, heart disease, etc.) are common health problems among the elderly. Such diseases often require long-term and continuous monitoring and management to ensure that the patient's condition is effectively controlled. The traditional health management model mainly relies on regular hospital visits and physical examinations, but this method is inefficient and cannot monitor the patient's health status in real time, resulting in some potential health risks that cannot be discovered and dealt with in a timely manner.

[0003] The existing smart health management system based on the Internet of Things has gradually become an important solution. The existing system collects patients' physiological data in real time by connecting smart health devices to the Internet, and uploads these data to the cloud platform for analysis and monitoring. Doctors and health managers can use the remote monitoring system to keep abreast of patients' health dynamics and provide personalized health advice and intervention measures, thereby significantly improving the efficiency and effectiveness of chronic disease management.

[0004] However, there are still some problems with the existing system: Although the existing health management system has begun to adopt basic security measures such as encryption technology and data access control, these measures are powerless in the face of complex and changeable network attacks and data abuse threats. The traditional centralized data storage and processing model also brings a high risk of data leakage. The level of data privacy protection should be further improved to prevent the occurrence of privacy leakage incidents. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a smart health management system and method based on the Internet of Things to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solutions: a smart health management system based on the Internet of Things, comprising:

[0007] Physiological health data collection module: Use smart devices to collect the physiological data of the elderly, and retrieve the elderly's health data and identity information from medical institutions after obtaining authorization;

[0008] Data homomorphic encryption module: Use homomorphic encryption technology to encrypt the physiological data, health data and identity information of the elderly. After encryption, the physiological data and health data of the elderly are integrated and separated from the identity information and stored in the smart device terminal;

[0009] Data access control module: Build a role access control model to set four roles: the elderly, the elderly's family guardians, the elderly's direct medical care staff, and the system administrators, and assign different data access rights to each role;

[0010] Federated learning model training module: The physiological health data of the elderly stored in smart devices is preprocessed as raw data, and then a comprehensive health warning model is constructed. The preprocessed data is used to train the node data of the constructed comprehensive health warning model based on federated learning technology. Once the training is completed and the model effect meets expectations, the model is put into application;

[0011] Health warning module: Use the comprehensive health warning model to analyze the physiological data of the elderly in real time, and send warning information and real-time location of the elderly to the elderly's direct medical staff and family guardians when the warning conditions are triggered;

[0012] Privacy risk detection module: used to detect whether there is a privacy risk in the system. If there is a privacy risk, a privacy risk alert will be sent to the system management center. If there is no privacy risk, no privacy risk alert will be triggered;

[0013] Privacy risk control module: used to formulate privacy risk control strategies and allow system administrators to adjust privacy protection plans based on privacy risk control strategies, and automatically back up privacy risk control records;

[0014] Health management effect evaluation module: collects health warning feedback data and privacy risk control data recorded by the system within a preset period, processes the data to evaluate the health management effect, and executes different instructions based on the evaluation results;

[0015] Database: used to store the data content of each module in the system.

[0016] Preferably, the physiological data of the elderly collected by the physiological health data collection module include heart rate, blood pressure, blood sugar and blood oxygen saturation, the health data of the elderly retrieved include medical history, medication, living habits and family genetic history, and the identity information of the elderly retrieved includes name, age, gender, ID number and contact information of family guardian.

[0017] Preferably, the data homomorphic encryption module includes a key generation unit, a data encryption unit, a ciphertext operation unit, a data decryption unit, a data integration unit and an anonymization processing unit. The key generation unit uses a key generation algorithm to generate an encryption key and a decryption key; the data encryption unit uses the generated encryption key to encrypt the physiological data, health data and identity information of the elderly to generate a ciphertext; the ciphertext operation unit is used to operate on the generated ciphertext; the data decryption unit uses the generated encryption key to decrypt the ciphertext of the physiological data, health data and identity information of the elderly to obtain a plaintext result; the data integration unit integrates the ciphertext of the physiological data and health data of the elderly to obtain the ciphertext of the physiological health data of the elderly; the anonymization processing unit separates the ciphertext of the physiological health data of the elderly from the ciphertext of the identity information and stores them in the smart device terminal.

[0018] Preferably, the federated learning model training module includes a node division unit, a raw data preprocessing unit, a health comprehensive early warning model construction unit, a model node distribution training unit, a training result aggregation unit, a model effect evaluation unit and a model application unit. The node division unit divides the physiological health data of the elderly stored in the smart device with the management agency to which the smart device belongs as the node to obtain the physiological health data of the elderly of multiple nodes; the raw data preprocessing unit marks the physiological health data of the elderly of multiple nodes as raw data and then performs data cleaning, data integration and data standardization in turn; the health comprehensive early warning model construction unit is used to construct a health comprehensive early warning model including heart rate monitoring, blood pressure monitoring, blood sugar monitoring and blood oxygen saturation monitoring; the model node distribution training unit sends the constructed health comprehensive early warning model to each node, Based on the federated learning technology, the comprehensive health warning model is trained using the physiological health data of the elderly pre-processed by each node; the training result aggregation unit aggregates the training results of the comprehensive health warning model of each node to obtain a global comprehensive health warning model; the model effect evaluation unit is used to evaluate the application accuracy, maximum concurrent processing volume and maximum analysis time of the global comprehensive health warning model. If they are greater than or equal to the corresponding expected values, the model effect is determined to be up to standard, otherwise the model effect is determined to be not up to standard; when the model effect meets the standard, the model application unit backs up the global comprehensive health warning model to the cloud database for storage, and at the same time sends the global comprehensive health warning model to each node for monitoring the physiological health of the elderly, the next round of model node distribution training and model update. When the model effect does not meet the standard, the model is optimized until the model effect meets the standard.

[0019] Preferably, the privacy risk detection module includes a data collection unit, a data processing unit, a privacy risk index calculation unit, a privacy risk determination unit and a privacy risk alarm unit. The data collection unit is used to collect the encryption algorithm strength k, the number of key bits my, the maximum ciphertext expansion rate αy, the minimum encryption rate αv, the maximum privacy loss εm, the privacy budget εy, the application accuracy of the comprehensive health warning model αz and the maximum analysis time Tfm of the comprehensive health warning model; the data processing unit calculates the data encryption security coefficient Xma based on the encryption algorithm strength and the number of key bits, and the specific formula is: Xma=lg(k×my+1); the data encryption efficiency coefficient Xmb is calculated based on the maximum ciphertext expansion rate and the minimum encryption rate, and the specific calculation formula is: The privacy loss coefficient Xs is calculated based on the maximum privacy loss εm and the privacy budget εy. The specific formula is: Based on the application accuracy rate αz of the comprehensive health early warning model and the maximum analysis time Tfm of the comprehensive health early warning model, the model application accuracy coefficient Xc is calculated. The specific formula is: The privacy risk index calculation unit is used to calculate the privacy risk index YF. The specific formula is: The privacy risk determination unit compares the calculated privacy risk index with the set privacy risk safety value. If the privacy risk index is greater than the privacy risk safety value, it is determined that there is a privacy risk; if the privacy risk is less than or equal to the safety value, it is determined that there is no privacy risk; the privacy risk alarm unit sends a privacy risk alarm to the system management center when it is determined that there is a privacy risk; if there is no privacy risk, the privacy risk alarm is not triggered.

[0020] Preferably, the health management effect evaluation module includes a data collection unit, a health management effect evaluation index calculation unit, a data comparison unit and an instruction output unit. The data collection unit is used to collect the health warning feedback data and privacy risk control data recorded by the system within a preset period, wherein the health warning feedback data includes the number of health warnings Na, the number of health rescues Nh, and the interval Tai between the elderly direct medical staff receiving the warning and setting off for rescue. The privacy risk control data includes the number of successful privacy risk regulation Nta, the number of failed privacy risk regulation Ntb, the privacy risk duration Txz and the preset period duration Te; the health management effect evaluation index calculation unit is used to calculate the health management effect evaluation index YP, and the specific formula is: The data comparison unit compares the calculated health management effect evaluation index with the preset health management effect evaluation qualified value. If the health management effect evaluation index is greater than or equal to the preset health management effect evaluation qualified value, the health management effect is qualified; if the health management effect evaluation index is less than the preset health management effect evaluation qualified value, the health management effect is unqualified; the instruction output unit sends a maintenance current management strategy instruction to the system management center when the health management effect is qualified; and sends an optimization current management strategy instruction to the system management center when the health management effect is unqualified.

[0021] To achieve the above object, the present invention provides the following technical solution: a smart health management method based on the Internet of Things, comprising the following steps:

[0022] S1: Use smart devices to collect the physiological data of the elderly, and retrieve the health data and identity information of the elderly from medical institutions after obtaining authorization;

[0023] S2: Use homomorphic encryption technology to encrypt the physiological data, health data and identity information of the elderly. After encryption, the physiological data and health data of the elderly are integrated and separated from the identity information and stored in the smart device terminal;

[0024] S3: Build a role-based access control model to set four roles: the elderly person, the elderly family guardian, the elderly direct medical care staff, and the system administrator, and assign different data access rights to each role;

[0025] S4: The physiological health data of the elderly stored in the smart device is preprocessed as raw data, and then a comprehensive health warning model is constructed. The preprocessed data is used to train the node data of the constructed comprehensive health warning model based on the federated learning technology. Once the training is completed and the model effect meets expectations, the model is put into application;

[0026] S5: Use the comprehensive health warning model to analyze the physiological data of the elderly in real time, and send warning information and real-time location of the elderly to the elderly's direct medical staff and family guardians when the warning conditions are triggered;

[0027] S6: Detect whether the system has privacy risks. If there are privacy risks, a privacy risk alert is sent to the system management center. If there are no privacy risks, no privacy risk alert is triggered.

[0028] S7: Formulate privacy risk control strategies and allow system administrators to adjust privacy protection plans based on the privacy risk control strategies, and automatically back up privacy risk control records;

[0029] S8: Collect the health warning feedback data and privacy risk control data recorded by the system within a preset period and perform data processing to evaluate the health management effect. When the health management effect is qualified, send an instruction to maintain the current management strategy to the system management center; when the health management effect is unsatisfactory, send an instruction to optimize the current management strategy to the system management center.

[0030] Technical effects and advantages of the present invention:

[0031] 1. The present invention sets a data homomorphic encryption module to apply homomorphic encryption technology to encrypt the physiological data, health data and identity information of the elderly. After the encryption is completed, the physiological data and health data of the elderly are integrated and separated from the identity information and stored in the smart device terminal. The data transmission and storage process always remains encrypted, which effectively reduces the risk of data leakage; a federated learning model training module is set to pre-process the physiological health data of the elderly stored in the smart device as original data, and then a comprehensive health early warning model is constructed and the pre-processed data is used based on the federated learning technology to perform node data training on the constructed comprehensive health early warning model. After the training is completed and the model effect meets the expectation, it is put into the application model, avoiding the large amount of data transmission and processing tasks caused by centralizing all data to a central server for processing, which can reduce Low data processing pressure. First, the comprehensive health warning model is trained at the node, and then the training results of the comprehensive health warning model of each node are aggregated to obtain the global comprehensive health warning model. This can not only reduce the possibility of privacy leakage in the process of large-scale data transmission and processing and ensure data privacy, but also use scattered data for machine learning modeling; a privacy risk detection module is set up to detect whether there is a privacy risk in the system. If there is a privacy risk, a privacy risk alert is sent to the system management center. If there is no privacy risk, the privacy risk alert is not triggered; a privacy risk control module is set up to formulate a privacy risk control strategy and allow system managers to adjust the privacy protection plan based on the privacy risk control strategy, which is conducive to timely clearing privacy risk events in the system, reducing the duration of privacy risks, and reducing the possibility of privacy data leakage.

[0032] 2. The present invention sets up a health warning module to use a comprehensive health warning model to perform real-time analysis on the physiological data of the elderly. After the warning conditions are triggered, warning information and real-time positioning of the elderly are sent to the elderly's direct medical staff and family guardians, thereby optimizing the elderly's health rescue process and being able to deal with the elderly's health risks in a timely manner; a health management effect evaluation module is set up to collect the health warning feedback data and privacy risk control data recorded by the system within a preset period and perform data processing to evaluate the health management effect. When the health management effect is qualified, an instruction to maintain the current management strategy is sent to the system management center. When the health management effect is unqualified, an instruction to optimize the current management strategy is sent to the system management center. The system can be continuously updated and iterated based on real-time data, and the priority and progress of the system can be maintained. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a system structure block diagram of the present invention.

[0034] Figure 2 It is a diagram of the method steps of the present invention. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] like Figure 1 The embodiment shown provides a smart health management system based on the Internet of Things, including a physiological health data acquisition module, a data homomorphic encryption module, a data access control module, a federated learning model training module, a health warning module, a privacy risk detection module, a privacy risk control module, a health management effect evaluation module and a database. The physiological health data acquisition module, the data homomorphic encryption module and the data access control module are connected sequentially, the data homomorphic encryption module, the federated learning model training module and the health warning module are connected sequentially, the data homomorphic encryption module, the federated learning model training module and the privacy risk detection module are connected, the privacy risk detection module, the privacy risk control module and the health management effect evaluation module are connected sequentially, and all modules in the system are connected to the database.

[0037] The physiological health data collection module uses smart devices to collect the physiological data of the elderly, and retrieves the health data and identity information of the elderly from medical institutions after obtaining authorization;

[0038] Furthermore, the physiological data of the elderly collected by the physiological health data acquisition module include heart rate, blood pressure, blood sugar and blood oxygen saturation, the health data of the elderly retrieved include medical history, medication, living habits and family genetic history, and the identity information of the elderly retrieved includes name, age, gender, ID number and contact information of family guardian.

[0039] The data homomorphic encryption module uses homomorphic encryption technology to encrypt the physiological data, health data and identity information of the elderly. After encryption, the physiological data and health data of the elderly are integrated and separated from the identity information and stored in the smart device terminal;

[0040] Furthermore, the data homomorphic encryption module includes a key generation unit, a data encryption unit, a ciphertext operation unit, a data decryption unit, a data integration unit and an anonymization processing unit. The key generation unit uses a key generation algorithm to generate an encryption key and a decryption key; the data encryption unit uses the generated encryption key to encrypt the physiological data, health data and identity information of the elderly to generate a ciphertext; the ciphertext operation unit is used to operate on the generated ciphertext; the data decryption unit uses the generated encryption key to decrypt the ciphertext of the physiological data, health data and identity information of the elderly to obtain a plaintext result; the data integration unit integrates the ciphertext of the physiological data and health data of the elderly to obtain the ciphertext of the physiological health data of the elderly; the anonymization processing unit separates the ciphertext of the physiological health data of the elderly from the ciphertext of the identity information and stores them in the smart device terminal.

[0041] The data access control module constructs a role access control model to set four roles: the elderly person, the elderly family guardian, the elderly direct medical care personnel and the system administrator, and assigns different data access rights to each role;

[0042] What needs to be specifically explained in this embodiment is that the construction of the role access control model is an existing technology and will not be elaborated here. In addition to the elderly person himself, the elderly person's family guardian, the elderly person's direct medical care staff and system administrators, the set roles can also include other identities such as smart device maintenance personnel. The roles can be expanded based on actual needs and are not listed here one by one.

[0043] The federated learning model training module pre-processes the physiological health data of the elderly stored in the smart device as raw data, then constructs a comprehensive health warning model and uses the pre-processed data based on the federated learning technology to perform node data training on the constructed comprehensive health warning model. Once the training is completed and the model effect meets expectations, the model is put into application;

[0044] Furthermore, the federated learning model training module includes a node division unit, a raw data preprocessing unit, a comprehensive health warning model construction unit, a model node distribution training unit, a training result aggregation unit, a model effect evaluation unit and a model application unit. The node division unit divides the physiological health data of the elderly stored in the smart device with the management agency to which the smart device belongs as the node to obtain the physiological health data of the elderly at multiple nodes; the raw data preprocessing unit marks the physiological health data of the elderly at multiple nodes as raw data and then performs data cleaning, data integration and data standardization in turn; the comprehensive health warning model construction unit is used to construct a comprehensive health warning model including heart rate monitoring, blood pressure monitoring, blood sugar monitoring and blood oxygen saturation monitoring; the model node distribution training unit sends the constructed comprehensive health warning model to each node Based on the federated learning technology, the comprehensive health warning model is trained using the physiological health data of the elderly pre-processed by each node; the training result aggregation unit aggregates the training results of the comprehensive health warning model of each node to obtain a global comprehensive health warning model; the model effect evaluation unit is used to evaluate the application accuracy, maximum concurrent processing volume and maximum analysis time of the global comprehensive health warning model. If they are greater than or equal to the corresponding expected values, the model effect is determined to be up to standard, otherwise the model effect is determined to be not up to standard; when the model effect meets the standard, the model application unit backs up the global comprehensive health warning model to the cloud database for storage, and at the same time sends the global comprehensive health warning model to each node for monitoring the physiological health of the elderly, the next round of model node distribution training and model update. When the model effect does not meet the standard, the model is optimized until the model effect meets the standard.

[0045] It should be specifically noted in this embodiment that the process of building a comprehensive health warning model is an existing technology and will not be elaborated on in detail here, and federated learning is also an existing technology and will not be elaborated on in detail here.

[0046] The health warning module uses a comprehensive health warning model to analyze the physiological data of the elderly in real time, and sends warning information and the elderly’s real-time location to the elderly’s direct medical staff and the elderly’s family guardians when the warning conditions are triggered;

[0047] The privacy risk detection module is used to detect whether there is a privacy risk in the system. If there is a privacy risk, a privacy risk alert is sent to the system management center. If there is no privacy risk, no privacy risk alert is triggered;

[0048] Furthermore, the privacy risk detection module includes a data collection unit, a data processing unit, a privacy risk index calculation unit, a privacy risk determination unit and a privacy risk alarm unit. The data collection unit is used to collect the encryption algorithm strength k, the number of key bits my, the maximum ciphertext expansion rate αy, the minimum encryption rate αv, the maximum privacy loss εm, the privacy budget εy, the application accuracy of the comprehensive health warning model αz and the maximum analysis time Tfm of the comprehensive health warning model; the data processing unit calculates the data encryption security coefficient Xma based on the encryption algorithm strength and the number of key bits, and the specific formula is: Xma=lg(k×my+1); the data encryption efficiency coefficient Xmb is calculated based on the maximum ciphertext expansion rate and the minimum encryption rate, and the specific calculation formula is: The privacy loss coefficient Xs is calculated based on the maximum privacy loss εm and the privacy budget εy. The specific formula is: Based on the application accuracy rate αz of the comprehensive health early warning model and the maximum analysis time Tfm of the comprehensive health early warning model, the model application accuracy coefficient Xc is calculated. The specific formula is: The privacy risk index calculation unit is used to calculate the privacy risk index YF. The specific formula is: The privacy risk determination unit compares the calculated privacy risk index with the set privacy risk safety value. If the privacy risk index is greater than the privacy risk safety value, it is determined that there is a privacy risk; if the privacy risk is less than or equal to the safety value, it is determined that there is no privacy risk; the privacy risk alarm unit sends a privacy risk alarm to the system management center when it is determined that there is a privacy risk; if there is no privacy risk, the privacy risk alarm is not triggered.

[0049] What needs to be specifically explained in this embodiment is that the encryption algorithm strength and the number of key bits can be directly determined by the selected encryption algorithm, the ciphertext expansion rate = encrypted data volume / original data volume, after the encryption is completed, the encrypted data volume and the original data volume are both known data, and the ciphertext expansion rate can be directly calculated, and the maximum value extracted therefrom is the maximum ciphertext expansion rate, if there is encryption behavior at the data collection time, the maximum ciphertext expansion rate is directly selected from the encrypted data, if there is no encryption behavior at the data collection time, the maximum ciphertext expansion rate is selected at the encryption behavior time closest to the data collection time, the subsequent minimum encryption rate, maximum privacy loss and maximum analysis time of the comprehensive health warning model are all consistent with the extraction rule of the maximum ciphertext expansion rate, encryption rate = encrypted data volume / encryption time, which can be directly obtained based on existing data, and the minimum value extracted therefrom is the minimum encryption rate, the maximum privacy loss can be directly obtained based on existing technology, and the privacy budget is custom data, which is selected based on actual needs.

[0050] The privacy risk control module is used to formulate privacy risk control strategies and allow system administrators to adjust privacy protection plans based on the privacy risk control strategies, and automatically back up privacy risk control records;

[0051] The health management effect evaluation module collects health warning feedback data and privacy risk control data recorded by the system within a preset period, processes the data and evaluates the health management effect, and executes different instructions based on the evaluation results;

[0052] Further, the health management effect evaluation module includes a data collection unit, a health management effect evaluation index calculation unit, a data comparison unit and an instruction output unit. The data collection unit is used to collect the health warning feedback data and privacy risk control data recorded by the system within a preset period, wherein the health warning feedback data includes the number of health warnings Na, the number of health rescues Nh, and the interval Tai between the elderly direct medical staff receiving the warning and setting off for rescue. The privacy risk control data includes the number of successful privacy risk regulation Nta, the number of failed privacy risk regulation Ntb, the duration of privacy risk Txz and the preset period duration Te; the health management effect evaluation index calculation unit is used to calculate the health management effect evaluation index YP, and the specific formula is:

[0053] The data comparison unit compares the calculated health management effect evaluation index with the preset health management effect evaluation qualified value. If the health management effect evaluation index is greater than or equal to the preset health management effect evaluation qualified value, the health management effect is qualified; if the health management effect evaluation index is less than the preset health management effect evaluation qualified value, the health management effect is unqualified; the instruction output unit sends a maintenance current management strategy instruction to the system management center when the health management effect is qualified; and sends an optimization current management strategy instruction to the system management center when the health management effect is unqualified.

[0054] The database is used to store the data content of each module in the system.

[0055] It should be specifically noted in this embodiment that the preset values ​​used are selected based on actual needs and are not limited to specific values ​​here.

[0056] like Figure 2 This embodiment provides a smart health management method based on the Internet of Things, comprising the following steps:

[0057] S1: Use smart devices to collect the physiological data of the elderly, and retrieve the health data and identity information of the elderly from medical institutions after obtaining authorization;

[0058] S2: Use homomorphic encryption technology to encrypt the physiological data, health data and identity information of the elderly. After encryption, the physiological data and health data of the elderly are integrated and separated from the identity information and stored in the smart device terminal;

[0059] S3: Build a role-based access control model to set four roles: the elderly person, the elderly family guardian, the elderly direct medical care staff, and the system administrator, and assign different data access rights to each role;

[0060] S4: The physiological health data of the elderly stored in the smart device is preprocessed as raw data, and then a comprehensive health warning model is constructed. The preprocessed data is used to train the node data of the constructed comprehensive health warning model based on the federated learning technology. Once the training is completed and the model effect meets expectations, the model is put into application;

[0061] S5: Use the comprehensive health warning model to analyze the physiological data of the elderly in real time, and send warning information and real-time location of the elderly to the elderly's direct medical staff and family guardians when the warning conditions are triggered;

[0062] S6: Detect whether the system has privacy risks. If there are privacy risks, a privacy risk alert is sent to the system management center. If there are no privacy risks, no privacy risk alert is triggered.

[0063] S7: Formulate privacy risk control strategies and allow system administrators to adjust privacy protection plans based on the privacy risk control strategies, and automatically back up privacy risk control records;

[0064] S8: Collect the health warning feedback data and privacy risk control data recorded by the system within a preset period and perform data processing to evaluate the health management effect. When the health management effect is qualified, send an instruction to maintain the current management strategy to the system management center; when the health management effect is unsatisfactory, send an instruction to optimize the current management strategy to the system management center.

[0065] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A smart health management system based on the Internet of Things, characterized by: include: Physiological health data collection module: Use smart devices to collect the physiological data of the elderly, and retrieve the elderly's health data and identity information from medical institutions after obtaining authorization; Data homomorphic encryption module: Use homomorphic encryption technology to encrypt the physiological data, health data and identity information of the elderly. After encryption, the physiological data and health data of the elderly are integrated and separated from the identity information and stored in the smart device terminal; Data access control module: Build a role access control model to set four roles: the elderly, the elderly's family guardians, the elderly's direct medical care staff, and the system administrators, and assign different data access rights to each role; Federated learning model training module: The physiological health data of the elderly stored in smart devices is preprocessed as raw data, and then a comprehensive health warning model is constructed. The preprocessed data is used to train the node data of the constructed comprehensive health warning model based on federated learning technology. Once the training is completed and the model effect meets expectations, the model is put into application; Health warning module: Use the comprehensive health warning model to analyze the physiological data of the elderly in real time, and send warning information and real-time location of the elderly to the elderly's direct medical staff and family guardians when the warning conditions are triggered; Privacy risk detection module: used to detect whether there is a privacy risk in the system. If there is a privacy risk, a privacy risk alert will be sent to the system management center. If there is no privacy risk, no privacy risk alert will be triggered; Privacy risk control module: used to formulate privacy risk control strategies and allow system administrators to adjust privacy protection plans based on privacy risk control strategies, and automatically back up privacy risk control records; Health management effect evaluation module: collects health warning feedback data and privacy risk control data recorded by the system within a preset period, processes the data to evaluate the health management effect, and executes different instructions based on the evaluation results.

2. According to claim 1, the smart health management system based on the Internet of Things is characterized by: The physiological data of the elderly collected by the physiological health data acquisition module include heart rate, blood pressure, blood sugar and blood oxygen saturation. The health data of the elderly retrieved include medical history, medication, living habits and family genetic history. The identity information of the elderly retrieved includes name, age, gender, ID number and contact information of family guardian.

3. According to claim 1, the smart health management system based on the Internet of Things is characterized by: The data homomorphic encryption module includes a key generation unit, a data encryption unit, a ciphertext operation unit, a data decryption unit, a data integration unit and an anonymization processing unit. The key generation unit generates an encryption key and a decryption key using a key generation algorithm; the data encryption unit encrypts the physiological data, health data and identity information of the elderly using the generated encryption key to generate a ciphertext; the ciphertext operation unit is used to operate on the generated ciphertext; The data decryption unit uses the generated encryption key to decrypt the ciphertext of the elderly's physiological data, health data and identity information to obtain the plaintext result; The data integration unit integrates the physiological data and health data ciphertext of the elderly to obtain the physiological health data ciphertext of the elderly; The anonymization processing unit separates the ciphertext of the elderly's physiological health data and the ciphertext of their identity information and stores them in the smart device terminal.

4. The smart health management system based on the Internet of Things according to claim 1 is characterized by: The federated learning model training module includes a node division unit, a raw data preprocessing unit, a health comprehensive early warning model construction unit, a model node distribution training unit, a training result aggregation unit, a model effect evaluation unit and a model application unit. The node division unit divides the physiological health data of the elderly stored in the smart device with the management agency to which the smart device belongs as a node to obtain the physiological health data of the elderly of multiple nodes; The raw data preprocessing unit marks the physiological health data of the elderly at multiple nodes as raw data and then performs data cleaning, data integration and data standardization processing in sequence; The health comprehensive early warning model construction unit is used to construct a health comprehensive early warning model including heart rate monitoring, blood pressure monitoring, blood sugar monitoring and blood oxygen saturation monitoring; the model node distribution training unit sends the constructed health comprehensive early warning model to each node, and trains the health comprehensive early warning model based on the federated learning technology using the physiological health data of the elderly pre-processed by each node; The training result aggregation unit aggregates the training results of the comprehensive health warning model of each node to obtain a global comprehensive health warning model; The model effect evaluation unit is used to evaluate the application accuracy, maximum concurrent processing volume and maximum analysis time of the global comprehensive health early warning model. If they are greater than or equal to the corresponding expected values, the model effect is judged to be up to standard; otherwise, the model effect is judged to be not up to standard. When the model effect meets the standard, the model application unit will back up the global comprehensive health early warning model to the cloud database for storage, and send the global comprehensive health early warning model to each node for monitoring the physiological health of the elderly, the next round of model node distribution training and model update. When the model effect does not meet the standard, the model will be optimized until the model effect meets the standard.

5. The smart health management system based on the Internet of Things according to claim 1 is characterized by: The privacy risk detection module includes a data collection unit, a data processing unit, a privacy risk index calculation unit, a privacy risk determination unit and a privacy risk alarm unit. The data collection unit is used to collect encryption algorithm strength k, key bit number my, maximum ciphertext expansion rate αy, minimum encryption rate αv, maximum privacy loss εm, privacy budget εy, health comprehensive early warning model application accuracy αz and health comprehensive early warning model maximum analysis time Tfm; the data processing unit calculates the data encryption security coefficient Xma based on the encryption algorithm strength and key bit number, and the specific formula is: Xma=lg(k×my+1); the data encryption efficiency coefficient Xmb is calculated based on the maximum ciphertext expansion rate and the minimum encryption rate, and the specific calculation formula is: The privacy loss coefficient Xs is calculated based on the maximum privacy loss εm and the privacy budget εy. The specific formula is: Based on the application accuracy rate αz of the comprehensive health early warning model and the maximum analysis time Tfm of the comprehensive health early warning model, the model application accuracy coefficient Xc is calculated. The specific formula is: The privacy risk index calculation unit is used to calculate the privacy risk index YF. The specific formula is: The privacy risk determination unit compares the calculated privacy risk index with the set privacy risk safety value. If the privacy risk index is greater than the privacy risk safety value, it is determined that there is a privacy risk; if the privacy risk is less than or equal to the safety value, it is determined that there is no privacy risk; the privacy risk alarm unit sends a privacy risk alarm to the system management center when it is determined that there is a privacy risk; if there is no privacy risk, the privacy risk alarm is not triggered.

6. The smart health management system based on the Internet of Things according to claim 1 is characterized by: The health management effect evaluation module includes a data collection unit, a health management effect evaluation index calculation unit, a data comparison unit and an instruction output unit. The data collection unit is used to collect health warning feedback data and privacy risk control data recorded by the system within a preset period, wherein the health warning feedback data includes the number of health warnings Na, the number of health rescues Nh, and the interval Tai between the elderly's direct medical staff receiving warnings and setting off for rescue. The privacy risk control data includes the number of successful privacy risk regulation Nta, the number of failed privacy risk regulation Ntb, the duration of privacy risk Txz and the preset period Te; The health management effect evaluation index calculation unit is used to calculate the health management effect evaluation index YP. The specific formula is: The data comparison unit compares the calculated health management effect evaluation index with the preset health management effect evaluation qualified value. If the health management effect evaluation index is greater than or equal to the preset health management effect evaluation qualified value, the health management effect is qualified; if the health management effect evaluation index is less than the preset health management effect evaluation qualified value, the health management effect is unqualified; The instruction output unit sends an instruction to maintain the current management strategy to the system management center when the health management effect is qualified, and sends an instruction to optimize the current management strategy to the system management center when the health management effect is unqualified.

7. A smart health management method based on the Internet of Things, using a smart health management system based on the Internet of Things as claimed in any one of claims 1 to 6, characterized in that: The following steps are involved: S1: Use smart devices to collect the physiological data of the elderly, and retrieve the health data and identity information of the elderly from medical institutions after obtaining authorization; S2: Use homomorphic encryption technology to encrypt the physiological data, health data and identity information of the elderly. After encryption, the physiological data and health data of the elderly are integrated and separated from the identity information and stored in the smart device terminal; S3: Build a role-based access control model to set four roles: the elderly person, the elderly family guardian, the elderly direct medical care staff, and the system administrator, and assign different data access rights to each role; S4: The physiological health data of the elderly stored in the smart device is preprocessed as raw data, and then a comprehensive health warning model is constructed. The preprocessed data is used to train the node data of the constructed comprehensive health warning model based on the federated learning technology. Once the training is completed and the model effect meets expectations, the model is put into application; S5: Use the comprehensive health warning model to analyze the physiological data of the elderly in real time, and send warning information and real-time location of the elderly to the elderly's direct medical staff and family guardians when the warning conditions are triggered; S6: Detect whether the system has privacy risks. If there are privacy risks, a privacy risk alert is sent to the system management center. If there are no privacy risks, no privacy risk alert is triggered. S7: Formulate privacy risk control strategies and allow system administrators to adjust privacy protection plans based on the privacy risk control strategies, and automatically back up privacy risk control records; S8: Collect the health warning feedback data and privacy risk control data recorded by the system within a preset period and perform data processing to evaluate the health management effect. When the health management effect is qualified, send an instruction to maintain the current management strategy to the system management center; when the health management effect is unsatisfactory, send an instruction to optimize the current management strategy to the system management center.

Images