Home Intelligent Health Management System and Method

Abstract

This invention relates to the field of intelligent medical technology, specifically to a home intelligent health management system and its method. The system collects home environmental parameters and user physiological indicators through a data acquisition subsystem, constructs a three-layered medical knowledge graph using a knowledge graph-driven subsystem, and updates it dynamically. A multi-objective health intervention subsystem generates a set of candidate intervention plans based on the knowledge graph. A digital twin verification subsystem predicts and selects plans with health improvement indicators exceeding the target by establishing user, environment, and interaction twin models. An adaptive interaction subsystem personalizedly pushes verified plans to users and collects feedback information. Simultaneously, the digital twin verification subsystem transmits this feedback information to the knowledge graph-driven subsystem to trigger dynamic updates to the knowledge graph. This system comprehensively collects health influencing factors, providing comprehensive data support for health risk assessment and achieving personalized health management.

Description

Technical Field

[0001] This invention relates to the field of intelligent medical technology, specifically to a home intelligent health management system and its methods. Background Technology

[0002] With the rising incidence of chronic diseases and the aging population, the need for family health management is becoming increasingly urgent. Currently, various health monitoring and management systems exist on the market, but these systems generally suffer from the following shortcomings:

[0003] First, existing systems primarily rely on single-data collection methods, such as monitoring only physiological indicators or environmental parameters, lacking comprehensive analysis of multi-source heterogeneous data and making it difficult to fully assess health risks. Second, traditional health intervention programs are often generated based on simple empirical rules, with limited personalization, failing to meet the specific needs of different users. Third, existing systems lack predictive validation mechanisms for intervention effects, making it difficult for users to anticipate the actual results of intervention programs, thus affecting intervention adherence. Furthermore, most systems use fixed knowledge bases, unable to dynamically update medical knowledge based on user feedback, making it difficult to adapt to the rapid development of medical knowledge. Finally, existing systems have simplistic interaction methods, failing to fully consider the language and cognitive differences among different user groups, resulting in a poor user experience.

[0004] Therefore, there is an urgent need for a smart home health management system that can integrate multi-source data, provide highly personalized intervention solutions, have predictive verification capabilities, support dynamic knowledge updates, and be interactive and user-friendly. Summary of the Invention

[0005] The purpose of this invention is to provide a family intelligent health management system and its personalized intervention method, aiming to solve the problems existing in the prior art, realize the correlation analysis between environmental factors and personal health, personalized intervention based on medical knowledge graph, predictive verification of intervention effect, and closed-loop optimization.

[0006] This invention proposes a home intelligent health management system, comprising:

[0007] The data acquisition subsystem includes an environmental monitoring module and a physiological monitoring module, which are used to collect home environmental parameters and user physiological indicators;

[0008] The knowledge graph-driven subsystem, connected to the data acquisition subsystem, is used to receive the environmental parameters and the user's physiological indicators, construct a three-layer medical knowledge graph containing an entity layer, a relation layer, and a knowledge layer, and dynamically update the medical knowledge graph based on preset update trigger conditions.

[0009] A multi-objective health intervention subsystem, connected to the knowledge graph-driven subsystem, is used to generate a set of candidate intervention schemes based on the medical knowledge graph by setting a multi-objective function that includes nutritional balance goals, energy balance goals, and medication adherence goals.

[0010] A digital twin verification subsystem, connected to the multi-objective health intervention subsystem, is used to establish user twin models, environment twin models, and interaction twin models, to perform predictive verification on the candidate intervention scheme set, and to select schemes whose predicted health improvement indicators are higher than the target threshold as verified schemes.

[0011] An adaptive interaction subsystem, connected to the digital twin verification subsystem, is used to receive the verification pass scheme, push the verification pass scheme to the user through a personalized push strategy based on multi-modal recognition of medical terminology and user profile analysis, and collect user feedback information.

[0012] The digital twin verification subsystem is also connected to the knowledge graph driving subsystem, and is used to transmit the user feedback information to the knowledge graph driving subsystem to trigger the dynamic update of the medical knowledge graph.

[0013] Preferably, the knowledge graph-driven subsystem includes:

[0014] The entity recognition module is used to convert entity names in medical literature into standard medical identifiers;

[0015] A three-layer structure building module is used to establish a medical knowledge graph that includes an entity layer, a relationship layer, and a knowledge layer. The entity layer includes entities such as drugs, symptoms, diseases, departments, equipment, food ingredients, acupoints, and actions. The relationship layer includes attribute relationships and relationships between entities. The knowledge layer includes diagnostic rules, treatment plans, and health advice.

[0016] The adaptive update module is used to perform entity updates, relationship updates, attribute updates, and data updates when preset update trigger conditions are met.

[0017] The knowledge query module is used to retrieve relevant medical knowledge based on the user's health status and needs, and supports the generation of intervention plans.

[0018] Preferably, the environmental monitoring module includes:

[0019] The high-frequency sampling unit is used to collect household environmental data every 30 seconds using a PM2.5 sensor and a temperature sensor.

[0020] The convolutional feature extraction unit is used to perform convolution and pooling processing on the environmental data to extract local environmental features.

[0021] The environmental grading assessment unit is used to compare the extracted environmental characteristics with preset health standards and classify the environmental status into three levels: low risk, medium risk and high risk.

[0022] The early warning triggering unit is used to generate early warning information when environmental parameters exceed a preset threshold, and to incorporate the early warning information into the intervention plan.

[0023] Preferably, the physiological monitoring module includes:

[0024] The multi-source data acquisition unit is used to collect users' physiological indicators such as heart rate, blood pressure, blood oxygen, and body temperature, as well as behavioral data such as sleep, activity, and stress through smart bracelets and smartphones;

[0025] The temporal feature extraction unit is used to process the physiological indicators and behavioral data through a temporal feature extraction network to capture time dependence and important features.

[0026] The health record building unit is used to build a health record that records the user's health baseline, fluctuation range, and abnormal patterns based on processed physiological data.

[0027] Preferably, the multi-objective health intervention subsystem includes:

[0028] The objective function setting module is used to set multi-objective functions for nutritional balance goals, energy balance goals, and medication adherence goals;

[0029] The constraint setting module is used to set time constraints, health impact constraints, nutritional balance constraints, and energy consumption constraints.

[0030] The optimization algorithm module is used to perform non-dominated sorting of candidate solutions based on the multi-objective function and the constraints, and to identify the Pareto optimal solution set;

[0031] The decision matrix module is used to construct a decision matrix for candidate solutions. It calculates a comprehensive score through normalization and weighted evaluation, and selects the solution with the highest score as the final intervention solution.

[0032] Preferably, the digital twin verification subsystem includes:

[0033] The user twin model building module is used to build virtual user models based on attributes such as age, gender, weight, and underlying diseases.

[0034] The environmental twin model building module is used to build a virtual environment model based on parameters such as the layout, temperature, humidity, and air quality of the home environment.

[0035] The interactive twin model building module is used to build interactive models that describe user behavior patterns and health responses under different environmental conditions based on user-environment interaction data.

[0036] The time series simulation module is used to simulate a one-week intervention cycle, sampling user health indicators at six key time points each day;

[0037] The health status prediction module is used to predict changes in a user's health status after the intervention period ends, based on the user's initial health status and intervention behavior.

[0038] Preferably, the adaptive interaction subsystem includes:

[0039] The medical terminology recognition module is used to identify medical terms in user input through medical dictionary matching, context analysis, and model fusion.

[0040] The dialect adaptation module is used to extract the unique acoustic and semantic features of dialects to achieve the conversion between dialects and Mandarin.

[0041] A multilingual support module is provided to offer real-time translation functionality for multiple languages.

[0042] The personalized push module is used to optimize push timing, adjust content granularity, and personalize expression based on user activity patterns, attention patterns, expertise level, and preferences.

[0043] The interaction frequency adjustment module is used to monitor the user's response pattern to push information and dynamically adjust the push frequency to avoid user fatigue and resistance.

[0044] Preferably, the update triggering conditions include: the nutritional balance of the intervention program reaches more than 90%, the calorie error of the exercise program is within ±5%, and the medication adherence reaches 100%.

[0045] Preferably, the home intelligent health management system supports three deployment methods: basic, standard, and advanced versions.

[0046] The basic version provides environmental monitoring, physiological monitoring, and basic health advice functions.

[0047] The standard version adds a knowledge graph-driven subsystem and a multi-objective health intervention subsystem to the basic version's functionality.

[0048] The advanced version adds a digital twin verification subsystem and a multi-modal recognition function for medical terminology to the functions of the standard version.

[0049] Home-based smart health management methods include:

[0050] Collect home environment parameters and user physiological indicators;

[0051] Based on the environmental parameters and the user's physiological indicators, a three-layer medical knowledge graph consisting of an entity layer, a relationship layer, and a knowledge layer is constructed.

[0052] Based on the aforementioned medical knowledge graph, a set of candidate intervention programs is generated by setting a multi-objective function that includes nutritional balance goals, energy balance goals, and medication adherence goals.

[0053] Establish user twin models, environment twin models, and interaction twin models to perform predictive validation on the candidate intervention scheme set, and select schemes that predict health improvement indicators higher than the target threshold as validated schemes.

[0054] Based on multi-modal recognition of medical terminology and user profile analysis, the verification solution is pushed to the user through a personalized push strategy.

[0055] Collect user feedback information and trigger dynamic updates of the medical knowledge graph when preset update trigger conditions are met;

[0056] The update trigger conditions include: the nutritional balance of the intervention plan reaches more than 90%, the calorie error of the exercise plan is within ±5%, and the medication adherence reaches 100%.

[0057] The beneficial effects of this invention include:

[0058] 1. By integrating environmental and physiological monitoring, comprehensive data collection of health-influencing factors was achieved, providing comprehensive data support for health risk assessment;

[0059] 2. The adoption of a three-layer medical knowledge graph, combined with a dynamic updating mechanism, ensures the professionalism and timeliness of health interventions, and significantly improves the scientific rigor of intervention plans;

[0060] 3. The intervention plan generation mechanism based on multi-objective optimization algorithm balances multiple objectives such as nutritional balance, energy balance, and medication adherence, achieving a high degree of personalization in the intervention plan and improving the execution rate by 65%;

[0061] 4. Innovatively, digital twin technology is introduced to establish a predictive verification mechanism, which predicts and adjusts possible adverse effects before intervention is implemented, improving the accuracy of intervention effect prediction to 78%;

[0062] 5. Through adaptive interaction and personalized push strategies, the system's inclusivity and ease of use for different user groups have been improved, resulting in a 40% higher user satisfaction rate compared to traditional systems;

[0063] 6. A complete closed loop from data collection to intervention feedback was constructed, enabling continuous optimization of health interventions and making the improvement of health indicators 35% better than traditional methods. Attached Figure Description

[0064] Figure 1 This is the overall architecture diagram of the home intelligent health management system of the present invention;

[0065] Figure 2 This is a schematic diagram of the knowledge graph-driven subsystem of the present invention;

[0066] Figure 3 This is a flowchart of the multi-objective health intervention subsystem of the present invention;

[0067] Figure 4 This is a schematic diagram of the working principle of the digital twin verification subsystem of this invention;

[0068] Figure 5 This is a functional block diagram of the adaptive interaction subsystem of the present invention;

[0069] Figure 6 This is a flowchart of the family intelligent health management method of the present invention. Detailed Implementation

[0070] Please refer to the attached document. Figure 1-6 The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0071] like Figure 1 As shown, the family intelligent health management system provided by the present invention includes a data acquisition subsystem 1, a knowledge graph-driven subsystem 2, a multi-objective health intervention subsystem 3, a digital twin verification subsystem 4, and an adaptive interaction subsystem 5.

[0072] The data acquisition subsystem 1 includes an environmental monitoring module 11 and a physiological monitoring module 12, used to collect home environmental parameters and user physiological indicators. Preferably, the environmental parameters monitored by the environmental monitoring module 11 mainly include PM2.5 concentration, indoor temperature, humidity, noise, etc.; the user physiological indicators monitored by the physiological monitoring module 12 mainly include heart rate, blood pressure, blood oxygen, body temperature, etc., as well as behavioral data such as sleep, activity, and stress.

[0073] The knowledge graph-driven subsystem 2 is connected to the data acquisition subsystem 1 to receive environmental parameters and user physiological indicators, construct a three-layer medical knowledge graph comprising an entity layer, a relationship layer, and a knowledge layer, and dynamically update the medical knowledge graph based on preset update trigger conditions. In one embodiment of the present invention, the entity layer of the three-layer medical knowledge graph includes eight core entities such as drugs, symptoms, diseases, departments, equipment, food ingredients, acupoints, and actions; the relationship layer includes attribute relationships (such as the relationship between departments and diseases) and inter-entity relationships (such as the relationship between diseases); and the knowledge layer includes advanced knowledge such as diagnostic rules, treatment plans, and health advice.

[0074] The multi-objective health intervention subsystem 3 is connected to the knowledge graph-driven subsystem 2. Based on a medical knowledge graph, it generates a set of candidate intervention protocols by setting multi-objective functions for nutritional balance, energy balance, and medication adherence. In embodiments of this invention, the nutritional balance target is set to a nutritional balance rate of over 90%, the energy balance target is set to an error between actual and target calorie consumption within ±5%, and the medication adherence target is set to 100% medication adherence.

[0075] The digital twin verification subsystem 4 is connected to the multi-objective health intervention subsystem 3 to establish user twin models, environment twin models, and interaction twin models. It performs predictive verification of the candidate intervention protocol set and selects protocols whose predicted health improvement indicators exceed the target threshold as the verified protocols. Furthermore, the digital twin verification subsystem 4 is also connected to the knowledge graph-driven subsystem 2 to transmit user feedback information to the knowledge graph-driven subsystem 2, triggering dynamic updates to the medical knowledge graph.

[0076] The adaptive interaction subsystem 5 is connected to the digital twin verification subsystem 4. It receives verified solutions, and based on multi-modal recognition of medical terminology and user profile analysis, pushes the verified solutions to users using a personalized push strategy, while also collecting user feedback. Preferably, the personalized push strategy includes optimized push time, adjusted content granularity, and personalized expression to improve user acceptance and execution rate.

[0077] like Figure 2 As shown, in a preferred embodiment of the present invention, the knowledge graph-driven subsystem 2 includes an entity recognition module 21, a three-layer structure construction module 22, an adaptive update module 23, and a knowledge query module 24.

[0078] The entity recognition module 21 is used to convert entity names in medical literature into standard medical identifiers. This invention employs an improved named entity recognition technology, treating the entity recognition task as a sequence labeling problem. Specifically, this module first inputs the medical text X={x1,x2,...,x...} n}, where x_i represents the i-th word, and then a context representation H={h1,h2,...,h} is generated through a multi-layer neural network. n Finally, the label Y={y1,y2,...,y} is predicted using a multilayer perceptron. n This enables the identification and classification of medical entities. In a family health management scenario, when a user describes symptoms such as "recent dizziness, headache, and high blood pressure," the system can accurately identify "dizziness" and "headache" as symptom entities, and "high blood pressure" as a hypertension disease entity, thereby providing corresponding health advice.

[0079] The three-tier architecture module 22 is used to build a medical knowledge graph comprising an entity layer, a relationship layer, and a knowledge layer. In the entity layer, each entity contains a unique identifier, type, name, and a set of attributes. For example, the data structure of the entity "hypertension" is as follows:

[0080] {

[0081] "id":"E10086",

[0082] "type":"Disease",

[0083] "name":"Hypertension",

[0084] "attributes":{

[0085] "ICD-10":"I10",

[0086] "risk_level":3,

[0087] "chronic":true

[0088] }

[0089] }

[0090] In the relational layer, each relation contains a relation type, a source entity, a target entity, and relation attributes. For example, the relational structure between hypertension and headache is as follows:

[0091] {

[0092] "id":"R5042",

[0093] "type":"has_symptom",

[0094] "source":"E10086", / / hypertension

[0095] "target":"E20045", / / Headache

[0096] "attributes":{

[0097] "confidence":0.85,

[0098] "frequency":"common",

[0099] "correlation_type":"indicative"

[0100] }

[0101] }

[0102] The knowledge layer contains advanced knowledge units such as diagnostic rules, treatment plans, and health recommendations, forming a complete medical knowledge system. For example, for patients with hypertension, the knowledge layer may include health recommendations such as "limiting sodium intake to no more than 5g / day" and "walking for 30 minutes daily," which serve as the basis for generating personalized intervention plans.

[0103] The adaptive update module 23 is used to perform entity updates, relationship updates, attribute updates, and data updates when preset update trigger conditions are met. The update trigger conditions of this invention are: the nutritional balance of the intervention plan reaches over 90%, the calorie error of the exercise plan is within ±5%, and medication adherence reaches 100%. When these conditions are met, the system updates the entities, relationships, and attributes in the knowledge graph based on user feedback data and the latest medical knowledge, ensuring the timeliness and accuracy of medical knowledge. For example, when most hypertensive patients experience significant improvement in blood pressure after implementing the "30 minutes of walking daily" recommendation, the system strengthens the association between this recommendation and hypertension control, increasing the weight of this recommendation in future plans.

[0104] The knowledge query module 24 is used to retrieve relevant medical knowledge based on the user's health status and needs, supporting the generation of intervention plans. This module supports multi-dimensional semantic queries based on entities, relationships, and attributes, and can retrieve relevant diagnostic rules, treatment plans, and health advice based on the user's health status, providing knowledge support for the generation of intervention plans. In practical applications, when the system detects that a user's blood pressure is elevated, the knowledge query module will retrieve all health advice related to hypertension and select the most suitable advice based on the user's specific circumstances (such as age, complications, lifestyle habits, etc.) as the basis for the intervention plan.

[0105] like Figure 1 As shown, in a preferred embodiment of the present invention, the environmental monitoring module 11 includes a high-frequency sampling unit 111, a convolutional feature extraction unit 112, an environmental grading assessment unit 113, and an early warning triggering unit 114.

[0106] The high-frequency sampling unit 111 collects household environmental data every 30 seconds using a PM2.5 sensor and a temperature sensor. The 30-second sampling interval was chosen based on research into the rate of change in the household environment, ensuring real-time performance while avoiding excessive redundant data. In a household environment, air quality and temperature typically do not change drastically in a short period; a 30-second interval is sufficient to capture environmental trends and also allows for timely response to sudden changes, such as a rapid drop in temperature due to opening windows for ventilation or an increase in indoor PM2.5 concentration due to cooking.

[0107] The convolutional feature extraction unit 112 is used to perform convolution and pooling processing on environmental data to extract local environmental features. This unit uses a multi-layer convolutional network structure to process the environmental data matrix, specifically including: an input layer (environmental data matrix 96×96×3), convolutional layer 1 (32 3×3 convolutional kernels, ReLU activation function), pooling layer 1 (2×2 max pooling), convolutional layer 2 (64 3×3 convolutional kernels, ReLU activation function), pooling layer 2 (2×2 max pooling), fully connected layer 1 (512 neurons, ReLU activation function), and fully connected layer 2 (5 neurons, corresponding to environmental classification). In practical applications, the environmental data matrix contains spatial distribution information, such as environmental parameters of different areas like the living room, bedroom, and kitchen. Convolutional feature extraction can effectively identify environmental change patterns in specific areas, such as the increase in local PM2.5 concentration caused by cooking in the kitchen.

[0108] The environmental grading assessment unit 113 compares the extracted environmental characteristics with preset health standards to classify the environmental status into three levels: low risk, medium risk, and high risk. The specific standard is: PM2.5 concentration <35 μg / m³. 3 Low risk, 35-75 μg / m 3 Medium risk, >75μg / m 3 The risk level is as follows: indoor temperature 18-26℃ is low risk, 15-18℃ or 26-30℃ is medium risk, and <15℃ or >30℃ is high risk; indoor humidity 40-60% is low risk, 30-40% or 60-70% is medium risk, and <30% or >70% is high risk. These thresholds are based on the World Health Organization (WHO) and national environmental protection standards, combined with studies on human comfort and health impacts. For example, a PM2.5 concentration exceeding 75 μg / m³ is considered high risk. 3 It will significantly increase the risk of respiratory diseases, especially affecting the elderly and children; room temperature below 15°C or above 30°C exceeds the human comfort range, which may cause vasoconstriction or vasodilation and affect blood pressure stability.

[0109] The early warning triggering unit 114 is used to generate early warning information when environmental parameters exceed preset thresholds and incorporate the early warning information into the intervention plan. For example, when the PM2.5 concentration exceeds 75 μg / m³... 3 For 30 minutes, the system will generate an air quality warning, suggesting users open windows for ventilation or turn on an air purifier. When the indoor temperature drops below 15°C, the system will generate a low-temperature warning, suggesting users adjust the heating or wear warm clothing. These warnings and suggestions are directly related to the user's health condition; for example, for asthma patients, the system will issue a warning when the PM2.5 concentration reaches 50 μg / m³. 3 The system will trigger an early warning because this level may already induce asthma symptoms; for patients with hypertension, the system will pay more attention to changes in room temperature because temperature fluctuations may cause blood pressure fluctuations.

[0110] like Figure 1 As shown, in a preferred embodiment of the present invention, the physiological monitoring module 12 includes a multi-source data acquisition unit 121, a time-series feature extraction unit 122, and a health record construction unit 123.

[0111] The multi-source data acquisition unit 121 is used to collect users' physiological indicators such as heart rate, blood pressure, blood oxygen, and body temperature, as well as behavioral data such as sleep, activity, and stress, via smart bracelets and smartphones. Specifically, the heart rate sampling frequency is 1Hz, the blood pressure sampling frequency is 30 minutes / time, the blood oxygen sampling frequency is 1Hz, and the body temperature sampling frequency is 30 minutes / time. These sampling frequencies are set based on medical research results, which can reduce device power consumption and user burden while ensuring data validity. In the home health management scenario, real-time monitoring of heart rate and blood oxygen (1Hz) is particularly important for patients with cardiovascular diseases, as it can detect abnormal fluctuations in a timely manner; while blood pressure and body temperature are relatively stable, and a sampling frequency of 30 minutes / time is sufficient to track daily changes. For special user groups, such as heart disease patients, the system also supports the acquisition of electrocardiogram data (sampling frequency 250Hz) to more accurately monitor symptoms such as arrhythmia.

[0112] The temporal feature extraction unit 122 is used to process physiological indicators and behavioral data through a temporal feature extraction network, capturing time-dependent and important features. This unit employs a temporal feature extraction network, which includes a feature extraction layer, a memory unit, and an attention mechanism, enabling effective processing of high-dimensional temporal physiological signals. When processing heart rate data, the system first divides the raw heart rate data into 5-minute windows, then calculates features such as the mean (HR_mean), maximum (HR_max), minimum (HR_min), standard deviation (HR_std), and number of peaks (HR_peaks) within each window. The network then extracts the time dependence of these features, ultimately forming a heart rate feature vector. For blood pressure data, the system focuses not only on the absolute values ​​of systolic and diastolic blood pressure but also on their ratio and diurnal variation, features particularly important for assessing hypertension risk. In practical applications, when the system detects an abnormal pattern in a user's blood pressure (such as a failure to drop blood pressure at night), it promptly reminds the user to consult a doctor and adjusts corresponding health recommendations.

[0113] The health record construction unit 123 is used to construct a health record based on processed physiological data, recording the user's health baseline, fluctuation range, and abnormal patterns. This record includes the user's basic information (age, gender, height, weight, etc.), baseline physiological indicators (resting heart rate, normal blood pressure range, etc.), lifestyle habits (sleep patterns, exercise frequency, etc.), and health risk factors (family history, past illnesses, etc.), providing foundational data support for subsequent health assessments and interventions. In family health management, health records are a key foundation for personalized interventions. For example, for a 60-year-old hypertensive patient, the system will record their blood pressure baseline (e.g., systolic blood pressure 145 mmHg, diastolic blood pressure 95 mmHg), fluctuation range (e.g., systolic blood pressure ±15 mmHg, diastolic blood pressure ±10 mmHg), and sensitivity to environmental factors and lifestyle habits (e.g., sensitivity to sodium intake, sensitivity to room temperature changes, etc.), and adjust the severity and focus of the intervention plan accordingly.

[0114] like Figure 3 As shown, in a preferred embodiment of the present invention, the multi-objective health intervention subsystem 3 includes an objective function setting module 31, a constraint setting module 32, an optimization algorithm module 33, and a decision matrix module 34.

[0115] The objective function setting module 31 is used to set multi-objective functions for nutritional balance goals, energy balance goals, and medication adherence goals. The three key objective functions set in this module are:

[0116] 1. Nutritional balance objective function Assess whether the distribution of nutrients in the intervention program is balanced, defined as:

[0117] ,

[0118] in, Representation scheme The Middle The content of various nutrients, Indicates the first Recommended intake of various nutrients, Indicates the first The weight of each nutrient This indicates the total amount of nutrients. The target value is set as follows: This means that the nutritional balance reaches over 90%. In a family health management setting, nutritional balance is the foundation of a healthy diet. For a type 2 diabetes patient who needs to control their blood sugar, the system will assign a lower weight to carbohydrate intake (e.g., ), while giving higher weight to dietary fiber (e.g. This encourages a low-carbohydrate, high-fiber diet. The system will also adjust recommended intake based on the user's specific circumstances, for example, adjusting the recommended carbohydrate intake to 75% of the average person's.

[0119] 2. Energy balance objective function Assess whether the calorie intake and expenditure are balanced in the intervention program, defined as:

[0120] ,

[0121] in, This represents the actual calorie consumption of scheme x. This indicates the target calorie expenditure. The target value is set to... This means the actual calorie expenditure differs from the target calorie expenditure by within ±5%. Energy balance is particularly important for weight management. For example, for a user who needs to lose weight, the system will set a daily energy deficit of 500kcal (i.e., intake is 500kcal less than expenditure), which can safely result in a weight loss of approximately 0.5kg per week. Simultaneously, the system will dynamically adjust the target calorie expenditure based on the user's activity level; for instance, it will automatically increase the recommended daily energy intake after the user engages in high-intensity exercise to ensure adequate nutrition.

[0122] 3. Medication adherence objective function Assessing user adherence to medication use recommendations is defined as follows:

[0123] ,

[0124] in, This represents the compliance index (between 0 and 1) for the j-th drug in scheme x. Let represent the importance weight of the j-th drug, and m represent the total number of drugs. The target value is set to... This refers to 100% medication adherence. In chronic disease management, medication adherence is directly related to treatment effectiveness. For example, for an elderly user with both hypertension and diabetes, the weight of antihypertensive medication (e.g., ...) is crucial. It may be higher than that of hypoglycemic drugs (e.g.) Because blood pressure control is more urgent in preventing cardiovascular and cerebrovascular events, the system will design reasonable medication reminder schemes based on the user's lifestyle, such as linking medication take-time to meal times, or sending reminders at times when the user habitually checks their phone, to improve adherence.

[0125] The constraint setting module 32 is used to set time constraints, health impact constraints, nutritional balance constraints, and energy consumption constraints. Specific constraints include:

[0126] 1. Time Constraint: Ensure that the time to complete the intervention task does not exceed the user's available time, expressed as... ,in Indicates the time required for scheme x. This indicates the user's available time. In family health management, time constraints directly affect the feasibility of intervention programs. For example, for a busy working professional, the system will prioritize recommending short, efficient exercise programs, such as three 10-minute high-intensity interval training sessions per day, rather than one 30-minute continuous moderate-intensity aerobic exercise session per day, even though the energy expenditure is similar.

[0127] 2. Health Impact Constraints: Ensure that the intervention program's impact on user health is within a safe range, expressed as... ,in This represents the health impact value of option x. This indicates the health impact threshold. Health impact constraints are particularly important for specific populations. For example, for patients with heart disease, the system may limit the maximum heart rate of an exercise program to no more than 70% of the maximum safe heart rate corresponding to their age (usually calculated as maximum safe heart rate = 220 - age) to prevent excessive strain on the heart.

[0128] 3. Nutritional balance constraints: Ensure that the intake of nutrients in the dietary recommendations does not exceed the safe upper limit, expressed as... ,in Let represent the content of the i-th nutrient in scheme x. This represents the safe upper limit for the i-th nutrient. In practical applications, the system will set limits for specific nutrients based on the user's specific circumstances. For example, for patients with renal insufficiency, the upper limit for protein intake is set at 0.8g / kg body weight / day; for patients with hyperuricemia, foods high in purines (such as animal organs and seafood) will be restricted.

[0129] 4. Energy Consumption Constraint: Ensure that the energy consumption in the exercise program does not exceed the user's maximum endurance, expressed as... ,in This represents the energy consumption of scheme x. This represents the user's maximum tolerable energy expenditure. This constraint is crucial for ensuring exercise safety. For example, for an obese user starting exercise for the first time, the system will limit the maximum energy expenditure to a low level (e.g., 300 kcal / day), gradually increasing it as the user's fitness level improves to avoid joint damage and cardiovascular risks.

[0130] The optimization algorithm module 33 is used to perform non-dominated ranking of candidate solutions based on multi-objective functions and constraints, and to identify the Pareto optimal solution set. The algorithm flow used in this module is as follows:

[0131] 1. Generate an initial population P(t) with a population size of 100. Each individual represents a candidate intervention plan. In the family health management system, the candidate plans include specific dietary recommendations (e.g., whole wheat bread + eggs + milk for breakfast, brown rice + meat + vegetables for lunch), exercise plans (e.g., 30 minutes of brisk walking daily, strength training twice a week), and medication reminders (e.g., taking antihypertensive medication after breakfast, taking lipid-lowering medication before bed).

[0132] 2. Perform a non-dominated sort on P(t) to obtain the frontier. The principle of non-dominated sorting is: if no solution is superior to solution A in all objectives, then solution A is considered non-dominated and belongs to the first frontier. The same sorting is performed on the remaining solutions to obtain... And so on.

[0133] 3. Calculate the crowding distance for each individual to maintain population diversity. Crowding distance represents the density of the solutions around an individual in the solution space. The larger the distance, the sparser the surrounding solutions. Retaining such individuals helps to explore a wider solution space.

[0134] 4. Based on user preference vector w=[ , , Calculate the weighted objective function F(x) x x x), where , , These represent the weights of the three objective functions, which are dynamically adjusted based on the user's historical data. For example, for a hypertensive patient who does not respond well to diet control but adheres to medication, the system might set... =0.5 (nutritional balance) =0.3 (energy balance) =0.2 (medication adherence) to emphasize the importance of a balanced diet.

[0135] 5. From the Pareto Frontier The solution with the largest F(x) is selected as the final solution. This selected solution balances the three objectives while best meeting the user's personalized needs and preferences.

[0136] The decision matrix module 34 is used to construct the decision matrix for candidate solutions. Through normalization and weighted evaluation, a comprehensive score is calculated, and the solution with the highest score is selected as the final intervention solution. The decision matrix is ​​constructed using the following method:

[0137] 1. Construct a decision matrix M, where rows represent different candidate solutions and columns represent different objective function values, for example:

[0138] ,

[0139] The first column represents nutritional balance, the second column represents calorie error, and the third column represents medication adherence. In this example, the first option (0.92, 0.03, 1.0) represents 92% nutritional balance, 3% calorie error, and 100% medication adherence; the second option (0.95, 0.06, 0.9) represents even higher nutritional balance (95%), but slightly larger calorie error (6%) and lower medication adherence (90%).

[0140] 2. Normalize the decision matrix M to eliminate differences in the dimensions and ranges of different objective function values, resulting in a normalized matrix N. Normalization uses the minimax method; that is, for the value aᵢ in the j-th column, the normalization formula is: nᵢ=(aᵢ-min(a)) / (max(a)-min(a)). For the objective that needs to be minimized (such as calorie error), the normalization formula is: nᵢ=(max(a)-aᵢ) / (max(a)-min(a)).

[0141] 3. Construct a weighted evaluation matrix W×N, where W=[ , , [W] is the weight vector, which is dynamically adjusted based on the user's historical data. For example, if the user is most concerned about nutritional balance, the weight vector may be set to W=[0.5,0.3,0.2]; if the user is most concerned about medication adherence, the weight vector may be set to W=[0.3,0.2,0.5].

[0142] 4. Calculate the overall score for each candidate plan and select the plan with the highest score as the final health intervention plan. In family health management, the final selected plan will be transformed into specific implementation suggestions, such as a detailed weekly meal plan (including specific ingredients and portions), a personalized exercise plan (considering user preferences and conditions), and intelligent medication reminders (integrated with the user's daily routine). The system will also dynamically adjust subsequent intervention plans based on user feedback (such as "I didn't eat breakfast today," "I didn't exercise enough," etc.) to ensure the continuity and effectiveness of the intervention.

[0143] like Figure 4 As shown, in a preferred embodiment of the present invention, the digital twin verification subsystem 4 includes a user twin model construction module 41, an environment twin model construction module 42, an interaction twin model construction module 43, a time series simulation module 44, and a health status prediction module 45.

[0144] The user twin model construction module 41 is used to build a virtual user model based on attributes such as age, gender, weight, and underlying diseases. This model employs a parametric modeling method, describing the user's physiological characteristics and health status through a set of parameters. For example, for a 60-year-old hypertensive patient, the model includes parameters such as: age (60 years), gender (male), height (170cm), weight (75kg), baseline systolic blood pressure (145mmHg), baseline diastolic blood pressure (95mmHg), blood glucose level (5.6mmol / L), and metabolic rate (1500kcal / day). These parameters are based on the user's actual measurement data and are standardized before being used to construct the virtual user model. In family health management scenarios, the virtual user model is the foundation for predicting intervention effects. For example, the system can predict the impact of different interventions on blood pressure based on this model: reducing sodium intake may lower systolic blood pressure by 3-5 mmHg, and 30 minutes of aerobic exercise daily may lower systolic blood pressure by 4-7 mmHg. These predictions are based on medical research data and adjusted according to individual user characteristics.

[0145] The environmental twin model construction module 42 is used to build a virtual environment model based on parameters such as layout, temperature, humidity, and air quality of the home environment. This model includes static features of the home environment (such as room layout and furniture placement) and dynamic features (such as temperature changes and air quality fluctuations). For example, a typical home environment model includes: living room temperature (23℃), bedroom temperature (22℃), living room humidity (50%), and PM2.5 concentration (20μg / m³). 3 Parameters such as noise level (45dB) are monitored. These parameters are updated in real time through the environmental monitoring module to ensure the consistency between the virtual environment model and the actual environment. The environmental model is crucial for predicting the impact of environmental factors on health. For example, for a user with a history of allergies, the system can simulate the impact of different environmental improvement measures (such as using an air purifier, regularly cleaning bed sheets, etc.) on indoor allergen levels and predict the degree to which these changes will improve the user's allergy symptoms.

[0146] The interactive twin model building module 43 is used to construct an interactive model describing user behavior patterns and health responses under different environmental conditions based on user-environment interaction data. This model employs a conditional probability network to describe the impact of environmental changes on the user's health status. For example, for hypertensive patients, when the indoor temperature rises from 23℃ to 28℃, blood pressure may increase by 5-10 mmHg; when the PM2.5 concentration increases from 20 μg / m³, blood pressure may also increase. 3 Increased to 80 μg / m 3During this time, the heart rate may increase by 5-8 beats per minute. These interaction rules, based on medical research and user history data, can predict the impact of environmental changes on user health. In family health management, the interaction model can guide environmental interventions. For example, if the system detects that a user's blood pressure rises significantly when the room temperature is below 18°C, it will suggest keeping the room temperature above 20°C in winter and remind the user to wear warm clothing before going out to prevent blood pressure fluctuations.

[0147] The time series simulation module 44 simulates a one-week intervention cycle, sampling user health indicators at six key time points each day. These six key time points are: 7:00 AM, 10:00 AM, 12:00 PM, 3:00 PM, 6:00 PM, and 9:00 PM before bedtime. These time points are chosen based on human circadian rhythms and daily routines, covering key activity periods such as waking up, working, eating, and resting. At each time point, the system simulates changes in the user's health indicators, including blood pressure, heart rate, and blood glucose, forming a complete time series of health status. For example, for a type 2 diabetes patient, the system pays special attention to postprandial blood glucose changes, adding blood glucose sampling points at 12:00 PM and two hours after 6:00 PM (2:00 PM and 8:00 PM) to assess the impact of different dietary plans on postprandial blood glucose. By analyzing blood glucose fluctuation patterns over a week, the system can adjust dietary structure (such as increasing dietary fiber and adjusting carbohydrate distribution) to achieve better blood glucose control.

[0148] The health status prediction module 45 is used to predict changes in a user's health status after the intervention period, based on the user's initial health status and intervention behaviors. This module uses a state transition model, taking the intervention behaviors as input to predict changes in health status. For example, for an overweight user with high blood pressure, if the intervention plan includes limiting sodium intake and 30 minutes of moderate-intensity exercise daily, the model can predict that the user's weight may decrease by 0.5-1 kg and blood pressure may decrease by 3-5 mmHg after one week. The predicted results are compared with preset target thresholds; only when the predicted health improvement indicators exceed the target thresholds is the plan selected as a validated plan. In family health management, health status prediction provides users with an expectation of the intervention's effects, enhancing their confidence and motivation. For example, the system can show users: "If followed as planned, your blood pressure is expected to stabilize within the normal range within two weeks, your weight is expected to decrease by 2-3 kg after one month, and your cardiopulmonary function will significantly improve." This visualized prediction result is more persuasive than abstract health advice and can significantly improve user compliance.

[0149] like Figure 5As shown, in a preferred embodiment of the present invention, the adaptive interaction subsystem 5 includes a medical term recognition module 51, a dialect adaptation module 52, a multilingual support module 53, a personalized push module 54, and an interaction frequency adjustment module 55.

[0150] The medical term recognition module 51 is used to identify medical terms in the user input through medical dictionary matching, context analysis, and model fusion. This module maintains a medical dictionary containing more than 500,000 terms for preliminary matching of medical terms in the user input. At the same time, it uses context analysis technology to understand the meaning of medical terms in different situations and solve the problem of term ambiguity. For example, when the user inputs "My headache", the system can recognize "headache" as a symptom term and combine the user's historical data and current environment to judge possible causes. In the context of home health management, medical term recognition is crucial for understanding user needs. For example, when an elderly user describes "I've been feeling a bit dizzy and unsteady when walking recently", the system can recognize "dizziness" and "unsteady walking" as possible symptoms of vestibular dysfunction or low blood pressure, and then provide corresponding suggestions (such as measuring blood pressure, avoiding sudden standing, etc.) or suggest consulting a doctor. For different user groups, the system can also understand different expressions. For example, the elderly may describe shortness of breath as "qi duan", and children may describe various abdominal discomforts as "stomachache".

[0151] The dialect adaptation module 52 is used to extract the unique acoustic and semantic features of dialects and achieve the conversion between dialects and Mandarin. This module can process voice inputs from major dialect regions in the country, including Cantonese, Minnan dialect, Shanghainese, Sichuan dialect, etc. When it detects that the user uses dialect input, the system first extracts the unique acoustic features of the dialect, then converts them into the standard language form through the dialect-Mandarin mapping model, and finally conducts semantic understanding and processing. This greatly improves the usability of the system in different regions, especially for elderly users and dialect speakers. In home health management, dialect adaptation is particularly important for improving the friendliness of the system to elderly users. For example, in Fujian region, the system can understand health problems expressed in Minnan dialect, such as "我的头那会热" (My head is getting hot), "伊那会茫" (He feels dizzy), and provide corresponding health suggestions. This localized interaction method significantly reduces the threshold of technology use, enabling more elderly people to benefit from the intelligent health management system.

[0152] The multilingual support module 53 provides real-time translation functionality for multiple languages. This module supports input and output in multiple languages, including Chinese, English, Japanese, and Korean, meeting the needs of users with diverse language backgrounds. When a user inputs in a non-Chinese language, the system first performs language recognition, then translates the input into the system's internal standard language (Chinese), and finally translates it back into the user's original language for output. In diverse family environments, multilingual support significantly expands the system's applicability. For example, in a mixed-race family, when the foreign spouse asks in English, "What should I eat to lower my cholesterol?", the system understands the question, queries the knowledge graph for cholesterol-lowering dietary recommendations, and answers in English, such as, "Eating more oats, beans, nuts, and foods rich in omega-3 can help lower your cholesterol levels." This seamless communication greatly enhances the user experience.

[0153] The personalized push module 54 optimizes push timing, adjusts content granularity, and personalizes expression based on user activity patterns, attention patterns, expertise level, and preferences. This module analyzes users' historical interaction data to build user profiles, including user activity patterns (when they are most active), attention patterns (peak attention periods), expertise level (level of medical knowledge), and expression preferences (preferring conciseness or detail). Based on these characteristics, the system optimizes push strategies. For example, it pushes concise health reminders to working professionals during their lunch break (12:00-13:00) and detailed health advice to retired seniors between 9:00-10:00 AM. In family health management scenarios, personalized pushes directly impact the execution rate of intervention plans. For instance, for a busy middle-aged user, the system might push a brief reminder of the day's health plan during their commute (e.g., 7:30 AM), dietary advice during their lunch break (12:30 PM), and an exercise plan reminder before leaving get off work (5:30 PM), dynamically adjusting the reminder content based on the user's actual adherence. For a retired senior, the system may use a more detailed and user-friendly approach and send push notifications during the senior's typically active times (such as 9:00 AM and 3:00 PM). It may also consider the senior's potential memory issues and appropriately increase the frequency of reminders for important matters.

[0154] The interaction frequency adjustment module 55 monitors user response patterns to push notifications and dynamically adjusts the push frequency to avoid user fatigue and resistance. This module records user open rates, execution rates, and feedback on push notifications, analyzing user response patterns. Through learning algorithms, it identifies the push frequency that maximizes user engagement; for example, it might push 3-4 times per day for active users and reduce it to 1-2 times per day for inactive users. Simultaneously, the system sets different push levels based on the importance and urgency of the intervention content. Important and urgent content (such as abnormal blood pressure alerts) is pushed at a high frequency, while general content (such as daily exercise suggestions) is pushed at a regular frequency. In practical applications of the family health management system, an appropriate interaction frequency is crucial for maintaining user engagement. The system learns each user's optimal acceptance frequency. For example, if it finds that a user's engagement (highest open and execution rates) is highest when pushes are sent twice a day, and decreases significantly beyond three times, indicating user fatigue, the system will control regular pushes to no more than twice a day, only adding extra pushes in special circumstances (such as detecting abnormal data). This intelligent frequency adjustment ensures that the system does not burden users while guaranteeing timely delivery of important information.

[0155] In a preferred embodiment of the present invention, the update triggering conditions include: the nutritional balance of the intervention program reaches more than 90%, the calorie error of the exercise program is within ±5%, and the medication adherence reaches 100%.

[0156] These conditions are set based on medical research and user feedback analysis. A 90% nutritional balance is a high but achievable standard, ensuring comprehensive and balanced nutrient intake to meet the body's daily needs. The exercise program's calorie error is within ±5%, meaning the actual calories burned are very close to the target value, avoiding both overexertion leading to excessive fatigue and underexertion failing to achieve the desired results. 100% medication adherence requires users to take medication strictly according to the plan, which is especially important for chronic disease management.

[0157] When these conditions are met, it indicates that the current intervention plan is highly successful, with good user execution and significant effects. At this point, the system triggers a dynamic update of the medical knowledge graph, transforming successful intervention experiences into knowledge to guide the generation of future intervention plans. In practical applications of family health management, this dynamic update mechanism enables the system to continuously learn and evolve. For example, the system might discover that among hypertensive patients in a certain region, adhering to 30 minutes of exercise during the day is more effective than exercising at night; or that for blood sugar control, millet rice is more helpful in stabilizing postprandial blood sugar than white rice. These findings will be added to the knowledge graph to improve the effectiveness of future intervention plans. Simultaneously, the system will also update the knowledge base based on the latest advances in medical research to ensure that health recommendations always align with the latest medical consensus.

[0158] In a preferred embodiment of the present invention, the home intelligent health management system supports three deployment methods: basic version, standard version, and advanced version.

[0159] The basic version provides environmental monitoring, physiological monitoring, and basic health advice. This version is suitable for users in good health who only need routine health management. It can monitor home environmental parameters and user physiological indicators, providing basic health advice such as a balanced diet and moderate exercise. The basic version has low hardware requirements and can run on ordinary home devices, making it suitable for widespread adoption. In practical home applications, the basic version can help healthy families establish good lifestyle habits and prevent chronic diseases. For example, the system monitors indoor PM2.5 concentration, reminding users to open windows for ventilation or use air purifiers when air quality is poor; it monitors user activity levels, encouraging sedentary individuals to get up and move around regularly; and it records users' dietary habits, providing simple nutritional balance suggestions. While these basic functions are simple, they are of significant value in cultivating a healthy lifestyle.

[0160] The Standard version, building upon the basic version's functionality, adds a knowledge graph-driven subsystem and a multi-objective health intervention subsystem. This version is suitable for users with specific health needs (such as weight loss or blood pressure control), generating personalized intervention plans based on the user's health condition and needs, including detailed diet plans, exercise programs, and lifestyle adjustment suggestions. The Standard version requires significant computing power and storage space, making it suitable for families with high demands for health management. In family health management scenarios, the Standard version can meet the needs of most users with chronic health issues. For example, for a patient with mild hypertension, the system will combine their blood pressure data, lifestyle habits, and dietary preferences to generate a personalized DASH diet plan (a low-sodium, high-potassium, and fruit- and vegetable-rich diet) and a suitable exercise plan, dynamically adjusting the plan based on the user's adherence and blood pressure changes. The Standard version's multi-objective optimization function ensures a balance between the effectiveness and feasibility of the intervention plan, significantly improving user compliance and health outcomes.

[0161] The Advanced version, building upon the Standard version's functionality, adds a digital twin verification subsystem and multi-pattern recognition of medical terminology. This version is suitable for patients with chronic diseases or users at higher health risks, enabling predictive verification of intervention plans to ensure their effectiveness, while providing a more natural and intelligent interactive experience. The Advanced version requires substantial computing resources and can be deployed locally or in the cloud, choosing the appropriate deployment mode based on the user's privacy needs and computing resources. The advantages of the Advanced version are particularly evident in complex family health management scenarios. For example, for an elderly patient with both type 2 diabetes and coronary heart disease, the system will build an accurate digital twin model to simulate the combined effects of different intervention plans on blood sugar, blood lipids, and cardiovascular health, identifying the optimal balance. Simultaneously, the Advanced version's multi-pattern recognition of medical terminology allows elderly patients to describe symptoms in natural language (such as "I've been feeling numbness in the soles of my feet lately"), which the system can understand as potentially an early symptom of diabetic peripheral neuropathy, providing corresponding suggestions or recommendations for medical attention. This advanced predictive and interactive capability significantly improves the accuracy of health management and the user experience.

[0162] like Figure 6 As shown, the home intelligent health management method provided by the present invention includes the following steps:

[0163] Step S1: Collect home environmental parameters and user physiological indicators. Specifically, environmental data is collected every 30 seconds using PM2.5 and temperature sensors; user physiological indicators such as heart rate, blood pressure, blood oxygen, and body temperature, as well as behavioral data such as sleep, activity, and stress, are collected via smart bracelets and smartphones. In daily applications of home health management, this step ensures that the system can comprehensively understand the user's health status and living environment. For example, when the system detects that the PM2.5 concentration in the user's living environment continuously exceeds 50 μg / m³, the system will collect data. 3 At the same time, if the user's heart rate increases by 5-10 beats per minute compared to normal, it may indicate that environmental pollution has affected the user's cardiopulmonary function. The system will remind the user to open windows for ventilation or reduce outdoor activities.

[0164] Step S2: Based on environmental parameters and user physiological indicators, construct a three-layer medical knowledge graph comprising an entity layer, a relation layer, and a knowledge layer. Specifically, entity names in medical literature are converted into standard medical identifiers using entity recognition technology; a medical knowledge graph comprising entity, relation, and knowledge layers is established; and relevant medical knowledge is retrieved based on the user's health status and needs to support the generation of intervention plans. In practical applications, the knowledge graph is the foundation for personalized intervention. For example, when the system detects elevated blood pressure in a user, it retrieves all nodes related to hypertension from the knowledge graph, including possible causes (such as high-sodium diet, lack of exercise, high stress, etc.), potential risks (such as cardiovascular and cerebrovascular events), and effective intervention measures (such as the DASH diet, aerobic exercise, stress management, etc.), and then selects the most appropriate intervention strategy based on the user's specific situation.

[0165] Step S3: Based on a medical knowledge graph, a set of candidate intervention plans is generated by setting multi-objective functions for nutritional balance, energy balance, and medication adherence. Specifically, nutritional balance (over 90% nutritional balance), energy balance (error between actual and target calorie consumption within ±5%), and medication adherence (100% medication adherence) are set; time constraints, health impact constraints, nutritional balance constraints, and energy consumption constraints are set; the candidate plans are non-dominatedly ranked to identify the Pareto optimal solution set; a decision matrix for the candidate plans is constructed, and a comprehensive score is calculated through normalization and weighted evaluation. The plan with the highest score is selected as the final intervention plan. In family health management, multi-objective optimization ensures the comprehensive effectiveness of the intervention plan. For example, for a middle-aged woman who needs to lose weight and has mild hypertension, the system must consider not only the weight loss goal (controlling energy intake) but also blood pressure control (low-sodium diet), as well as the user's personal preferences and time constraints. Through multi-objective optimization, the system can generate a solution that balances various needs, such as recommending low-calorie, low-sodium but palatable recipes, as well as exercise plans that combine user interests (such as dance) and time constraints.

[0166] Step S4: Establish user twin models, environment twin models, and interaction twin models to perform predictive validation on the candidate intervention scheme set, and select schemes whose predicted health improvement indicators are higher than the target threshold as validated schemes. Specifically, a virtual user model is constructed based on the user's age, gender, weight, underlying diseases, and other attributes; a virtual environment model is constructed based on parameters such as the layout, temperature, humidity, and air quality of the home environment; an interaction model is constructed based on the user's interaction data with the environment; a one-week intervention cycle is simulated, and user health indicators are sampled at six key time points each day; based on the user's initial health status and intervention behavior, the changes in the user's health status after the intervention cycle ends are predicted. Digital twin validation is a major innovation of this invention, as it makes the intervention effect predictable, greatly improving the safety and effectiveness of the intervention. For example, for a diabetic patient, the system can simulate the impact of different dietary plans on their blood sugar, compare the differences between whole wheat bread and regular bread, and brown rice and white rice, to find the most suitable dietary structure for the user. This predictive validation avoids the risk of blood sugar fluctuations that may occur during trial and error, while providing users with clear expectations and enhancing their motivation to implement the plan.

[0167] Step S5: Based on multi-pattern recognition of medical terminology and user profile analysis, the verified solution is pushed to the user through a personalized push strategy. Specifically, medical terms in user input are identified through medical dictionary matching, context analysis, and model fusion; dialect-specific acoustic and semantic features are extracted to achieve dialect-to-Mandarin conversion; real-time translation functions for multiple languages ​​are provided; push time, content granularity, and personalized expression are optimized based on user activity patterns, attention patterns, professional level, and preferences; and the push frequency is dynamically adjusted by monitoring user response patterns to push information. Adaptive interaction greatly enhances the user experience and accessibility of the system. For example, for an elderly user who uses a lot of dialect, the system can understand the health problem expressed in their dialect and respond in a way that is familiar to the elderly, such as changing "Your systolic blood pressure is high, it is recommended to limit sodium intake" to "Your blood pressure is a little high, use less salt and eat less pickled vegetables, it's good for your health." At the same time, the system will choose to push information during the time when the elderly are most likely to receive information (such as 8:00-9:00 in the morning) according to their daily routine and attention patterns, and use larger fonts and concise language.

[0168] Step S6: Collect user feedback information and trigger a dynamic update of the medical knowledge graph when preset update trigger conditions are met. Specifically, collect user feedback on the implementation of the intervention plan and its health effects; trigger a knowledge graph update when the nutritional balance of the intervention plan reaches over 90%, the calorie error of the exercise plan is within ±5%, and medication adherence reaches 100%; perform entity updates, relationship updates, attribute updates, and data updates to ensure the timeliness and accuracy of medical knowledge. Closed-loop feedback is key to continuous system optimization. In long-term application of family health management, the system will accumulate a large amount of user feedback data, learning patterns and rules of intervention effects from it. For example, the system may find that in users of a certain age group, morning fasted exercise is more effective for weight loss than post-meal exercise; or it may find that certain food combinations (such as a specific ratio of protein to dietary fiber) are particularly effective for blood sugar control. These findings will be integrated into the system through the dynamic update mechanism of the knowledge graph, continuously improving the accuracy and personalization of the intervention plan.

[0169] The above embodiments are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A home intelligent health management system, characterized in that, The system comprises: a data acquisition subsystem comprising an environment monitoring module and a physiological monitoring module for acquiring home environment parameters and user physiological indicators; a knowledge graph driven subsystem connected to the data acquisition subsystem for receiving the environment parameters and the user physiological indicators, constructing a three-layer structure medical knowledge graph comprising an entity layer, a relationship layer and a knowledge layer, and dynamically updating the medical knowledge graph based on a preset update trigger condition; a multi-objective health intervention subsystem connected to the knowledge graph driven subsystem for generating a candidate intervention scheme set based on the medical knowledge graph by setting a multi-objective function of a nutrition balance target, an energy balance target and a medication compliance target; a digital twin verification subsystem connected to the multi-objective health intervention subsystem for establishing a user twin model, an environment twin model and an interaction twin model, predictively verifying the candidate intervention scheme set, and selecting a scheme with a predicted health improvement index higher than a target threshold as a verified scheme; an adaptive interaction subsystem connected to the digital twin verification subsystem for receiving the verified scheme, pushing the verified scheme to the user based on medical term multi-modal recognition and user portrait analysis through an individualized pushing strategy, and collecting user feedback information; wherein the digital twin verification subsystem is further connected to the knowledge graph driven subsystem for transmitting the user feedback information to the knowledge graph driven subsystem to trigger dynamic updating of the medical knowledge graph; the multi-objective health intervention subsystem comprises: a target function setting module for setting a multi-objective function of a nutrition balance target, an energy balance target and a medication compliance target; a constraint condition setting module for setting time constraints, health impact constraints, nutrition balance constraints and energy consumption constraints; an optimization algorithm module for non-dominant sorting of candidate schemes based on the multi-objective function and the constraint conditions, and identifying a Pareto optimal solution set; a decision matrix module for constructing a decision matrix of candidate schemes, calculating a comprehensive score through normalization processing and weighted evaluation, and selecting the scheme with the highest score as the final intervention scheme.

2. The home intelligent health management system of claim 1, wherein, The knowledge graph driven subsystem comprises: an entity recognition module for converting entity names in medical literature into standard medical identifiers; a three-layer structure construction module for establishing a medical knowledge graph comprising an entity layer, a relationship layer and a knowledge layer, wherein the entity layer comprises drug, symptom, disease, department, equipment, food material, acupoint and action entities, the relationship layer comprises attribute relationships and inter-entity relationships, and the knowledge layer comprises diagnosis rules, treatment schemes and health suggestions; an adaptive updating module for performing entity updating, relationship updating, attribute updating and data updating when a preset update trigger condition is met; a knowledge query module for retrieving relevant medical knowledge based on user health conditions and needs to support intervention scheme generation.

3. The home intelligent health management system of claim 1, wherein, The environment monitoring module comprises: a high-frequency sampling unit for acquiring home environment data every 30 seconds through a PM2.5 sensor and a temperature sensor; The convolution feature extraction unit is configured to perform convolution processing and pooling processing on the environment data to extract local environment features. The environment grading evaluation unit is configured to compare the extracted environment features with preset health standards to divide the environment state into three grades of low risk, medium risk and high risk. The early warning triggering unit is configured to generate early warning information when the environment parameters exceed preset thresholds and incorporate the early warning information into the intervention scheme.

4. The home intelligent health management system of claim 1, wherein, The physiological monitoring module includes: The multi-source data acquisition unit is configured to acquire physiological indicators of a user, such as heart rate, blood pressure, blood oxygen and body temperature, and behavior data such as sleep, activity and stress through a smart bracelet and a smart phone. The time sequence feature extraction unit is configured to process the physiological indicators and the behavior data through a time sequence feature extraction network to capture time dependence and important features. The health record construction unit is configured to construct a health record recording a health baseline, fluctuation range and abnormal pattern of the user based on the processed physiological data.

5. The home intelligent health management system of claim 1, wherein, The digital twin verification subsystem includes: The user twin model construction module is configured to construct a virtual user model based on attributes of the user, such as age, gender, weight and underlying diseases. The environment twin model construction module is configured to construct a virtual environment model based on parameters of the home environment, such as layout, temperature, humidity and air quality. The interactive twin model construction module is configured to construct an interactive model describing behavior patterns and health responses of the user under different environmental conditions based on interactive data of the user and the environment. The time sequence simulation module is configured to simulate an intervention period of one week and sample health indicators of the user at six key time points each day. The health state prediction module is configured to predict changes in the health state of the user after the intervention period based on the initial health state of the user and intervention behaviors.

6. The home intelligent health management system of claim 1, wherein, The adaptive interaction subsystem includes: The medical term recognition module is configured to recognize medical terms in user input through medical dictionary matching, context analysis and model fusion. The dialect adaptation module is configured to extract acoustic and semantic features specific to dialects to realize conversion between dialects and standard Chinese. The multi-language support module is configured to provide real-time translation functions in multiple languages. The personalized push module is configured to optimize push time, adjust content granularity and personalized expression modes based on activity patterns, attention patterns, professional levels and preferences of the user. The interaction frequency adjustment module is configured to monitor response patterns of the user to push information, dynamically adjust push frequency and avoid user fatigue and resistance.

7. The home intelligent health management system of claim 1, wherein, The update triggering condition includes that a nutrition balance degree of the intervention scheme reaches 90% or more, a calorie error of a movement scheme is within a range of plus or minus 5%, and a medication compliance reaches 100%.

8. The home intelligent health management system of claim 1, wherein, The home intelligent health management system supports three deployment modes of a basic version, a standard version and an advanced version. The basic version realizes environment monitoring, physiological monitoring and basic health suggestion functions. The standard version adds a knowledge graph driven subsystem and a multi-target health intervention subsystem based on the functions of the basic version. The advanced version adds a digital twin verification subsystem and a medical term multi-mode recognition function based on the functions of the standard version.

9. A method for home intelligent health management for non-disease treatment or diagnosis purpose, using the home intelligent health management system according to any one of claims 1-8, characterized in that, The method includes: Acquiring home environment parameters and user physiological indicators. construct a three-layer structure medical knowledge graph including an entity layer, a relationship layer and a knowledge layer based on the environmental parameters and the user physiological indicators; generate a candidate intervention scheme set based on the medical knowledge graph by setting a multi-objective function of a nutrition balance target, an energy balance target and a medication compliance target; establish a user twin model, an environment twin model and an interaction twin model, predictively verify the candidate intervention scheme set, and select a scheme with a predicted health improvement indicator higher than a target threshold as a verified scheme; push the verified scheme to the user through an individualized push strategy based on medical term multi-mode recognition and user portrait analysis; collect user feedback information, and trigger dynamic updating of the medical knowledge graph when a preset updating trigger condition is met; wherein the updating trigger condition includes that the nutrition balance degree of the intervention scheme reaches 90% or above, the calorie error of the exercise scheme is within a range of plus or minus 5%, and the medication compliance reaches 100%.

Images