Intelligent health diet guidance system for the elderly

Abstract

The application discloses a kind of wisdom old-age health diet guidance system, comprising: for through multi-device cooperation collection old person's full-dimension health data, and the multi-source data acquisition module of the data collected is preprocessed;For using the multi-source health data of preprocessed diet demand analysis module is analyzed old person's dietary needs;For the individualized recipe generation module of old person's dietary needs input using pre-trained recipe generation model, generate personalized recipe in line with the needs of old people;For real-time data feedback, dynamically optimize the recipe optimization adjustment module of the personalized recipe.Its remarkable effect is: the overall perception of the health status of old people, accurate positioning of dietary needs, intelligent generation of personalized recipes, dynamic optimization of the scheme and convenient guidance output can be realized, which can solve the problem of one-sidedness, poor adaptability and insufficient dynamics of existing system data.

Description

Technical Field

[0001] This invention relates to the field of smart elderly care and health management technology, specifically to a smart elderly care healthy diet guidance system. Background Technology

[0002] With the increasing aging of the population, smart elderly care has become an important development direction, and healthy diet management is key to ensuring the physical and mental health of the elderly. Current elderly care diet guidance systems often suffer from the following problems: health data sources are limited, failing to integrate multi-dimensional information such as medical history, medication, and physical function, resulting in a lack of comprehensiveness in diet guidance; diet plans lack personalization, ignoring practical factors such as the elderly's dietary preferences, regional food differences, and religious variations, leading to low compliance; diet plan adjustments are lagging, unable to be updated in real time according to health status and seasonal changes; nutritional calculations are not precise enough, ignoring the impact of micronutrients and cooking methods; and the interaction methods are complex, making them inconvenient for the elderly to operate.

[0003] It is evident that existing systems largely fail to provide dietary guidance solutions that integrate multi-source data, construct personalized models, dynamically adapt and adjust, and offer convenient interaction. They also lack targeted system designs for the specific needs of the elderly. Therefore, there is an urgent need for a system capable of providing precise, personalized, and dynamic dietary guidance. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the purpose of this invention is to provide a smart elderly health and diet guidance system that can achieve comprehensive perception of the elderly's health status, accurate positioning of dietary needs, intelligent generation of personalized recipes, dynamic optimization of plans, and convenient guidance output. This system can solve the problems of incomplete data, poor adaptability, and insufficient dynamism in existing systems.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A smart elderly care health diet guidance system, the key of which includes:

[0007] The multi-source data acquisition module is used to collect comprehensive health data of the elderly through multiple devices and to preprocess the collected data.

[0008] The dietary needs analysis module is used to assess health status using preprocessed multi-source health data and analyze the dietary needs of the elderly in conjunction with dietary-related rules.

[0009] The personalized recipe generation module is used to input the dietary needs of the elderly into a pre-trained recipe generation model to generate personalized recipes that meet the needs of the elderly.

[0010] The recipe optimization and adjustment module is used to dynamically optimize the personalized recipe based on real-time data feedback.

[0011] The interactive output module is used to provide various forms of portable interaction and personalized healthy eating guidance based on the elderly's usage habits and personalized recipes.

[0012] Furthermore, the multi-source data acquisition module includes:

[0013] A health monitoring unit is used to collect core health monitoring data;

[0014] The body function detection unit is used to obtain data on the elderly's chewing ability and digestive function;

[0015] The environmental and behavioral data acquisition unit is used to acquire food reserve data and consumption data, as well as the local seasonal food supply situation;

[0016] The data transmission unit is used to build a communication network to aggregate and upload data collected by the health monitoring unit, the physical function detection unit, and the environmental and behavioral data acquisition unit.

[0017] Furthermore, the core health monitoring data includes basic physical data, disease data, exercise data, dietary data, and medication data.

[0018] Furthermore, the dietary needs analysis module includes:

[0019] The health assessment indicator determination unit is used to determine the health indicators used to assess the health status of the elderly, and the weight of each health indicator is determined by the analytic hierarchy process.

[0020] The health status assessment unit is used to construct a health status assessment model based on the determined health indicators and their corresponding weights, and output the health status score of the elderly.

[0021] The personalized dietary needs extraction unit is used to calculate nutritional needs based on health status scores, extract dietary constraint information, match regional needs, and generate personalized dietary needs.

[0022] Furthermore, the health indicators include physiological indicators, physical function indicators, past medical history indicators, and medication indicators. The expression of the health status assessment model is as follows:

[0023]

[0024] in, As the indicator weight, These are the standardized indicator values ​​for each health indicator.

[0025] Furthermore, the personalized recipe generation module includes:

[0026] The recipe generation model building unit is used to build a recipe generation model based on the elderly’s health data, personalized dietary needs, and regional ingredient information.

[0027] The recipe generation model training unit is used to train the recipe generation model using elderly diet recipe data, optimize the loss function, and obtain the pre-trained recipe generation model.

[0028] The recipe generation unit is used to input the target elderly person's health data, personalized dietary needs, and regional ingredient information into the pre-trained recipe generation model, and output personalized recipes with nutritional annotations.

[0029] Furthermore, the recipe generation unit needs to perform the following steps before generating personalized recipes:

[0030] Ingredient selection: Based on nutritional needs and a regional ingredient database, suitable ingredients are selected;

[0031] Meal plan design: Design the meal plan structure for each day's meals and snacks, taking into account both nutritional balance and dietary diversity;

[0032] Cooking method selection: Each personalized recipe includes information on calories, macronutrients, core micronutrients, and how it is suited to your health needs.

[0033] Furthermore, the recipe optimization and adjustment module includes:

[0034] The feedback data collection unit is used to collect data on the target elderly person's health status, dietary adherence, changes in local food supply, and environmental parameters.

[0035] The recipe dynamic optimization unit is used to adjust personalized recipes based on data collected by the feedback data collection unit.

[0036] Furthermore, the recipe dynamic optimization unit adjusts the personalized recipe in the following way:

[0037] Health-driven adjustments: If health indicators are abnormal, the diet will be automatically adjusted with an adjustment response time of ≤1 hour;

[0038] Execution-driven adjustment: If a certain food ingredient is not consumed twice consecutively, it will be replaced with a nutritionally similar alternative food ingredient;

[0039] Weekly optimization and adjustment: The menu is comprehensively optimized weekly based on health data and performance feedback;

[0040] Emergency Adaptation: If an elderly person suddenly falls ill, a temporary diet plan will be automatically generated.

[0041] Furthermore, the interactive output module includes:

[0042] A multimodal interaction unit is used to enable voice and visual interaction between the elderly, their families, and the system.

[0043] The recipe content output unit is used to output personalized recipes and their corresponding ingredient purchase lists and dietary reminders on a daily schedule.

[0044] The significant effects of this invention are:

[0045] 1. This invention, through multi-source data fusion and personalized recipe generation, significantly improves the nutritional accuracy and health indicator matching rate of recipes compared to traditional technologies, effectively enhancing the accuracy and personalization of dietary guidance.

[0046] 2. This invention effectively enhances elderly people's compliance and program execution rate by adapting to dietary preferences, regional ingredients, and physical functions;

[0047] 3. This invention ensures nutritional balance by dynamically optimizing personalized recipes, enabling real-time dynamic adjustment of recipes and timely response to changes in health.

[0048] 4. The system's interaction method is age-friendly and easy to operate; it is suitable for deployment in various elderly care scenarios such as home, community, and institutions, and has strong scalability. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the structure of the present invention;

[0050] Figure 2 This is a schematic diagram of the multi-source data acquisition module;

[0051] Figure 3 This is a structural diagram of the dietary needs analysis module;

[0052] Figure 4 This is a structural diagram of the personalized recipe generation module. Detailed Implementation

[0053] The specific embodiments and working principles of the present invention will be further described in detail below with reference to the accompanying drawings.

[0054] Example:

[0055] like Figure 1 As shown, a smart elderly care health and diet guidance system is described. This system, through a full-process architecture comprising a multi-source health data perception module, an intelligent dietary needs analysis module, an AIGC personalized recipe generation module, a dynamic adaptation and adjustment module, and a convenient interactive output module, achieves comprehensive perception of the elderly's health status, precise positioning of their dietary needs, intelligent generation of personalized recipes, dynamic optimization of plans, and convenient guidance output. Specifically, it includes:

[0056] The multi-source data acquisition module is used to collect comprehensive health data of the elderly through multiple devices and to preprocess the collected data.

[0057] The dietary needs analysis module is used to assess health status using preprocessed multi-source health data and analyze the dietary needs of the elderly in conjunction with dietary-related rules.

[0058] The personalized recipe generation module is used to input the dietary needs of the elderly into a pre-trained recipe generation model to generate personalized recipes that meet the needs of the elderly.

[0059] The recipe optimization and adjustment module is used to dynamically optimize the personalized recipe based on real-time data feedback.

[0060] The interactive output module is used to provide various forms of portable interaction and personalized healthy eating guidance based on the elderly's usage habits and personalized recipes.

[0061] In some implementations, the system also includes a data security management module, which is used to implement encrypted data storage, access control of the system, encrypted communication between the system and client terminals and between the system and cloud servers, and to protect the privacy of the elderly.

[0062] Based on the above description, it can be seen that the system described in this embodiment achieves comprehensive perception of the elderly's health status, accurate positioning of dietary needs, intelligent generation of personalized recipes, dynamic optimization of plans, and convenient guidance output, which can solve the problems of incomplete data, poor adaptability, and insufficient dynamism in existing systems. Furthermore, it protects the privacy and data security of the elderly through end-to-end encryption and access control.

[0063] In this embodiment, the multi-source data acquisition module collects comprehensive health data of the elderly through multi-device collaboration, providing data support for dietary guidance. (See Appendix) Figure 2 This module specifically includes:

[0064] A health monitoring unit is used to collect core health monitoring data;

[0065] In practical implementation, the core health monitoring data includes basic physical data, disease data, exercise data, dietary data, and medication data, specifically:

[0066] Basic physical data: heart rate, blood pressure, blood sugar, weight, body fat percentage, muscle mass, etc.

[0067] Patient data: medical record data;

[0068] Activity data: Steps;

[0069] Dietary data: Dietary restrictions

[0070] Medication data: type of medication, dosage, and time of administration.

[0071] The body function detection unit is used to detect the elderly's chewing ability through a bite force sensor and to monitor the elderly's digestive function data through intestinal flora test strips and portable detectors.

[0072] The environmental and behavioral data acquisition unit is used to acquire food reserve data and consumption data, and to obtain local seasonal food supply information through the regional food database.

[0073] The data transmission unit is used to build a Bluetooth + 5G communication network. Near-field devices collect data via Bluetooth, while long-distance devices upload data to the cloud server via 5G, realizing the aggregation and uploading of data collected by the health monitoring unit, body function detection unit, and environmental and behavioral data acquisition unit.

[0074] A data preprocessing unit is used to preprocess the received data. In this embodiment, the data preprocessing includes:

[0075] Data cleaning: Outlier data (such as sudden changes in blood pressure) was removed using the 3σ criterion and the Isolation Forest algorithm, with an anomaly detection rate of ≥98%;

[0076] It should be noted that the process of removing outliers using the 3σ criterion and the Isolation Forest algorithm is as follows: Verify whether the data conforms to a normal distribution using the Shapiro-Wilk test (sample size ≤ 5000) or the Kolmogorov-Smirnov test (sample size > 5000); for normally distributed data, calculate the mean μ and standard deviation σ; set a threshold interval [μ-3σ, μ+3σ]; iterate through the data, marking samples outside the interval as outliers and removing them; if the data has a slight skewness (e.g., right-skewed blood glucose data), a logarithmic transformation can be performed to convert the data to an approximate normal distribution before applying the 3σ criterion. Furthermore, the core of Isolation Forest is to quickly "isolate" outliers (a minority of samples deviating from the group) by randomly partitioning the feature space, without assuming a data distribution, making it suitable for anomaly detection in non-normal / complex correlated data. Normal data processed by the 3σ criterion is integrated with non-normal data that has not undergone 3σ into a feature matrix D (shape = (number of samples, feature dimension)). An isolated forest model is constructed, and core parameters are set. The model is trained and sample labels are predicted (1 = normal, -1 = abnormal). Abnormal samples marked as -1 are removed to obtain the final cleaned data. The outlier indices identified by 3σ and isolated forest are merged to avoid duplicate removal. For key outliers (such as sudden blood pressure changes), the actual health status of the elderly is verified (e.g., whether it is a measurement error or a sudden illness) to ensure that valid data is not mistakenly deleted. The mean, standard deviation, and distribution characteristics of the cleaned data are calculated to confirm that there are no obvious abnormal fluctuations. The health background of the elderly (e.g., whether there is a history of hypertension) is considered to verify whether the outliers represent the true health status (e.g., sudden hypertension). If they are valid data, they are re-included. The algorithm's effectiveness is verified through labeled outlier samples (e.g., known measurement error data) to ensure that the outlier identification rate is ≥98% (meeting system requirements).

[0077] The above steps quickly screen out obvious anomalies using 3σ and accurately identify latent anomalies using isolated forest, achieving an anomaly identification rate of ≥98%, which meets system requirements. The algorithm automatically adapts to different health data distribution characteristics, reducing the manual verification costs for elderly care institutions / family members. Through the business verification process, it avoids mistakenly deleting abnormal values ​​of the elderly's true health status (such as sudden hypertension), ensuring the accuracy of dietary guidance.

[0078] Missing value completion: K-nearest neighbor algorithm (K=5) is used to complete randomly missing data, and LSTM model is used to complete continuously missing data (missing rate ≤30%), with completion accuracy ≥92%;

[0079] In practice, the missing value completion steps described above are as follows:

[0080] First, the collected health time-series data (such as blood pressure, heart rate, and blood sugar collected hourly) are standardized in format, missing values ​​are marked, and the missing types are distinguished.

[0081] Then, each sample is treated as a point in the feature space. For a sample with missing values, the K nearest neighbors of the missing values ​​are found, and the missing values ​​are filled with the mean / median of the corresponding features of these K samples. Dimensions that are strongly correlated with the missing features are selected (e.g., when filling in missing blood pressure values, features such as heart rate, blood sugar, age, and body fat percentage are selected), and normalization is performed to eliminate the influence of dimensions. The value of K is set to 5. For each feature column containing random missing values, a KNN model is trained and the missing values ​​are filled.

[0082] Finally, the LSTM model captures the short-term and long-term dependencies of time series data through a gating mechanism, making it suitable for completing continuously missing health time series data: convert one-dimensional time series data into a sliding window input format (e.g., use heart rate data from the previous 6 hours to predict the heart rate in the 7th hour), with the recommended window length being twice the duration of continuous missing data; calculate the mean squared error (MSE) and mean absolute error (MAE) between the completed value and the true value (if any); and combine this with the health status of the elderly (e.g., whether the blood pressure completed value for elderly people with hypertension is within a reasonable range) to ensure that the completed value conforms to medical common sense and avoids obviously unreasonable values.

[0083] The missing value completion process described above takes into account the temporal characteristics and differences in missing value types of elderly health data, and achieves "categorized and high-precision" missing data completion, ensuring the accuracy of subsequent dietary needs analysis and recipe generation.

[0084] Standardization processing: Convert data from different devices into a unified format and store them according to health indicator categories (physiological indicators, physical functions, medication information, etc.).

[0085] In this embodiment, the structural diagram of the dietary needs analysis module is as follows: Figure 3 As shown, it specifically includes:

[0086] The health assessment indicator determination unit is used to determine the health indicators used to assess the health status of the elderly, and the weight of each health indicator is determined by the analytic hierarchy process.

[0087] In practice, the health indicators include physiological indicators, physical function indicators, past medical history indicators, and medication indicators, with corresponding weights of 0.4, 0.3, 0.2, and 0.1, respectively.

[0088] The health status assessment unit is used to construct a health status assessment model based on the determined health indicators and their corresponding weights, and output the health status score of the elderly. The health status score is 0-100 points, with a score <60 indicating a high-risk state, 60-80 indicating a normal state, and >80 indicating a healthy state.

[0089] The expression for the health status assessment model is as follows:

[0090]

[0091] in, As the indicator weight, These are the standardized indicator values ​​for each health indicator.

[0092] The personalized dietary needs extraction unit is used to calculate nutritional needs based on health status scores, extract dietary constraint information, match regional needs, and generate personalized dietary needs.

[0093] It should be noted that the specific steps for personalized dietary needs are as follows:

[0094] Nutritional requirements calculation: Based on health score, age, gender, and activity level, the modified Mifflin-StJeor formula is used to calculate the daily calorie requirement. The macronutrient ratio (e.g., carbohydrate ratio ≤45% for elderly people with diabetes) and micronutrient requirements (e.g., calcium intake ≥1000mg / day for elderly people with osteoporosis) are determined in combination with health status.

[0095] It should be noted that the specific process of calculating daily calorie requirements using the modified Mifflin-StJeor formula is as follows:

[0096] The traditional Mifflin-St. Jeor formula calculates basal metabolic rate (BMR) based solely on age, sex, height, weight, and activity level, then multiplies it by an activity factor to obtain total daily energy requirement (TDEE). The improved Mifflin-St. Jeor formula is optimized for the physiological characteristics of older adults (60+ years old), with the following core additions:

[0097] Body function correction items (chewing / digestion ability, muscle mass): correct the difference in calorie consumption caused by functional decline;

[0098] Health status correction items (chronic diseases, medication): adapted to the metabolic characteristics of elderly people with hypertension / diabetes, etc.

[0099] Regional climate correction (e.g., higher heat demand in northern winters): adapts to differences in energy consumption across different regions.

[0100] The improved Mifflin-StJeor formula consists of two steps:

[0101] Step 1: Calculate Basal Metabolic Rate (BMR):

[0102] Male: BMR=(10×W+6.25×H-5×A+5)×α×β×γ;

[0103] Female: BMR = (10 × W + 6.25 × H - 5 × A − 161) × α × β × γ;

[0104] Where: W is weight (kg), H is height (cm), A is age (years); α is the body function correction coefficient (0.85~1.1); β is the health status correction coefficient (0.9~1.05); γ is the regional climate correction coefficient (0.95~1.05).

[0105] Step 2: Calculate your total daily energy requirement (TDEE):

[0106] TDEE = BMR × AF

[0107] AF is the activity coefficient for the elderly, which is used to optimize for elderly care scenarios and avoid the traditional coefficient being too high. Its values ​​are as follows: 1.0 for completely bedridden (such as disabled elderly); 1.2 for light activity (less walking at home); 1.375 for moderate activity (1 hour of walking per day); and 1.55 for heavy activity (more activity in community elderly care).

[0108] Dietary restrictions include: disease restrictions (e.g., elderly people with hypertension should have a low salt intake of ≤5g / day), medication restrictions (e.g., elderly people taking antihypertensive drugs should avoid high-potassium foods), physical function restrictions (e.g., elderly people with weak chewing ability should have food softness ≥8), and preference and taboo restrictions (e.g., elderly vegetarians should avoid animal protein foods).

[0109] Regional adaptation requirements: By combining regional food databases, priority is given to local seasonal ingredients (abundant supply and good nutrient retention) to reduce the difficulty of implementing the diet plan.

[0110] This embodiment achieves the adaptation of the recipe to dietary preferences, regional ingredients and physical functions through the above steps, effectively enhancing the compliance and implementation rate of the program among the elderly.

[0111] In this embodiment, the structure of the personalized recipe generation module is as follows: Figure 4 As shown, it specifically includes:

[0112] The recipe generation model building unit is used to build a recipe generation model based on the elderly’s health data, personalized dietary needs, and regional ingredient information.

[0113] In practice, the model architecture of the recipe generation model is based on the GPT-4V+Stable Diffusion hybrid model. It takes health data, dietary needs, and regional ingredient information as input and outputs a complete solution including text recipes and visual images.

[0114] The recipe generation model training unit is used to train the recipe generation model using elderly diet recipe data, optimize the loss function, and obtain the pre-trained recipe generation model.

[0115] In practice, the model is trained using 100,000 elderly diet recipes (including annotations on disease compatibility, nutritional ratios, cooking methods, etc.) and the loss function is optimized (cross-entropy loss + nutritional bias loss) to ensure that the nutritional accuracy of the recipes is ≥95%.

[0116] The recipe generation unit is used to input the target elderly person's health data, personalized dietary needs, and regional ingredient information into the pre-trained recipe generation model, and output personalized recipes with nutritional annotations.

[0117] In specific implementation, the process by which the recipe generation unit generates personalized recipes with nutritional annotations is as follows:

[0118] Ingredient selection: Based on nutritional needs and a regional ingredient database, suitable ingredients are selected, with priority given to seasonal ingredients.

[0119] Meal plan design: The daily meal plan includes breakfast, lunch, dinner and snacks, taking into account nutritional balance and dietary diversity, ensuring that there are ≥30 kinds of ingredients per week, and avoiding repeating the same ingredients for 3 consecutive days;

[0120] Cooking methods should be appropriate for the elderly: Choose gentle cooking methods such as steaming, boiling, stewing, and braising, based on their chewing and digestion abilities. Clearly define the requirements for food preparation, such as cutting meat into small pieces ≤1cm and cooking vegetables until soft.

[0121] Nutrition labeling: Each recipe includes the content of calories, macronutrients, and core micronutrients, as well as instructions on its suitability for health needs, such as this recipe being low in salt and fat and suitable for elderly people with high blood pressure.

[0122] The embodiments of the present invention, through the above-mentioned personalized recipe generation, significantly improve the nutritional accuracy and health indicator matching rate of the recipes compared with traditional technologies, effectively enhancing the accuracy and personalization of dietary guidance.

[0123] In this embodiment, the recipe optimization and adjustment module includes:

[0124] The feedback data collection unit is used to collect data on the target elderly person's health status, dietary adherence, changes in local food supply, and environmental parameters.

[0125] Health status feedback: Real-time monitoring of changes in health indicators through smart wearable devices;

[0126] Dietary adherence feedback: The smart kitchen devices record the adherence to the recipes, such as whether a certain ingredient was not consumed; seniors can provide feedback on their dining experience via voice assistant, such as "too tough" or "tastes bad."

[0127] External factor feedback: Real-time synchronization of changes in regional food supply: such as food shortages; seasonal changes: such as increasing the intake of cooling foods in summer.

[0128] The recipe dynamic optimization unit is used to adjust personalized recipes based on data collected by the feedback data collection unit.

[0129] In practical implementation, the recipe dynamic optimization unit adjusts personalized recipes in the following way:

[0130] Health-driven adjustments: If health indicators are abnormal, the diet will be automatically adjusted, such as reducing high-sodium foods and increasing blood pressure-lowering foods such as celery and black fungus. The adjustment response time is ≤1 hour.

[0131] Execution-driven adjustment: If a certain food ingredient is not consumed twice in a row, it will be replaced with a nutritionally similar alternative food ingredient, such as using tofu instead of eggs to supplement protein.

[0132] Weekly optimization and adjustment: The menu is comprehensively optimized weekly based on health data and performance feedback;

[0133] Emergency Adaptation: If an elderly person suddenly falls ill, a temporary dietary plan will be automatically generated, such as light and easily digestible porridge or soup.

[0134] The embodiments of the present invention dynamically optimize personalized recipes through the above-described method, ensuring nutritional balance and enabling real-time dynamic adjustment of recipes, which can respond promptly to changes in health.

[0135] In this embodiment, the interactive output module includes:

[0136] A multimodal interaction unit is used to enable voice and visual interaction between the elderly, their families, and the system.

[0137] Specifically, the multimodal interaction unit supports dialect recognition, enabling the elderly to query recipes and provide feedback through voice; the visual interaction uses a large font and high contrast interface, and supports enlarging recipe images to make it easier for the elderly to see the recipe content. It also allows family members to view the elderly's diet plan and implementation status on the APP or in the cloud, and remotely adjust the recipe.

[0138] The recipe content output unit is used to generate personalized recipes and their corresponding ingredient shopping lists and dietary reminders daily at set times. For example, it can push the recipe for the next day at 8:00 PM every day, supporting three formats: text, images, and video tutorials (visualized cooking steps); push the ingredient shopping list weekly, categorized by supermarket section, such as vegetable section and meat section, indicating the amount and alternative ingredients; remind you to eat 30 minutes before meals, and remind you of dietary restrictions after taking medication, such as "Avoid alcohol and foods containing alcohol after taking cephalosporin drugs."

[0139] As can be seen from the above, this system has an age-friendly interaction method and is easy to operate; it is suitable for deployment in various elderly care scenarios such as home, community, and institutions, and has strong scalability.

[0140] In this embodiment, the data security management module includes:

[0141] The data encryption storage unit is configured to perform the following functions:

[0142] Transmission encryption: The AES-256 encryption algorithm is used to encrypt transmitted data to prevent data leakage;

[0143] Storage encryption: Health data is stored using a "partition encryption + key management" model, and core health data (such as medical records and genetic data) is additionally encrypted using the national cryptographic algorithm SM4;

[0144] Access control: The RBAC permission model is adopted, and different permissions are set for the elderly, family members, medical staff and system administrators. Access logs are recorded throughout the process and are traceable.

[0145] The privacy protection unit is configured to perform the following functions:

[0146] Data anonymization: When sharing data externally, sensitive information such as the elderly person's name and ID number is hidden, and a virtual ID is used for identification;

[0147] Authorization Management: Seniors can authorize the use of data for purposes such as dietary guidance only and not for other commercial purposes, and the authorization can be revoked at any time.

[0148] In summary, the multi-source data acquisition module of this invention integrates multi-dimensional data such as physiological, functional, medication, and environmental data, avoiding the one-sidedness of single data sources and providing comprehensive support for dietary guidance. The dietary needs analysis module performs personalized dietary needs analysis from multiple dimensions, including nutritional requirements, physical functions, dietary preferences, and regional ingredients, improving the implementation rate of the plan and the elderly's compliance. The personalized recipe generation module achieves rapid recipe generation and visualization through an improved AIGC model, reducing the cost of professional recipe design. The recipe optimization and adjustment module continuously adjusts the plan based on real-time feedback, adapting to changes in health status and external environment, ensuring long-term guidance effectiveness. The interactive output module optimizes the interaction method for the elderly's usage habits, reducing the difficulty of operation and improving the user experience. Thus, it achieves comprehensive perception of the elderly's health status, accurate positioning of dietary needs, intelligent generation of personalized recipes, dynamic optimization of plans, and convenient guidance output, solving the problems of limited data, poor adaptability, and insufficient dynamism in existing systems.

[0149] The technical solution provided by this invention has been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. It should be noted that those skilled in the art can make several improvements and modifications to this invention without departing from the principles of this invention, and these improvements and modifications also fall within the protection scope of the claims of this invention.

Claims

1. A smart elderly care health diet guidance system, characterized in that, include: The multi-source data acquisition module is used to collect comprehensive health data of the elderly through multiple devices and to preprocess the collected data. The dietary needs analysis module is used to assess health status using preprocessed multi-source health data and analyze the dietary needs of the elderly in conjunction with dietary-related rules. The personalized recipe generation module is used to input the dietary needs of the elderly into a pre-trained recipe generation model to generate personalized recipes that meet the needs of the elderly. The recipe optimization and adjustment module is used to dynamically optimize the personalized recipe based on real-time data feedback. The interactive output module is used to provide various forms of portable interaction and personalized healthy eating guidance based on the elderly's usage habits and personalized recipes.

2. The intelligent elderly care health diet guidance system according to claim 1, characterized in that: The multi-source data acquisition module includes: A health monitoring unit is used to collect core health monitoring data; The body function detection unit is used to obtain data on the elderly's chewing ability and digestive function; The environmental and behavioral data acquisition unit is used to acquire food reserve data and consumption data, as well as the local seasonal food supply situation; The data transmission unit is used to build a communication network to aggregate and upload data collected by the health monitoring unit, the physical function detection unit, and the environmental and behavioral data acquisition unit.

3. The intelligent elderly care health diet guidance system according to claim 2, characterized in that: The core health monitoring data includes basic physical data, disease data, exercise data, diet data, and medication data.

4. The intelligent elderly care health diet guidance system according to claim 1, characterized in that: The dietary needs analysis module includes: The health assessment indicator determination unit is used to determine the health indicators used to assess the health status of the elderly, and the weight of each health indicator is determined by the analytic hierarchy process. The health status assessment unit is used to construct a health status assessment model based on the determined health indicators and their corresponding weights, and output the health status score of the elderly. The personalized dietary needs extraction unit is used to calculate nutritional needs based on health status scores, extract dietary constraint information, match regional needs, and generate personalized dietary needs.

5. The intelligent elderly care health diet guidance system according to claim 4, characterized in that: The health indicators include physiological indicators, physical function indicators, past medical history indicators, and medication indicators. The expression of the health status assessment model is as follows: in, As the indicator weight, These are the standardized indicator values ​​for each health indicator.

6. The intelligent elderly care health diet guidance system according to claim 1, characterized in that: The personalized recipe generation module includes: The recipe generation model building unit is used to build a recipe generation model based on the elderly’s health data, personalized dietary needs, and regional ingredient information. The recipe generation model training unit is used to train the recipe generation model using elderly diet recipe data, optimize the loss function, and obtain the pre-trained recipe generation model. The recipe generation unit is used to input the target elderly person's health data, personalized dietary needs, and regional ingredient information into the pre-trained recipe generation model, and output personalized recipes with nutritional annotations.

7. The intelligent elderly care health diet guidance system according to claim 6, characterized in that: Before generating personalized recipes, the recipe generation unit needs to perform the following steps: Ingredient selection: Based on nutritional needs and a regional ingredient database, suitable ingredients are selected; Meal plan design: Design the meal plan structure for each day's meals and snacks, taking into account both nutritional balance and dietary diversity; Cooking method selection: Each personalized recipe includes information on calories, macronutrients, core micronutrients, and how it is suited to your health needs.

8. The intelligent elderly care health diet guidance system according to claim 1, characterized in that: The recipe optimization and adjustment module includes: The feedback data collection unit is used to collect data on the target elderly person's health status, dietary adherence, changes in local food supply, and environmental parameters. The recipe dynamic optimization unit is used to adjust personalized recipes based on data collected by the feedback data collection unit.

9. The intelligent elderly care health diet guidance system according to claim 1, characterized in that: The recipe dynamic optimization unit adjusts personalized recipes in the following way: Health-driven adjustments: If health indicators are abnormal, the diet will be automatically adjusted with an adjustment response time of ≤1 hour; Execution-driven adjustment: If a certain food ingredient is not consumed twice consecutively, it will be replaced with a nutritionally similar alternative food ingredient; Weekly optimization and adjustment: The menu is comprehensively optimized weekly based on health data and performance feedback; Emergency Adaptation: If an elderly person suddenly falls ill, a temporary diet plan will be automatically generated.

10. The intelligent elderly care health diet guidance system according to claim 1, characterized in that: The interactive output module includes: A multimodal interaction unit is used to enable voice and visual interaction between the elderly, their families, and the system. The recipe content output unit is used to output personalized recipes and their corresponding ingredient purchase lists and dietary reminders on a daily schedule.

Images