Intelligent old-age care health monitoring method and system fusing AI and big data

Abstract

The invention relates to the technical field of smart medical services, and discloses a smart old-age care health monitoring method and system fusing AI and big data, and the method comprises the steps: setting a dynamic health baseline of an old-age user in a smart old-age care platform; after continuously-collected multi-source real-time data streams of the elderly users are uploaded to the smart pension platform, physiological deviation indexes and behavior abnormity indexes of the elderly users are output through a built-in feature extraction unit of the smart pension platform; and identifying a collaborative degradation mode between the physiological deviation index and the behavior abnormality index, and calculating a comprehensive risk index of the elderly user in the collaborative degradation mode to output a health monitoring report of the elderly user. The method can improve the accuracy of the health monitoring result of smart old-age care.

Description

Technical Field

[0001] This invention relates to a smart elderly care health monitoring method and system that integrates AI and big data, belonging to the field of smart medical service technology. Background Technology

[0002] Smart elderly care monitoring refers to the process of continuously collecting and quantitatively analyzing data on the elderly and their living environment to dynamically assess their health and safety. This process aims to improve the predictability and targeted nature of elderly care services, providing core data support for early warning of health anomalies and personalized care for the elderly.

[0003] Traditional smart elderly care monitoring mainly relies on preset rules and basic sensing technology. By setting static thresholds and environmental triggering mechanisms, it can initially identify abnormal physiological indicators or environmental risks. Although this can achieve basic risk alarm functions in specific scenarios, its rigid judgment logic and lack of in-depth analysis of complex behaviors and health trends can easily lead to a high false alarm rate in smart elderly care health monitoring. Summary of the Invention

[0004] This invention provides a smart elderly care health monitoring method and system that integrates AI and big data, with the main purpose of improving the accuracy of smart elderly care health monitoring results.

[0005] To achieve the above objectives, this invention provides a smart elderly care health monitoring method integrating AI and big data, comprising: Retrieve the multimodal historical data of elderly users stored on the smart elderly care platform to set the dynamic health baseline of the elderly users. The multimodal historical data includes physiological data, behavioral data and environmental data. After the multi-source real-time data streams of the elderly users are continuously collected and uploaded to the smart elderly care platform, the feature extraction unit built into the smart elderly care platform compares and analyzes the multi-source real-time data streams with the dynamic health baseline, and outputs the physiological deviation indicators and behavioral abnormality indicators of the elderly users. Identify the co-degeneration pattern between the physiological deviation indicators and the behavioral abnormality indicators, calculate the comprehensive risk index of the elderly user under the co-degeneration pattern, and output the health monitoring report of the elderly user based on the comprehensive risk index.

[0006] Optionally, the feature extraction unit includes a real-time feature calculation layer and a dynamic comparison layer.

[0007] Optionally, the feature extraction unit built into the smart elderly care platform compares and analyzes the multi-source real-time data stream with the dynamic health baseline to output the physiological deviation indicators and behavioral abnormality indicators of the elderly users, including: The real-time feature vector of the multi-source real-time data stream is calculated through the real-time feature calculation layer. Based on the dynamic comparison layer, the real-time feature value is compared with the corresponding benchmark range and normal behavior pattern in the dynamic health baseline to obtain the comparison result; Based on the comparison results, the physiological deviation indicators and behavioral abnormality indicators of the elderly users are output.

[0008] Optionally, calculating the real-time feature vector of the multi-source real-time data stream through the real-time feature calculation layer includes: After receiving the multi-source real-time data stream, the real-time feature calculation layer performs invalid value removal and noise filtering on the multi-source real-time data stream to obtain a standardized data stream. Extract multidimensional features corresponding to the core health indicators in the dynamic health baseline from the standardized data stream; After normalizing the multidimensional features, the normalized multidimensional features are used as the real-time feature vector of the multi-source real-time data stream.

[0009] Optionally, the step of comparing the real-time feature values ​​with the corresponding baseline range and normal behavior patterns in the dynamic health baseline based on the dynamic comparison layer to obtain the comparison results includes: Calculate the degree of deviation of each feature value in the real-time feature values ​​relative to the reference range; Extract the set of behavioral features corresponding to the real-time feature values, and calculate the degree of difference between the set of behavioral features and the normal behavioral pattern; The comparison result is obtained by combining the degree of deviation and the degree of difference.

[0010] Optionally, identifying the co-degradation pattern between the physiological deviation indicator and the behavioral abnormality indicator includes: The abnormal physiological state level corresponding to the physiological deviation index and the abnormal behavioral state level corresponding to the behavioral abnormality index are respectively analyzed. By utilizing the temporal correlation between the abnormal physiological state level and the abnormal behavioral state level, the co-degeneration pattern between the physiological deviation index and the abnormal behavioral index is determined.

[0011] Optionally, calculating the comprehensive risk index of the elderly user under the collaborative degradation mode includes: Extract the physiological and behavioral abnormality time sequences from the collaborative degradation pattern; Feature enhancement calculations are performed on the physiological abnormality time series and the behavioral abnormality time series respectively to obtain physiological modality risk values ​​and behavioral modality risk values; By combining the physiological modal risk value and the behavioral modal risk value, a comprehensive risk index for the elderly user under the co-degeneration mode is calculated.

[0012] Optionally, retrieving the multimodal historical data of elderly users stored on the smart elderly care platform to set a dynamic health baseline for the elderly users includes: Core health indicators were selected from the multimodal historical data, and the specific physiological characteristics of the elderly users were extracted. Based on the specific physiological characteristics and the core health indicators, a dynamic health baseline is set for the elderly users.

[0013] Optionally, the step of outputting a health monitoring report for the elderly user based on the comprehensive risk index includes: Identify the key collaborative degradation patterns corresponding to the current comprehensive risk index, and locate the initial abnormal indicators in the key collaborative degradation patterns; Based on the initial abnormal indicators, a health event tracing chain for the elderly user is constructed; Based on the health event tracing chain, the corresponding graded intervention plan is matched from the preset health intervention knowledge graph; Calculate the slope of the historical composite risk index series; When the slope of change exceeds a preset threshold, a risk trend warning signal for the elderly user is generated. By integrating the health event tracing chain, the tiered intervention plan, and the risk trend early warning signal, a health monitoring report for the elderly user is output.

[0014] To address the aforementioned problems, this invention also provides a smart elderly care health monitoring system integrating AI and big data, the system comprising: The baseline construction module is used to retrieve the multimodal historical data of elderly users stored in the smart elderly care platform in order to set the dynamic health baseline of the elderly users. The multimodal historical data includes physiological data, behavioral data and environmental data. An anomaly detection module is used to upload the continuously collected multi-source real-time data stream of the elderly user to the smart elderly care platform, and then compare and analyze the multi-source real-time data stream with the dynamic health baseline through the feature extraction unit built into the smart elderly care platform, and output the physiological deviation index and behavioral abnormality index of the elderly user. The report output module is used to identify the co-degeneration pattern between the physiological deviation index and the behavioral abnormality index, calculate the comprehensive risk index of the elderly user under the co-degeneration pattern, and output the health monitoring report of the elderly user based on the comprehensive risk index.

[0015] Compared to the problems described in the background technology, this invention, by retrieving multimodal historical data stored by elderly users on a smart elderly care platform to establish a dynamic health baseline for them, can construct a health benchmark reference system that fits individual differences. This provides accurate core reference for subsequent real-time data comparison, abnormal indicator identification, and quantitative assessment of health risks, ensuring the targeting and reliability of smart elderly care health monitoring. Furthermore, this invention, through the feature extraction unit built into the smart elderly care platform, compares and analyzes the multi-source real-time data stream with the dynamic health baseline, outputting the elderly user's physiological deviation indicators and behavioral abnormality indicators. This transforms the raw data into quantifiable abnormal indicators with clear health orientation, providing criteria for subsequent risk fusion judgment. By identifying the synergistic degradation patterns between the physiological deviation indicators and the behavioral abnormality indicators, this invention can effectively capture the intrinsic correlation between multi-dimensional health signals, thereby discovering complex abnormal conditions that are difficult to identify with single indicator monitoring. Finally, based on the comprehensive risk index, this invention outputs a health monitoring report for the elderly user, enabling a systematic and structured presentation of the elderly user's health status and matching differentiated coping strategies for different levels of health risks. Therefore, the smart elderly health monitoring method and system that integrates AI and big data provided in this embodiment of the invention can improve the accuracy of smart elderly health monitoring results. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a smart elderly care health monitoring method that integrates AI and big data, provided as an embodiment of the present invention. Figure 2 This is a system structure diagram of a smart elderly care platform that integrates AI and big data for smart elderly care health monitoring, provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of a module for implementing a smart elderly care and health monitoring system that integrates AI and big data, provided as an embodiment of the present invention. Figure 4 This is a schematic diagram of a computer device for a smart elderly care health monitoring method that integrates AI and big data, as provided in an embodiment of the present invention.

[0017] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] This application provides a smart elderly care health monitoring method integrating AI and big data. The executing entity of this smart elderly care health monitoring method integrating AI and big data includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application embodiment: a server, a terminal, etc. In other words, the smart elderly care health monitoring method integrating AI and big data can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0020] Reference Figure 1 The diagram shown is a flowchart illustrating a smart elderly care health monitoring method integrating AI and big data, according to an embodiment of the present invention. In this embodiment, the smart elderly care health monitoring method integrating AI and big data includes: S1. Retrieve the multimodal historical data of the elderly user stored on the smart elderly care platform to set the dynamic health baseline of the elderly user.

[0021] This invention retrieves multimodal historical data stored by elderly users on a smart elderly care platform to set a dynamic health baseline for the elderly users. This allows for the construction of a health benchmark reference system that fits individual differences, providing accurate core reference for subsequent real-time data comparison, abnormal indicator identification, and quantitative assessment of health risks, thus ensuring the relevance and reliability of smart elderly care health monitoring.

[0022] The elderly users referred to here are individuals receiving smart elderly care health monitoring services. For example, elderly people living alone or those with hypertension. The smart elderly care platform is a hardware and software system integrating data collection, storage, computation, and service functions. It can connect to various monitoring terminal devices to receive, integrate, and apply data. It mainly consists of a data management unit, a dynamic baseline modeling unit, a feature extraction unit, a risk perception unit, a decision center, and an interactive interface. The multimodal historical data refers to the collection of various data related to elderly users uploaded by the smart elderly care platform through various types of collection devices during past monitoring periods. Specifically, the data collection consists of three types of structured records: the first type is physiological parameter records, including time-series measurements of heart rate and blood pressure; the second type is behavioral activity records, including the time and duration of daily activities, diet, and other activities; and the third type is environmental status records, including measurements of physical quantities such as temperature, humidity, and illuminance in the living space. The dynamic health baseline refers to a health status reference standard set based on multimodal historical data and combined with the individual characteristics of the elderly user. For example, a heart rate reference interval set based on the average and fluctuation range of a certain elderly user's heart rate over the past three months.

[0023] As an embodiment of the present invention, retrieving the multimodal historical data of elderly users stored in the smart elderly care platform to set the dynamic health baseline of the elderly users includes: selecting core health indicators from the multimodal historical data and extracting the specific physiological characteristics of the elderly users; and setting the dynamic health baseline of the elderly users based on the specific physiological characteristics and the core health indicators.

[0024] The core health indicators refer to key physiological parameters that directly reflect the physiological function, health level, and potential health risks of elderly users. For example, core health indicators for cardiovascular health monitoring include specific parameters such as resting heart rate, diurnal blood pressure fluctuation range, and heart rate variability. The specific physiological characteristics refer to physiological state-related features unique to individual elderly users, distinguishing them from the general population. These characteristics are determined by factors such as age, underlying medical history, physical differences, and lifestyle habits, and possess long-term stable individual identifiability. For example, an elderly user may have a long-term characteristic of maintaining elevated diastolic blood pressure.

[0025] Optionally, the specific physiological characteristics of the elderly user can be extracted by a variational autoencoder; the dynamic health baseline of the elderly user can be set using a Gaussian mixture model, such as modeling the multimodal historical data using a Gaussian mixture model, fitting its probability distribution, and setting the central trend and normal fluctuation range of the probability distribution as the dynamic health baseline of the elderly user.

[0026] S2. After uploading the continuously collected multi-source real-time data streams of the elderly users to the smart elderly care platform, the feature extraction unit built into the smart elderly care platform compares and analyzes the multi-source real-time data streams with the dynamic health baseline, and outputs the physiological deviation indicators and behavioral abnormality indicators of the elderly users.

[0027] This invention provides real-time data for personalized health risk monitoring by uploading continuously collected multi-source real-time data streams from elderly users to the smart elderly care platform. The multi-source real-time data streams refer to heterogeneous data sets collected at high-frequency intervals using various types of sensors deployed in the elderly user's living environment and on their body, and which have a temporal sequence. Examples include heart rate reading sequences collected and uploaded once per minute by wearable smart bracelets, and door opening and closing event records triggered in real-time by magnetic sensors on kitchen and bathroom doors.

[0028] Furthermore, this embodiment of the invention utilizes the feature extraction unit built into the smart elderly care platform to compare and analyze the multi-source real-time data stream with the dynamic health baseline, outputting physiological deviation indicators and behavioral abnormality indicators for the elderly user. This transforms the raw data into quantifiable abnormal indicators with clear health implications, providing criteria for subsequent risk fusion judgment. The physiological deviation indicator is a numerical measure used to quantify the degree of deviation of the user's current physiological state from the normal fluctuation range of their own health benchmark. For example, it can be reflected as the standard deviation multiple of the currently continuously monitored heart rate data relative to the user's personal heart rate benchmark interval. The behavioral abnormality indicator is a numerical measure used to quantify the degree of deviation of the user's daily behavioral activity pattern from their own routine behavioral habits. For example, it can be reflected as the percentage change in the number of times the kitchen was used on the current day relative to the user's average daily kitchen usage over the past month.

[0029] Specifically, the feature extraction unit includes a real-time feature calculation layer and a dynamic comparison layer.

[0030] The real-time feature calculation layer refers to a functional module used to perform data preprocessing and feature engineering on the input multi-source real-time data stream, thereby transforming the original, heterogeneous sensor data into a set of standardized, numerical feature vectors; the dynamic comparison layer refers to a functional module used to quantitatively compare the real-time feature vectors with a pre-stored dynamic health baseline, and output one or more abnormal indicators that can characterize the degree to which the current state deviates from the normal pattern.

[0031] As an embodiment of the present invention, the step of comparing and analyzing the multi-source real-time data stream with the dynamic health baseline through the feature extraction unit built into the smart elderly care platform, and outputting the physiological deviation indicators and behavioral abnormality indicators of the elderly user, includes: calculating the real-time feature vector of the multi-source real-time data stream through the real-time feature calculation layer; comparing the real-time feature value with the corresponding benchmark range and normal behavior pattern in the dynamic health baseline based on the dynamic comparison layer to obtain the comparison result; and outputting the physiological deviation indicators and behavioral abnormality indicators of the elderly user based on the comparison result.

[0032] The real-time feature vector refers to a structured data set composed of multiple numerical elements arranged sequentially. Each element represents a quantified value of a specific health or behavioral indicator extracted from multi-source real-time data streams within the current time period. The baseline range refers to the numerical intervals within which various physiological and behavioral indicators should fall when an elderly user is in a normal health state, calculated using statistical learning methods. The normal behavioral pattern refers to a typical behavioral sequence or state transition pattern of elderly users that exhibits repetitiveness and regularity, extracted from historical behavioral data. For example, a typical daytime activity pattern can be characterized as a sequence of behaviors with high activity intensity from 9:00 to 10:00 AM, followed by a period of rest. The comparison result refers to a set of quantified difference values ​​obtained after calculating the deviation between each element in the real-time feature vector and its corresponding baseline range, and calculating the similarity between the real-time behavioral sequence and the historical behavioral pattern.

[0033] Optionally, the physiological deviation indicators and behavioral abnormality indicators of the elderly users can be output by a classifier based on threshold rules.

[0034] As another embodiment of the present invention, the step of calculating the real-time feature vector of the multi-source real-time data stream through the real-time feature calculation layer includes: after receiving the multi-source real-time data stream, the real-time feature calculation layer performs invalid value removal and noise filtering on the multi-source real-time data stream to obtain a standardized data stream; extracts multi-dimensional features corresponding to the core health indicators in the dynamic health baseline from the standardized data stream; and after normalizing the multi-dimensional features, uses the normalized multi-dimensional features as the real-time feature vector of the multi-source real-time data stream.

[0035] The invalid value removal process refers to the process of identifying and removing abnormal data points that clearly do not conform to physiological or physical common sense based on preset data rationality rules; the noise filtering process refers to the process of using digital signal processing algorithms to suppress random and high-frequency fluctuations in the data in order to restore the effective underlying trend signal; the multidimensional features refer to a set of multiple indicators that can quantify and characterize the user's health and behavioral status. Specifically, this includes: for the physiological data substream, extracting time-domain features such as mean, variance, maximum, minimum, and slope of the trend line within a specified time window; for the behavioral data substream, extracting statistical features such as the frequency, duration, and intensity of specific behaviors; for the environmental data substream, extracting periodic features such as the fluctuation range and average value of indoor temperature within a daily cycle; and the normalization process refers to the data processing step of mapping multiple indicators with different dimensions and numerical ranges in the multidimensional features to the same numerical range through mathematical transformation.

[0036] As another embodiment of the present invention, the step of comparing the real-time feature values ​​with the corresponding benchmark range and normal behavior pattern in the dynamic health baseline based on the dynamic comparison layer to obtain the comparison result includes: calculating the degree of deviation of each feature value in the real-time feature values ​​relative to the benchmark range; extracting the behavioral feature set corresponding to the real-time feature values ​​and calculating the degree of difference between the behavioral feature set and the normal behavior pattern; and fusing the degree of deviation and the degree of difference to obtain the comparison result.

[0037] The deviation degree refers to the degree to which a single physiological or environmental feature value deviates from its baseline range; the behavioral feature set refers to all feature subsets specifically used to describe the user's behavioral state, separated from the real-time feature vector; and the difference degree refers to an indicator used to quantify the dissimilarity between the real-time behavioral feature set and historical normal behavioral patterns.

[0038] Optionally, the deviation of each feature value in the real-time feature values ​​from the benchmark range can be calculated using the standard score method; the difference between the behavioral feature set and the normal behavioral pattern can be calculated using the dynamic time warping algorithm.

[0039] S3. Identify the co-degeneration pattern between the physiological deviation index and the behavioral abnormality index, calculate the comprehensive risk index of the elderly user under the co-degeneration pattern, and output the health monitoring report of the elderly user based on the comprehensive risk index.

[0040] This invention, by identifying the synergistic degradation pattern between the physiological deviation indicators and the behavioral abnormality indicators, can effectively capture the intrinsic correlation between multidimensional health signals, thereby discovering complex abnormal conditions that are difficult to identify by single indicator monitoring. The synergistic degradation pattern refers to the statistically significant correlation deterioration trend between two or more physiological deviation indicators and behavioral abnormality indicators within a preset time window.

[0041] As an embodiment of the present invention, identifying the co-degradation pattern between the physiological deviation index and the behavioral abnormality index includes: parsing out the abnormal physiological state level corresponding to the physiological deviation index and the abnormal behavioral state level corresponding to the behavioral abnormality index; and determining the co-degradation pattern between the physiological deviation index and the behavioral abnormality index by utilizing the temporal correlation between the abnormal physiological state level and the abnormal behavioral state level.

[0042] The abnormal physiological state level refers to a scalar value, obtained by quantifying and classifying the physiological deviation indicators, used to characterize the severity of a user's physiological function deviating from their normal state. The abnormal behavioral state level refers to a scalar value, obtained by quantifying and classifying the behavioral abnormality indicators, used to characterize the significant degree to which a user's behavioral pattern deviates from their daily habits. The temporal correlation refers to the interdependence between two or more time series data in terms of their changing trends. This relationship is manifested as a measurable statistical correlation between the numerical change of one series at a specific time point or time period and the numerical change of another series at the same or lagging time point. For example, when the system detects an abnormal behavioral state level in an elderly user, specifically manifested as a significant decrease in daytime activity for three consecutive days, and simultaneously, an abnormal physiological state level, specifically manifested as a synchronous and statistically significant upward trend in nighttime average heart rate, then the two abnormal levels show a significant negative correlation in the direction of change, and the system may determine that a co-degeneration pattern exists.

[0043] Optionally, the abnormal physiological state level corresponding to the physiological deviation index and the abnormal behavioral state level corresponding to the behavioral deviation index can be analyzed using the isolated forest algorithm; the temporal correlation between the abnormal physiological state level and the abnormal behavioral state level can be determined using the Granger causality test method.

[0044] Furthermore, by calculating the comprehensive risk index of the elderly user under the synergistic degeneration mode, the embodiments of the present invention can achieve quantitative integration and severity classification of multi-dimensional abnormal health states, thereby providing a precise basis for differentiated health intervention. The comprehensive risk index refers to a scalar value used to quantitatively characterize the severity of abnormalities in the overall health status of the elderly user under the current synergistic degeneration mode.

[0045] As an embodiment of the present invention, calculating the comprehensive risk index of the elderly user under the co-degeneration mode includes: extracting the physiological abnormality time series and the behavioral abnormality time series from the co-degeneration mode; performing feature enhancement calculations on the physiological abnormality time series and the behavioral abnormality time series respectively to obtain physiological modal risk values ​​and behavioral modal risk values; and combining the physiological modal risk values ​​and the behavioral modal risk values ​​to calculate the comprehensive risk index of the elderly user under the co-degeneration mode.

[0046] The physiological abnormality time series refers to a series of continuous data points arranged chronologically, recording the degree to which various physiological indicators of elderly users deviate from their personal normal baseline; the behavioral abnormality time series refers to a series of continuous data points arranged chronologically, recording the degree to which various behavioral patterns of elderly users deviate from their personal daily habits; the feature enhancement calculation refers to a mathematical processing procedure that extracts key information from the original time series data that can more profoundly reveal its inherent risk characteristics. Specifically, for the physiological abnormality time series, the feature enhancement calculation can be specifically as follows: first, identify all abnormal peaks in the sequence that exceed a preset threshold; then calculate the product of the average amplitude of these peaks and their duration; finally, weight and fuse this product with the frequency of occurrence of abnormal events, and the result is the physiological modality risk value. The feature enhancement calculation of the behavioral abnormality time series can be specifically as follows: first, calculate the longest duration of a single behavioral abnormality state; then, calculate the variance of the series within a certain period to measure its volatility; finally, multiply the duration by the volatility index, and the result is the behavioral modal risk value; the physiological modal risk value refers to a scalar value obtained by performing feature enhancement calculation on the physiological abnormality time series; the behavioral modal risk value refers to a scalar value obtained by performing feature enhancement calculation on the behavioral abnormality time series.

[0047] Optionally, a peak detection algorithm can be used to perform feature enhancement calculations on the physiological abnormality time series and the behavioral abnormality time series, respectively.

[0048] For example, the comprehensive risk index is calculated using the following formula. It should be noted that this calculation method is only one possible method and does not affect the implementation of the basic scheme above:

[0049] in, This represents the comprehensive risk index. The modality weighting coefficient represents the physiological modality risk value. Indicates the physiological modality risk value. Modal weighting coefficients represent behavioral modal risk values. Indicates the behavioral modality risk value. Indicates the mode magnification factor. This indicates the degree of synergy between time series of physiological abnormalities and time series of behavioral abnormalities. This indicates a trend-reinforcing factor for the co-degradation pattern.

[0050] In detail, the modal weighting coefficients and the modal weighting coefficients , is a preset weighting coefficient, and satisfies The value of this factor can be determined based on the relative importance of physiological and behavioral data in health assessment, and is mainly used to adjust the contribution ratio of different modal risk values ​​in the basic risk calculation; the mode amplification factor K represents a configurable, positive constant used to adjust the overall amplification intensity of the synergy and trend reinforcement factor on the comprehensive risk index, and its specific value can be calibrated through historical data verification; the synergy M between the physiological abnormality time series and the behavioral abnormality time series can be obtained by calculating the maximum mutual information value of the two time series within the sliding time window, and this value is mainly used to quantify the statistical interdependence and nonlinear correlation strength between physiological and behavioral abnormalities; the trend reinforcement factor of the synergistic degradation mode Linear regression analysis can be performed on the weighted fusion sequence of the physiological abnormality time series and the behavioral abnormality time series, and the regression coefficient can be taken as a quantitative value to characterize the rate of aggravation or deceleration of the collaborative degradation mode per unit time.

[0051] This invention, through its embodiment, outputs a health monitoring report for elderly users based on the comprehensive risk index. This enables a systematic and structured presentation of the health status of elderly users and allows for the matching of differentiated response strategies for different levels of health risks. The health monitoring report is a structured electronic document automatically generated by the smart elderly care platform, containing basic user information, report generation time, comprehensive risk index value and level, description of major abnormal patterns, health trend judgment, and treatment suggestions.

[0052] As an embodiment of the present invention, the step of outputting a health monitoring report for the elderly user based on the comprehensive risk index includes: identifying the key collaborative degradation pattern corresponding to the current comprehensive risk index and locating the initial abnormal indicators in the key collaborative degradation pattern; constructing a health event tracing chain for the elderly user based on the initial abnormal indicators; matching the corresponding graded intervention plan from a preset health intervention knowledge graph according to the health event tracing chain; calculating the slope of the historical comprehensive risk index sequence of the comprehensive risk index; generating a risk trend warning signal for the elderly user when the slope exceeds a preset threshold; and integrating the health event tracing chain, the graded intervention plan, and the risk trend warning signal to output a health monitoring report for the elderly user.

[0053] The key synergistic degradation pattern refers to the specific pattern that contributes the most to the index value among all synergistic degradation patterns constituting the current comprehensive risk index; the initial abnormal indicator refers to the physiological or behavioral indicator that deviates from the individual's dynamic health baseline first on the timeline among the multiple abnormal indicators included in the key synergistic degradation pattern; the health event tracing chain refers to a logical sequence arranged in chronological order, which records the complete event development path from the appearance of the initial abnormal indicator to the subsequent abnormalities of related indicators, ultimately forming the key synergistic degradation pattern; the health intervention knowledge graph refers to a structured domain knowledge database that stores the relationships between different health abnormality patterns, abnormal indicators, and corresponding intervention measures in a graph structure; and the tiered intervention program refers to a step-by-step management strategy based on the severity of health risks. The risk profile typically includes immediate emergency measures, medium-term recommendations requiring short-term observation, and fundamental solutions aimed at long-term improvement. The historical comprehensive risk index sequence refers to a set of historical comprehensive risk index values ​​arranged chronologically prior to the current comprehensive risk index. The slope of change refers to the slope of the regression line obtained through linear regression analysis of the historical comprehensive risk index sequence, used to quantify the average rate and direction of change of the index per unit time. The preset threshold is a value pre-set based on expert experience, used to determine whether the deteriorating trend represented by the slope of change has reached a severity level requiring a warning. The risk trend warning signal is an automatically generated alert indicating that the overall health status of elderly users is showing a statistically significant deteriorating trend when the slope of change exceeds the preset threshold.

[0054] Optionally, the health event tracing chain of the elderly user can be generated by constructing a directed acyclic graph model, where nodes represent abnormal health events and edges represent the temporal and causal relationships between events; the risk trend warning signal of the elderly user can be generated using an exponential smoothing state space model, by fitting the long-term trend components of the historical comprehensive risk index sequence, and triggering a warning signal when the lower limit of the prediction interval continuously exceeds a preset risk threshold.

[0055] It should be noted that the health monitoring report is presented simultaneously in a machine-readable standard data format and a human-readable visualization format. The visualization format includes at least one of the following: trend charts, risk level color-coded charts, and key indicator radar charts. For example, when the system identifies a synergistic deterioration pattern of "continuous decline in blood oxygen saturation at night and slowed reaction speed during the day," the generated health monitoring report will include the following typical content: a "high risk" level marked in red, a correlation analysis of the blood oxygen saturation decline trend chart and the reaction speed test result change table, and finally, a targeted recommendation to "conduct sleep apnea monitoring and neurological function examination."

[0056] See Figure 2 The diagram shown illustrates the system architecture of a smart elderly care platform that integrates AI and big data for health monitoring, according to an embodiment of the present invention. The data management unit receives real-time streaming data and historical batch data from wearable devices, environmental sensors, etc., and provides reliable data support for subsequent calculations and analyses through data cleaning and format standardization. The dynamic baseline modeling unit generates and continuously updates a computational model representing a user's normal health status, which is the core of personalized monitoring. Utilizing historical data and machine learning methods such as Gaussian mixture models, it constructs a dynamic, probabilistic health benchmark for each user. The feature extraction unit compares real-time data with the individual baseline. It can extract features through built-in real-time data extraction... The symptom calculation layer and dynamic comparison layer transform cleaned real-time data into structured features and compare them with a dynamic baseline, ultimately outputting quantifiable physiological deviation indicators and behavioral abnormality indicators. The risk perception unit identifies collaborative degradation patterns by analyzing the temporal correlation between different abnormal indicators. The decision center is responsible for constructing a health event tracing chain, matching graded intervention strategies from a knowledge graph, and generating early warnings based on risk trend predictions, ultimately driving report generation. The interaction interface serves as an information exchange hub between the system and internal and external users and services. On the one hand, it can present the reports and early warnings generated by the decision center to caregivers and families in a visual form. On the other hand, it can connect with other health service platforms through APIs to achieve seamless linkage between monitoring information and care services.

[0057] like Figure 3 The diagram shown is a functional block diagram of a smart elderly care and health monitoring system that integrates AI and big data according to the present invention.

[0058] The intelligent elderly care and health monitoring system 200 integrating AI and big data described in this invention can be installed in an electronic device. Depending on the functions implemented, the intelligent elderly care and health monitoring system integrating AI and big data may include a baseline construction module 201, an anomaly identification module 202, and a report output module 203. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and is stored in the memory of the electronic device.

[0059] In this embodiment of the invention, the functions of each module / unit are as follows: The baseline construction module 201 is used to retrieve the multimodal historical data of elderly users stored in the smart elderly care platform in order to set the dynamic health baseline of the elderly users. The multimodal historical data includes physiological data, behavioral data and environmental data. The anomaly identification module 202 is used to upload the continuously collected multi-source real-time data stream of the elderly user to the smart elderly care platform, and then compare and analyze the multi-source real-time data stream with the dynamic health baseline through the feature extraction unit built into the smart elderly care platform, and output the physiological deviation index and behavioral abnormality index of the elderly user. The report output module 203 is used to identify the co-degeneration pattern between the physiological deviation index and the behavioral abnormality index, calculate the comprehensive risk index of the elderly user under the co-degeneration pattern, and output the health monitoring report of the elderly user based on the comprehensive risk index.

[0060] In detail, the modules in the smart elderly care and health monitoring system 200 integrating AI and big data described in this embodiment of the invention adopt the same approach as described above when in use. Figure 1 The method used is the same as the smart elderly care health monitoring method that integrates AI and big data described in the article, and can produce the same technical effect, so it will not be elaborated here.

[0061] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used for communication with external clients via a network connection. When the computer program is executed by the processor, it implements the functions or steps of a server-side smart elderly care health monitoring method integrating AI and big data.

[0062] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps: Retrieve the multimodal historical data of elderly users stored on the smart elderly care platform to set the dynamic health baseline of the elderly users. The multimodal historical data includes physiological data, behavioral data and environmental data. After the multi-source real-time data streams of the elderly users are continuously collected and uploaded to the smart elderly care platform, the feature extraction unit built into the smart elderly care platform compares and analyzes the multi-source real-time data streams with the dynamic health baseline, and outputs the physiological deviation indicators and behavioral abnormality indicators of the elderly users. Identify the co-degeneration pattern between the physiological deviation indicators and the behavioral abnormality indicators, calculate the comprehensive risk index of the elderly user under the co-degeneration pattern, and output the health monitoring report of the elderly user based on the comprehensive risk index.

[0063] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor: Retrieve the multimodal historical data of elderly users stored on the smart elderly care platform to set the dynamic health baseline of the elderly users. The multimodal historical data includes physiological data, behavioral data and environmental data. After the multi-source real-time data streams of the elderly users are continuously collected and uploaded to the smart elderly care platform, the feature extraction unit built into the smart elderly care platform compares and analyzes the multi-source real-time data streams with the dynamic health baseline, and outputs the physiological deviation indicators and behavioral abnormality indicators of the elderly users. Identify the co-degeneration pattern between the physiological deviation indicators and the behavioral abnormality indicators, calculate the comprehensive risk index of the elderly user under the co-degeneration pattern, and output the health monitoring report of the elderly user based on the comprehensive risk index.

[0064] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or computer device described above can be referred to the relevant descriptions on the server side and client side in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0065] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0066] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0067] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0068] Finally, it should be noted that in the above embodiments, each embodiment can be combined with each other or independent. Deleting any one of them will not affect the technical implementation of other embodiments. The above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A smart elderly care health monitoring method integrating AI and big data, characterized in that, The method includes: Retrieve the multimodal historical data of elderly users stored on the smart elderly care platform to set the dynamic health baseline of the elderly users. The multimodal historical data includes physiological data, behavioral data and environmental data. After the multi-source real-time data streams of the elderly users are continuously collected and uploaded to the smart elderly care platform, the feature extraction unit built into the smart elderly care platform compares and analyzes the multi-source real-time data streams with the dynamic health baseline, and outputs the physiological deviation indicators and behavioral abnormality indicators of the elderly users. Identify the co-degeneration pattern between the physiological deviation indicators and the behavioral abnormality indicators, calculate the comprehensive risk index of the elderly user under the co-degeneration pattern, and output the health monitoring report of the elderly user based on the comprehensive risk index.

2. The smart elderly care health monitoring method integrating AI and big data as described in claim 1, characterized in that, The feature extraction unit includes a real-time feature calculation layer and a dynamic comparison layer.

3. The smart elderly care health monitoring method integrating AI and big data as described in claim 2, characterized in that, The feature extraction unit built into the smart elderly care platform compares and analyzes the multi-source real-time data stream with the dynamic health baseline, and outputs the physiological deviation indicators and behavioral abnormality indicators of the elderly users, including: The real-time feature vector of the multi-source real-time data stream is calculated through the real-time feature calculation layer. Based on the dynamic comparison layer, the real-time feature value is compared with the corresponding benchmark range and normal behavior pattern in the dynamic health baseline to obtain the comparison result; Based on the comparison results, the physiological deviation indicators and behavioral abnormality indicators of the elderly users are output.

4. The smart elderly care health monitoring method integrating AI and big data as described in claim 3, characterized in that, The step of calculating the real-time feature vector of the multi-source real-time data stream through the real-time feature calculation layer includes: After receiving the multi-source real-time data stream, the real-time feature calculation layer performs invalid value removal and noise filtering on the multi-source real-time data stream to obtain a standardized data stream. Extract multidimensional features corresponding to the core health indicators in the dynamic health baseline from the standardized data stream; After normalizing the multidimensional features, the normalized multidimensional features are used as the real-time feature vector of the multi-source real-time data stream.

5. The smart elderly care health monitoring method integrating AI and big data as described in claim 3, characterized in that, The process, based on the dynamic comparison layer, compares the real-time feature values ​​with the corresponding baseline range and normal behavior patterns in the dynamic health baseline to obtain the comparison results, including: Calculate the degree of deviation of each feature value in the real-time feature values ​​relative to the reference range; Extract the set of behavioral features corresponding to the real-time feature values, and calculate the degree of difference between the set of behavioral features and the normal behavioral pattern; The comparison result is obtained by combining the degree of deviation and the degree of difference.

6. The smart elderly care health monitoring method integrating AI and big data as described in claim 1, characterized in that, The identification of the co-degeneration pattern between the physiological deviation indicators and the behavioral abnormality indicators includes: The abnormal physiological state level corresponding to the physiological deviation index and the abnormal behavioral state level corresponding to the behavioral abnormality index are respectively analyzed. By utilizing the temporal correlation between the abnormal physiological state level and the abnormal behavioral state level, the co-degeneration pattern between the physiological deviation index and the abnormal behavioral index is determined.

7. The smart elderly care health monitoring method integrating AI and big data as described in claim 1, characterized in that, The calculation of the comprehensive risk index of the elderly user under the collaborative degradation mode includes: Extract the physiological and behavioral abnormality time sequences from the collaborative degradation pattern; Feature enhancement calculations are performed on the physiological abnormality time series and the behavioral abnormality time series respectively to obtain physiological modality risk values ​​and behavioral modality risk values; By combining the physiological modal risk value and the behavioral modal risk value, a comprehensive risk index for the elderly user under the co-degeneration mode is calculated.

8. The smart elderly care health monitoring method integrating AI and big data as described in claim 1, characterized in that, The step of retrieving multimodal historical data of elderly users stored on the smart elderly care platform to set a dynamic health baseline for the elderly users includes: Core health indicators were selected from the multimodal historical data, and the specific physiological characteristics of the elderly users were extracted. Based on the specific physiological characteristics and the core health indicators, a dynamic health baseline is set for the elderly users.

9. The smart elderly care health monitoring method integrating AI and big data as described in claim 1, characterized in that, The health monitoring report for the elderly user, based on the comprehensive risk index, includes: Identify the key collaborative degradation patterns corresponding to the current comprehensive risk index, and locate the initial abnormal indicators in the key collaborative degradation patterns; Based on the initial abnormal indicators, a health event tracing chain for the elderly users is constructed. Based on the health event tracing chain, the corresponding graded intervention plan is matched from the preset health intervention knowledge graph; Calculate the slope of the historical composite risk index series; When the slope of change exceeds a preset threshold, a risk trend warning signal for the elderly user is generated. By integrating the health event tracing chain, the tiered intervention plan, and the risk trend early warning signal, a health monitoring report for the elderly user is output.

10. A smart elderly care health monitoring system integrating AI and big data, used to execute the smart elderly care health monitoring method integrating AI and big data as described in any one of claims 1-9, characterized in that, The system includes: The baseline construction module is used to retrieve the multimodal historical data of elderly users stored in the smart elderly care platform in order to set the dynamic health baseline of the elderly users. The multimodal historical data includes physiological data, behavioral data and environmental data. An anomaly detection module is used to upload the continuously collected multi-source real-time data stream of the elderly user to the smart elderly care platform, and then compare and analyze the multi-source real-time data stream with the dynamic health baseline through the feature extraction unit built into the smart elderly care platform, and output the physiological deviation index and behavioral abnormality index of the elderly user. The report output module is used to identify the co-degeneration pattern between the physiological deviation index and the behavioral abnormality index, calculate the comprehensive risk index of the elderly user under the co-degeneration pattern, and output the health monitoring report of the elderly user based on the comprehensive risk index.

Images