Home-based old-age care intelligent health monitoring system based on Internet of Things

Abstract

The invention discloses a home-based old-age care intelligent health monitoring system based on the Internet of Things, and relates to the technical field of intelligent monitoring, and the system comprises a multi-dimensional sensing module, a data preprocessing module, a space correlation risk pre-judgment module, an intelligent linkage avoidance execution module and a bidirectional information interaction module. The multi-dimensional sensing module is deployed in a home kitchen, a bathroom, a balcony and a public area, and is used for collecting microenvironment data and action track data of the old; according to the invention, on the aspect of time correlation processing through the space correlation risk pre-judgment module, the time characteristics of the action track and the microenvironment data of the old can be accurately extracted, and a time correlation analysis matrix is constructed, so that the accurate matching of the two types of characteristics is realized.

Description

Technical Field

[0001] This invention relates to the field of intelligent monitoring technology, specifically to an intelligent health monitoring system for home-based elderly care based on the Internet of Things. Background Technology

[0002] With the accelerating aging of the global population, home-based care has become the choice for more and more elderly people. How to ensure the safety and health of elderly people living alone or of advanced age in their home environment has become a focus of social concern. The rapid development of Internet of Things (IoT) technology has brought new opportunities and solutions to the field of home-based care. By building a smart health monitoring system for home-based care through IoT technology, it is possible to collect real-time data on the elderly's living environment and behavioral information, and achieve effective monitoring of the elderly's home safety and health status, thereby providing the elderly with a safer, more convenient and comfortable home-based care environment.

[0003] However, most existing home-based elderly care intelligent monitoring systems are based on a post-event alarm model, which has many technical limitations. First, the sensing dimension is limited, focusing only on collecting basic environmental data such as gas and smoke, failing to combine it with the elderly's movement trajectory for comprehensive analysis. This results in the inability to identify the risk correlation between environment and behavior, such as the inability to determine whether the elderly's behavior in a specific environment will cause danger. Second, the data processing stage lacks professional optimization. Sensors are prone to false triggering due to various factors during data collection, leading to misjudgments of alarms and causing unnecessary trouble for the elderly and their families. Third, risk intervention is passive, only pushing alarm information after danger occurs, lacking the ability to predict and proactively eliminate risks in advance, and failing to effectively prevent accidents. Fourth, the interaction design has an imbalance problem. Some systems only push alarm information to family members in one direction without synchronously intervening in the results, making it impossible for family members to fully understand the situation. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide an IoT-based intelligent health monitoring system for home-based elderly care. This system utilizes a multi-dimensional sensing module, deploying dedicated sensors in key areas of the home to simultaneously collect micro-environmental data and the elderly person's movement trajectory data, providing a rich data source for comprehensive monitoring. A data preprocessing module filters outliers, standardizes formats, and completes data to ensure data quality and continuity. A spatial correlation risk prediction module accurately determines risk levels using a three-dimensional time-space-behavior correlation model, providing a scientific basis for risk intervention. An intelligent linkage avoidance execution module automatically triggers targeted avoidance measures based on risk information, proactively eliminating risks. A two-way information interaction module establishes an information exchange channel between the elderly person and their family, balancing the family's need for information with the elderly person's operating habits, while also protecting privacy through data scope limitations. This provides more comprehensive, accurate, and intelligent home safety protection and health monitoring services for elderly people living alone or of advanced age.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: an IoT-based home-based elderly care intelligent health monitoring system, which includes: a multi-dimensional perception module, a data preprocessing module, a spatial correlation risk prediction module, an intelligent linkage avoidance execution module, and a two-way information interaction module; The multi-dimensional sensing module is deployed in the home kitchen, bathroom, balcony and public areas to collect micro-environment data and elderly movement trajectory data; The data preprocessing module performs outlier filtering, format standardization, and data completion on the raw data. The spatial correlation risk prediction module determines the risk level through a three-dimensional correlation model of time, space, and behavior. The intelligent linkage avoidance execution module triggers risk elimination measures in conjunction with intelligent devices. The two-way information interaction module enables two-way synchronization of safety information between the elderly and their families.

[0006] Furthermore, the multi-dimensional sensing module deploys dedicated sensors in key areas of the home to simultaneously collect micro-environment data and elderly movement trajectory data. Specifically, the kitchen area deploys gas concentration sensors and temperature and humidity sensors; the bathroom area deploys a floor slippage sensor based on pressure and humidity dual parameters and a waterproof human presence sensor; the balcony area deploys an infrared ranging object location sensor and a door and window status sensor; and public areas, including corridors, bedroom doors, and bathroom doors, deploy non-contact millimeter-wave radar sensors that do not collect image information, while also attaching sensors to monitor exposed electrical wires in public areas.

[0007] Furthermore, the data preprocessing module receives all the collected raw data and uses a preset data cleaning algorithm to identify and remove instantaneous fluctuation data caused by sensor mis-triggers and invalid data that exceeds the reasonable collection range. After filtering outliers, the data format is standardized, and a three-dimensional structured identifier of timestamp-region label-data type is automatically added to each valid data. The format difference of the output data of different sensors is eliminated by the unified format. When a temporary fault of a sensor is detected, which causes data interruption, a basic data model is built based on the historical data of the same period in the area where the faulty sensor is located. Combined with the real-time collection data of adjacent sensors of the same type, the missing data during the fault period is temporarily filled in by the data interpolation algorithm.

[0008] Furthermore, the spatial correlation risk prediction module matches the temporal characteristics of the elderly's movement trajectory with the temporal changes of micro-environmental data through temporal correlation. It binds the spatial route of the elderly's movement to the micro-environmental risk of the corresponding area through spatial correlation. Combined with behavioral correlation, it dynamically adjusts the monitoring frequency of micro-environmental data according to the elderly's daily behavioral habits. Based on the three-dimensional correlation results, it determines the three levels of risk: low, medium, and high.

[0009] Furthermore, the spatial correlation risk prediction module matches the time characteristics of the elderly's movement trajectory with the time changes of the microenvironment data through time correlation. The specific steps are as follows: After receiving standardized data with three-dimensional labels output by the data preprocessing module, the module parses the time attribute information from the elderly's movement trajectory data, extracts the time characteristics of the elderly's actions through time series analysis technology, including high-frequency activity periods, the start and end times of a single action, and the periodic parameters of the action interval, forming a structured action time feature set; at the same time, it performs continuous analysis on the timestamp information of the microenvironment data, captures the change patterns, change rates, and change nodes of various microenvironment parameters over time, generates a microenvironment time change feature set, and then calls the time axis alignment engine to map the action time feature set and the microenvironment time change feature set onto the same time axis, constructs a time correlation analysis matrix, and quantitatively analyzes the overlapping areas, sequential relationships, and continuous matching duration of the two types of features on the time axis based on a preset time correlation rule library, thus completing the accurate matching of the elderly's movement trajectory time characteristics with the time changes of the microenvironment data.

[0010] Furthermore, the spatial association risk prediction module binds the spatial routes of elderly people's movements with the micro-environmental risks of corresponding areas through spatial association. The specific steps are as follows: extracting spatial attribute information of the elderly people's movement trajectories from standardized data; relying on the location markers of millimeter-wave radar sensors deployed in public areas and sensors in various areas; using spatial coordinate transformation technology to parse the elderly people's movement data into a continuous spatial route coordinate sequence containing the starting point, intermediate points, and the ending point; constructing an elderly people's movement spatial route dataset; and synchronously associating the timestamp information corresponding to each coordinate point; simultaneously, based on the regional labels of the micro-environmental data and combined with the preset home space zoning model, associating the risk status corresponding to various micro-environmental parameters with specific spatial areas; generating a micro-environmental risk spatial distribution dataset containing regional coordinate range, risk type, and risk status; then, based on the pre-constructed home space coordinate system, performing spatial overlay analysis on the elderly people's movement spatial route dataset and the micro-environmental risk spatial distribution dataset to identify all spatial areas traversed by the elderly people's movement routes; and then, based on the preset spatial association rule library, establishing a mapping relationship between the areas traversed by the route and the micro-environmental risks of those areas; completing the binding of the elderly people's movement spatial routes with the corresponding regional micro-environmental risks; and generating a spatial association result containing route-region-risk.

[0011] Furthermore, the spatial correlation risk prediction module dynamically adjusts the micro-environment data monitoring frequency based on the elderly's daily behavioral habits, in conjunction with behavioral correlation. The specific steps are as follows: From the pre-processed elderly movement trajectory data, the type, time period, and corresponding spatial area information of the elderly's fixed behaviors are extracted. The trajectory data within a continuous period is analyzed using a behavioral pattern mining algorithm to construct a feature library of the elderly's daily behavioral habits, including behavioral identifiers, time ranges, and spatially correlated areas. Simultaneously, a behavior-micro-environment correlation rule library is established, pre-setting the mapping relationship between different behavior types and key micro-environment parameters within the corresponding spatial area. Key parameters for monitoring the micro-environment for various behavioral correlations are clarified. Real-time time information is acquired and matched with the behavior time range in the behavioral habit feature library to determine if the current time is within the occurrence window of the elderly's habitual behavior. If a match is successful, the key micro-environment parameters for the corresponding spatial area are determined based on the correlation rule library, and a frequency increase command is sent to the multi-dimensional perception module. If it is within a non-habitual behavior period or there is no matching behavior, a frequency decrease command is sent. Simultaneously, the elderly's movement trajectory data is periodically updated and analyzed. When a change in behavioral habits is detected, the behavioral habit feature library and correlation rules are automatically updated.

[0012] Furthermore, the spatial correlation risk prediction module analyzes trajectory data within a continuous period using a behavioral pattern mining algorithm. The algorithm formula is as follows: Where C represents the confidence score of the behavior habit, which measures the probability that a certain behavior will become a daily habit of the elderly. It refers to the frequency of behavior within a continuous period, specifically the total number of times the target behavior is identified from trajectory data collected by millimeter-wave radar within a preset continuous analysis period. It is the deviation rate of the time of occurrence of a behavior, which refers to the degree of relative deviation between the actual time of occurrence of a target behavior and the average time of occurrence of that behavior within a continuous period. It is the deviation rate of the behavior occurrence space, which refers to the degree of relative deviation between the actual occurrence space of the target behavior and the average occurrence space of the behavior within a continuous period. It is the time deviation weighting coefficient. It is the spatial deviation weighting coefficient.

[0013] Furthermore, the spatial association risk prediction module determines low / medium / high risk levels based on the three-dimensional association results. The specific steps are as follows: The temporal association results, spatial association results, and behavioral association results are fused to construct a three-dimensional association comprehensive dataset containing temporal matching degree, spatial overlap degree, behavioral association strength, and corresponding micro-environmental risk parameters. A three-level risk judgment rule base is built-in, which clearly defines the combination conditions for time matching degree thresholds, spatial overlap thresholds, behavioral association coefficients, and micro-environmental risk parameter thresholds. It also defines the mapping relationship between different combinations of conditions and low, medium, and high risk levels, and configures corresponding weight coefficients for each association dimension parameter based on preset... The weighting coefficients are used to calculate the comprehensive risk assessment value by weighting the parameters in the three-dimensional associated comprehensive dataset. The comprehensive assessment value is then compared with the level classification threshold in the rule base. At the same time, it is verified whether each associated parameter meets the combination conditions of the corresponding risk level. If the comprehensive assessment value and the parameter combination conditions both match a certain risk level, the level judgment result is directly output. If the assessment value and the combination conditions do not completely match, the priority judgment mechanism is activated. The status of the micro-environment risk parameters is used as the core priority basis, and the final risk level is determined by combining the temporal and spatial correlation tightness. Finally, a risk assessment report containing the risk level, associated dimension parameters, and judgment basis is generated.

[0014] Compared with existing technologies, this IoT-based smart health monitoring system for home-based elderly care has the following advantages: I. This invention, through a spatial correlation risk prediction module, can accurately extract the temporal features of elderly people's movement trajectories and microenvironment data in temporal correlation processing, and construct a temporal correlation analysis matrix to achieve accurate matching of the two types of features. In spatial correlation processing, the elderly people's movement data is parsed into a continuous spatial route coordinate sequence, and spatial overlay analysis is performed with the microenvironment risk spatial distribution dataset to establish a mapping relationship between routes and regional microenvironment risks. Behavioral correlation processing mines the elderly people's behavioral habits, constructs a behavioral habit feature library, and establishes a behavior-microenvironment correlation rule library. Finally, the three-dimensional correlation results are integrated to determine the risk level, effectively avoiding the limitations of single-dimensional judgment, and can more accurately identify potential risks, providing a reliable basis for subsequent avoidance measures.

[0015] Second, this invention continuously updates the elderly's behavioral habit feature database through a spatial correlation risk prediction module, while simultaneously activating a dynamic monitoring frequency adjustment engine. Based on the current time and the behavioral time range, it adjusts the monitoring frequency of the multi-dimensional sensing module in real time, increasing the frequency during habitual behavior periods and decreasing the frequency during non-habitual periods. This ensures the effectiveness of monitoring and improves resource utilization efficiency. In the risk level determination process, it calls upon the built-in three-level risk determination rule library, performs weighted calculations on the comprehensive dataset parameters based on preset weight coefficients, obtains a comprehensive risk assessment value, and determines the final risk level by combining micro-environment risk parameters and the tightness of time and spatial correlation. This allows for timely capture of risk changes, providing more real-time and effective safety guarantees for home-based elderly care.

[0016] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0018] Figure 1 This is a structural block diagram of an IoT-based smart health monitoring system for home-based elderly care. Figure 2 A flowchart of an IoT-based smart health monitoring system for home-based elderly care; Figure 3 This is a flowchart illustrating the steps of the spatial correlation risk prediction module in an IoT-based smart health monitoring system for home-based elderly care. Detailed Implementation

[0019] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0020] This invention provides an IoT-based intelligent health monitoring system for home-based elderly care, such as... Figure 1 As shown, the system comprises a multi-dimensional perception module, a data preprocessing module, a spatial correlation risk prediction module, an intelligent linkage avoidance execution module, and a two-way information interaction module. The multi-dimensional perception module deploys dedicated sensors in key areas of the home to simultaneously collect micro-environmental data and the elderly person's movement trajectory data, providing a rich data source for comprehensive monitoring. The data preprocessing module filters outliers, standardizes formats, and completes data to ensure data quality and continuity. The spatial correlation risk prediction module accurately determines the risk level using a three-dimensional correlation model of time, space, and behavior, providing a scientific basis for risk intervention. The intelligent linkage avoidance execution module automatically triggers targeted avoidance measures based on risk information, proactively eliminating risks. The two-way information interaction module establishes an information exchange channel between the elderly person and their family, balancing the family's need for information with the elderly person's operating habits, while also protecting privacy through data scope limitations. This provides more comprehensive, accurate, and intelligent home safety protection and health monitoring services for elderly people living alone or of advanced age.

[0021] In one embodiment, the multi-dimensional sensing module is activated, and the gas concentration sensor and temperature and humidity sensor deployed in the kitchen, the floor slippage sensor and waterproof human presence sensor in the bathroom, the infrared ranging object location sensor and door and window status sensor on the balcony, and the non-contact millimeter-wave radar sensor and exposed wire monitoring sensor in public areas (corridor, bedroom doorway, bathroom doorway) simultaneously enter working state. Each sensor collects micro-environmental data and elderly movement trajectory data of the corresponding area according to the initially set collection frequency. The micro-environmental data includes gas concentration, temperature and humidity, floor slippage, distance between objects and balcony edges, door and window status, and wire insulation status. The elderly movement trajectory data includes walking route, dwell time, and activity period. All collected raw data are continuously sent to the data preprocessing module through a real-time data transmission link.

[0022] After receiving the raw data from the multi-dimensional sensing module, the data preprocessing module initiates its internal data processing flow: It employs a pre-set professional data cleaning algorithm to identify and remove transient fluctuations caused by sensor malfunctions and invalid data exceeding the reasonable collection range. After outlier filtering, it automatically adds a three-dimensional structured identifier—timestamp, region label, and data type—to each valid data point (the timestamp is accurate to the moment of data collection, the region label corresponds to the specific residential area where the data was collected, and the data type clearly distinguishes between micro-environment data and elderly person's movement trajectory data). If a temporary sensor malfunction causes data interruption, a basic data model is constructed based on historical data from the same period in the area where the malfunctioning sensor is located. Combined with real-time data from adjacent sensors of the same type, a data interpolation algorithm is used to temporarily fill in the missing data during the malfunction period. After the above processing, the data preprocessing module pushes the standardized and continuous data to the spatial correlation risk prediction module in real time through a dedicated data interface.

[0023] After receiving standardized data from the data preprocessing module, the spatial correlation risk prediction module initiates a three-dimensional correlation analysis process of "time-space-behavior": First, time correlation processing is performed. The time feature extraction unit parses time attribute information from the elderly's movement trajectory data, extracting core time parameters such as high-frequency activity periods, start and end times of single actions, and action intervals through time series analysis technology, forming a set of elderly movement time features. Simultaneously, based on the timestamp information of the micro-environment data, the module captures the patterns, rates of change, and nodes of change of various micro-environment parameters over time, generating a set of micro-environment time change features. Subsequently, the time axis alignment engine, using millisecond-level unified time granularity as a benchmark, maps the elderly's movement time feature set and the micro-environment time change feature set onto the same time axis, constructing a time correlation analysis matrix. Based on a preset time correlation rule library, it quantitatively analyzes the overlapping areas, sequential relationships, and continuous matching durations of the two types of features on the time axis, achieving accurate matching between the two. Next, spatial correlation processing is performed, relying on the location of millimeter-wave radar sensors in public areas and sensors in various areas. The system identifies elderly movement data by using spatial coordinate transformation technology to parse it into a continuous spatial route coordinate sequence containing a start point, waypoints, and end point, thus constructing an elderly movement spatial route dataset. Simultaneously, based on the regional labels of the micro-environment data and a pre-defined home space zoning model, it associates the risk status corresponding to various micro-environment parameters with specific spatial areas, generating a micro-environment risk spatial distribution dataset containing regional coordinate ranges, risk types, and risk statuses. Using a pre-constructed home space coordinate system as a benchmark, the elderly movement spatial route dataset and the micro-environment risk spatial distribution dataset are spatially overlaid. A coordinate matching algorithm identifies all spatial areas traversed by the elderly's movement routes, and based on a pre-defined spatial association rule base, a one-to-one mapping relationship is established between the route-traversed areas and the micro-environment risk of those areas, completing the spatial association binding. Then, behavioral association processing is performed. From the pre-processed elderly movement trajectory data, core information such as the type of fixed behavior, the time period, and the corresponding spatial area are extracted. A behavioral pattern mining algorithm is used to analyze the trajectory data within a continuous period. The algorithm formula is as follows: Where C represents the confidence score of the behavior habit, which measures the probability that a certain behavior will become a daily habit of the elderly. It refers to the frequency of behavior within a continuous period, specifically the total number of times the target behavior is identified from trajectory data collected by millimeter-wave radar within a preset continuous analysis period. It is the deviation rate of the time of occurrence of a behavior, which refers to the degree of relative deviation between the actual time of occurrence of a target behavior and the average time of occurrence of that behavior within a continuous period. It is the deviation rate of the behavior occurrence space, which refers to the degree of relative deviation between the actual occurrence space of the target behavior and the average occurrence space of the behavior within a continuous period. It is the time deviation weighting coefficient. This involves a spatial deviation weighting coefficient, constructing a feature library of elderly daily behavioral habits that includes behavioral identifiers, time ranges, and spatially related areas. Simultaneously, a behavior-microenvironment association rule library is established, pre-setting the mapping relationship between different behavior types and key microenvironment parameters within corresponding spatial areas, clarifying the key parameters for monitoring the microenvironment associated with various behaviors. A dynamic monitoring frequency adjustment engine acquires current time information in real time and matches it with the behavior time range in the behavioral habit feature library to determine if the current time is within the occurrence window of the elderly's habitual behavior. If a match is successful, the key microenvironment parameters for the corresponding spatial area are determined based on the behavior-microenvironment association rule library, and a monitoring frequency increase command is sent to the multi-dimensional perception module. If it is outside the habitual behavior period or there is no matching behavior, a monitoring frequency decrease command is sent to the multi-dimensional perception module. The module also periodically updates and analyzes the elderly's movement trajectory data, automatically updating the behavioral habit feature library and the behavior-microenvironment association rule library when changes in the elderly's behavioral habits are detected. Finally, a risk level determination is performed. The module integrates the time association results, spatial association results, and behavioral association results to construct a feature library including time matching degree, spatial overlap, behavioral association strength, and risk level assessment. The system generates a three-dimensional correlation comprehensive dataset corresponding to micro-environment risk parameters. Then, it calls the built-in three-level risk judgment rule base (which specifies the combination conditions of time matching threshold, spatial overlap threshold, behavioral correlation coefficient, and micro-environment risk parameter thresholds, defines the mapping relationship between different combinations of conditions and low, medium, and high risk levels, and configures corresponding weight coefficients for each correlation dimension parameter). Based on the preset weight coefficients, the risk assessment engine performs weighted calculations on each parameter in the three-dimensional correlation comprehensive dataset to obtain a comprehensive risk assessment value. The comprehensive risk assessment value is compared with the level classification thresholds in the rule base, while simultaneously verifying whether each correlation parameter meets the combination conditions for the corresponding risk level. If both the comprehensive assessment value and the parameter combination conditions match a certain risk level, the level judgment result is directly output. If the assessment value and combination conditions do not completely match, a priority judgment mechanism is activated, using the micro-environment risk parameter status as the core priority basis, combined with the tightness of time and spatial correlation to determine the final risk level (low / medium / high). Finally, the spatial correlation risk prediction module generates a risk assessment report containing the risk level, correlation dimension parameters, and judgment basis, and pushes it to the intelligent linkage avoidance execution module.

[0024] After receiving the risk assessment report output by the spatial correlation risk prediction module, the intelligent linkage avoidance execution module extracts the risk level and risk type information from it, and automatically links the corresponding smart home devices according to the preset risk-measure mapping rules: If a high risk of gas leakage is determined, the module immediately sends a shut-off command to the smart gas valve and a start command to the kitchen exhaust fan; if a high / medium risk of slippery ground is determined, the module sends a start command to the heating and dehumidification module built into the bathroom anti-slip mat and a shut-off command to the bathroom floor faucet; if a high risk of falling objects is determined, the module first sends a safety reminder playback command to the balcony voice prompt; if no manual adjustment of the object's position is detected within a preset time, it sends an action command to the balcony micro-robotic arm to control it to move the object to a safe area inside the balcony; if a high / medium risk of route safety is determined, the module sends a light-on command to the floor LED strips in the corridor and bathroom and a shut-off command to the lights in the obstacle area along the route; during the execution of the above avoidance measures, the intelligent linkage avoidance execution module collects the operating status information of each device in real time and synchronously feeds this status information back to the two-way information interaction module.

[0025] After receiving the device operation status information from the intelligent linkage avoidance execution module, the two-way information interaction module initiates the information interaction process: the family member's dedicated APP receives the information transmitted by the module in real time, automatically organizes it, and pushes complete risk information to the family member, including risk type, occurrence area, intervention measures, and current status; the family member can view historical risk records through the APP, and can also customize risk warning thresholds within the APP; at the same time, the voice prompt device on the elderly's end receives the risk notification command sent by the module and plays the preset risk notification content; if the elderly press the confirmation button on the simple panel, the two-way information interaction module immediately captures the feedback signal and pushes the information that the elderly are aware of the risk to the family member's APP; if the elderly trigger the emergency call button, a distress signal is simultaneously sent to the family member's APP and the preset emergency contact, completing the risk intervention and information interaction process of the entire system.

[0026] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A smart health monitoring system for home-based elderly care based on the Internet of Things, characterized in that: The system includes: a multi-dimensional perception module, a data preprocessing module, a spatial correlation risk prediction module, an intelligent linkage avoidance execution module, and a two-way information interaction module; The multi-dimensional sensing module is deployed in the home kitchen, bathroom, balcony and public areas to collect micro-environment data and elderly movement trajectory data; The data preprocessing module performs outlier filtering, format standardization, and data completion on the raw data. The spatial correlation risk prediction module determines the risk level through a three-dimensional correlation model of time, space, and behavior. The intelligent linkage avoidance execution module triggers risk elimination measures in conjunction with intelligent devices. The two-way information interaction module enables two-way synchronization of safety information between the elderly and their families.

2. The IoT-based intelligent health monitoring system for home-based elderly care according to claim 1, characterized in that, The multi-dimensional sensing module collects micro-environmental data and elderly movement trajectory data simultaneously by deploying dedicated sensors in key areas of the home. Specifically, the kitchen area is equipped with gas concentration sensors and temperature and humidity sensors; the bathroom area is equipped with a floor slippage sensor based on pressure and humidity dual parameters and a waterproof human presence sensor; the balcony area is equipped with an infrared ranging object location sensor and a door and window status sensor; and public areas, including corridors, bedroom doors, and bathroom doors, are equipped with non-contact millimeter-wave radar sensors that do not collect image information. At the same time, exposed wire monitoring sensors are attached to public areas.

3. The IoT-based intelligent health monitoring system for home-based elderly care according to claim 1, characterized in that, The data preprocessing module receives all the collected raw data and uses a preset data cleaning algorithm to identify and remove instantaneous fluctuation data caused by sensor mis-triggers and invalid data that exceeds the reasonable collection range. After filtering outliers, the data format is standardized, and a three-dimensional structured identifier of timestamp, region label, and data type is automatically added to each valid data. The format difference of the output data of different sensors is eliminated by the unified format. When a temporary fault of a sensor is detected, which causes data interruption, a basic data model is built based on the historical data of the same period in the area where the faulty sensor is located. Combined with the real-time collection data of adjacent sensors of the same type, the missing data during the fault period is temporarily filled in by the data interpolation algorithm.

4. The IoT-based intelligent health monitoring system for home-based elderly care according to claim 1, characterized in that, The spatial correlation risk prediction module matches the temporal characteristics of the elderly's movement trajectory with the temporal changes of micro-environment data through temporal correlation. It binds the spatial route of the elderly's movement to the micro-environment risk of the corresponding area through spatial correlation. Combined with behavioral correlation, it dynamically adjusts the monitoring frequency of micro-environment data according to the elderly's daily behavioral habits. Based on the three-dimensional correlation results, it determines the risk level as low / medium / high.

5. The IoT-based intelligent health monitoring system for home-based elderly care according to claim 4, characterized in that, The spatial correlation risk prediction module matches the time characteristics of the elderly's movement trajectory with the time changes of micro-environment data through time correlation. The specific steps are as follows: After receiving standardized data with three-dimensional labels output by the data preprocessing module, the module parses the time attribute information from the elderly's movement trajectory data. It extracts the time characteristics of the elderly's actions through time series analysis technology, including high-frequency activity periods, the start and end times of a single action, and the periodic parameters of the action interval, forming a structured action time feature set. At the same time, it performs continuous analysis on the timestamp information of the micro-environment data, captures the change patterns, change rates, and change nodes of various micro-environment parameters over time, and generates a micro-environment time change feature set. Then, it calls the time axis alignment engine to map the action time feature set and the micro-environment time change feature set onto the same time axis, constructs a time correlation analysis matrix, and performs quantitative analysis on the overlapping areas, sequential relationships, and continuous matching duration of the two types of features on the time axis based on a preset time correlation rule library, thus completing the accurate matching of the elderly's movement trajectory time characteristics with the time changes of micro-environment data.

6. The IoT-based intelligent health monitoring system for home-based elderly care according to claim 4, characterized in that, The spatial association risk prediction module binds the spatial routes of elderly people's movements with the micro-environmental risks of corresponding areas through spatial association. The specific steps are as follows: Spatial attribute information of the elderly person's movement trajectory is extracted from standardized data. Relying on millimeter-wave radar sensors deployed in public areas and the location markers of sensors in each area, spatial coordinate transformation technology is used to parse the elderly person's movement data into a continuous spatial route coordinate sequence containing the start point, intermediate points, and end point, constructing an elderly person's movement spatial route dataset. Simultaneously, the timestamp information corresponding to each coordinate point is associated. At the same time, based on the regional labels of the micro-environmental data and combined with a preset home space zoning model, the risk status corresponding to various micro-environmental parameters is associated with specific spatial areas, generating a micro-environmental risk spatial distribution dataset containing regional coordinate ranges, risk types, and risk status. Then, using the pre-constructed home space coordinate system as a benchmark, spatial overlay analysis is performed on the elderly person's movement spatial route dataset and the micro-environmental risk spatial distribution dataset to identify all spatial areas traversed by the elderly person's movement route. Finally, based on a preset spatial association rule base, a mapping relationship is established between the areas traversed by the route and the micro-environmental risks of those areas, completing the binding of the elderly person's movement spatial route with the corresponding regional micro-environmental risks, generating a spatial association result containing route-area-risk.

7. The IoT-based intelligent health monitoring system for home-based elderly care according to claim 4, characterized in that, The spatial correlation risk prediction module dynamically adjusts the monitoring frequency of microenvironment data based on the elderly's daily behavioral habits, combining behavioral correlation. The specific steps are as follows: From the pre-processed elderly movement trajectory data, the type, time period, and corresponding spatial area information of the elderly's fixed behaviors are extracted. The trajectory data within a continuous period is analyzed using a behavioral pattern mining algorithm to construct a feature library of the elderly's daily behavioral habits, including behavioral identifiers, time ranges, and spatially correlated areas. Simultaneously, a behavior-microenvironment correlation rule library is established, pre-setting the mapping relationship between different behavior types and key microenvironment parameters within the corresponding spatial area. Key parameters for monitoring the microenvironment for various behavioral correlations are clarified. Current time information is acquired in real time and matched with the behavior time range in the behavioral habit feature library to determine if the current period falls within the occurrence window of the elderly's habitual behavior. If a match is successful, the key microenvironment parameters for the corresponding spatial area are determined based on the correlation rule library, and a frequency increase command is sent to the multi-dimensional perception module. If the period is outside of habitual behavior or there is no matching behavior, a frequency decrease command is sent. Simultaneously, the elderly's movement trajectory data is periodically updated and analyzed. When a change in behavioral habits is detected, the behavioral habit feature library and correlation rules are automatically updated.

8. The IoT-based intelligent health monitoring system for home-based elderly care according to claim 1, characterized in that, The spatial correlation risk prediction module analyzes trajectory data within a continuous period using a behavioral pattern mining algorithm. The algorithm formula is as follows: Where C represents the confidence score of the behavior habit, which measures the probability that a certain behavior will become a daily habit of the elderly. It refers to the frequency of behavior within a continuous period, specifically the total number of times the target behavior is identified from trajectory data collected by millimeter-wave radar within a preset continuous analysis period. It is the deviation rate of the time of occurrence of a behavior, which refers to the degree of relative deviation between the actual time of occurrence of a target behavior and the average time of occurrence of that behavior within a continuous period. It is the deviation rate of the behavior occurrence space, which refers to the degree of relative deviation between the actual occurrence space of the target behavior and the average occurrence space of the behavior within a continuous period. It is the time deviation weighting coefficient. It is the spatial deviation weighting coefficient.

9. The IoT-based intelligent health monitoring system for home-based elderly care according to claim 1, characterized in that, The spatial association risk prediction module determines low / medium / high risk levels based on the three-dimensional association results. The specific steps are as follows: the temporal association results, spatial association results, and behavioral association results are fused to construct a three-dimensional association comprehensive dataset containing temporal matching degree, spatial overlap degree, behavioral association strength, and corresponding micro-environmental risk parameters. A three-level risk judgment rule base is built-in. The rule base specifies the combination conditions of temporal matching degree threshold, spatial overlap threshold, behavioral association coefficient, and micro-environmental risk parameter threshold. At the same time, the mapping relationship between different combination conditions and low, medium, and high risk levels is defined, and corresponding weight coefficients are configured for each association dimension parameter. Based on the preset weight coefficients, the parameters in the three-dimensional association comprehensive dataset are weighted and calculated to obtain a comprehensive risk assessment value. The comprehensive assessment value is then compared with the level classification threshold in the rule base. At the same time, it is verified whether each association parameter meets the combination conditions of the corresponding risk level. If the comprehensive assessment value and parameter combination conditions both match a certain risk level, the judgment result of that level is directly output. If the assessed value does not completely match the combined conditions, a priority determination mechanism is activated. The status of microenvironmental risk parameters is used as the core priority basis, and the final risk level is determined by combining the tightness of temporal and spatial correlation. Finally, a risk assessment report containing the risk level, correlation dimension parameters, and determination basis is generated.

Images