Home-based care system integrating traditional Chinese medicine knowledge and Internet of Things technology

Abstract

The invention discloses a home-based care system fusing traditional Chinese medicine knowledge and an internet of things technology. The system comprises a comprehensive data acquisition module used for acquiring multi-dimensional data of a user; the data transmission module is used for transmitting the multi-dimensional data to the edge calculation module; the edge calculation module performs data preprocessing, feature coding and data fusion on the multi-dimensional data to obtain complementary data and cooperative data; complementary data and cooperative data are processed through a standard time axis establishing sub-module, a home minimum event data set establishing sub-module and a complementary data and cooperative data time axis repositioning sub-module, and the processed data are output to a health condition prediction module; and the health condition prediction module performs key feature extraction by using the processing data, and generates a future health risk result through a prediction model. According to the invention, deep integration of the traditional Chinese medicine symptom time dimension and the Internet of Things data is realized, and the home health risk early warning precision is significantly improved.

Description

Technical Field

[0001] This invention relates to the field of smart elderly care technology, and in particular to a home-based elderly care system that integrates traditional Chinese medicine knowledge and Internet of Things technology. Background Technology

[0002] With the accelerating aging process, the demand for home-based health monitoring for the elderly is becoming increasingly urgent. Existing IoT health monitoring systems mainly rely on wearable devices to collect physiological indicators (such as heart rate and blood pressure). Although they can achieve real-time data upload, they have significant shortcomings.

[0003] In the prior art, patent publication number "CN115719141A" discloses a home-based elderly care system and method based on the Internet of Things. This system collects and trains data on both service providers and elderly individuals to obtain service scores for service providers and nutritional scores for elderly individuals. Based on these scores, it assesses and categorizes the service capabilities of service providers and the quality of life of elderly individuals, enabling service providers with different capabilities to efficiently serve elderly individuals at corresponding levels, effectively improving the overall service effectiveness. This invention addresses the problem in existing solutions that lack assessment and categorization of service providers and elderly individuals, preventing service providers with different capabilities from providing targeted services and failing to assess the relationship between service providers and elderly individuals, resulting in poor overall effectiveness of home-based elderly care.

[0004] Patent publication number "CN118502581A" discloses a home-based elderly care system based on virtual reality technology, including a virtual reality headset, a data acquisition module, a human-computer interaction interface device, a terminal device, a medical service module, a comparison and analysis module, a smart home controller, and intelligent home devices. The virtual reality headset is worn on the head of the elderly, utilizing virtual reality technology to enable the wearer to engage in audio-visual communication, virtual games, and social entertainment activities. This invention, by setting up a home-based elderly care system based on virtual reality technology, can provide intelligent home-based elderly care services. Simultaneously, by incorporating the virtual reality headset, data acquisition module, human-computer interaction interface device, terminal device, medical service module, comparison and analysis module, smart home controller, and intelligent home devices, the system can provide convenient medical services, social entertainment, and intelligent home control functions for the elderly, improving their quality of life.

[0005] While the above two methods utilize IoT and virtual reality technologies to implement home-based elderly care systems, they do not consider how to prevent diseases through these systems. In particular, the traditional Chinese medicine concept of "prevention of disease" has a clear natural advantage in elderly health management. Therefore, this invention combines IoT technology with traditional Chinese medicine knowledge to construct a home-based elderly care system. Summary of the Invention

[0006] This invention provides a home-based elderly care system that integrates traditional Chinese medicine knowledge and Internet of Things technology to overcome the problem that existing home-based elderly care systems cannot play a role in prevention and monitoring.

[0007] To achieve the above objectives, the technical solution of the present invention is as follows:

[0008] A home-based elderly care system that integrates traditional Chinese medicine knowledge and Internet of Things technology includes: a comprehensive data acquisition module, a data transmission module, an edge computing module, and a health status prediction module;

[0009] The integrated data acquisition module is used to acquire multi-dimensional data of home-based elderly care users based on Internet of Things (IoT) technology; the multi-dimensional data includes: environmental data, physiological data, dietary data, daily routine data, and activity data.

[0010] The data transmission module is used to transmit the collected multi-dimensional data to the edge computing module;

[0011] The edge computing module is used to process the received multi-dimensional data, construct a standard data axis, and generate a minimum home event dataset. The edge computing module includes a data processing submodule, a standard time axis establishment submodule, a minimum home event dataset establishment submodule, and a complementary data and collaborative data time axis relocation submodule.

[0012] The data processing submodule is used to perform data preprocessing, feature encoding, and data fusion on the multi-dimensional data to obtain complementary data and collaborative data; the complementary data is used to obtain a home event through the mutual supplementation of multiple complementary data; the collaborative data is used to infer the user's TCM symptoms through the collaboration and analysis of multiple collaborative data and based on TCM diagnostic theory.

[0013] The standard time axis establishment submodule is used to group and analyze the complementary and collaborative data using the K-Means clustering algorithm, and to construct a standard time axis based on the grouping analysis results and the time span of TCM symptoms in TCM diagnostic theory.

[0014] The Home Minimal Events Dataset Establishment Submodule is used to extract home events related to health status from five dimensions—environment, meals, daily routines, sleep, and activities—based on the grouping analysis results. For different health statuses, a home minimum event evidence chain is constructed on the standard timeline, and the home minimum event evidence chains are summarized to form a corresponding home minimum event dataset.

[0015] The complementary and collaborative data time axis repositioning submodule is used to map the complementary and collaborative data onto the standard time axis; perform time and frequency comparison calibration with the grouped analysis data of the standard time axis; re-standardize and reposition the calibrated complementary and collaborative data to obtain reconstructed complementary and collaborative data; and fuse the reconstructed complementary and collaborative data on the standard time axis to form time-aligned fused complementary and fused collaborative data.

[0016] The health status prediction module uses environmental data collected by the edge computing module and health and behavioral data from the minimum home event dataset to extract key features; it then constructs a prediction model and combines the key features to predict health risks, obtaining the user's future health risk prediction results; the key features are representative information reflecting environmental conditions, health status, and behavioral patterns.

[0017] Furthermore, the data transmission module includes a Wi-Fi communication module, a Bluetooth transmission module, and a ZigBee networking module; the Wi-Fi communication module is used to transmit the collected video and audio data; the Bluetooth transmission module is used to transmit physiological characteristic information collected through the wearable device; and the ZigBee networking module is used to transmit environmental data collected by the sensor nodes.

[0018] Furthermore, the specific steps of the data processing submodule include:

[0019] S11, Data preprocessing step, preprocessing the multi-dimensional data, including: outlier handling and missing data filling;

[0020] The outlier handling step involves setting upper and lower limits for the multi-dimensional data based on the data acquisition range of each sensor, and identifying and removing data that exceeds the corresponding upper and lower limits as outliers.

[0021] The missing data completion step involves performing a stationarity statistical test on the data. If the test result meets the preset stationarity conditions, the missing data is completed using the linear interpolation method. If the preset stationarity conditions are not met, the data is subjected to difference transformation or logarithmic transformation. At the same time, the nonlinear trend of the data is determined by the discontinuity saliency method. If the data is nonlinear, the missing data is completed using polynomial interpolation or spline interpolation.

[0022] S12, data feature encoding step, involves labeling the multi-dimensional data preprocessed in step S11 and encoding the labeled data according to preset encoding rules; the encoding rules include: ID card number, topic classification code, acquisition device code, home device number, event number, and sequence code;

[0023] S13, Data fusion step: Based on the principles of traditional Chinese medicine diagnostics, the encoded data is fused and divided into complementary data and collaborative data.

[0024] Furthermore, the specific steps for constructing the standard timeline using the standard timeline creation module are as follows:

[0025] S21. Based on the distance between the devices that collect the multi-dimensional data, the K-Means clustering algorithm is used to group and analyze the complementary data and the collaborative data.

[0026] S22. Based on the results of the grouping analysis and the time span of TCM symptoms according to the principles of TCM diagnostics, a standard time axis is constructed; wherein, the time span of TCM symptoms is divided according to any of the following cycles, including: natural cycles, the physiological or pathological change cycle patterns described in TCM health preservation theories, and the chronic disease conditioning cycle determined by the TCM case mechanism; the standard time axis is used to provide a unified time reference for users' activity data in different indoor spaces.

[0027] Furthermore, the specific execution steps of the complementary data and collaborative data time axis relocation submodule are as follows:

[0028] S31. Map the complementary data and collaborative data to the corresponding time points on the standard time axis to obtain mapped data; the mapped data includes the mapped complementary data and the mapped collaborative data.

[0029] S32. On the standard time axis, the mapped data is compared and calibrated with the grouped analysis data of the standard time axis in terms of time and frequency. For the mapped data whose time deviates from the standard time axis, it is adjusted according to the time pattern of the grouped analysis data of the standard time axis. For the mapped data with abnormal frequency, the sensor acquisition settings are changed or calibration is performed by data supplementation method so that the frequency of the mapped data reaches the reasonable range of the grouped analysis data of the standard time axis.

[0030] S33. Remap the calibrated mapping data onto the standard time axis and synchronize it with the grouped analysis data of the standard time axis; associate the time-synchronized mapping data with the standard minimum home events on the standard time axis to obtain reconstructed complementary and collaborative data.

[0031] S34. The reconstructed complementary and collaborative data are fused on a standard timeline to obtain the minimum home events. The fusion process includes time alignment fusion, information complementarity fusion, and information collaboration fusion. Time alignment fusion refers to aligning the complementary and collaborative data in the mapped data at the same point in time on the standard timeline. Information complementarity fusion refers to quantifying the reconstructed complementary data through regression analysis to obtain the minimum home events. Information collaboration fusion uses the minimum home events and the reconstructed collaborative data, combined with TCM diagnostic theories, to infer TCM symptoms and diseases.

[0032] Furthermore, the specific implementation process of the health status prediction module is as follows:

[0033] S41. Extract features from the environmental data collected by the edge computing module and the health and behavioral data in the minimum home events to obtain environmental data features, environmental impact features, health data features, behavioral data features and behavioral pattern features.

[0034] The extracted environmental data features are as follows:

[0035]

[0036] In the formula, t0 and t n These represent the start and end times of data collection, respectively; E j Let be the j-th environmental data point; μE be the mean of the environmental data; and γ be the weighting coefficient of the environmental data features.

[0037] The extracted environmental impact characteristics are:

[0038]

[0039] In the formula, Temp(t) represents the temperature data at time t, and μ Temp σ Temp , respectively, represent the mean and standard deviation of the temperature data; Hum(t) represents the humidity data at time t; μ Hum σ Hum , respectively, represent the mean and standard deviation of humidity data; Light(t) represents the light concentration data at time t, indicating indoor pollution levels; μ Light σ Light , respectively, represent the mean and standard deviation of the light intensity concentration data; Smoke(t) represents the smoke concentration data at time t, indicating indoor pollution levels; μ Smoke σ Smoke Here, μ represents the mean and standard deviation of the smoke concentration data; Noise(t) represents the noise level at time t; μ Noise σ Noiseα1, α2, α3, α4, and α5 are the mean and standard deviation of the noise data, respectively; α1, α2, α3, α4, and α5 are the weighting coefficients of the environmental data features.

[0040] The extracted health data features are:

[0041]

[0042] In the formula, HR(t) represents the heart rate data at time t; μ HR and σ HR These represent the mean and standard deviation of the heart rate data, respectively; BP(t) represents the blood pressure data at time t; μ BP and σ BP These represent the mean and standard deviation of the blood pressure data, respectively; BS(t) represents the blood glucose data at time t; μ BS and σ BS These represent the mean and standard deviation of blood glucose data, respectively; BT(t) represents the body temperature data at time t; μ BT and σ BT Here, BW(t) represents the mean and standard deviation of body temperature data, respectively; BW(t) represents body weight data at time t; μ BW and σ BW These represent the mean and standard deviation of blood glucose data, respectively; α6, α7, α8, α9, α 10 Weighting coefficients for health data features;

[0043] The extracted behavioral data features are:

[0044]

[0045] In the formula, S i This represents the data for the i-th step. A is the step-count feature extraction function; i This represents the i-th motion time data; L is the motion feature extraction function; i This is the i-th GPS coordinate data; β1, β2, and β3 are the location feature extraction functions; β1, β2, and β3 are the weight coefficients of the behavioral data features.

[0046] The extracted behavioral pattern features are:

[0047]

[0048] In the formula, D k For the kth daily activity data; ω daily (D k ) is the feature extraction function for daily activities; Ab k For the kth abnormal behavior data; ω abnormal (Ab k) is the abnormal behavior feature extraction function; β4 and β5 are the weight coefficients of abnormal and daily behavior data features;

[0049] S42. Based on the environmental data features, environmental impact features, health data features, behavioral data features, and behavioral pattern features, and combined with the decision tree algorithm, a prediction model is constructed. The expression of the prediction model is:

[0050]

[0051] In the formula, K is the final predicted health index, i.e., the predicted value; E is the environmental data characteristic; N(t) is the health data characteristic; B is the behavioral data characteristic; d(t) is the environmental impact characteristic; and D is the behavioral pattern characteristic.

[0052] S43. Based on the health index, behavioral data characteristics, and behavioral pattern characteristics, output a health report; manage the user's health according to the health report and TCM diseases and symptoms in the minimum home events; the health report includes health indicator change trends, health risk assessment, and health recommendations.

[0053] Beneficial effects:

[0054] 1. This invention constructs a Traditional Chinese Medicine (TCM) home-based health system based on the Internet of Things (IoT). Through its integrated data acquisition module, it obtains five-dimensional data on home-based elderly care users, overcoming the limitations of traditional systems that only collect physiological data. By utilizing a hybrid transmission protocol of Wi-Fi, Bluetooth, and ZigBee, it achieves efficient networking of various sensors, smart home appliances, wearable devices, and small medical devices, providing a comprehensive data foundation for TCM health analysis.

[0055] 2. This invention addresses the problem that home health sensing layer devices generate large volumes of data with multiple dimensions, and that this data has low value in assisting TCM clinical diagnosis. First, it introduces edge computing to preprocess the data. Second, based on the principles of TCM diagnostics, it divides home health data into complementary data and collaborative data, and proposes a solution for data fusion at three levels: device data, home events, and TCM symptoms.

[0056] 3. This invention extracts key features of home-based elderly care users through a health status prediction module and makes predictions based on a prediction model, thereby achieving disease prevention and control throughout the entire life cycle in a home setting. It can perform tasks such as health monitoring and chronic disease prevention and control for the elderly at low cost and high efficiency. Attached Figure Description

[0057] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0058] Figure 1 This is a flowchart illustrating the home-based elderly care system of the present invention.

[0059] Figure 2 This is a schematic diagram of the first architecture of the home-based elderly care system of the present invention;

[0060] Figure 3 This is a schematic diagram of the second architecture of the home-based elderly care system of the present invention;

[0061] Figure 4 This is a schematic diagram illustrating the time axis relocation of complementary and collaborative data in a home setting according to the present invention. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] This embodiment provides a home-based elderly care system that integrates traditional Chinese medicine knowledge and Internet of Things technology, such as Figure 1 As shown, it includes: a comprehensive data acquisition module, a data transmission module, an edge computing module, and a health status prediction module;

[0064] The integrated data acquisition module is used to acquire multi-dimensional data of home-based elderly care users based on Internet of Things (IoT) technology; the multi-dimensional data includes: environmental data, physiological data, dietary data, daily routine data, and activity data.

[0065] The data transmission module is used to transmit the collected multi-dimensional data to the edge computing module;

[0066] The edge computing module is used to process the received multi-dimensional data, construct a standard data axis, and generate a minimum home event dataset. The edge computing module includes a data processing submodule, a standard time axis establishment submodule, a minimum home event dataset establishment submodule, and a complementary data and collaborative data time axis relocation submodule.

[0067] The data processing submodule is used to perform data preprocessing, feature encoding, and data fusion on the multi-dimensional data to obtain complementary data and collaborative data; the complementary data is used to obtain a home event through the mutual supplementation of multiple complementary data; the collaborative data is used to infer the user's TCM symptoms through the collaboration and analysis of multiple collaborative data and based on TCM diagnostic theory.

[0068] The standard time axis establishment submodule is used to group and analyze the complementary and collaborative data using the K-Means clustering algorithm, and to construct a standard time axis based on the grouping analysis results and the time span of TCM symptoms in TCM diagnostic theory.

[0069] The Home Minimal Events Dataset Establishment Submodule is used to extract home events related to health status from five dimensions—environment, meals, daily routines, sleep, and activities—based on the grouping analysis results. For different health statuses, a home minimum event evidence chain is constructed on the standard timeline, and the home minimum event evidence chains are summarized to form a corresponding home minimum event dataset.

[0070] The complementary and collaborative data time axis repositioning submodule is used to map the complementary and collaborative data onto the standard time axis; perform time and frequency comparison calibration with the grouped analysis data of the standard time axis; re-standardize and reposition the calibrated complementary and collaborative data to obtain reconstructed complementary and collaborative data; and fuse the reconstructed complementary and collaborative data on the standard time axis to form time-aligned fused complementary and fused collaborative data.

[0071] The health status prediction module uses environmental data collected by the edge computing module and health and behavioral data from the minimum home event dataset to extract key features; it then constructs a prediction model and combines the key features to predict health risks, obtaining the user's future health risk prediction results; the key features are representative information reflecting environmental conditions, health status, and behavioral patterns.

[0072] Specifically, such as Figure 2 , Figure 3 As shown, a comprehensive data acquisition module is first built based on Internet of Things technology to acquire multi-dimensional data of home-based elderly care users in real time, including environmental data, physiological data, dietary data, daily routine data and activity data;

[0073] Secondly, the data transmission module aggregates multi-dimensional data to the edge computing module. The edge computing module executes a hierarchical processing flow: the data processing submodule preprocesses, encodes, and fuses the multi-dimensional data to generate complementary data (used to reconstruct complete home events) and collaborative data (inferring TCM symptoms based on TCM diagnostic theory). The complementary and collaborative data are used to establish a correlation mapping between TCM symptoms and home events. The standard timeline establishment submodule uses the K-Means clustering algorithm to group and analyze the complementary and collaborative data, and constructs a standardized time series framework based on the time span of TCM symptoms. The home minimum event dataset establishment submodule extracts key events from five dimensions: environment, meals, work and rest, sleep, and activities. It constructs a home minimum event evidence chain on the standard timeline and summarizes it into a dataset to compress the data size and retain key evidence for TCM health diagnosis. The complementary and collaborative data timeline relocation submodule maps the data to the standard timeline, performs time-frequency calibration, restandardization, and fusion alignment to form time-unified fused data, which is used to solve the time asynchrony problem of multi-source heterogeneous data.

[0074] Finally, the health status prediction module extracts features from the environmental data output by edge computing and the minimum home event dataset to build a prediction model to achieve dynamic early warning of health risks in traditional Chinese medicine.

[0075] Specifically, the data in the integrated data acquisition module is acquired through the following sub-modules:

[0076] The environmental data acquisition submodule is used to collect environmental data, including: air humidity, indoor temperature, indoor light intensity, CO2 concentration, smoke concentration, special sounds of the elderly, and indoor noise.

[0077] The physiological data acquisition submodule is used to collect user physiological data, including heart rate, blood oxygen saturation, and blood pressure.

[0078] The dietary data collection submodule is used to collect user dietary data; the dietary data includes: meal time, food color, food properties, and food taste;

[0079] The daily routine data collection submodule is used to collect user daily routine data; the daily routine data includes: sleep time, sleep duration, awake time, and awake duration;

[0080] The activity data collection submodule is used to collect users' home activity data; the home activity data includes: exercise duration, exercise time and exercise type.

[0081] In this embodiment, guided by the theory of traditional Chinese medicine diagnostics, and focusing on the elements required for traditional Chinese medicine diagnosis, four types of intelligent devices were selected, including basic sensors, smart home wearable devices, and small medical devices, to monitor five types of data: environment, physiology, diet, rest, and activity, forming a 360-degree digital profile of the elderly's home health.

[0082] Based on health appliance standards and traditional Chinese medicine (TCM) metadata standards, 20 TCM home health data elements were defined, and 21 datasets were defined around environment, physiology, diet, rest, and activities. A coding method was developed for the data elements and datasets to guide the collection of TCM home health scenario data. This embodiment defines the data type representation format and value range of the TCM home health scenario data elements, as shown in Table 1:

[0083] Table 1 Data Type Definitions

[0084]

[0085] Specifically, the data transmission module includes a Wi-Fi communication module, a Bluetooth transmission module, and a ZigBee networking module;

[0086] The Wi-Fi module transmits video and audio data, and provides extensive coverage by directly connecting to edge computing nodes;

[0087] The wearable device transmits physiological characteristic information through the Bluetooth module; the interconnection between portable smart devices adopts Bluetooth Low Energy (BLE) technology, which significantly reduces power consumption while maintaining the same communication range, and is suitable for short-range data transmission.

[0088] The ZigBee networking module transmits environmental data collected by sensor nodes; other terminal nodes use ZigBee wireless networking to aggregate the collected information to the ZigBee coordinator for comprehensive aggregation. However, the coordinator, based on the CC2530 processor, has limited data processing capabilities, so the ZigBee coordinator connects to the edge computing nodes via a wired connection.

[0089] In this embodiment, control is achieved through an STM32 chip, and the communication module supports three short-range data transmission protocols: ZigBee, Bluetooth, and Wi-Fi, enabling synchronous data collection and uploading of multiple IoT nodes in the elderly's home environment.

[0090] Specifically, the specific steps of the data processing submodule include:

[0091] S11, Data preprocessing step, preprocessing the multi-dimensional data, including: outlier handling and missing data filling;

[0092] The outlier handling step involves setting upper and lower limits for the multi-dimensional data based on the data acquisition range of each sensor, and identifying and removing data that exceeds the corresponding upper and lower limits as outliers.

[0093] The missing data completion step involves performing a stationarity statistical test on the data. If the test result meets the preset stationarity conditions, the missing data is completed using the linear interpolation method. If the preset stationarity conditions are not met, the data is subjected to difference transformation or logarithmic transformation. At the same time, the nonlinear trend of the data is determined by the discontinuity saliency method. If the data is nonlinear, the missing data is completed using polynomial interpolation or spline interpolation.

[0094] S12, the data feature encoding step, involves labeling the multi-dimensional data preprocessed in step S11 and encoding the labeled data according to preset encoding rules; the encoding rules include: ID card number, topic classification code, data acquisition device code, home device number, event number, and sequence code; the encoding rules are as follows:

[0095] The ID number, an 18-digit number, is used to identify the user's region, age, and gender.

[0096] The subject-based classification coding includes five categories: environment, physiology, diet, activity, and sleep, which are coded with numbers 1 to 5 respectively;

[0097] The data acquisition devices are coded into categories such as smart home appliances, basic sensors, portable devices, and small medical devices, and are coded with numbers 1 to 4 respectively.

[0098] Home device serial number, a three-digit number, used to locate the data acquisition device, such as: 001 thermometer, 002 blood pressure monitor, etc.

[0099] Event number, a four-digit number, used to locate the data collection time;

[0100] The sequence code, consisting of three digits, represents the Nth data report submitted by a single device on a single day.

[0101] S13, Data fusion step, based on the principles of traditional Chinese medicine diagnostics, the encoded data is fused and divided into complementary data and collaborative data;

[0102] The complementary data is defined as data from different observation dimensions that are independent of each other and can complement and verify each other to provide more comprehensive information. In a specific embodiment, the home-based events obtained through complementary data are shown in Table 2.

[0103] Table 2 Complementary Data

[0104]

[0105] In the table, home behavior events are listed as multi-dimensional data collected from devices. A home event is composed of multiple complementary data. For example, the nighttime wake-up event is composed of data collected by sound sensors, sleep aids, motion sensors, millimeter-wave sensors, smart mattresses, and smart bracelets.

[0106] The collaborative data is defined as heterogeneous data in which different sensors have interdependent relationships and synergistic effects during the observation and information processing process; in a specific embodiment, the TCM symptoms obtained through collaborative data are shown in Table 3:

[0107] Table 3 Collaboration Data

[0108]

[0109]

[0110] In the table, behavioral TCM symptoms are listed as TCM symptoms of home-based events that are derived from the combination of multi-dimensional data. These symptoms are composed of multiple home-based events. For example, the symptoms of cold and heat are derived from dietary preferences, water intake, insomnia, drowsiness, frequent urination at night, diarrhea, indoor activities, air humidity, abnormal room temperature, and abnormal physiological indicators.

[0111] Specifically, the steps for constructing a standard timeline using the standard timeline creation module are as follows:

[0112] S21. Based on the distance between the devices that collect the multi-dimensional data, the K-Means clustering algorithm is used to group and analyze the complementary data and the collaborative data.

[0113] S22. Based on the results of the grouping analysis and the time span of TCM symptoms according to the principles of TCM diagnostics, a standard time axis is constructed; wherein, the time span of TCM symptoms is divided according to any of the following cycles, including: natural cycles, the physiological or pathological change cycle patterns described in TCM health preservation theories, and the chronic disease conditioning cycle determined by the TCM case mechanism; the standard time axis is used to provide a unified time reference for users' activity data in different indoor spaces.

[0114] Specifically, based on the impact of different home events on health, this invention extracts health-related home events from five perspectives: environment, physiology, dining, work and rest, sleep, and activity. Table 4 shows the minimum home events summarized by this invention.

[0115] Table 4 Minimum Home-Based Events

[0116]

[0117]

[0118] Taking a dining process as an example, a timeline is constructed from the start of cooking to the end of the meal. The devices involved are activated sequentially and exchange information in the following order: "induction cooker - range hood - refrigerator - camera." The information content is: "induction cooker starts and reports its working status - range hood starts and reports its working status - refrigerator reports being accessed due to food retrieval - camera captures and records dining behavior and specific time." In this process, the data provided by each device is considered complementary, collectively constituting complete data on a dining activity. Furthermore, the data collected from this dining event plays the role of collaborative data for analyzing patterns in dining habits; that is, these data work together to support the understanding and analysis of dining patterns.

[0119] Specifically, such as Figure 4 As shown, the specific execution steps of the complementary data and collaborative data time axis relocation submodule are as follows:

[0120] S31. Map the complementary data and collaborative data to the corresponding time points on the standard time axis to obtain mapped data; the mapped data includes the mapped complementary data and the mapped collaborative data.

[0121] S32. On the standard time axis, the mapped data is compared and calibrated with the grouped analysis data of the standard time axis in terms of time and frequency. For the mapped data whose time deviates from the standard time axis, it is adjusted according to the time pattern of the grouped analysis data of the standard time axis. For the mapped data with abnormal frequency, the sensor acquisition settings are changed or calibration is performed by data supplementation method so that the frequency of the mapped data reaches the reasonable range of the grouped analysis data of the standard time axis.

[0122] S33. Remap the calibrated mapping data onto the standard time axis and synchronize it with the grouped analysis data of the standard time axis; associate the time-synchronized mapping data with the standard minimum home events on the standard time axis to obtain reconstructed complementary and collaborative data.

[0123] S34. The reconstructed complementary and collaborative data are fused on a standard timeline to obtain the minimum home events. The fusion process includes time alignment fusion, information complementarity fusion, and information collaboration fusion. Time alignment fusion refers to aligning the complementary and collaborative data in the mapped data at the same point in time on the standard timeline. Information complementarity fusion refers to quantifying the reconstructed complementary data through regression analysis to obtain the minimum home events. Information collaboration fusion uses the minimum home events and the reconstructed collaborative data, combined with TCM diagnostic theories, to infer TCM symptoms and diseases.

[0124] In this embodiment, data is transmitted from the edge computing module to the health status prediction module based on the MQTT protocol. Specifically, the real-time process is as follows: the edge computing module uploads the fused data to the health status prediction module on the cloud server in the form of a JSON string using the MQTT protocol. MQTT is a publish / subscribe messaging protocol. It operates on the TCP / IP protocol suite and is designed for remote devices with low hardware performance and poor network conditions. The MQTT protocol is lightweight, simple, open, and easy to implement, making it widely applicable.

[0125] Specifically, the implementation process of the health status prediction module is as follows:

[0126] S41. Extract features from the environmental data collected by the edge computing module and the health and behavioral data in the minimum home events to obtain environmental data features, environmental impact features, health data features, behavioral data features and behavioral pattern features.

[0127] The extracted environmental data features are as follows:

[0128]

[0129] In the formula, t0 and t n These represent the start and end times of data collection, respectively; E j Let be the j-th environmental data point; μE be the mean of the environmental data; and γ be the weighting coefficient of the environmental data features.

[0130] The extracted environmental impact characteristics are:

[0131]

[0132] In the formula, Temp(t) represents the temperature data at time t, and μ Temmp σ Temp , respectively, represent the mean and standard deviation of the temperature data; Hum(t) represents the humidity data at time t; μ Hum σ Hum , respectively, represent the mean and standard deviation of humidity data; Light(t) represents the light concentration data at time t, indicating indoor pollution levels; μ Light σ Light , respectively, represent the mean and standard deviation of the light intensity concentration data; Smoke(t) represents the smoke concentration data at time t, indicating indoor pollution levels; μ Smoke σ Smoke , respectively, represent the mean and standard deviation of the smoke concentration data; Noise(t) represents the noise level at time t; μ Noise σ Noiseα1, α2, α3, α4, and α5 are the mean and standard deviation of the noise data, respectively; α1, α2, α3, α4, and α5 are the weighting coefficients of the environmental data features.

[0133] The extracted health data features are:

[0134]

[0135] In the formula, HR(t) represents the heart rate data at time t; μ HR and σ HR These represent the mean and standard deviation of the heart rate data, respectively; BP(t) represents the blood pressure data at time t; μ BP and σ BP These represent the mean and standard deviation of the blood pressure data, respectively; BS(t) represents the blood glucose data at time t; μ BS and σ BS These represent the mean and standard deviation of blood glucose data, respectively; BT(t) represents the body temperature data at time t; μ BT and σ BT Here, BW(t) represents the mean and standard deviation of body temperature data, respectively; BW(t) represents body weight data at time t; μ BW and σ BW These represent the mean and standard deviation of blood glucose data, respectively; α6, α7, α8, α9, α 10 Weighting coefficients for health data features;

[0136] The extracted behavioral data features are:

[0137]

[0138] In the formula, S i This represents the data for the i-th step. A is the step-count feature extraction function; i This represents the i-th motion time data; L is the motion feature extraction function; i This is the i-th GPS coordinate data; β1, β2, and β3 are the location feature extraction functions; β1, β2, and β3 are the weight coefficients of the behavioral data features.

[0139] The extracted behavioral pattern features are:

[0140]

[0141] In the formula, D k For the kth daily activity data; ω daily (D k ) is the feature extraction function for daily activities; Ab k For the kth abnormal behavior data; ω abnormal (Ab k) is the abnormal behavior feature extraction function; β4 and β5 are the weight coefficients of abnormal and daily behavior data features;

[0142] S42. Based on the environmental data features, environmental impact features, health data features, behavioral data features, and behavioral pattern features, and combined with the decision tree algorithm, a prediction model is constructed. The expression of the prediction model is:

[0143]

[0144] In the formula, K is the final predicted health index, i.e., the predicted value; E is the environmental data characteristic; N(t) is the health data characteristic; B is the behavioral data characteristic; E(t) is the environmental impact characteristic; and D is the behavioral pattern characteristic.

[0145] Among them, the user's health prediction is determined by the K value. The larger the K value, the better the health status and behavioral characteristics of the elderly; the smaller the K value, the worse the health status and behavioral characteristics of the elderly.

[0146] S43. Based on the health index, behavioral data characteristics, and behavioral pattern characteristics, output a health report; manage the user's health according to the health report and TCM diseases and symptoms in the minimum home events; the health report includes health indicator change trends, health risk assessment, and health recommendations.

[0147] In optional implementations, health recommendations cover aspects such as diet, exercise, medication reminders, and regular check-ups. For example, for older adults at risk of hypertension, the system can recommend a low-sodium diet, moderate exercise, and regular blood pressure monitoring; for older adults with low activity levels, a daily exercise plan can be developed to improve their physical fitness.

[0148] The community management system can provide corresponding elderly care service resources based on the health and behavioral data of seniors, such as medical support, psychological counseling, emergency response services, and daily living assistance. For example, when an elderly person's health deteriorates, the system will recommend nearby medical institutions or doctors; in the event of an accident, the system can quickly activate emergency response services to ensure safety.

[0149] In addition, the system can regularly adjust health management plans and service recommendations based on changes in the health and behavioral patterns of the elderly. For example, by analyzing recent health data, if improvements are found in health status, exercise and dietary recommendations can be adjusted appropriately; if the elderly person's activity range or habits change, the system will re-match and recommend suitable service resources.

[0150] The present invention has the following beneficial effects:

[0151] 1. This invention constructs a Traditional Chinese Medicine (TCM) home-based health system based on the Internet of Things (IoT). Through its integrated data acquisition module, it obtains five-dimensional data on home-based elderly care users, overcoming the limitations of traditional systems that only collect physiological data. By utilizing a hybrid transmission protocol of Wi-Fi, Bluetooth, and ZigBee, it achieves efficient networking of various sensors, smart home appliances, wearable devices, and small medical devices, providing a comprehensive data foundation for TCM health analysis.

[0152] 2. This invention addresses the problem that home health sensing layer devices generate large volumes of data with multiple dimensions, and that this data has low value in assisting TCM clinical diagnosis. First, it introduces edge computing to preprocess the data. Second, based on the principles of TCM diagnostics, it divides home health data into complementary data and collaborative data, and proposes a solution for data fusion at three levels: device data, home events, and TCM symptoms.

[0153] 3. This invention extracts key features of home-based elderly care users through a health status prediction module and makes predictions based on a prediction model, thereby achieving disease prevention and control throughout the entire life cycle in a home setting. It can perform tasks such as health monitoring and chronic disease prevention and control for the elderly at low cost and high efficiency.

[0154] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A home-based elderly care system integrating traditional Chinese medicine knowledge and Internet of Things technology, characterized in that: include: It integrates a data acquisition module, a data transmission module, an edge computing module, and a health status prediction module; The integrated data acquisition module is used to acquire multi-dimensional data of home-based elderly care users based on Internet of Things technology; The multi-dimensional data includes: environmental data, physiological data, dietary data, daily routine data, and activity data; The data transmission module is used to transmit the collected multi-dimensional data to the edge computing module; The edge computing module is used to process the received multi-dimensional data, construct a standard data axis, and generate a minimum home event dataset. The edge computing module includes a data processing submodule, a standard time axis establishment submodule, a minimum home event dataset establishment submodule, and a complementary data and collaborative data time axis relocation submodule. The data processing submodule is used to perform data preprocessing, feature encoding, and data fusion on the multi-dimensional data to obtain complementary data and collaborative data; the complementary data is used to obtain a home event through the mutual supplementation of multiple complementary data; the collaborative data is used to infer the user's TCM symptoms through the collaboration and analysis of multiple collaborative data and based on TCM diagnostic theory. The standard time axis establishment submodule is used to group and analyze the complementary and collaborative data using the K-Means clustering algorithm, and to construct a standard time axis based on the grouping analysis results and the time span of TCM symptoms in TCM diagnostic theory. The Home Minimal Events Dataset Establishment Submodule is used to extract home events related to health status from five dimensions—environment, meals, daily routines, sleep, and activities—based on the grouping analysis results. For different health statuses, a home minimum event evidence chain is constructed on the standard timeline, and the home minimum event evidence chains are summarized to form a corresponding home minimum event dataset. The complementary and collaborative data time axis repositioning submodule is used to map the complementary and collaborative data onto the standard time axis; perform time and frequency comparison calibration with the grouped analysis data of the standard time axis; re-standardize and reposition the calibrated complementary and collaborative data to obtain reconstructed complementary and collaborative data; and fuse the reconstructed complementary and collaborative data on the standard time axis to form time-aligned fused complementary and fused collaborative data. The health status prediction module uses environmental data collected by the edge computing module and health and behavioral data from the minimum home event dataset to extract key features; it then constructs a prediction model and combines the key features to predict health risks, obtaining the user's future health risk prediction results; the key features are representative information reflecting environmental conditions, health status, and behavioral patterns.

2. The home-based elderly care system integrating traditional Chinese medicine knowledge and Internet of Things technology according to claim 1, characterized in that, The data transmission module includes a Wi-Fi communication module, a Bluetooth transmission module, and a ZigBee networking module; The Wi-Fi communication module is used to transmit the collected video and audio data; the Bluetooth transmission module is used to transmit the physiological characteristic information collected by the wearable device; and the ZigBee networking module is used to transmit the environmental data collected by the sensor nodes.

3. The home-based elderly care system integrating traditional Chinese medicine knowledge and Internet of Things technology according to claim 1, characterized in that, The specific steps of the data processing submodule include: S11, Data preprocessing step, preprocessing the multi-dimensional data, including: outlier handling and missing data filling; The outlier handling step involves setting upper and lower limits for the multi-dimensional data based on the data acquisition range of each sensor, and identifying and removing data that exceeds the corresponding upper and lower limits as outliers. The missing data completion step involves performing a stationarity statistical test on the data. If the test result meets the preset stationarity conditions, the missing data is completed using the linear interpolation method. If the preset stationarity conditions are not met, the data is subjected to difference transformation or logarithmic transformation. At the same time, the nonlinear trend of the data is determined by the discontinuity saliency method. If the data is nonlinear, the missing data is completed using polynomial interpolation or spline interpolation. S12, data feature encoding step, involves labeling the multi-dimensional data preprocessed in step S11 and encoding the labeled data according to preset encoding rules; the encoding rules include: ID card number, topic classification code, acquisition device code, home device number, event number, and sequence code; S13, Data fusion step: Based on the principles of traditional Chinese medicine diagnostics, the encoded data is fused and divided into complementary data and collaborative data.

4. The home-based elderly care system integrating traditional Chinese medicine knowledge and Internet of Things technology according to claim 1, characterized in that, The specific steps for constructing a standard timeline using the standard timeline creation module are as follows: S21. Based on the distance between the devices that collect the multi-dimensional data, the K-Means clustering algorithm is used to group and analyze the complementary data and the collaborative data. S22. Based on the results of the grouping analysis and the time span of TCM symptoms according to the principles of TCM diagnostics, a standard time axis is constructed; wherein, the time span of TCM symptoms is divided according to any of the following cycles, including: natural cycles, the physiological or pathological change cycle patterns described in TCM health preservation theories, and the chronic disease conditioning cycle determined by the TCM case mechanism; the standard time axis is used to provide a unified time reference for users' activity data in different indoor spaces.

5. The home-based elderly care system integrating traditional Chinese medicine knowledge and Internet of Things technology according to claim 1, characterized in that, The specific execution steps of the complementary data and collaborative data time axis relocation submodule are as follows: S31. Map the complementary data and collaborative data to the corresponding time points on the standard time axis to obtain mapped data; the mapped data includes the mapped complementary data and the mapped collaborative data. S32. On the standard time axis, perform a time and frequency comparison calibration operation between the mapped data and the grouped analysis data of the standard time axis; For mapped data whose time deviates from the standard time axis, adjustments are made based on the time patterns of the grouped analysis data on the standard time axis. For mapping data with abnormal frequencies, change the sensor acquisition settings or perform calibration through data retransmission to bring the frequency of the mapping data into a reasonable range for grouped analysis data of the standard time axis. S33. Remap the calibrated mapping data onto the standard time axis and synchronize it with the grouped analysis data of the standard time axis; associate the time-synchronized mapping data with the standard minimum home events on the standard time axis to obtain reconstructed complementary and collaborative data. S34. The reconstructed complementary data and collaborative data are fused on the standard timeline to obtain the minimum home-based events. The fusion process includes time alignment fusion, information complementarity fusion, and information collaboration fusion. Time alignment fusion refers to aligning the complementary data and collaborative data in the mapped data at the same time point on the standard timeline. Information complementarity fusion refers to quantifying the reconstructed complementary data through regression analysis to obtain the minimum home-based events. Information collaboration and integration, based on the minimum home-based events and reconstructed collaborative data, combined with TCM diagnostic theories, can infer TCM symptoms and diseases.

6. The home-based elderly care system integrating traditional Chinese medicine knowledge and Internet of Things technology according to claim 1, characterized in that, The specific implementation process of the health status prediction module is as follows: S41. Extract features from the environmental data collected by the edge computing module and the health and behavioral data in the minimum home events to obtain environmental data features, environmental impact features, health data features, behavioral data features and behavioral pattern features. The extracted environmental data features are as follows: In the formula, t0 and t n These represent the start and end times of data collection, respectively; E j Let be the j-th environmental data point; μE be the mean of the environmental data; and γ be the weighting coefficient of the environmental data features. The extracted environmental impact characteristics are: In the formula, Temp(t) represents the temperature data at time t, and μ Temp σ Temp , respectively, represent the mean and standard deviation of the temperature data; Hum(t) represents the humidity data at time t; μ Hum σ Hum , respectively, represent the mean and standard deviation of humidity data; Light(t) represents the light concentration data at time t, indicating indoor pollution levels; μ Light σ Light , respectively, represent the mean and standard deviation of the light intensity concentration data; Smoke(t) represents the smoke concentration data at time t, indicating indoor pollution levels; μ Smoke σ Smoke Here, μ represents the mean and standard deviation of the smoke concentration data; Noise(t) represents the noise level at time t; μ Noise σ Noise α1, α2, α3, α4, and α5 are the mean and standard deviation of the noise data, respectively; α1, α2, α3, α4, and α5 are the weighting coefficients of the environmental data features. The extracted health data features are: In the formula, HR(t) represents the heart rate data at time t; μ HR and μ HR Here, b represents the mean and standard deviation of the heart rate data; bP(t) represents the blood pressure data at time t; μ represents the mean and standard deviation of the heart rate data. BP and σ BP These represent the mean and standard deviation of the blood pressure data, respectively; BS(t) represents the blood glucose data at time t; μ BS and σ BS These represent the mean and standard deviation of blood glucose data, respectively; BT(t) represents the body temperature data at time t; μ BT and σ BT Here, BW(t) represents the mean and standard deviation of body temperature data, respectively; BW(t) represents body weight data at time t; μ BW and σ BW These represent the mean and standard deviation of blood glucose data, respectively; α6, α7, α8, α9, α 10 Weighting coefficients for health data features; The extracted behavioral data features are: In the formula, S i This represents the data for the i-th step. A is the step-count feature extraction function; i This represents the i-th motion time data; L is the motion feature extraction function; i This is the i-th GPS coordinate data; β1, β2, and β3 are the location feature extraction functions; β1, β2, and β3 are the weight coefficients of the behavioral data features. The extracted behavioral pattern features are: In the formula, D k For the kth daily activity data; ω daily (D k ) is the feature extraction function for daily activities; Ab k For the kth abnormal behavior data; ω abnormal (Ab k ) is the abnormal behavior feature extraction function; β4 and β5 are the weight coefficients of abnormal and daily behavior data features; S42. Based on the environmental data features, environmental impact features, health data features, behavioral data features, and behavioral pattern features, and combined with the decision tree algorithm, a prediction model is constructed. The expression of the prediction model is: In the formula, K is the final predicted health index, i.e., the predicted value; E is the environmental data characteristic; N(t) is the health data characteristic; B is the behavioral data characteristic; E(t) is the environmental impact characteristic; and D is the behavioral pattern characteristic. S43. Based on the health index, behavioral data characteristics, and behavioral pattern characteristics, output a health report; manage the user's health according to the health report and TCM diseases and symptoms in the minimum home events; the health report includes health indicator change trends, health risk assessment, and health recommendations.

Images