Internet of things-based preschool child health data management system and method

Abstract

The invention relates to the technical field of health management, in particular to a preschool child health data management system and method based on the Internet of Things. A preschool child health data management system based on the Internet of Things comprises a data acquisition module which is used for acquiring physiological data and daily activity data of preschool children read along with class through a sensor; the data acquisition module is used for acquiring basic health data and daily activity data of the preschool children who are read along with the class through a sensor; and the data processing module is used for processing the daily activity data based on an artificial intelligence technology to obtain a daily state sequence. The method overcomes the defects that in the prior art, the behavior recognition dimension is single, and the psychological behavior state of the child is difficult to comprehensively reflect, can continuously monitor the physiological data and daily behaviors of the child before class reading and dynamically assess the health state, provides decision support for teachers, and improves the teaching efficiency. And a teacher is helped to make a more targeted education and intervention plan.

Description

Technical Field

[0001] This invention relates to the field of health management technology, and in particular to an Internet of Things-based health data management system and method for preschool children. Background Technology

[0002] The preschool stage is a critical period for children's physical and mental development. Including preschool children with health problems such as attention deficit hyperactivity disorder (ADHD) and developmental delays in mainstream kindergartens is of great significance for their early intervention and social integration. However, existing child health data management systems are mainly designed for mainstream children and have significant limitations and technical obstacles when applied to preschool children in mainstream kindergartens. Existing technologies typically rely on routine physiological and basic activity data, lacking monitoring of core disability dimensions in children in mainstream kindergartens. These children's health problems are often not reflected in traditional physiological indicators, but are more significantly manifested as social interaction deficits, stereotyped and repetitive behaviors, and learning and daily life difficulties. Existing technologies cannot comprehensively and objectively characterize these complex behavioral patterns, resulting in superficial health assessments that fail to capture specific health problems or identify deterioration trends relative to their baseline. Therefore, there is an urgent need for an IoT-based preschool child health data management system and method that can continuously monitor and dynamically assess the physiological data and daily behaviors of preschool children in mainstream kindergartens, thereby providing decision support for teachers and helping to develop more targeted education and intervention plans. Summary of the Invention

[0003] To overcome the shortcomings of existing technologies in assessing the health status of preschool children attending mainstream schools, this invention provides a preschool children's health data management system and method based on the Internet of Things.

[0004] The technical implementation scheme of the present invention is: a preschool children's health data management system based on the Internet of Things, comprising: The data acquisition module is used to collect basic health data and daily activity data of preschool children attending regular classes through sensors; The data processing module is used to process the daily activity data based on artificial intelligence technology to obtain a daily state sequence; The basic health status module is used to obtain historical normal basic health data of preschool children attending regular classes, construct an individual basic health status model for each preschool child attending regular classes, and obtain the distribution of individual basic health status models based on the individual basic health status models. The status assessment module is used to obtain teacher record evaluation data through the support and care record form for preschool children attending regular classes, and to conduct status assessment on preschool children attending regular classes based on the basic health data, individual basic health status model distribution, daily status sequence and teacher record evaluation data, and obtain status assessment results. The report generation module is used to periodically generate health status reports for preschool children attending regular classes based on the basic health data, daily status sequences, and status assessment results.

[0005] Preferably, the data acquisition module is used to collect basic health data and daily activity data of preschool children attending regular classes through sensors, including: collecting basic health data and daily activity data of preschool children attending regular classes within a preset period through sensors deployed in the environment, wherein the basic health data includes vital sign data, dietary data, activity data and sleep data.

[0006] Preferably, the data processing module is used to process the daily activity data based on artificial intelligence technology to obtain a daily state sequence, including: processing the daily activity data using emotion recognition technology to obtain the duration of resistance emotions; using a stereotyped action recognition model to identify the daily activity data to obtain the frequency and average duration of stereotyped actions; processing the daily activity data using action recognition technology and semantic recognition technology to obtain the number of social withdrawals and constructing an undirected graph of class activities; performing activity path analysis on the undirected graph to obtain activity centrality reflecting the importance of preschool children attending mainstream classes in activities; and aligning and standardizing the duration of resistance emotions, frequency of stereotyped actions, average duration of stereotyped actions, number of social withdrawals, and activity centrality according to time to generate a daily state sequence.

[0007] Preferably, the basic health status module is used to acquire historical normal basic health data of preschool children attending regular classes, construct an individual basic health status model for each preschool child attending regular classes, and obtain an individual basic health status model distribution based on the individual basic health status model, including: standardizing the historical normal basic health data into historical normal basic health vectors, constructing an individual basic health status model for each preschool child attending regular classes using a variational autoencoder, training the individual basic health status model based on the normal basic health vectors, obtaining the encoder probability distribution based on the individual basic health status model, and periodically standardizing the normal basic health data into normal basic health vectors to continuously update the individual basic health status model.

[0008] Preferably, the step of periodically standardizing normal basic health data into normal basic health vectors and continuously updating the individual basic health status model includes: continuously updating the individual basic health status model using a sliding window re-evaluation method, pre-setting a sliding window and a sliding step size, performing gradient descent optimization based on the parameters of the individual basic health status model and the normal basic health vectors, and updating the individual physiological status model.

[0009] Preferably, the status assessment module is used to obtain teacher record evaluation data through the support and care record form for preschool children attending regular classes, and to conduct a status assessment of the preschool children attending regular classes based on the basic health data, the individual basic health status model distribution, the daily status sequence, and the teacher record evaluation data to obtain a status assessment result. This includes: obtaining and quantifying the teacher record evaluation data from the support and care record form for preschool children attending regular classes using natural language processing technology; the teacher record evaluation data includes self-care ability and independent task completion rate; standardizing the basic health data into a basic health vector; calculating the Mahalanobis distance between the basic health vector and the encoder probability distribution as the basic health deviation; calculating the daily performance index of the preschool children attending regular classes based on the daily status sequence and the teacher record evaluation data; and conducting a status assessment based on the basic health deviation and the daily performance index using fuzzy inference to obtain a status assessment result.

[0010] Preferably, the step of calculating the daily performance index of preschool children attending regular classes based on the daily state sequence and the teacher's recorded evaluation data includes: normalizing the duration of resistance, frequency of stereotyped actions, average duration of stereotyped actions, number of social withdrawals, activity centrality, self-care ability, and task independent completion rate, and then calculating the daily performance index using the daily performance calculation formula, wherein the daily performance calculation formula is: ; In the formula, This is a daily performance index. For the centrality of the event intermediary, For the ability to take care of oneself, For task completion rate, To resist the duration of the emotion, For the frequency of stereotyped movements, The average duration of the stereotyped action. For the number of times of social withdrawal, It is a non-zero constant. , , , , and These are the weighting coefficients.

[0011] Preferably, the process of obtaining a state assessment result through fuzzy inference based on the physiological deviation and the daily performance index includes: determining the universe of discourse of the relevant fuzzy variables, the basic health deviation and the daily performance index; selecting a membership function and determining the membership degree; formulating fuzzy rules based on expert domain experience and historical data; performing inference based on the fuzzy rules and membership degrees; using the MIN operation to calculate the premise strength of each rule; using the MAX operation to synthesize the output of all activated rules; and performing defuzzification using the centroid method to obtain the state assessment result. The state assessment result includes excellent, good, requires attention, and requires intervention.

[0012] Preferably, the report generation module is used to periodically generate health status reports for preschool children attending regular classes based on the basic health data, daily status sequence, and status assessment results. This includes: performing multi-dimensional fusion analysis on the basic health data, daily status sequence, and status assessment results to generate health status reports, reminding teachers to take intervention measures, and assisting teachers in developing educational plans.

[0013] A method for managing preschool children's health data based on the Internet of Things, comprising: S1: Collect basic health data and daily activity data of preschool children attending regular classes through sensors; S2: Process the daily activity data based on artificial intelligence technology to obtain a daily state sequence; S3: Obtain historical normal basic health data of preschool children attending regular classes, construct an individual basic health status model for each preschool child attending regular classes, and obtain the distribution of the individual basic health status model based on the individual basic health status model; S4: Obtain teacher record evaluation data through the support and care record form for preschool children attending regular classes. Based on the basic health data, individual basic health status model distribution, daily status sequence and teacher record evaluation data, conduct a status assessment of the preschool children attending regular classes and obtain the status assessment results. S5: Generate health status reports for preschool children attending regular classes periodically based on the basic health data, daily status sequences, and status assessment results.

[0014] The beneficial effects of this invention are as follows: The data processing module of this invention uses artificial intelligence technology to comprehensively apply emotion recognition, posture estimation, action recognition and semantic recognition methods to intelligently analyze children's daily activity data. It automatically extracts key features such as the duration of resistance, frequency of stereotyped actions, activity mediator centrality and the number of social withdrawals, and generates daily state sequences. This enables fine-grained quantitative analysis of children's emotional, social and behavioral states, overcoming the shortcomings of existing technologies that have a single behavioral recognition dimension and cannot fully reflect children's psychological and behavioral states.

[0015] The individual dynamic profiling module of this invention constructs an individual basic health status model for each preschool child attending regular classes based on a variational autoencoder. It forms an individual physiological characteristic distribution by standardizing historical normal physiological data and continuously updates the model using a sliding window re-evaluation method. This allows the model to dynamically adjust with the child's physiological changes, thereby more accurately reflecting the child's individual differences and long-term health trends, avoiding the problem that existing static models cannot adapt to individual growth changes.

[0016] The status assessment module of this invention integrates basic health data, individual health status model distribution, daily activity sequence, and teacher-recorded evaluation data to calculate basic health deviation and daily performance index. It also combines fuzzy reasoning methods for comprehensive assessment, enabling intelligent classification of children's status in both physiological and psychological dimensions. This module has high interpretability and helps teachers quickly identify abnormal states and take targeted interventions.

[0017] The report generation module of this invention automatically generates a health status report based on basic health deviation, daily activity sequence, and status assessment results. The report comprehensively displays the child's overall status in terms of physiology, emotion, and social aspects, and provides intervention reminders. It realizes closed-loop management of the entire process from data collection to intelligent analysis, and from assessment and diagnosis to intervention reminders, thereby improving the work efficiency and scientific decision-making ability of teachers in inclusive education scenarios. Attached Figure Description

[0018] Figure 1 This is a structural diagram of the Internet of Things-based preschool children's health data management system of the present invention; Figure 2 This is a flowchart of the Internet of Things-based method for managing preschool children's health data according to the present invention. Detailed Implementation

[0019] 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, and 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.

[0020] Example 1: An Internet of Things-based preschool children's health data management system, such as Figure 1 As shown, it includes: The data acquisition module is used to collect basic health data and daily activity data of preschool children attending regular classes through sensors; Basic health data and daily activity data of preschool children attending regular classes are collected by sensors deployed in the environment within a preset period. The basic health data includes vital signs data, dietary data, activity data, and sleep data.

[0021] It should be further explained that during actual operation, based on a preset time period, sensors deployed in the environment record the basic health data and daily activity data of the target children. Typically, the preset time period is set to one school day. The sensor network covers important spaces such as classrooms, rest areas, and activity areas to ensure the integrity of the acquired information in both time and space. Basic health data, as a key indicator of children's physiological condition, covers four main categories. Vital signs data are acquired by smart bracelets worn by children or flexible sensor patches integrated into their clothing, specifically involving parameters such as heart rate, body surface temperature, and skin conductance. Dietary data is collected using visual sensing devices placed in the dining area to record the start and end times of each meal, food intake, and eating rhythm. Activity data comes from wide-angle cameras and inertial measurement units set up in the environment, capable of capturing children's movement paths, exercise intensity, and typical movement patterns. Sleep data is collected through pressure sensors built into the mattresses placed in the rest area, including the time of falling asleep, total sleep duration, and nighttime body movement frequency.

[0022] In addition to basic health parameters, data on children's daily activities are simultaneously acquired. This data is primarily recorded using high-definition cameras and directional audio acquisition arrays deployed in the environment. These sensing units continuously capture children's behavioral patterns and social interactions in various scenarios, including free play, group collaboration, and group teaching, providing initial data for subsequent behavioral sequence modeling. All collected basic health and daily activity data undergo preliminary encryption and de-identification processing before being transmitted securely to the central data processing unit via a wireless transmission channel for in-depth analysis by subsequent modules. Through this collaborative mechanism, the data acquisition module achieves multi-dimensional recording of the health status and behavioral characteristics of preschool children attending mainstream education, providing data support for the construction of an overall assessment system.

[0023] The data processing module is used to process the daily activity data based on artificial intelligence technology to obtain a daily state sequence; Emotion recognition technology is used to process the daily activity data to obtain the duration of resistance emotions. A stereotyped action recognition model is used to identify the daily activity data to obtain the frequency and average duration of stereotyped actions. Action recognition technology and semantic recognition technology are used to process the daily activity data to obtain the number of social withdrawals and construct an undirected graph of class activities. Activity path analysis is performed on the undirected graph to obtain the activity mediator centrality, which reflects the importance of preschool children attending mainstream classes in activities. The duration of resistance emotions, frequency of stereotyped actions, average duration of stereotyped actions, number of social withdrawals, and activity mediator centrality are aligned and standardized according to time to generate a daily state sequence.

[0024] It should be further explained that the core task of the data processing module is to use artificial intelligence technology to conduct in-depth analysis of the collected daily activity data, and then generate a daily state sequence that can reflect the behavioral characteristics of children.

[0025] Emotion recognition technology is used to process input daily activity data. A pre-trained deep learning model identifies negative emotional states exhibited by children in social or command-based situations and accurately records the duration of each instance of resistance, from its onset to its resolution. A stereotyped action recognition model is also used to process the daily activity data. This model is built upon the YOLO series of target recognition models and is trained on a dataset of typical stereotyped actions, such as repetitive clapping and body shaking. The frequency of these stereotyped actions per unit time is statistically analyzed, and the duration of a single stereotyped action from start to finish is accurately measured. The average duration of the stereotyped action is calculated based on the durations of multiple stereotyped actions.

[0026] Action recognition and semantic recognition technologies are used collaboratively to analyze daily activity data. Action recognition technology identifies children's active avoidance of social contact by analyzing their spatial displacement patterns and body orientation changes in a group environment, and records the number of social withdrawals. Based on the real-time location and interaction relationships of all children in the class during free activities, an undirected graph of class activities is constructed, with children as nodes and their interaction relationships as edges. Based on this, a deeper analysis of activity paths is performed on the undirected graph, targeting preschool children attending regular classes. The betweenness centrality of activities is calculated using the shortest path betweenness metric. First, the shortest path between all pairs of nodes in the graph is calculated. A breadth-first search algorithm is used to find the shortest path between any two nodes; the length of these paths is determined by the number of edges traversed. Then, the number of shortest paths passing through the target node is counted. For example, for any three different nodes, suppose there are several shortest paths of equal length between node a and node b, where the number of shortest paths passing through node c is denoted as x. The total number of shortest paths between nodes a and b is denoted as y. Then, the betweenness centrality of node c is the ratio of x to y. The higher the value of this indicator, the more important the child's role as a bridge for information transmission in the class's social network.

[0027] After extracting the above indicators, the data processing module enters the data integration stage. In this stage, the acquired duration of resistance, frequency of stereotyped actions, average duration of stereotyped actions, number of social withdrawals, and activity centrality parameters are aligned and standardized in chronological order to ultimately generate a structured sequence of daily states.

[0028] The basic health status module is used to obtain historical normal basic health data of preschool children attending regular classes, construct an individual basic health status model for each preschool child attending regular classes, and obtain the distribution of individual basic health status models based on the individual basic health status models. The historical normal basic health data is standardized into a historical normal basic health vector. A variational autoencoder is used to construct an individual basic health status model for each preschool child attending regular classes. The individual basic health status model is obtained by training a dataset based on the normal basic health vector. The encoder probability distribution is obtained based on the individual basic health status model. The normal basic health data is periodically standardized into normal basic health vectors to continuously update the individual basic health status model.

[0029] The individual basic health status model is continuously updated using the sliding window re-evaluation method. The sliding window and sliding step size are preset, and gradient descent optimization is performed based on the parameters of the individual basic health status model and the normal basic health vector to update the individual physiological status model.

[0030] It should be further explained that, in this embodiment, the basic health status module is used to acquire historical normal basic health data of preschool children attending regular classes, and to construct an individual basic health status model for each child. First, the collected historical normal basic health data is standardized to unify the scale and units of various indicators, eliminating biases caused by differences in measurement equipment, sampling frequency, and environment. The standardized data is defined as a historical normal basic health vector. This vector can accurately represent the comprehensive characteristics of a child's health status in a multi-dimensional space and serves as the basic input for subsequent model construction and calculation.

[0031] In the model construction process, a variational autoencoder (VAE) algorithm is used to establish an individual basic health status model for each preschool child attending mainstream education. The VAE learns features from the input historical normal basic health vector through an encoder and decoder structure. The encoder maps the input health vector to a probability distribution in the latent space to characterize the latent patterns of individual health characteristics, while the decoder generates corresponding health data samples based on the latent variables, achieving data reconstruction and optimization. This approach automatically extracts deep features of individual health status from a large amount of historical health data, thus forming a basic health status model specific to each child. The basic health status module further constructs a dataset based on normal basic health vectors and trains the VAE to obtain the individual basic health status model. The training process uses historical normal basic health data as input samples. After multiple iterations and optimizations, it automatically learns the distribution of individual health features, forming the encoder probability distribution. The encoder probability distribution reflects the statistical distribution characteristics of an individual's health characteristics under normal conditions and is used to subsequently assess the degree of deviation from the individual's health status. During status assessment, the difference between the real-time collected health data and this probability distribution is compared to determine abnormal fluctuations in the individual's health status.

[0032] To ensure the timeliness and accuracy of the individual basic health status model, the basic health status module periodically standardizes newly collected normal basic health data to generate new normal basic health vectors, which are then used to continuously update the individual basic health status model. This periodic model update ensures the model continuously reflects the latest health characteristics and behavioral patterns of children, avoiding feature drift due to time changes and ensuring the model's effectiveness in long-term monitoring. The basic health status module continuously updates the individual basic health status model using a sliding window re-evaluation method. Preset sliding windows and step sizes control the time range and frequency of model updates. The sliding window limits the range of data samples participating in the update, and the step size determines the speed and interval of window movement. When new data enters the window, older data is gradually removed, and the model is retrained based on the latest normal basic health vector within the window. This sliding window mechanism ensures that the model balances historical stability and real-time responsiveness during the update process. Gradient descent optimization is performed based on the individual basic health status model parameters and the normal basic health vectors. The optimization process continuously calculates the gradient direction and magnitude of the loss function, adjusting the model parameters to gradually reduce the reconstruction error of the health data output by the model. After multiple rounds of iterative optimization, the parameters are updated, thereby achieving continuous improvement of the individual physiological state model. Through this mechanism, the accuracy and robustness of the model's representation of individual health status can be continuously improved, enabling it to more accurately reflect the health change trends of preschool children attending mainstream education over different time periods.

[0033] The status assessment module is used to obtain teacher record evaluation data through the support and care record form for preschool children attending regular classes, and to conduct status assessment on preschool children attending regular classes based on the basic health data, individual basic health status model distribution, daily status sequence and teacher record evaluation data, and obtain status assessment results. Natural language processing technology is used to obtain and quantify teacher evaluation data from the support and care record form for preschool children attending regular classes. The teacher evaluation data includes self-care ability and independent task completion rate. The basic health data is standardized into a basic health vector, and the Mahalanobis distance between the basic health vector and the encoder probability distribution is calculated as the basic health deviation. The daily performance index of the preschool children attending regular classes is calculated based on the daily state sequence and the teacher evaluation data. Based on the basic health deviation and the daily performance index, a state assessment is performed through fuzzy inference to obtain the state assessment result.

[0034] After normalizing the duration of resistance, frequency of stereotyped actions, average duration of stereotyped actions, number of social withdrawals, activity centrality, self-care ability, and independent task completion rate, the daily performance index is calculated using the daily performance calculation formula, where the daily performance calculation formula is: ; In the formula, This is a daily performance index. For the centrality of the event intermediary, For the ability to take care of oneself, For task completion rate, To resist the duration of the emotion, For the frequency of stereotyped movements, The average duration of the stereotyped action. For the number of times of social withdrawal, It is a non-zero constant. , , , , and These are the weighting coefficients.

[0035] The universe of discourse for the relevant fuzzy variables, basic health deviation and daily performance index, is determined. Membership functions are selected and membership degrees are determined. Fuzzy rules are formulated based on expert domain experience and historical data. Reasoning is performed based on fuzzy rules and membership degrees. The premise strength of each rule is calculated using the MIN operation. The output of all activated rules is synthesized using the MAX operation. Defuzzification is performed using the centroid method to obtain the state assessment results, which include excellent, good, need attention, and need intervention.

[0036] It should be further explained that, in this embodiment, the status assessment module is used to obtain teacher evaluation data through the support and care record form for preschool children attending regular classes. The module automatically extracts textual records related to the daily behavior, psychological reactions, task completion, and self-care abilities of the preschool children attending regular classes from the teacher-filled support and care record form. Natural language processing technology is used to identify, segment, extract features, and perform sentiment analysis on the textual data recorded by the teacher, extracting representative evaluation data. The teacher evaluation data includes two key indicators: self-care ability and independent task completion rate. These two data reflect the child's self-care ability, activity independence, and task execution stability in daily learning and life. The teacher evaluation data is then converted into quantified numerical data to form a teacher evaluation feature vector.

[0037] A comprehensive health assessment of preschool children attending mainstream schools is conducted based on basic health data, individual basic health status model distribution, daily status sequences, and teacher-recorded evaluation data. The basic health data is standardized and transformed into a basic health vector with uniform dimensions. To assess the difference between individual health status and the model distribution, the Mahalanobis distance between the basic health vector and the encoder probability distribution is calculated, and this distance is defined as the basic health deviation. By using the probability distribution parameters output by the encoder as the basis for deviation calculation, the basic health deviation has a clear probabilistic statistical meaning, reflecting the degree of deviation of an individual's health status relative to the group's health distribution, thereby improving the interpretability of the results. Introducing Mahalanobis distance as a measure of health deviation incorporates covariance information between features in the distance calculation, thus accurately reflecting the intrinsic correlation structure between health features and improving the discriminative power and robustness of health status assessment. The basic health deviation characterizes the degree of deviation between a child's current health status and the historical health model; a larger deviation indicates a more significant gap between the individual's health status and its baseline model.

[0038] In daily performance assessments, a daily performance index is calculated for preschool children attending mainstream schools based on a daily performance sequence and teacher-recorded evaluation data. The daily performance sequence comprises continuously collected behavioral and emotional data, encompassing multidimensional behavioral indicators such as activity mediator centrality, duration of resistance, frequency of stereotyped actions, average duration of stereotyped actions, and frequency of social withdrawal. These indicators, along with self-care abilities and independent task completion rates extracted from teacher-recorded evaluation data, are normalized to eliminate the influence of dimensions. The normalized data is then input into the daily performance calculation formula to obtain the daily performance index. This formula comprehensively assesses children's social interaction level, emotional stability, and task performance ability. Social communication-related indicators have a positive effect in the numerator, representing the degree of active communication; duration of resistance, frequency of stereotyped actions, and frequency of social withdrawal have a negative moderating effect in the denominator, representing the intensity of negative behavior. A higher daily performance index indicates a more positive overall state for the child, while a lower index suggests potential emotional or behavioral risks.

[0039] After obtaining the baseline health deviation and daily performance index, the status assessment module performs a comprehensive status assessment based on fuzzy inference. First, the domain of the fuzzy variables baseline health deviation and daily performance index is determined. The baseline health deviation is divided into three fuzzy subsets: low deviation, medium deviation, and high deviation. The daily performance index is divided into four fuzzy subsets: excellent, good, average, and poor. Then, membership functions are selected and the membership values ​​for each variable are determined. The shape and parameters of the membership functions are determined by expert experience and historical data statistics, reflecting the strength of the membership relationship between the variable and its subsets.

[0040] Based on the aforementioned fuzzy variables and membership degrees, fuzzy rules are formulated by combining expert domain experience and historical behavioral data. These fuzzy rules express the relationship between baseline health deviation and daily performance index. For example, when the baseline health deviation is low and the daily performance index is high, the child's condition tends to be excellent; when the baseline health deviation is moderate and the daily performance index is at a medium level, the condition is judged as good; when the baseline health deviation is high and the daily performance index is low, the condition is judged as requiring attention; and when the baseline health deviation is extremely high and the daily performance index is significantly low, the condition is judged as requiring intervention. Reasoning is performed according to the fuzzy rules and membership degrees. MIN is used to calculate the premise strength of each rule, and MAX is used to synthesize the output of all activated rules. Finally, the synthesized output is defuzzified using the centroid method to obtain the final state assessment result.

[0041] The status assessment results are categorized into four types: Excellent, Good, Needs Attention, and Needs Intervention. An Excellent result indicates that the child performs well in basic health, social interaction, emotional regulation, and task performance. A Good result indicates that the child's overall health is stable, with occasional deviations that do not affect overall performance. A Needs Attention result indicates that the child has mild health or emotional problems, requiring appropriate guidance and observation from teachers or parents. A Needs Intervention result indicates that the child is currently experiencing significant deviations, requiring further intervention or support measures from teachers and professionals. This status assessment method enables dynamic quantitative analysis and intelligent grading of the individual status of preschool children in inclusive education, providing educational administrators and teachers with a scientific basis for decision-making.

[0042] The report generation module is used to periodically generate health status reports for preschool children attending regular classes based on the basic health data, daily status sequences, and status assessment results.

[0043] The basic health data, daily status sequences, and status assessment results are integrated and analyzed from multiple dimensions to generate a health status report, which reminds teachers to take intervention measures and assists teachers in developing educational plans.

[0044] It should be further explained that this embodiment elaborates on the specific working mechanism and output format of the report generation module. This module is responsible for the final output, and its core function is to periodically generate a comprehensive health status report for preschool children attending regular classes based on basic health data, daily status sequences, and status assessment results.

[0045] The report generation process begins with a multi-dimensional fusion analysis program. This program collaboratively processes basic health data from various sources, daily state sequences reflecting behavioral characteristics, and comprehensive state assessment results. It doesn't simply list this information; instead, it delves into their inherent connections, such as cross-referencing fluctuations in vital signs at specific moments with behavioral performance during the same period, thereby revealing potential links between physiological indicators and behavioral patterns. After data analysis, a structured health status report is automatically generated. This report has a clear hierarchical structure, first presenting a summary of the core state assessment results, followed by a detailed presentation of the periodic trends in basic health data, key behavioral events in the daily state sequences, and statistical characteristics. The report uses a combination of objective descriptions and charts to clearly outline the child's overall condition within the assessment period.

[0046] Based on the information revealed in the report, specific action recommendations are generated. If the report indicates that the status assessment results are at a level requiring attention or intervention, the module will generate clear prompts to remind teachers that necessary intervention measures need to be taken for the child. These prompts are usually directly related to specific abnormal data points in the report, providing clues for the direction of intervention. In addition, the health status report assists teachers in developing the next stage of educational plans. By analyzing changes in children's long-term baseline health data and the evolution trends of daily status sequences, the report helps teachers identify children's stable traits, areas of progress, and persistent challenges, thereby providing data support and decision-making basis for designing more targeted and personalized educational programs. Through the above process, the report generation module successfully transforms the analysis and assessment results into practical tools that educators can directly use, realizing a complete closed loop from data collection to educational intervention, and significantly improving the scientific rigor and effectiveness of inclusive education.

[0047] Example 2: Based on Example 1, a method for managing preschool children's health data based on the Internet of Things, such as... Figure 2 As shown, it includes: S1: Collect basic health data and daily activity data of preschool children attending regular classes through sensors; S2: Process the daily activity data based on artificial intelligence technology to obtain a daily state sequence; S3: Obtain historical normal basic health data of preschool children attending regular classes, construct an individual basic health status model for each preschool child attending regular classes, and obtain the distribution of the individual basic health status model based on the individual basic health status model; S4: Obtain teacher record evaluation data through the support and care record form for preschool children attending regular classes. Based on the basic health data, individual basic health status model distribution, daily status sequence and teacher record evaluation data, conduct a status assessment of the preschool children attending regular classes and obtain the status assessment results. S5: Generate health status reports for preschool children attending regular classes periodically based on the basic health data, daily status sequences, and status assessment results.

[0048] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A preschool children's health data management system based on the Internet of Things, characterized in that, include: The data acquisition module is used to collect basic health data and daily activity data of preschool children attending regular classes through sensors; The data processing module is used to process the daily activity data based on artificial intelligence technology to obtain a daily state sequence; The basic health status module is used to obtain historical normal basic health data of preschool children attending regular classes, construct an individual basic health status model for each preschool child attending regular classes, and obtain the distribution of individual basic health status models based on the individual basic health status models. The status assessment module is used to obtain teacher record evaluation data through the support and care record form for preschool children attending regular classes, and to conduct status assessment on preschool children attending regular classes based on the basic health data, individual basic health status model distribution, daily status sequence and teacher record evaluation data, and obtain status assessment results. The report generation module is used to periodically generate health status reports for preschool children attending regular classes based on the basic health data, daily status sequences, and status assessment results.

2. The preschool children's health data management system based on the Internet of Things according to claim 1, characterized in that, The data acquisition module is used to collect basic health data and daily activity data of preschool children attending regular classes through sensors, including: collecting basic health data and daily activity data of preschool children attending regular classes within a preset period through sensors deployed in the environment. The basic health data includes vital signs data, dietary data, activity data and sleep data.

3. The preschool children's health data management system based on the Internet of Things according to claim 2, characterized in that, The data processing module is used to process the daily activity data based on artificial intelligence technology to obtain a daily state sequence, including: using emotion recognition technology to process the daily activity data to obtain the duration of resistance emotions; using a stereotyped action recognition model to identify the daily activity data to obtain the frequency and average duration of stereotyped actions; using action recognition technology and semantic recognition technology to process the daily activity data to obtain the number of social withdrawals and construct an undirected graph of class activities; performing activity path analysis on the undirected graph to obtain activity centrality reflecting the importance of preschool children attending mainstream classes in activities; and aligning and standardizing the duration of resistance emotions, frequency of stereotyped actions, average duration of stereotyped actions, number of social withdrawals, and activity centrality according to time to generate a daily state sequence.

4. The preschool children's health data management system based on the Internet of Things according to claim 1, characterized in that, The basic health status module is used to acquire historical normal basic health data of preschool children attending regular classes, construct an individual basic health status model for each child, and obtain the individual basic health status model distribution based on the individual basic health status model. This includes: standardizing the historical normal basic health data into historical normal basic health vectors; constructing an individual basic health status model for each child using a variational autoencoder; training the individual basic health status model based on the normal basic health vectors; obtaining the encoder probability distribution based on the individual basic health status model; and periodically standardizing the normal basic health data into normal basic health vectors to continuously update the individual basic health status model.

5. A preschool children's health data management system based on the Internet of Things according to claim 4, characterized in that, The step of periodically standardizing normal basic health data into normal basic health vectors and continuously updating the individual basic health status model includes: continuously updating the individual basic health status model through a sliding window re-evaluation method, pre-setting a sliding window and a sliding step size, performing gradient descent optimization based on the parameters of the individual basic health status model and the normal basic health vectors, and updating the individual physiological status model.

6. A preschool children's health data management system based on the Internet of Things according to claim 4, characterized in that, The status assessment module is used to obtain teacher record evaluation data from the support and care record form for preschool children attending regular classes, and to conduct a status assessment of the preschool children attending regular classes based on the basic health data, individual basic health status model distribution, daily status sequence, and teacher record evaluation data to obtain a status assessment result. This includes: obtaining and quantifying teacher record evaluation data from the support and care record form for preschool children attending regular classes using natural language processing technology; the teacher record evaluation data includes self-care ability and independent task completion rate; standardizing the basic health data into a basic health vector; calculating the Mahalanobis distance between the basic health vector and the encoder probability distribution as the basic health deviation; calculating the daily performance index of the preschool children attending regular classes based on the daily status sequence and the teacher record evaluation data; and conducting a status assessment based on the basic health deviation and the daily performance index using fuzzy inference to obtain a status assessment result.

7. A preschool children's health data management system based on the Internet of Things according to claim 6, characterized in that, The step of calculating the daily performance index of preschool children attending mainstream classes based on the daily state sequence and the teacher's recorded evaluation data includes: normalizing the duration of resistance, frequency of stereotyped movements, average duration of stereotyped movements, number of social withdrawals, activity centrality, self-care ability, and independent task completion rate, and then using the daily performance calculation formula to calculate the daily performance index, wherein the daily performance calculation formula is: ； In the formula, This is a daily performance index. For the centrality of the event intermediary, For the ability to take care of oneself, For task completion rate, To resist the duration of the emotion, For the frequency of stereotyped movements, The average duration of the stereotyped action. For the number of times of social withdrawal, It is a non-zero constant. , , , , and These are the weighting coefficients.

8. A preschool children's health data management system based on the Internet of Things according to claim 6, characterized in that, The process of obtaining a state assessment result by performing fuzzy inference based on the baseline health deviation and the daily performance index includes: determining the universe of discourse of the relevant fuzzy variables, the baseline health deviation and the daily performance index; selecting a membership function and determining the membership degree; formulating fuzzy rules based on expert domain experience and historical data; performing inference based on the fuzzy rules and membership degrees; calculating the premise strength of each rule using the MIN operation; synthesizing the output of all activated rules using the MAX operation; and defuzzifying the results using the centroid method to obtain the state assessment result. The state assessment result includes excellent, good, requires attention, and requires intervention.

9. A preschool children's health data management system based on the Internet of Things according to claim 1, characterized in that, The report generation module is used to periodically generate health status reports for preschool children attending regular classes based on the basic health data, daily status sequence, and status assessment results. This includes: performing multi-dimensional fusion analysis on the basic health data, daily status sequence, and status assessment results to generate health status reports, reminding teachers to take intervention measures, and assisting teachers in developing educational plans.

10. A method for managing preschool children's health data based on the Internet of Things (IoT), used to implement the IoT-based preschool children's health data management system as described in any one of claims 1-9, characterized in that, include: S1: Collect basic health data and daily activity data of preschool children attending regular classes through sensors; S2: Process the daily activity data based on artificial intelligence technology to obtain a daily state sequence; S3: Obtain historical normal basic health data of preschool children attending regular classes, construct an individual basic health status model for each preschool child attending regular classes, and obtain the distribution of the individual basic health status model based on the individual basic health status model; S4: Obtain teacher record evaluation data through the support and care record form for preschool children attending regular classes. Based on the basic health data, individual basic health status model distribution, daily status sequence and teacher record evaluation data, conduct a status assessment of the preschool children attending regular classes and obtain the status assessment results. S5: Generate health status reports for preschool children attending regular classes periodically based on the basic health data, daily status sequences, and status assessment results.

Images