Prenatal Health Care Control System Based on Data Feedback Analysis

Abstract

The present invention relates to the field of data feedback control technology, and discloses a pregnancy health care control system based on data feedback analysis, comprising: a data acquisition module, a dynamic physiological characteristic mapping module, an implicit feedback channel module, a self-stabilizing control module, and a nursing execution module. The present invention deeply integrates the multi-dimensional physiological data of pregnant women with historical trends by constructing a dynamic physiological characteristic mapping matrix to generate a continuous health state vector, breaking through the limitations of the existing technology that relies on a single threshold and lacks dynamic adaptability. At the same time, the system introduces implicit feedback and self-stabilizing control mechanisms, which can automatically adjust the nursing strategy when data anomalies or environmental changes are detected, realizing personalized, continuous, and robust pregnancy health management, and significantly improving the sensitivity, stability, and personalized nursing capabilities of the monitoring system.

Description

Technical Field

[0001] The present invention relates to the technical field of data feedback control, and in particular to a pregnancy health care control system based on data feedback analysis. Background Art

[0002] Currently, pregnancy health monitoring technology has been widely used in clinical and home care fields. It usually assists in assessing the health status of pregnant women and fetuses by periodically monitoring physiological indicators such as the pregnant woman's heart rate, fetal movement, and blood pressure. However, existing pregnancy health monitoring systems generally have the following technical limitations:

[0003] First, existing systems mostly rely on periodic testing or single data collection, lacking real-time and continuous monitoring of pregnant women's physiological data, which can easily lead to key health risk signals being ignored in a short period of time, and the timeliness of monitoring is insufficient. Secondly, most existing technologies use fixed threshold models or empirical rules to judge health risks, failing to fully consider the historical health data and physiological characteristics of individual pregnant women. They have limited personalized adaptability and are at risk of misjudgment or missed judgments. In addition, traditional systems usually only focus on a single or a few physiological parameters, lack effective integration and correlation analysis of pregnant women's daily activity pattern data (such as gait, sleep quality) and environmental data (such as temperature, humidity, and air quality), making it difficult to comprehensively assess the dynamic changes in the health status of pregnant women.

[0004] Existing technologies are also deficient in their ability to handle abnormal data. During the monitoring process, sensor signals may be affected by factors such as environmental noise, equipment interference, or user behavior, resulting in data anomalies. Existing systems often lack effective anomaly identification and correction mechanisms, which can easily lead to false alarms or excessive intervention, affecting the monitoring experience. To address the above issues, some technical solutions have attempted to improve monitoring capabilities by increasing sensor types and introducing simple data analysis algorithms, but these solutions usually only stay at the data aggregation level and lack dynamic fusion and feedback control mechanisms for multi-dimensional data, making it impossible to achieve continuous tracking and real-time response to the individualized health status of pregnant women. Summary of the Invention

[0005] The technical problem solved by the present invention is to address the defects existing in the above-mentioned existing technologies and provide a pregnancy health care control system based on data feedback analysis to solve the problems raised in the above-mentioned background technologies, such as insufficient continuity of data collection, limited individualized risk assessment capabilities, lack of multi-dimensional data integration and analysis, insufficient data anomaly processing capabilities, and poor dynamic adaptability of nursing strategies.

[0006] To solve the above technical problems, the present invention adopts the following technical solutions: a pregnancy health care control system based on data feedback analysis, comprising:

[0007] The data acquisition module is used to collect real-time physiological data of pregnant women, including heart rate, fetal movement, blood pressure, as well as data on daily activity patterns of pregnant women, including gait and sleep quality, and environmental data, including temperature, humidity, and air quality;

[0008] The dynamic physiological characteristic mapping module is in communication with the data acquisition module and is used to construct a dynamic physiological characteristic mapping matrix based on the collected real-time physiological data and historical health trend data, and convert the discrete physiological data stream into a continuous health status vector H:

[0009]

[0010] Where n is the number of physiological data types, w i is the weight of the i-th physiological data determined based on historical data analysis, P i is the real-time value of the i-th physiological data, N(P i ) is the range of P according to the historical data of the pregnant woman i The normalized value is used to match the initial care strategy in real time by calculating the similarity between the current health status vector and the historical health status vector.

[0011] an implicit feedback channel module, communicatively connected to the data acquisition module and the dynamic physiological characteristic mapping module, configured to analyze the association between daily activity pattern data and the physiological data, perform analysis based on the environmental data, automatically modify the priority of the preliminary nursing strategy according to the analysis results, and generate a final nursing strategy;

[0012] a self-stabilizing control module, in communication with the dynamic physiological feature mapping module and the implicit feedback channel module, configured to monitor the health state vector and, upon detecting data anomalies, automatically switch to a robust mode based on the historical fluctuation range of the health state vector, wherein fuzzy logic inference is used in place of precise calculation to generate a robust nursing strategy in the robust mode;

[0013] A nursing execution module is communicatively connected with the implicit feedback channel module and the self-stabilizing control module, and is used for executing the final nursing strategy or the robust nursing strategy.

[0014] As a further solution of the present invention, the implicit feedback channel module further includes: an activity pattern-physiological response relationship library, which is used to store personalized activity-response models established based on the historical activity data of pregnant women and changes in corresponding physiological indicators; a real-time strategy fine-tuning unit, which is used to automatically reduce the weight of the associated nursing strategy when it is detected that the current activity pattern deviates from the historical model in the activity pattern-physiological response relationship library, and give priority to the conservative intervention plan as the final nursing strategy.

[0015] As a further solution of the present invention, the dynamic physiological feature mapping module further includes: a cross-modal consistency verification unit, which is used to determine whether the anomaly is a real risk or noise interference by analyzing the consistency in trend between other modal data and the abnormal data when receiving an indication of an abnormality in single sensor data; an asymmetric correction unit, which is used to only locally downgrade the data source that generates the anomaly when the cross-modal consistency verification unit determines that it is noise interference, without adjusting the global care strategy.

[0016] As a further solution of the present invention, the implicit feedback channel module further includes: an environmental stress index calculation unit, which is used to generate an environmental stress index ESI based on real-time environmental data and sensitivity parameters in the maternal health file:

[0017]

[0018] Among them, m is the number of environmental data types, s j is the sensitivity parameter of the pregnant woman to the jth environmental data determined based on her health records, E j is the real-time value of the jth type of environmental data; a nursing strategy elasticity correction unit is used to automatically increase the monitoring frequency of physiological indicators related to the environment when the environmental stress index exceeds a preset threshold, and trigger active intervention measures, including playing soothing music.

[0019] As a further solution of the present invention, the self-stabilizing control module further includes: a fuzzy health status partitioning unit, used to map the health status vector to a pre-set safety zone, warning zone and dangerous fuzzy interval, wherein the boundaries of the fuzzy interval are dynamically adjusted according to the individual differences of pregnant women; a dual threshold triggering unit, used to initiate progressive intervention measures, including pushing reminder information, when the health status vector continuously deviates from the safety zone but does not reach the dangerous threshold.

[0020] As a further solution of the present invention, the dynamic physiological feature mapping module adopts an incremental update algorithm to update the dynamic physiological feature mapping matrix.

[0021] As a further solution of the present invention, the data acquisition module is compatible with mainstream wearable devices, including smart bracelets and fetal heart monitors.

[0022] As a further solution of the present invention, the fuzzy logic inference in the self-stabilizing control module sets fuzzy rules based on the historical fluctuation range of the health status vector of the pregnant woman.

[0023] As a further solution of the present invention, the nursing execution module generates nursing suggestions according to the final nursing strategy or the robust nursing strategy and sends the suggestions to the user terminal.

[0024] As a further solution of the present invention, the system is deployed on an edge computing device.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] 1. In traditional pregnancy health monitoring, physiological data is not updated in a timely manner and cannot fully reflect individual differences. This invention introduces a dynamic physiological feature mapping mechanism to construct a continuous health status vector. It combines historical data for similarity matching, achieving real-time, personalized assessment of the health status of pregnant women. This approach no longer relies on discrete threshold judgments, but can more delicately capture the dynamic changes in health status, thus providing a more accurate basis for the formulation of subsequent nursing strategies.

[0027] 2. By analyzing daily activity patterns such as pregnant women's gait and sleep quality, as well as data such as ambient temperature, humidity, and air quality, the system can automatically adjust the priority of nursing strategies. This multi-dimensional data fusion approach, without the need for additional sensors or manual intervention, can more comprehensively assess the health risks of pregnant women and improve the accuracy of early warnings. This avoids the one-sidedness that can result from relying solely on single-dimensional physiological data, and avoids focusing solely on physiological indicators while ignoring the potential impact of the environment and daily activities on pregnant women's health. Furthermore, when data anomalies are detected, drastic intervention measures are not immediately taken. Instead, the system switches to robust mode based on the fluctuation range of historical data and uses fuzzy logic for inference. This mechanism effectively reduces the misjudgment rate, ensures the stability and reliability of the system in complex dynamic environments, and better meets the dual needs of pregnancy care for safety and psychological comfort.

[0028] 3. It can adapt to the pregnant woman's behavioral habits and issue more inappropriate intervention instructions. Specifically, it further integrates the activity pattern-physiological response relationship library through implicit feedback channels. By learning a personalized model that links the pregnant woman's historical activity data with changes in her physiological indicators, the system can identify whether the current activity pattern deviates from the normal range and adjust the weight of the nursing strategy accordingly, prioritizing conservative intervention plans. This dynamic adaptive approach effectively reduces the risk of excessive intervention and improves the precision and humanization of nursing care. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0030] Figure 1This is a schematic diagram of the functional module structure of the pregnancy health care control system based on data feedback analysis of the present invention.

[0031] Figure 2 Schematic diagram of the data processing flow of the dynamic physiological feature mapping module of the present invention.

[0032] Figure 3 This is a schematic diagram of the health status zoning and intervention mechanism of the present invention.

[0033] Figure 4 Schematic diagram of the functional structure of the implicit feedback channel module of the present invention.

[0034] Figure 5 This is a flow chart for the data feedback analysis and dynamic generation of health status vectors of the present invention.

[0035] Figure 6 This is a flowchart of the robust control mode switching and anomaly detection of the present invention. DETAILED DESCRIPTION

[0036] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present invention.

[0037] See also Figure 1 — Figure 6 , a pregnancy health care control system based on data feedback analysis, comprising:

[0038] The data acquisition module is used to collect real-time physiological data of pregnant women, including heart rate, fetal movement, blood pressure, as well as data on daily activity patterns of pregnant women, including gait and sleep quality, and environmental data, including temperature, humidity, and air quality;

[0039] The dynamic physiological characteristic mapping module is in communication with the data acquisition module and is used to construct a dynamic physiological characteristic mapping matrix based on the collected real-time physiological data and historical health trend data, and convert the discrete physiological data stream into a continuous health status vector H:

[0040]

[0041] Where n is the number of physiological data types, w i is the weight of the i-th physiological data determined based on historical data analysis, P i is the real-time value of the i-th physiological data, N(P i ) is the range of P according to the historical data of the pregnant woman iThe normalized value is used to match the initial care strategy in real time by calculating the similarity between the current health status vector and the historical health status vector.

[0042] an implicit feedback channel module, in communication with the data acquisition module and the dynamic physiological characteristic mapping module, for analyzing the implicit association between the daily activity pattern data of the pregnant woman and the environmental data, and automatically correcting the priority of the preliminary nursing strategy based on the analysis results to generate a final nursing strategy;

[0043] a self-stabilizing control module, in communication with the dynamic physiological feature mapping module and the implicit feedback channel module, configured to monitor the health state vector and, upon detecting data anomalies, automatically switch to a robust mode based on the historical fluctuation range of the health state vector, wherein fuzzy logic inference is used in place of precise calculation to generate a robust nursing strategy in the robust mode;

[0044] A nursing execution module is communicatively connected with the implicit feedback channel module and the self-stabilizing control module, and is used for executing the final nursing strategy or the robust nursing strategy.

[0045] As a further solution of the present invention, the implicit feedback channel module further includes: an activity pattern-physiological response relationship library, which is used to store personalized activity-response models established based on the historical activity data of pregnant women and changes in corresponding physiological indicators; a real-time strategy fine-tuning unit, which is used to automatically reduce the weight of the associated nursing strategy when it is detected that the current activity pattern deviates from the historical model in the activity pattern-physiological response relationship library, and give priority to the conservative intervention plan as the final nursing strategy.

[0046] As a further solution of the present invention, the dynamic physiological feature mapping module further includes: a cross-modal consistency verification unit, which is used to determine whether the anomaly is a real risk or noise interference by analyzing the consistency in trend between other modal data and the abnormal data when receiving an indication of an abnormality in single sensor data; an asymmetric correction unit, which is used to only locally downgrade the data source that generates the anomaly when the cross-modal consistency verification unit determines that it is noise interference, without adjusting the global care strategy.

[0047] As a further solution of the present invention, the implicit feedback channel module further includes: an environmental stress index calculation unit, which is used to generate an environmental stress index ESI based on real-time environmental data and sensitivity parameters in the maternal health file:

[0048]

[0049] Among them, m is the number of environmental data types, s j is the sensitivity parameter of the pregnant woman to the jth environmental data determined based on her health records, Ej is the real-time value of the jth type of environmental data; a nursing strategy elasticity correction unit is used to automatically increase the monitoring frequency of physiological indicators related to the environment when the environmental stress index exceeds a preset threshold, and trigger active intervention measures, including playing soothing music.

[0050] As a further solution of the present invention, the self-stabilizing control module further includes: a fuzzy health status partitioning unit, used to map the health status vector to a pre-set safety zone, warning zone and dangerous fuzzy interval, wherein the boundaries of the fuzzy interval are dynamically adjusted according to the individual differences of pregnant women; a dual threshold triggering unit, used to initiate progressive intervention measures, including pushing reminder information, when the health status vector continuously deviates from the safety zone but does not reach the dangerous threshold.

[0051] As a further solution of the present invention, the dynamic physiological feature mapping module adopts an incremental update algorithm to update the dynamic physiological feature mapping matrix.

[0052] As a further solution of the present invention, the data acquisition module is compatible with mainstream wearable devices, including smart bracelets and fetal heart monitors.

[0053] As a further solution of the present invention, the fuzzy logic inference in the self-stabilizing control module sets fuzzy rules based on the historical fluctuation range of the health status vector of the pregnant woman.

[0054] As a further solution of the present invention, the nursing execution module generates nursing suggestions according to the final nursing strategy or the robust nursing strategy and sends the suggestions to the user terminal.

[0055] As a further solution of the present invention, the system is deployed on an edge computing device.

[0056] Example 1: In this example, a pregnancy health care control system is deployed in a smart health terminal device with edge computing capabilities. This device can receive and process multi-dimensional data collected by smart bracelets and fetal heart rate monitors worn by pregnant women in real time, including heart rate, fetal movement, blood pressure, gait, sleep quality, and environmental data (temperature, humidity, and air quality). All data is wirelessly transmitted to the system's data acquisition module.

[0057] Specifically, the data acquisition module pre-processes the above data to generate a real-time physiological data sequence and activity pattern data sequence with time stamp alignment, while also acquiring and archiving external environmental data. On this basis, the system uses the dynamic physiological feature mapping module to perform weight mapping and normalization on various types of physiological data to form a continuous health state vector H. In this embodiment, the calculation formula for the health state vector is:

[0058]

[0059] The variable n represents the number of physiological data types monitored by the system, specifically the three indicators of heart rate, fetal movement, and blood pressure, so n = 3. i The weight of the i-th physiological data is determined by the system through statistical analysis of historical data. The weight is determined based on the pregnant woman's personal health record and risk sensitivity assessment during the pregnancy stage. For example, if historical data shows that the pregnant woman's blood pressure fluctuations in the second trimester have a greater impact on her overall health status, the system will appropriately increase the corresponding weight of blood pressure, w i Variable P i represents the real-time collection value of the i-th physiological data, which comes from the sensor data of the smart bracelet and fetal heart monitor. Function N(P i ) indicates the i The normalized results use the minimum-maximum normalization method of the same type of historical data for the pregnant woman to ensure that each indicator is within the interval [0,1], enhancing data comparability across indicators. Furthermore, to prevent abnormal fluctuations in the health state vector from misleading nursing strategies, the system introduces a cross-modal consistency verification unit in the dynamic physiological feature mapping module. When a mutation or abnormal signal appears in a sensor data, the system automatically compares the trend of other modal data in the same time window. If the trends are inconsistent, the system determines that the anomaly is noise interference and locally downgrades the data source through an asymmetric correction unit without affecting the weight distribution of the overall health state vector H, thereby reducing the risk of misjudgment.

[0060] The system further explores the potential correlation between the activity pattern of pregnant women and environmental data through the implicit feedback channel module. In this embodiment, the activity pattern-physiological response relationship library establishes a personalized activity response model based on the activity data of pregnant women in the past 30 days and the changes in corresponding physiological indicators. When the system detects that the current activity pattern (such as gait rhythm or sleep time) deviates significantly from the historical pattern, and this deviation trend is associated with abnormal fluctuations in the health status vector H, the system automatically reduces the intervention priority related to the activity pattern in the current nursing strategy and adopts a more conservative nursing plan. In terms of environmental data, the system uses the environmental stress index calculation unit to dynamically weight environmental parameters such as temperature, humidity and air quality to generate an environmental stress index ESI. In this embodiment, the weight parameter s j According to the sensitivity parameter setting in the health record of pregnant women, if pregnant women have basic respiratory diseases, the weight of the air quality index s j The calculation formula of environmental stress index is:

[0061]

[0062] Among them, m is the number of environmental data types, E j is the real-time value of the j-th environmental data, sj is the corresponding sensitivity weight. When the ESI exceeds the system-set threshold, the system automatically increases the monitoring frequency of environmental physiological indicators (such as heart rate fluctuations) and pushes soothing music to the user terminal, forming a flexible intervention mechanism. For the system's self-stabilizing control module, this embodiment details the triggering conditions and implementation mechanism of the robust mode. When the health state vector H deviates from the historical fluctuation range for 5 consecutive minutes and the cross-modal consistency check determines that the data anomaly is non-noise, the system activates the robust mode. In this mode, the system no longer relies on precise calculations, but instead uses preset fuzzy logic rules to map the H value to three fuzzy intervals: safe, warning, and dangerous, and adopts different intervention strategies based on the mapping results. The boundaries of the fuzzy intervals are dynamically adjusted based on the individual historical data of the pregnant woman to ensure that the intervention measures have personalized adaptability. To ensure the timeliness and dynamic adaptability of the health state mapping matrix, the dynamic physiological feature mapping module adopts an incremental update algorithm. Whenever new data is collected, the system locally updates the historical statistical data in the mapping matrix without the need for full reconstruction, thereby achieving low-latency, high-real-time dynamic adjustment of nursing strategies. These are all extended implementation methods known to those skilled in the art.

[0063] Example 2:

[0064] In the dynamic physiological feature mapping module, the calculation formula of the health state vector H is: The variable n represents the number of physiological data types monitored by the system, including heart rate, fetal movement, and blood pressure, so n = 3. i The weight of the i-th physiological data is determined based on the historical health records and individual physiological characteristics of the pregnant woman. In this embodiment, the system uses a statistical analysis of historical data, combined with the sensitivity of the pregnancy stage to various physiological data, to determine the weight w i For example, if a pregnant woman's blood pressure fluctuations in the second trimester have a greater impact on her overall health, the system will determine the corresponding weight w for blood pressure by analyzing the standard deviation of the physiological data over the past 30 days. i Higher, the specific value range is [0.2, 0.5], and it is updated in real time in the data acquisition module.

[0065] variable P i represents the real-time collection value of the i-th physiological data. The data sources are the smart bracelet and fetal heart rate monitor worn by pregnant women, ensuring that the data sources are clear and verifiable. Function N(P i ) indicates the i The result after normalization processing, this embodiment adopts the minimum-maximum normalization method, and the specific calculation method is:

[0066]

[0067] Among them, Pi,min and P i,max Respectively represent the minimum and maximum values of the i-th physiological data recorded in the pregnant woman's historical health records. The data is derived from historical monitoring data within the past 90 days. This normalization process ensures that all types of physiological data are integrated and calculated at a unified scale, avoiding the impact of data dimension differences on the accuracy of the health state vector. In the implicit feedback channel module, the environmental stress index ESI is calculated as follows:

[0068]

[0069] In this embodiment, the parameter m represents the number of environmental data types, specifically temperature, humidity, and air quality, so m=3. j It represents the sensitivity weight of pregnant women to the jth type of environmental data. Its source is the individual sensitivity record in the health file of pregnant women, and its value range is [0.1, 0.9]. For example, when a pregnant woman has respiratory diseases, the corresponding sensitivity weight of air quality is s j Set to 0.8. Parameter E j Represents the real-time value of the jth type of environmental data, and the data source is the environmental monitoring equipment deployed in the living environment of pregnant women. When the environmental stress index ESI exceeds the system-set threshold (set to 0.7 in this embodiment), the system automatically increases the monitoring frequency of relevant physiological indicators (such as heart rate), and at the same time triggers flexible intervention measures in the nursing execution module, such as pushing soothing music or prompting pregnant women to adjust their rest environment. In addition, in the self-stabilizing control module, the system sets the switching condition of the robust mode as follows: the health state vector H deviates from its historical fluctuation range for 5 consecutive minutes, and the cross-modal consistency verification unit determines that the abnormal data is non-noise interference. In this embodiment, the historical fluctuation range is dynamically calculated based on the mean and standard deviation of the health state vector H in the past 14 days, and the specific range is:

[0070] [H mean -2×H std ,H mean +2×H std ],

[0071] Among them, H mean and H std Respectively represent the mean and standard deviation of the health state vector H over the past 14 days. The data comes from the historical records of the dynamic physiological feature mapping module. In robust mode, the system uses a fuzzy logic inference mechanism to map the current health state vector H to three fuzzy intervals: safe zone, warning zone, and danger zone. The specific interval boundaries are dynamically adjusted based on individual differences of pregnant women. In this embodiment, the fuzzy interval boundaries are set based on the health risk sensitivity parameters in the pregnant woman's historical health file and the historical fluctuation characteristics of the health state vector H, ensuring that the intervention measures have individual adaptability. These are all extended implementation methods known to ordinary technicians in this field.

[0072] Example 3: In this example, the structure of each functional module and the data processing flow in the example of the present invention are further described in detail with reference to the accompanying drawings.

[0073] like Figure 1 As shown, the data feedback analysis-based pregnancy health care control system in this embodiment includes a data acquisition module, a dynamic physiological feature mapping module, an implicit feedback channel module, a self-stabilizing control module, and a care execution module. The data acquisition module is used to collect real-time physiological data, activity pattern data, and environmental data, forming a data stream that is transmitted to the dynamic physiological feature mapping module. The dynamic physiological feature mapping module constructs a health state vector for the pregnant woman based on the received data stream and outputs the health state vector. The health state vector is then passed to the implicit feedback channel module for policy fine-tuning, calculation of the environmental stress index, and generation of the final care strategy. Simultaneously, the health state vector is also input into the self-stabilizing control module for anomaly detection and robust mode determination, resulting in a robust care strategy. Once the final care strategy output by the implicit feedback channel module or the robust care strategy output by the self-stabilizing control module is generated, it is passed to the care execution module for policy delivery and intervention, achieving personalized, dynamic, and continuous pregnancy health care control for the pregnant woman. Through this system structure, information interaction of data flow, health state vector, final nursing strategy and robust nursing strategy is realized among the data acquisition module, dynamic physiological feature mapping module, implicit feedback channel module, self-stabilizing control module and nursing execution module, which improves the real-time performance, individual adaptability and robustness of the system.

[0074] like Figure 2As shown in Figure 2, this embodiment further details the data processing flow within the dynamic physiological feature mapping module. First, the system receives multi-source physiological data, including multi-dimensional real-time physiological data such as heart rate, fetal movement, and blood pressure. This multi-source physiological data undergoes historical data normalization using a minimum-maximum normalization method based on the data range in the pregnant woman's historical health records to ensure comparability of various physiological indicators. The normalized data then enters a dynamic weighting step. Based on the pregnant woman's historical health trends and current pregnancy stage, the system dynamically determines weight parameters for each physiological data type through statistical analysis of historical data, reflecting the impact of different physiological indicators on health status. Subsequently, the system calculates a health state vector based on the normalized data and dynamic weight parameters, generating a health state vector for subsequent care decisions. After the health state vector calculation is complete, the system initiates a cross-modal validation phase, identifying and validating abnormal data within the current data using a trend consistency analysis method. If the cross-modal validation determines that the data is normal, it is directly used for policy matching. If the validation determines that the data is abnormal, a local weighting process is triggered, locally adjusting the weight of the data source that generated the abnormal data to prevent the abnormal data from misleading the overall health status assessment. After cross-modal validation and local weight reduction, the data enters the strategy matching step. Based on the health status vector, the system compares the similarity with the maternal historical nursing strategy library to determine the appropriate initial nursing strategy, which serves as the basis for subsequent nursing decisions. Through this process, the system achieves standardized processing of multi-source physiological data, individualized weight assignment, abnormal data identification and correction, and dynamic matching of initial nursing strategies, improving the accuracy of maternal health status assessment and the scientific nature of nursing strategy generation.

[0075] like Figure 3 As shown, this embodiment further clarifies the specific structure of the health status zoning and intervention mechanism. The mechanism includes a fuzzy health status zoning unit, a dual-threshold trigger unit, and a progressive intervention strategy. The fuzzy health status zoning unit is used to map the health status vector of the pregnant woman to preset zoning information, specifically three intervals: safe zone, warning zone, and danger zone. The zoning information is used as a basis for judgment and transmitted to the dual-threshold trigger unit, which is used to trigger intervention based on the distribution of the health status vector between the safe zone, warning zone, and danger zone. When the health status vector continuously deviates from the safe zone but does not reach the danger zone threshold, the system activates the progressive intervention strategy through the dual-threshold trigger unit. Intervention methods include reminders, music, and intervention measures. Through flexible zoning and graded intervention, this mechanism effectively reduces the risk of over-intervention or omissions in pregnancy care. It realizes the dynamic division of safe zone, warning zone, and danger zone, and the organic synergy between dual-threshold triggering based on zoning information and the progressive intervention strategy, thereby improving the personalized adaptability and real-time intervention capabilities of the nursing system.

[0076] like Figure 4As shown, this embodiment further clarifies the functional structure of the implicit feedback channel module. This module includes an activity pattern-physiological response relationship library, an environmental stress index calculation unit, a real-time strategy fine-tuning unit, and a nursing strategy elasticity correction unit. The activity pattern-physiological response relationship library is used to store the historical activity model of individual pregnant women and establish a correspondence between activity patterns and physiological responses. The environmental stress index calculation unit is used to calculate the environmental stress index of the pregnant woman's environment in real time, specifically based on the sensitivity parameters in the pregnant woman's health record and current environmental data. The real-time strategy fine-tuning unit fine-tunes the priority of the current nursing strategy based on the degree of deviation between the historical activity model and the current activity pattern. After receiving the fine-tuning strategy and environmental stress index, the nursing strategy elasticity correction unit dynamically adjusts the current nursing strategy to ensure that the nursing strategy can be adjusted in real time based on the pregnant woman's individual historical behavior patterns and current environmental conditions. Through the above structure, the implicit feedback channel module achieves collaborative operation between the activity pattern-physiological response relationship library, the environmental stress index calculation unit, the real-time strategy fine-tuning unit, and the nursing strategy elasticity correction unit, effectively enhancing the individual adaptability and environmental sensitivity of the nursing decision-making process, and improving the dynamic optimization capability and robustness of the nursing strategy.

[0077] Figure 5This paper demonstrates the complete process for dynamic health state vector generation and anomaly detection in a prenatal health care control system based on data feedback analysis. The input layer first acquires real-time sensor data through a real-time sensor data acquisition module and temporarily stores the collected data using a sliding window cache mechanism to ensure the continuity and real-time nature of the data stream. Furthermore, the input layer also includes a historical feature library for storing historical health trend data for individual pregnant women. Based on this historical feature library, a covariance matrix generation module dynamically generates a covariance matrix, providing historical reference data for subsequent health status assessments. The processing layer includes a data normalization module and a dynamic weight calculator. The data normalization module normalizes the real-time sensor data in the sliding window cache to generate a feature vector, ensuring comparability of physiological data across different dimensions. The dynamic weight calculator dynamically updates the weight vector based on individual difference information in the historical feature library to form a weight vector, enhancing the system's personalized adaptability. Subsequently, the system performs a vector dot product operation on the feature vector and the weight vector at the processing layer to obtain the health state vector H. The output layer performs anomaly detection based on the health state vector H. If an abnormal signal is detected, the system activates the feedback update mechanism, transmits the current abnormal situation back to the input layer, and updates the data in the historical feature library and covariance matrix generation module to enhance the model's adaptability and robustness to dynamic changes in the health status of pregnant women. Through the above process, the present invention realizes the dynamic fusion of multi-dimensional health data during pregnancy, the real-time generation and anomaly detection of the health state vector H, and through the collaborative work between the sliding window cache, historical feature library, covariance matrix generation, data normalization module, feature vector, dynamic weight calculator, weight vector, vector dot product operation, health state vector H, anomaly detection and feedback update modules, a complete data feedback analysis and dynamic nursing strategy support system is formed.

[0078] Figure 6The present invention shows the processing flow of the self-stabilizing control module in the case of abnormal fluctuations in the health state vector H in the pregnancy health care control system based on data feedback analysis. The system continuously monitors the health state vector H in normal mode. When it detects that H exceeds historical fluctuations, it immediately triggers a cross-modal verification operation. The cross-modal verification is used to compare the trend consistency of the current abnormal data with other modal data. When the cross-modal verification result shows that the data is consistent, the system determines that the current abnormality is a real abnormality, and then immediately initiates the emergency protocol, takes active nursing intervention measures, and switches to the robust control mode to ensure the safety and continuity of care for pregnant women and fetuses; when the cross-modal verification result shows that the data is inconsistent, the system determines that the abnormality is noise interference, and then performs a sensor de-weighting operation to reduce the weight of the sensor from which the abnormal data comes, to avoid the noise data from misleading the overall health status assessment. At the same time, the system switches to the robust control mode to continue to maintain health monitoring. When the health state vector H is subsequently monitored to return to normal, the system automatically exits the robust control mode and returns to normal mode operation. The present invention achieves accurate identification and dynamic response to abnormal fluctuations in health status by setting specific process nodes such as H exceeding historical fluctuations, cross-modal verification, data consistency, real abnormality, initiation of emergency protocol, noise interference, sensor demotion, robust control mode, H returning to normal, and normal mode, thereby improving the robustness, stability and personalized adaptability of the pregnancy health care control system, all of which are extended implementation methods known to ordinary technicians in this field.

[0079] Example 4: In this example, during the modeling of the health status during pregnancy, the system uses the dynamic physiological feature mapping module to standardize and weight the collected multi-dimensional physiological parameters to generate a continuous health status vector H. In this example, n is the number of physiological data types, which is fixed to the three core physiological data types currently accessed by the system: heart rate (HR), fetal movement (FM), and blood pressure (BP), so n = 3. i : The real-time values of the i-th category physiological data are directly collected by the wearable smart bracelet and fetal heart monitoring equipment to ensure that the data source is verifiable; N(P i ):P i The results of the minimum-maximum normalization process are set based on the extreme values of the pregnant woman's historical data in the past 90 days. Specifically, the normalization process uses:

[0080]

[0081] Among them, P i,min With P i,max Represent the minimum and maximum values in historical observations, respectively, to ensure that the normalized data fluctuates within the interval [0,1] and enhance the comparability of different data dimensions; w i: Dynamic weight of the i-th category of physiological data, with a value range of [0.1, 0.6]. The weight calculation is based on the contribution of each physiological parameter to the system state fluctuation in the historical health trend, and is adjusted by comparing the standard deviation results in the last 30 days. For example, if a pregnant woman's blood pressure fluctuations in the second trimester significantly affect the health status assessment results, the system will increase the blood pressure weight w accordingly. BP After constructing the health status vector H, the system does not directly use the result to generate a nursing strategy, but uses it as an intermediate state and combines it with the vector H in the historical health status record database. his Perform similarity matching. The matching method uses the cosine similarity measurement method:

[0082]

[0083] The dot product and norm operations are both based on the basic linear algebra definition of three-dimensional vectors. The matching results are directly used to determine whether the current state is within a historically known health pattern or a deviation range. In terms of adjusting the priority of nursing strategies, the implicit feedback channel module introduces an activity pattern-physiological response relationship library mechanism. The construction of this mechanism is based on the following path: the system combines daily activity pattern data such as pregnant women's daily gait data (cadence, stride length changes) and sleep structure (deep sleep duration, number of sleep interruptions) with physiological data of the corresponding time period for multivariate linear regression modeling. For each type of activity pattern, a corresponding physiological response function is established, resulting in a set of personalized mapping models to determine whether the current activity state falls within the normal variation range. When the current activity state deviates significantly from the model prediction result, and this deviation is accompanied by a sharp fluctuation in the health state vector, the system will automatically trigger a strategy fine-tuning process, reducing the execution priority of the strategy associated with the abnormal activity and guiding it to a conservative nursing path. To further enhance the robustness of the system in highly uncertain environments, this embodiment standardizes the fuzzy logic inference mechanism in the self-stabilizing control module. The system constructs a dynamic fuzzy interval based on the standard deviation and mean of the H vector within 5 minutes of continuous sampling, and divides it into three levels: safe, warning, and dangerous. The mapping rules are as follows:

[0084] Safe zone: H∈[H mean -σ,H mean +σ];

[0085] Warning area: H∈[H mean ±2σ]\safe zone;

[0086] Danger zone: values exceeding the warning zone;

[0087] Among them H meanand σ are the mean and standard deviation of the health status vector over the last 14 days, respectively. The system employs a hierarchical strategy to trigger logic in different intervals to avoid misidentifying brief fluctuations as major anomalies and improve the accuracy of nursing recommendations. These are all extended implementations known to those skilled in the art.

[0088] Example 5: In the dynamic physiological feature mapping module, the health state vector H is calculated using a weighted normalized summation formula, which is expressed as follows:

[0089]

[0090] The variable n represents the number of physiological data types collected by the system, which is currently set to 3, corresponding to the indicators of heart rate (HR), fetal movement (FM) and blood pressure (BP). i Represents the real-time collection value of the i-th type of physiological data. The data source is the smart bracelet and fetal heart monitoring equipment worn by pregnant women to ensure the authenticity and verifiability of the data. Function N(P i ) represents the real-time acquisition value P i The normalization method used for the result after normalization is minimum-maximum normalization, and the specific calculation method is:

[0091]

[0092] Among them, P i,min With P i,max They respectively represent the minimum and maximum values of the i-th category physiological data in the historical health records of the pregnant woman in the past 90 days, with clear sources and traceability.

[0093] Weight parameter w i The sensitivity of each physiological indicator to fluctuations in the individual health records of pregnant women is dynamically set. Specifically, the standard deviation of each physiological data of the pregnant woman in the past 30 days is statistically analyzed, and the weight of each indicator's impact on the overall health status is determined in combination with the risk assessment results of the pregnancy stage. i The value range of is set to [0.1, 0.6]. This range is determined to avoid imbalances in the health state vector caused by excessively high or low weights for any one indicator, ensuring a reasonable distribution of the contributions of each indicator. The design logic behind the above health state vector calculation formula is that, unlike traditional methods that rely on fixed thresholds or single-indicator monitoring, this system establishes a continuous and interpretable health state assessment model by introducing historical individual data normalization and dynamic weight allocation. This enhances the system's adaptability to individual differences in pregnant women and avoids the risk of misjudgment due to a single abnormal indicator.

[0094] In the implicit feedback channel module, the system further enhances the personalized adaptability of the nursing strategy by establishing an activity pattern-physiological response relationship library. This relationship library is based on the daily activity pattern data (including gait frequency, sleep structure, etc.) of pregnant women in the past 30 days and the physiological data of the corresponding time period. Through multivariate linear regression analysis, an individualized mapping model between each activity pattern and the fluctuation of physiological indicators is established. The current activity pattern data comes from the real-time monitoring results of the data acquisition module. The system compares the degree of deviation between the current activity pattern and the historical model in real time. When it is detected that there is a significant deviation between the current activity pattern and the historical model, and the deviation trend is associated with the abnormal fluctuation of the health state vector H, the system automatically reduces the weight of the nursing strategy related to the abnormal activity pattern and gives priority to conservative intervention plans. The design logic of this fine-tuning process is to avoid excessive intervention due to short-term behavioral abnormalities and ensure the rationality and stability of the nursing strategy. In addition, the calculation formula of the environmental stress index ESI in the system is:

[0095]

[0096] The variable m represents the number of environmental data types, which is currently set to 3, corresponding to the indicators of temperature, humidity, and air quality. j is the real-time collection value of the jth type of environmental data, which comes from the environmental monitoring equipment in the living environment of pregnant women. Sensitivity parameter s j It is derived from the records of environmental sensitivity in the health files of pregnant women, and the value range is set to [0.1, 0.9], which is dynamically adjusted according to the individual health status and previous medical history of the pregnant woman. When the calculated environmental stress index ESI exceeds the system preset threshold (the threshold is specifically set to 0.7 in the embodiment), the system automatically increases the monitoring frequency of physiological indicators related to the environment (such as heart rate), and pushes flexible intervention measures to the user terminal through the nursing execution module, such as playing soothing music or prompting to adjust the rest environment. The original intention of the design of this mechanism is to dynamically quantify the potential impact of environmental factors on the health status of pregnant women and provide real-time feedback to avoid delayed response of nursing strategies due to ignoring environmental changes.

[0097] In the self-stabilizing control module, in order to effectively distinguish between real anomalies and noise interference, the system introduces a cross-modal consistency verification unit and an asymmetric correction unit. When a single sensor data anomaly is detected, the system compares the trend consistency of other modal data in the same period. If there is trend consistency, it is determined to be a real anomaly and the robust mode is activated; if the trend is inconsistent, it is determined to be noise interference, and only the weight of the abnormal data source is locally reduced to avoid misjudgment from spreading to the overall care strategy. The switching condition of the robust mode is set as follows: the health state vector H deviates from the historical fluctuation range for 5 consecutive minutes, and the cross-modal verification result is a real anomaly. The historical fluctuation range is dynamically calculated based on the mean and standard deviation of H in the past 14 days. The specific range is [Hmean -2×H std ,H mean +2×H std ], where H mean With H std represent the historical mean and standard deviation respectively.

[0098] In robust mode, the system uses preset fuzzy logic rules to map the health status vector H into three intervals: safe, warning, and dangerous. The interval boundaries are dynamically set based on the risk sensitivity in the pregnant woman's historical health records to ensure that the intervention measures are individualized and adaptable, and effectively avoid excessive intervention due to short-term abnormalities. These are all extended implementation methods that can be known to ordinary technicians in this field.

[0099] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations that come within the meaning and range of equivalents of the claims be embraced therein.

Claims

1. A pregnancy health care control system based on data feedback analysis, characterized in that: include: The data acquisition module is used to collect real-time physiological data of pregnant women, including heart rate, fetal movement, blood pressure, as well as data on daily activity patterns of pregnant women, including gait and sleep quality, and environmental data, including temperature, humidity, and air quality; The dynamic physiological characteristic mapping module is in communication with the data acquisition module and is used to construct a dynamic physiological characteristic mapping matrix based on the collected real-time physiological data and historical health trend data, and convert the discrete physiological data stream into a continuous health status vector H: Where n is the number of physiological data types, w i is the weight of the i-th physiological data determined based on historical data analysis, P i is the real-time value of the i-th physiological data, N(P i ) is the range of P according to the historical data of the pregnant woman i The normalized value is used to match the initial care strategy in real time by calculating the similarity between the current health status vector and the historical health status vector. an implicit feedback channel module, communicatively connected to the data acquisition module and the dynamic physiological characteristic mapping module, configured to analyze the association between daily activity pattern data and the physiological data, perform analysis based on the environmental data, automatically modify the priority of the preliminary nursing strategy according to the analysis results, and generate a final nursing strategy; a self-stabilizing control module, in communication with the dynamic physiological feature mapping module and the implicit feedback channel module, configured to monitor the health state vector and, upon detecting data anomalies, automatically switch to a robust mode based on the historical fluctuation range of the health state vector, wherein fuzzy logic inference is used in place of precise calculation to generate a robust nursing strategy in the robust mode; a nursing execution module, communicatively connected to the implicit feedback channel module and the self-stabilizing control module, and configured to execute the final nursing strategy or the robust nursing strategy; Furthermore, the implicit feedback channel module further includes: an environmental stress index calculation unit, configured to generate an environmental stress index ESI based on real-time environmental data and sensitivity parameters in the maternal health file: Among them, m is the number of environmental data types, s j is the sensitivity parameter of the pregnant woman to the jth environmental data determined based on her health records, E j is the real-time value of the jth type of environmental data; a nursing strategy elasticity correction unit is used to automatically increase the monitoring frequency of physiological indicators related to the environment when the environmental stress index exceeds a preset threshold, and trigger active intervention measures, including playing soothing music.

2. The pregnancy health care control system based on data feedback analysis according to claim 1, characterized in that: The implicit feedback channel module further includes: an activity pattern-physiological response relationship library, which is used to store personalized activity-response models established based on the historical activity data of pregnant women and changes in corresponding physiological indicators; a real-time strategy fine-tuning unit, which is used to automatically reduce the weight of the associated nursing strategy when it is detected that the current activity pattern deviates from the historical model in the activity pattern-physiological response relationship library, and give priority to adopting a conservative intervention plan as the final nursing strategy.

3. The pregnancy health care control system based on data feedback analysis according to claim 1, characterized in that: The dynamic physiological feature mapping module further includes: a cross-modal consistency verification unit, which is used to determine whether the anomaly is a real risk or noise interference by analyzing the consistency in trend between other modal data and the abnormal data when receiving an indication of an abnormality in single sensor data; an asymmetric correction unit, which is used to only locally downgrade the data source that generated the anomaly when the cross-modal consistency verification unit determines that it is noise interference, without adjusting the global care strategy.

4. The pregnancy health care control system based on data feedback analysis according to claim 1, characterized in that: The self-stabilizing control module further includes: a fuzzy health status partitioning unit, used to map the health status vector to a pre-set safety zone, warning zone and dangerous fuzzy interval, wherein the boundary of the dangerous fuzzy interval is dynamically adjusted according to the individual differences of pregnant women; a dual-threshold triggering unit, used to initiate progressive intervention measures, including pushing reminder information, when the health status vector continuously deviates from the safety zone but does not reach the dangerous fuzzy interval.

5. The pregnancy health care control system based on data feedback analysis according to claim 1, characterized in that: The dynamic physiological feature mapping module updates the dynamic physiological feature mapping matrix using an incremental update algorithm.

6. The pregnancy health care control system based on data feedback analysis according to claim 1, characterized in that: The data acquisition module is compatible with mainstream wearable devices, including smart bracelets and fetal heart monitors.

7. The pregnancy health care control system based on data feedback analysis according to claim 1, characterized in that: The fuzzy logic inference in the self-stabilizing control module sets fuzzy rules based on the historical fluctuation range of the health status vector of the pregnant woman.

8. The pregnancy health care control system based on data feedback analysis according to claim 1, characterized in that: The nursing execution module generates nursing suggestions according to the final nursing strategy or the robust nursing strategy and sends the suggestions to the user terminal.

9. The pregnancy health care control system based on data feedback analysis according to claim 1, characterized in that: The system is deployed on edge computing devices.

Images