A health status assessment method and system based on multi-point body surface temperature and vital sign data during sleep

Abstract

The present application relates to the field of artificial intelligence and intelligent sleep technology, and discloses a health state evaluation model generation method and system based on multi-point body surface temperature and vital sign data during sleep. The method continuously collects body surface temperature data of multiple body contact areas during the user's sleep process, and synchronously collects data such as heart rate, respiratory rate, body movement, environmental temperature and bed surface temperature, to construct a multi-dimensional sleep physiological data set. On this basis, the multi-dimensional sleep physiological data is preprocessed, feature extraction and model analysis are performed, and the health state evaluation result and personalized health insight information of the user during sleep are generated. The staged evaluation can also be carried out in combination with the change of sleep stage, and the model is continuously optimized based on the long-term temperature adjustment behavior of the user. By using the present application, the accuracy, continuity and individual adaptability of sleep health state evaluation can be improved, and more accurate sleep health analysis basis is provided for the user.

Description

Technical Field

[0001] This invention relates to the fields of artificial intelligence and intelligent control technology, specifically to a method and system for assessing health status using multi-point body surface temperature and vital signs data during sleep. Background Technology

[0002] With the rapid development of smart sleep and digital health technologies, analyzing health status based on human physiological data during sleep has become an important application area in the smart bedding field. In existing technologies, the collection of nighttime human physiological data mainly relies on wearable devices such as smart bracelets and smartwatches. These devices typically use physiological parameters such as heart rate, respiratory rate, and body movement as core analytical criteria, combining only local single-point temperature data to complete sleep health assessments. However, the temperature collection locations of these wearable devices are mostly limited to local areas such as the wrist, making them susceptible to interference from wearing tightness, skin contact conditions, and local ambient temperature. This makes it difficult to obtain continuous, stable, and zonal body surface temperature change data throughout the night, resulting in significant limitations in reflecting the overall thermoregulation state of the human body during sleep.

[0003] Furthermore, the metabolic level, stress recovery status, and menstrual cycle changes during sleep are not only closely related to core vital signs such as heart rate and respiration, but also highly correlated with the distribution pattern of body surface temperature throughout the night and the heat dissipation characteristics of different body areas. Existing wearable devices such as smart bracelets and smartwatches lack continuous, multi-point, and zoned body surface temperature data input, making it impossible to accurately support in-depth inference and analysis of these health states. In contrast, smart bedding systems can continuously collect body surface temperature data from multiple contact areas of the body without the user's awareness, and combine this data with relevant information such as heart rate, respiratory rate, and ambient temperature to construct a more complete sleep physiological dataset. Therefore, multi-point body surface temperature data possesses unique technical value in estimating human metabolism, assessing stress status and recovery level, and inferring women's menstrual cycles, which is difficult for ordinary wearable devices to replace. Based on the shortcomings of existing technologies and practical application needs, there is an urgent need to provide a method and system for generating personalized health insight models based on multi-point body surface temperature and vital sign data during sleep, to solve the technical problems of insufficient accuracy and limited applicability of existing assessment methods. Summary of the Invention

[0004] To overcome the shortcomings of the existing technology and address the issues of limited temperature control zones, insufficient adjustment precision, and lack of coordinated control across multiple bedding components, this invention provides a health status assessment method based on multi-point body surface temperature and vital sign data during sleep. The specific solution is as follows: A method for sleep health assessment and target temperature generation based on multi-point body surface temperature and vital sign data during sleep, characterized by the following steps: Step S101: Input basic model parameters, obtain user basic information, and establish a basic model parameter set based on the user basic information; Step S102: Input standard values ​​of sleep stage temperature, and determine the corresponding standard values ​​or target temperature ranges for the user in different sleep stages; Step S103: Access third-party auxiliary data, accessing third-party auxiliary data related to the user's sleep state; Step S104: Establish a long-term learning model, establishing a long-term learning model that includes at least a sleep habit health model and a personalized sleep temperature health model; Step S105: Perform state recognition and multi-dimensional data collection, identifying the user's state and collecting multi-dimensional sleep physiological data in real time during the user's sleep process, the multi-dimensional sleep physiological data including temperature data of multiple body contact areas, vital sign data, environmental data, user's bed-in state, and user's current sleep state; Step S106: Perform data preprocessing and special processing. The process involves several steps: Step S107: Preprocessing and extracting features from the multidimensional sleep physiological data to obtain structured feature information; Step S108: Constructing a health status assessment model and outputting a target temperature. Based on the structured feature information, a health status assessment model is constructed or invoked, outputting the health status assessment result and the target temperature for the current or next stage; Step S109: Executing target temperature control and recording user fine-tuning feedback. Target temperature control is executed on the corresponding zones of the smart bedding according to the target temperature, and user fine-tuning feedback information is recorded; Step S110: Updating the personalized model. Based on the user fine-tuning feedback information, the current sleep health assessment result, and long-term accumulated sleep-temperature-health correlation data, the personalized sleep temperature health model is updated; Step S110: Performing a comprehensive sleep health evaluation. After the user completes a full night's sleep, a comprehensive sleep health evaluation result is output; Step S111: Outputting a comprehensive sleep report and personalized health insight information. Based on the comprehensive sleep health evaluation result, a comprehensive sleep report and personalized health insight information are output.

[0005] Furthermore, in step S101, the user basic information includes one or more of gender, age, BMI, height, weight, region, physical characteristics, and sleep preferences; the personalized model initialization includes: performing feature encoding on the categorical data in the user basic information, normalizing the continuous data, and inputting the processed data into the basic model to construct the initial user profile of the current user.

[0006] Furthermore, in step S102, the standard temperature value or target temperature range corresponding to different sleep stages is determined based on at least one of user presets, historical learning results, basic model recommended values, or output results of similar population models; in step S103, the third-party auxiliary data includes exercise data, female menstrual cycle data, daily activity data, and other auxiliary data related to sleep state.

[0007] Furthermore, in step S104, the sleep habit health model is used to record the correlation between temperature, vital signs and health outcomes of multiple users during sleep, and the personalized sleep temperature health model is used to record the temperature control preferences, body temperature response patterns, health assessment results and feedback information of specific users; the long-term learning model is updated using an online learning method.

[0008] Furthermore, in step S105, the temperature data of the multiple body contact areas are collected through the bed surface partitioning method of the smart bedding; the vital signs data include one or more of heart rate, respiratory rate, heart rate variability, body movement data, and pulse data; the environmental data includes one or more of ambient temperature, bed surface temperature, ambient humidity, air quality, and day / night information; the user's bed status is identified through pressure distribution detection, and the user's current sleep status is identified based on heart rate, respiratory rate, and body movement intensity.

[0009] Furthermore, in step S106, the preprocessing includes time synchronization processing, outlier removal, noise suppression, missing value correction, and sleep stage segmentation; the feature extraction includes multi-point body surface temperature change features, regional temperature gradient features, heart rate variability features, respiratory stability features, body movement features, and environmental change correlation features.

[0010] Furthermore, in step S107, the health status assessment model adopts an architecture that integrates a basic model and a personalized model, and performs joint modeling based on the sleep temperature basic model and the current user's individual data; the target temperature is generated according to the current sleep stage, vital signs, body surface temperature change pattern and environmental changes, and the target temperature is a single area target temperature or a target temperature of multiple zones.

[0011] Furthermore, in step S108, the target temperature control includes heating, cooling, constant temperature maintenance, or zoned differentiated temperature control; during the execution of the target temperature control, the actual temperature of the bed surface is collected in real time, and closed-loop adjustment is performed based on the actual temperature of the bed surface; the user fine-tuning feedback information includes user operation information for adjusting the target temperature upwards, downwards, switching modes, or terminating control.

[0012] Furthermore, in step S109, the update of the personalized sleep temperature health model includes one or more of the following: preference parameter correction, temperature generation weight adjustment, phased temperature curve revision, feature correlation strength update, and user individual profile remodeling; the update is based on the temperature deviation corresponding to the user's fine-tuning instruction, the current sleep feature data, and the current sleep health assessment results.

[0013] Furthermore, in step S110, the comprehensive sleep health evaluation results include sleep duration, sleep temperature performance, sleep recovery status, and evaluation indicators reflecting sleep quality and physiological state; in step S111, the comprehensive sleep report includes sleep duration, sleep stage distribution, sleep temperature changes, sleep recovery status, and health status assessment results; the personalized health insight information includes one or more of the following: metabolic estimation, stress recovery assessment, menstrual cycle prediction, and other health analysis results related to sleep and body temperature changes.

[0014] Meanwhile, this invention provides a sleep health assessment and target temperature generation device based on multi-point body surface temperature and vital sign data during sleep. The device comprises: a basic model parameter input module for acquiring basic user information and establishing a basic model parameter set based on that information; a sleep stage temperature standard value input module for determining the corresponding temperature standard value or target temperature range for the user at different sleep stages; a third-party auxiliary data access module for accessing third-party auxiliary data related to the user's sleep state; a long-term learning model establishment module for establishing a long-term learning model including at least a sleep habit health model and a personalized sleep temperature health model; a state recognition and multi-dimensional data acquisition module for recognizing the user's state during sleep and acquiring multi-dimensional sleep physiological data in real time, including temperature data from multiple body contact areas, vital sign data, environmental data, the user's in-bed state, and the user's current sleep state; and data preprocessing and feature extraction. The system comprises the following modules: a data acquisition module for preprocessing and feature extraction of the multidimensional sleep physiological data to obtain structured feature information; a health status assessment and target temperature output module for constructing or calling a health status assessment model based on the structured feature information, outputting the health status assessment result and the target temperature for the current or next stage; a target temperature control and user fine-tuning feedback recording module for implementing target temperature control on the corresponding zones of the smart bedding according to the target temperature, and recording user fine-tuning feedback information; a personalized model update module for updating the personalized sleep temperature health model based on the user fine-tuning feedback information, the current sleep health assessment result, and long-term accumulated sleep-temperature-health correlation data; a comprehensive sleep health evaluation module for outputting a comprehensive sleep health evaluation result after the user completes a full night's sleep; and a comprehensive sleep report and personalized health insight information output module for outputting a comprehensive sleep report and personalized health insight information based on the comprehensive sleep health evaluation result.

[0015] This invention constructs a dedicated health status assessment model. Relying on the sensing module integrated into a smart bedding system, it collects cardiac impact signals imperceptibly during the user's natural sleep state. The model then performs real-time analysis and calculation of these signals, accurately extracting three core vital signs: heart rate, respiratory rate, and heart rate variability, while simultaneously quantifying real-time body movement intensity. Simultaneously, through the smart bedding's multi-point temperature sensing units, it continuously collects surface temperature data from multiple contact areas, including the chest, abdomen, and extremities. This data, combined with environmental parameters such as ambient temperature and bed surface temperature, forms a multi-dimensional, full-cycle set of sleep physiological data. The extracted vital signs, body movement intensity data, and multi-point surface temperature data are input into the assessment model. The model uses a preset algorithm to perform feature matching, weight analysis, and anomaly identification, completing an accurate assessment of the user's sleep outcomes (including sleep stage division and sleep quality level determination). Furthermore, by incorporating dynamic changes in sleep stages, it outputs a health status assessment result tailored to the user's individual characteristics.

[0016] Compared with existing technologies, this invention has the following advantages and beneficial effects: This invention relies on an intelligent bedding system to continuously collect multi-point body surface temperature data during the user's sleep process. Compared to the single-point collection method of traditional wearable devices, it can more realistically reflect the overall thermoregulation state of the human body. Since data collection is completed imperceptibly during natural sleep, without the need for additional devices, it reduces sleep disturbance and improves data stability and accuracy. Simultaneously, this invention integrates multimodal data such as body surface temperature, heart rate, respiration, and environmental parameters, which can improve the accuracy and reliability of health status assessment. Furthermore, this invention can dynamically analyze the physiological characteristics of different sleep stages, achieving phased assessment and personalized analysis, and can be extended to various health scenarios such as metabolic estimation, stress recovery assessment, and menstrual cycle prediction. In addition, this invention can continuously update personalized models by recording users' sleep, temperature, and health data over a long period, thereby continuously optimizing health assessment and thermoregulation decisions. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a health status assessment method based on multi-point body surface temperature and vital signs data during sleep, provided by an embodiment of the present invention.

[0018] Figure 2 This is a flowchart illustrating a health status assessment method based on multi-point body surface temperature and vital signs data during sleep, provided in an embodiment of the present invention. Detailed Implementation

[0019] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0020] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0021] First, combine Figure 1 The illustration provides a detailed description of a health status assessment method based on multi-point body surface temperature and vital signs data during sleep, as provided in this embodiment.

[0022] This invention provides a health status assessment method based on multi-point body surface temperature and vital sign data during sleep, applicable to intelligent bedding systems with multiple temperature control zones. The intelligent bedding system may include multiple temperature control zones distributed along the longitudinal and / or transverse direction of the human body. Each temperature control zone can receive control commands and execute heating, cooling, or heat preservation adjustments. It should be noted that this embodiment does not limit the specific temperature adjustment execution method; the temperature control zones can employ electric heating, fluid circulation, air supply heat exchange, phase change temperature regulation, or other technologies capable of temperature regulation. Therefore, the control method described in this embodiment has good versatility.

[0023] The smart bedding with multiple temperature control zones includes: a main unit and a bed temperature control structure connected to the main unit, wherein: The main unit includes a storage tank, a distribution tank, a circulating power component, a semiconductor temperature control module, a heat exchange component, and a control processing unit. The storage tank is used to store the circulating medium, preferably water or a heat-conducting liquid. The circulating power component is used to drive the circulating medium to flow in the system. The semiconductor temperature control module is used to heat or cool the circulating medium. The heat exchange component is used to realize heat exchange between the circulating medium and the outside air. The control processing unit is used to execute temperature control algorithms and generate control commands.

[0024] The bed temperature control structure includes a skin-contact layer, a heat-conducting layer, a temperature control layer, and a bottom layer. The temperature control layer has multiple conduits inside, which are connected to the main unit through inlet and outlet water pipes to form a circulation loop. The temperature control layer is divided into multiple temperature control zones along the longitudinal and / or transverse direction of the human body, preferably four temperature control zones, each of which corresponds to an independent fluid channel or flow control path.

[0025] Furthermore, the main unit and the bed temperature control structure are connected by an inlet pipe and a return pipe. The circulating medium is output from the main unit, flows through each temperature control zone, and returns to the main unit to form a closed loop. In the implementation, each temperature control zone corresponds to an independent branch or is controlled by a diversion structure, so that different zones can receive circulating media of different temperatures or different flow rates.

[0026] The control processing unit is connected to the circulating power component, the temperature control module, and each zone control unit, respectively. It is used to control the working state of the semiconductor temperature control module and the flow output of the water pump component according to the target temperature of each temperature control zone obtained by the multi-zone independent collaborative temperature control method, and to adjust the fluid distribution of each zone, thereby realizing independent control of the temperature of each temperature control zone.

[0027] The device also includes an air duct structure for dissipating heat from the heat exchange components in the main unit. The air duct is preferably configured as a downward air outlet structure to reduce thermal interference to the user on the bed and avoid affecting the ambient temperature detection.

[0028] In this embodiment, the method may specifically include the following steps: Step S101: Input basic model parameters The system first acquires basic information about different users and then establishes a basic model parameter set based on this information. This basic information includes parameters such as gender, age, BMI, height, weight, region, physical characteristics, and sleep preferences. In the initial stage, the system collects, tests, and classifies a large amount of basic data from different users to construct a sleep-temperature basic dataset. This dataset records the temperature change patterns, vital signs, and health status of different types of users during sleep, providing data support for subsequent feature mining. Based on this basic dataset, a basic model is trained using a logistic regression algorithm. Through training on massive amounts of sample data, the system determines the correlation weights between different basic information and sleep temperature and vital signs, providing quantitative support for the initialization of subsequent personalized models.

[0029] For each specific user, the system constructs an initial user profile based on their basic information. This initial user profile serves as the foundational input for subsequent multi-point body surface temperature analysis, sleep health status assessment, and target temperature generation, providing support for personalized analysis and model initialization. During the initial user profile construction process, feature encoding is used to convert categorical data such as gender and physical characteristics into numerical features. These features are then normalized together with continuous features such as age and BMI. After processing, the data is input into the basic model to complete the initialization of the personalized model.

[0030] Step S102: Input the standard temperature value for the sleep stage. Based on user presets, historical learning results, recommended values ​​from the base model, and output results from models targeting similar populations, the system determines the standard temperature value or target temperature range for the user at different sleep stages. These sleep stages may include light sleep, deep sleep, REM sleep, and wakefulness, with different sleep stages corresponding to different target temperature ranges, temperature baselines, or temperature change trends.

[0031] By establishing temperature standard values ​​corresponding to different sleep stages, the system can provide a reference for subsequent dynamic analysis of sleep cycles, phased assessment of health status, and personalized target temperature output, enabling temperature control logic and health assessment logic to work together around changes in sleep stages. Specifically, the temperature standard values ​​for different sleep stages are obtained through clustering analysis of sleep temperature data from similar populations using a clustering algorithm, ensuring the rationality and universality of the standard values.

[0032] Step S103: Access third-party auxiliary data The system integrates third-party auxiliary data from external devices, platforms, and databases as supplementary inputs for personalized sleep health assessments and target temperature generation, expanding the data dimensions of sleep health analysis. This third-party auxiliary data includes exercise data, women's menstrual cycle data, daily activity data, and other auxiliary data related to sleep status.

[0033] Among them, exercise data can reflect a user's daily activity level, fatigue level, and recovery needs, while menstrual cycle data can capture the fluctuations in body temperature and differences in sleep temperature requirements during specific physiological stages. Integrating third-party auxiliary data can improve the completeness of information in assessment and target temperature generation, breaking the limitations of a single data source and making the output results more consistent with the user's current actual state. After standardized preprocessing, the third-party auxiliary data is integrated with the sleep physiological data collected in real time by the system to ensure the rationality and accuracy of data utilization.

[0034] Step S104: Establish a long-term learning model The system utilizes a sleep habit health model and a personalized sleep temperature health model to continuously record, analyze, and learn from users' long-term sleep behavior, temperature preferences, and changes in health status. The sleep habit health model, built using a deep learning architecture, records the correlations between temperature, vital signs, and health outcomes during sleep for a large number of users. Through multi-dimensional data correlation analysis, it captures long-term data dependency characteristics, forming a general-purpose basic learning capability. The personalized sleep temperature health model, built using machine learning algorithms, records specific users' temperature control preferences, body temperature response patterns, health assessment results, and feedback information during sleep, and is continuously updated over long-term use.

[0035] By constructing the aforementioned long-term learning model, the system can not only continuously assess users' sleep health status but also continuously correct and optimize model parameters based on the closed-loop relationship of users' long-term sleep-temperature-health data. This gradually forms a more accurate, personalized, and iterative user-specific model. Model optimization adopts an online learning approach. After each sleep session, the sleep data, temperature control feedback, and health assessment results are used as new samples to incrementally update the model parameters. This eliminates the need to retrain the entire model, effectively improving optimization efficiency and ensuring that the model adapts to individual user characteristics over the long term.

[0036] Step S105: Perform status recognition and multi-dimensional data collection. During the user's sleep, the system identifies the user's state through the smart bedding system and collects multi-dimensional sleep physiological data in real time to build a multi-dimensional sleep physiological dataset for health assessment and temperature control decisions.

[0037] The collected data includes at least: temperature data of multiple body contact areas, vital sign data, environmental data, user's bed-being status, and user's current sleep status.

[0038] Temperature data from multiple body contact areas reflects changes in the surface temperature of different body regions. Preferably, these multiple body contact areas can be set up by dividing the bed surface into zones, such as upper body, lower body, left side, right side, or other areas corresponding to the body contact points, to achieve continuous multi-point data collection. Vital signs data reflects changes in the user's physiological state during sleep. This vital signs data may include one or more of heart rate, respiratory rate, body movement data, and pulse data. Heart rate, respiratory rate, and heart rate variability are obtained through cardiac impact signal analysis. After denoising the cardiac impact signal using a wavelet transform algorithm, feature parameters are extracted using a peak detection algorithm. Body movement data is obtained by quantifying the signal intensity collected by a pressure sensor, and the signal variance within the window is calculated using a sliding window method as an indicator of body movement intensity. Environmental data reflects changes in the sleep environment. This environmental data may include one or more of ambient temperature, bed surface temperature, ambient humidity, air quality, or diurnal time information. Environmental data is collected using integrated sensors at a fixed sampling frequency for continuous acquisition. After collection, outlier values ​​are initially filtered out to remove data exceeding reasonable limits, ensuring the validity and reliability of the environmental data. User bed-being status is used to determine whether the user is in bed, out of bed, or preparing for sleep. A piezoresistive sensor array detects the pressure distribution of the human body on the bed surface, and a threshold judgment method is used to accurately identify the user's bed-being status, ensuring accurate status assessment. The user's current sleep state is used to determine whether the user is in the sleep onset, light sleep, deep sleep, or REM sleep stage. Based on three core data points—heart rate, respiratory rate, and body movement intensity—a machine learning classification algorithm is used to identify the user's current sleep state. This algorithm, trained and optimized with a large amount of sample data, can accurately distinguish sleep states, meeting the practical needs of sleep stage analysis.

[0039] Compared to the single-point body temperature collection method of traditional wearable devices, this embodiment uses a smart bedding system to complete the non-intrusive continuous collection of multi-point body surface temperature during the user's natural sleep state. This reduces the interference of wearing devices on the sleep process, improves the stability, authenticity and continuity of data collection, and more realistically reflects the overall thermoregulation state of the human body.

[0040] Step S106: Perform data preprocessing and feature extraction. After obtaining the aforementioned multidimensional data, the system preprocesses the collected data and extracts feature information related to sleep health status assessment. The preprocessing process may include time synchronization, outlier removal, noise suppression, missing value correction, and sleep stage segmentation to improve the accuracy and stability of subsequent analysis and modeling. Specific processing details are as follows: Step S1061, Time Synchronization Processing: Based on the system time, timestamps are aligned for temperature data, vital signs data, and environmental data to ensure time consistency of different types of data, with synchronization errors controlled within ±100ms.

[0041] Step S1062, Outlier Removal: Statistical anomaly detection methods are used to detect outliers in various types of data and remove data that exceeds the reasonable statistical range. For temperature data, an additional screening is performed based on the normal human body surface temperature range (35.5-37.5℃) to ensure the rationality of the data.

[0042] Step S1063, Noise Suppression: Temperature data is filtered using a filtering algorithm to suppress environmental interference noise, and cardiac impact signals are denoised using a denoising processing method to retain effective signal components.

[0043] Step S1064, Missing Value Correction: For a small amount of missing data (missing rate ≤ 5%), interpolation is used for imputation. For data with a high missing rate (5% < missing rate ≤ 10%), a data imputation algorithm is used to ensure the integrity of the dataset.

[0044] Step S1065, Sleep Stage Division: Based on the sleep state identified in step S105, the entire sleep cycle is divided into stages. The data after division is grouped according to the sleep stage to provide a basis for subsequent staged assessment.

[0045] Based on this, the system extracts various feature information, including: multi-point body surface temperature change features, zone temperature gradient features, heart rate variability features, respiratory stability features, body movement features, and environmental change correlation features. Among these, multi-point body surface temperature change features can be used to characterize the dynamic changes in body surface temperature in different areas during sleep. Zone temperature gradient features can be used to characterize the temperature distribution relationship between different body regions. Heart rate variability features and respiratory stability features can be used to reflect the user's autonomic nervous activity state and sleep stability during sleep. Feature extraction specifically includes: for temperature features, extracting the mean, standard deviation, maximum, minimum, rate of temperature change, and temperature fluctuation amplitude of each zone temperature, and using correlation analysis to calculate the correlation between different zone temperatures as zone temperature correlation features. For vital sign features, heart rate variability features extract time-domain and frequency-domain indicators; respiratory stability features extract the standard deviation and coefficient of variation of respiratory frequency; and body movement features extract the number of body movements, mean body movement intensity, and maximum body movement intensity. For environmental correlation features, extracting the difference between ambient temperature and bed surface temperature, the rate of change of ambient temperature, and the correlation between ambient humidity and body surface temperature.

[0046] This step transforms the raw collected data into structured features more suitable for modeling and analysis, providing fundamental support for subsequent health status assessment, sleep stage analysis, and target temperature generation. All extracted features undergo normalization to eliminate the influence of dimensions and ensure feature comparability.

[0047] Step S107: Construct a health status assessment model and output the target temperature. Based on the aforementioned feature information, the system constructs or invokes a health status assessment model to analyze the user's physiological state during sleep and outputs corresponding assessment results. The health status assessment model adopts a fusion architecture of a "basic model + personalized model," jointly modeling based on a sleep-temperature big data basic model and the current user's individual data. The outputs of the two models are then fused using a weighted fusion method, with the fusion weights dynamically adjusted according to the user's usage time, thereby achieving a health assessment that balances general patterns with individual differences.

[0048] Furthermore, the health status assessment model not only outputs health status analysis results during sleep, but also, by combining the current sleep stage, vital signs, body surface temperature change patterns, and environmental changes, outputs the target temperature for the current or next stage. The target temperature can be a single-area target temperature or a target temperature for multiple zones, thus linking the health status assessment results with the temperature control strategy. The target temperature generation employs a multi-objective optimization algorithm, with "body surface temperature conforming to the stage standard value, stable vital signs, and environmental adaptation" as optimization objectives, and calculates the optimal target temperature by combining feature weights.

[0049] The health status assessment model can support phased modeling and assessment of different sleep stages. For example, it can establish corresponding analysis logic and feature weights for the sleep onset period, light sleep period, deep sleep period and REM sleep period, and determine the weight of each feature in different stages through weight analysis method, thereby improving the precision and personalization of phased assessment.

[0050] Step S108: Perform target temperature control and record user fine-tuning feedback. The control device performs target temperature control on corresponding zones of the smart bedding based on the target temperature. The target temperature control may include heating, cooling, constant temperature maintenance, or zone-specific temperature control. For different zones, the system can execute corresponding temperature control actions to achieve independent or coordinated temperature control across multiple areas. During temperature control execution, the actual temperature of the bed surface is collected in real time, and a PID control algorithm is used to perform closed-loop adjustment of the temperature control device to ensure that the deviation between the actual temperature and the target temperature is controlled within ±0.5℃.

[0051] During execution, the system records the output target temperature, the actual executed temperature, and user feedback on fine-tuning of the target temperature. This feedback may include user actions such as adjusting the target temperature (upwards, downwards, switching modes, or terminating control) via the app, bedding temperature control buttons, or other interactive methods. This feedback reflects the user's actual acceptance of the model's output target temperature and their individual preferences. The feedback is recorded using an encoded method, with the deviation between the upward / downward adjustment and the target temperature serving as a preference feature, stored in the personalized model database.

[0052] Step S109: Update the personalized model The system updates the personalized sleep temperature health model based on the user's fine-tuning instructions based on the target temperature, the results of the current sleep health assessment, and the long-term accumulated sleep-temperature-health correlation data, so as to fine-tune and iteratively optimize the model parameters.

[0053] The update process may include preference parameter correction, temperature generation weight adjustment, phased temperature curve revision, feature association strength update, and user profile remodeling. The specific update logic is as follows: using the temperature deviation corresponding to the user's fine-tuning command as a label, and combining it with the feature data of the current sleep session, the personalized model is incrementally trained, adjusting the weights of temperature features and vital sign features in the model. Simultaneously, based on the current sleep health assessment results, the deviation of temperature standard values ​​for different sleep stages is corrected, and the personalized temperature curve is revised. The user profile is remodeled every 7 days based on recent sleep data, updating the basic feature parameters. By long-term recording of user feedback behavior and sleep results at the target temperature given by the sleep-temperature model, the system can gradually form a more accurate, personalized, and continuously iteratively improved user-specific model, thus achieving a more accurate result with continued use.

[0054] Step S110: Conduct a comprehensive sleep health assessment. After a user completes a full night's sleep, the system outputs a comprehensive health status assessment result based on a health status evaluation model. This assessment result may include sleep duration, sleep temperature, sleep recovery status, and other evaluation indicators reflecting sleep quality and physiological state. The comprehensive evaluation uses a weighted summation method, with the weight of each indicator scientifically determined based on the correlation between each parameter and sleep health. Different health evaluation levels are assigned based on the comprehensive score, achieving a quantitative assessment of sleep quality.

[0055] Furthermore, the system can comprehensively analyze a user's sleep throughout the night by combining multi-point changes in body surface temperature, heart rate, respiratory stability, sleep stage distribution, and environmental changes to form a more comprehensive assessment of sleep health. This comprehensive evaluation not only reflects the user's sleep state but also provides a basis for subsequent health insights and model optimization. During the evaluation process, anomaly detection algorithms are used to detect abnormal physiological signals during sleep, marking abnormal periods and potential health risks to support health insights.

[0056] Step S111: Output a comprehensive sleep report and personalized health insights. The system outputs a comprehensive sleep report based on the comprehensive sleep health assessment results and generates corresponding personalized health insights. The comprehensive sleep report may include sleep duration, sleep stage distribution, sleep temperature changes, sleep recovery status, health status assessment results, and other analyses related to sleep quality.

[0057] The personalized health insights can be extended to various health scenarios based on the data characteristics of different users, such as metabolism estimation, stress recovery assessment, menstrual cycle prediction, or other health analysis scenarios related to sleep and body temperature changes. Specifically, metabolism estimation is based on the amplitude of body surface temperature changes and mean heart rate, with adjustments made using the basal metabolic rate formula. Stress recovery assessment is based on the frequency domain index of heart rate variability and the amplitude of body surface temperature fluctuations, using a logistic regression model for determination. Menstrual cycle prediction is based on long-term body surface temperature data, using an LSTM model to capture temperature change cycles and predict the timing of the next menstrual period. By outputting comprehensive sleep reports and personalized health insights, the application value and scalability of this invention in the field of health management can be enhanced.

[0058] In this embodiment, the multi-point body surface temperature acquisition, vital sign acquisition, environmental data acquisition, health status assessment, target temperature output, user fine-tuning feedback recording, and model update process constitute a closed-loop execution mechanism. That is, the system continuously records the user's sleep-temperature-health closed-loop logic during long-term use and continuously optimizes the model output results. Compared with traditional static rule control or single-assessment schemes, this embodiment not only achieves a more accurate assessment of sleep health status but also continuously improves the ability to generate personalized target temperatures and provide health insights over long-term use. During the closed-loop process, all data is stored in a time-series database, supporting data backtracking and model traceability to ensure the traceability of model optimization.

[0059] Secondly, this invention provides a sleep health assessment and target temperature generation device based on multi-point body surface temperature and vital sign data during sleep. The device comprises: a basic model parameter input module for acquiring user basic information and establishing a basic model parameter set based on that information; a sleep stage temperature standard value input module for determining the corresponding temperature standard value or target temperature range for the user at different sleep stages; a third-party auxiliary data access module for accessing third-party auxiliary data related to the user's sleep state; a long-term learning model establishment module for establishing a long-term learning model including at least a sleep habit health model and a personalized sleep temperature health model; a state recognition and multi-dimensional data acquisition module for recognizing the user's state during sleep and acquiring multi-dimensional sleep physiological data in real time, including temperature data from multiple body contact areas, vital sign data, environmental data, the user's in-bed state, and the user's current sleep state; and data preprocessing and feature extraction. The system comprises the following modules: a data acquisition module for preprocessing and feature extraction of the multidimensional sleep physiological data to obtain structured feature information; a health status assessment and target temperature output module for constructing or calling a health status assessment model based on the structured feature information, outputting the health status assessment result and the target temperature for the current or next stage; a target temperature control and user fine-tuning feedback recording module for implementing target temperature control on the corresponding zones of the smart bedding according to the target temperature, and recording user fine-tuning feedback information; a personalized model update module for updating the personalized sleep temperature health model based on the user fine-tuning feedback information, the current sleep health assessment result, and long-term accumulated sleep-temperature-health correlation data; a comprehensive sleep health evaluation module for outputting a comprehensive sleep health evaluation result after the user completes a full night's sleep; and a comprehensive sleep report and personalized health insight information output module for outputting a comprehensive sleep report and personalized health insight information based on the comprehensive sleep health evaluation result.

[0060] In the several embodiments provided in this application, it should be understood that the disclosed health status assessment system, apparatus, and method based on multi-point body surface temperature and vital sign data during sleep can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, other division methods can be used according to the technical requirements of the present invention. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed can be achieved through indirect coupling or communication connection of some interfaces, devices, or units, and can be electrical, mechanical, or other suitable forms, adapting to the data transmission and command interaction requirements between the intelligent bedding system and the sensing module and control device in this invention.

[0061] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Based on the actual functional requirements of this invention, such as health status assessment, data acquisition, and temperature control execution, some or all of the units can be selected to achieve the purpose of this embodiment, adapting to the modular deployment requirements of intelligent bedding systems.

[0062] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit, adapting to the flexible deployment requirements of different functions such as model calculation, data processing, and temperature control in the present invention.

[0063] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the health status assessment method based on multi-point body surface temperature and vital sign data during sleep as described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0064] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in this application should all be included within the protection scope of this invention.

Claims

1. A method for sleep health assessment and target temperature generation based on multi-point body surface temperature and vital sign data during sleep, characterized in that, The process includes the following steps: Step S101: Input basic model parameters, obtain basic user information, and establish a basic model parameter set based on the basic user information; Step S102: Input standard values ​​for sleep stage temperature, and determine the standard temperature values ​​or target temperature ranges corresponding to different sleep stages for the user; Step S103: Access third-party auxiliary data, and access third-party auxiliary data related to the user's sleep state; Step S104: Establish a long-term learning model, and establish a long-term learning model that includes at least a sleep habit health model and a personalized sleep temperature health model; Step S105: Perform state recognition and multi-dimensional data collection, and during the user's sleep process, identify the user's state and collect multi-dimensional sleep physiological data in real time. The multi-dimensional sleep physiological data includes temperature data of multiple body contact areas, vital sign data, environmental data, user's bed-in state, and user's current sleep state; Step S106: Perform data preprocessing and feature extraction, and perform preprocessing and feature extraction on the multi-dimensional sleep physiological data to obtain structured feature information; Step S107: Construct a health status assessment model and output the target temperature. Based on the structured feature information, construct or call the health status assessment model, and output the health status assessment result and the target temperature for the current or next stage. Step S108: Execute target temperature control and record user fine-tuning feedback. Execute target temperature control on the corresponding zones of the smart bedding according to the target temperature, and record user fine-tuning feedback information. Step S109: Update the personalized model. Update the personalized sleep temperature health model based on the user fine-tuning feedback information, the current sleep health assessment result, and long-term accumulated sleep-temperature-health correlation data. Step S110: Perform a comprehensive sleep health evaluation. After the user completes a full night's sleep, output the comprehensive sleep health evaluation result. Step S111: Output a comprehensive sleep report and personalized health insight information. Output a comprehensive sleep report and personalized health insight information based on the comprehensive sleep health evaluation result.

2. The method according to claim 1, characterized in that, In step S101, the user's basic information includes one or more of the following: gender, age, BMI, height, weight, region, physical characteristics, and sleep preferences. The personalized model initialization includes: performing feature encoding on the categorical data in the user's basic information, normalizing the continuous data, and inputting the processed data into the basic model to construct the initial user profile of the current user.

3. The method according to claim 1, characterized in that, In step S102, the standard temperature value or target temperature range corresponding to different sleep stages is determined based on at least one of user preset, historical learning results, basic model recommended values, or output results of similar population models; in step S103, the third-party auxiliary data includes exercise data, female menstrual cycle data, daily activity data, and other auxiliary data related to sleep state.

4. The method according to claim 1, characterized in that, In step S104, the sleep habit health model is used to record the correlation between temperature, vital signs and health outcomes of multiple users during sleep, and the personalized sleep temperature health model is used to record the temperature control preferences, body temperature response patterns, health assessment results and feedback information of specific users; the long-term learning model is updated using an online learning method.

5. The method according to claim 1, characterized in that, In step S105, the temperature data of the multiple body contact areas are collected through the bed surface partitioning method of the smart bedding; the vital signs data include one or more of heart rate, respiratory rate, heart rate variability, body movement data, and pulse data; the environmental data includes one or more of ambient temperature, bed surface temperature, ambient humidity, air quality, and day / night information; the user's bed status is identified through pressure distribution detection, and the user's current sleep status is identified based on heart rate, respiratory rate, and body movement intensity.

6. The method according to claim 1, characterized in that, In step S106, the preprocessing includes time synchronization processing, outlier removal, noise suppression, missing value correction, and sleep stage segmentation; the feature extraction includes multi-point body surface temperature change features, regional temperature gradient features, heart rate variability features, respiratory stability features, body movement features, and environmental change correlation features.

7. The method according to claim 1, characterized in that, In step S107, the health status assessment model adopts an architecture that integrates a basic model and a personalized model, and performs joint modeling based on the sleep temperature basic model and the current user's individual data; the target temperature is generated according to the current sleep stage, vital signs, body surface temperature change pattern and environmental changes, and the target temperature is a single area target temperature or a target temperature of multiple zones.

8. The method according to claim 1, characterized in that, In step S108, the target temperature control includes heating, cooling, constant temperature maintenance, or zoned differentiated temperature control; during the execution of the target temperature control, the actual temperature of the bed surface is collected in real time, and closed-loop adjustment is performed based on the actual temperature of the bed surface; the user fine-tuning feedback information includes user operation information for adjusting the target temperature upwards, downwards, switching modes, or terminating control.

9. The method according to claim 1, characterized in that, In step S109, the update of the personalized sleep temperature health model includes one or more of the following: preference parameter correction, temperature generation weight adjustment, phased temperature curve revision, feature association strength update, and user individual profile remodeling; the update is based on the temperature deviation corresponding to the user's fine-tuning instruction, the current sleep feature data, and the current sleep health assessment results.

10. The method according to claim 1, characterized in that, In step S110, the comprehensive sleep health evaluation results include sleep duration, sleep temperature performance, sleep recovery status, and evaluation indicators reflecting sleep quality and physiological state; in step S111, the comprehensive sleep report includes sleep duration, sleep stage distribution, sleep temperature changes, sleep recovery status, and health status assessment results; the personalized health insight information includes one or more of the following: metabolic estimation, stress recovery assessment, menstrual cycle prediction, and other health analysis results related to sleep and body temperature changes.

11. A sleep health assessment and target temperature generation device based on multi-point body surface temperature and vital sign data during sleep, characterized in that, include: The basic model parameter input module is used to obtain basic user information and establish a basic model parameter set based on the basic user information; the sleep stage temperature standard value input module is used to determine the corresponding temperature standard value or target temperature range for the user in different sleep stages. The third-party auxiliary data access module is used to access third-party auxiliary data related to the user's sleep status; The long-term learning model building module is used to build a long-term learning model that includes at least a sleep habit health model and a personalized sleep temperature health model. The status recognition and multi-dimensional data acquisition module is used to identify the user's status during sleep and collect multi-dimensional sleep physiological data in real time. The multi-dimensional sleep physiological data includes temperature data of multiple body contact areas, vital sign data, environmental data, user's status in bed, and user's current sleep status. The data preprocessing and feature extraction module is used to preprocess and extract features from the multidimensional sleep physiological data to obtain structured feature information. The health status assessment and target temperature output module is used to construct or call a health status assessment model based on the structured feature information, and output the health status assessment results and the target temperature of the current stage or the next stage. The target temperature control and user fine-tuning feedback recording module is used to perform target temperature control on the corresponding zones of the smart bedding according to the target temperature and record user fine-tuning feedback information; the personalized model update module is used to update the personalized sleep temperature health model according to the user fine-tuning feedback information, the current sleep health assessment results, and the long-term accumulated sleep-temperature-health correlation data. The comprehensive sleep health assessment module is used to output a comprehensive sleep health assessment result after the user has completed a full night's sleep; The module for outputting comprehensive sleep reports and personalized health insights is used to output comprehensive sleep reports and personalized health insights based on the comprehensive sleep health evaluation results.

12. An electronic device, comprising: One or more processors; A storage device on which one or more programs are stored; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-10.

13. A computer-readable medium having a computer program stored thereon, wherein, When the program is executed by the processor, it implements the method as described in any one of claims 1-10.

14. A computer program product comprising a computer program, wherein, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-10.

Images