Sleep quality analysis method and system for unmanned nursing

Abstract

The invention discloses an unattended sleep quality analysis method and system, and the method comprises the steps: obtaining sleep related signals from the surrounding environment of a user through a non-contact sensor technology, carrying out the real-time capturing of body movement frequency characteristics and breathing rhythm characteristics, and obtaining a preliminary sleep behavior data record; according to the initial sleep behavior data record, filtering the body movement frequency signal and the breathing rhythm signal by adopting a signal denoising processing method, eliminating environmental noise interference, and determining an optimized sleep signal data set; aiming at the optimized sleep signal data set, performing feature extraction by utilizing a pre-established sleep stage classification model, analyzing regular changes of a body movement mode and a breathing law, and judging a sleep stage distribution condition of the user; according to the sleep stage distribution condition, the sleep quality of the user is graded and evaluated in combination with the preset threshold range, and accurate sleep stage judgment and sleep quality evaluation can be achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of health management, and particularly discloses a sleep quality analysis method and system for unattended care. BACKGROUND

[0002] Sleep quality analysis, as an important field of health management, has an undeniable value in improving individual quality of life and preventing chronic diseases. With the acceleration of life pace and the enhancement of health awareness, more and more people begin to pay attention to their sleep conditions, and related monitoring technologies have become a research hotspot.

[0003] However, the current sleep monitoring methods still have significant limitations in practical application. Traditional monitoring methods often rely on specific environments and devices, requiring users to receive detection in fixed locations. This approach not only limits the flexibility of monitoring, but also easily disturbs the natural sleep state of users due to the complexity of device wearing and the unfamiliarity of the environment, thereby affecting the authenticity of the data. In addition, existing methods often require a large amount of manual intervention in data processing, which is inefficient and difficult to meet the needs of large-scale and routine monitoring. Under this background, the core challenges in this field gradually emerge.

[0004] The primary problem is how to break away from the dependence on specific environments and achieve a more free and natural monitoring method, which directly relates to user experience and data reliability. This problem further extends to how to obtain accurate sleep data without affecting user comfort, especially in scenarios where users do not need to wear cumbersome devices or change their daily habits. These two factors are closely related, with the former determining the universality of the monitoring scene and the latter affecting the acceptability of the technology in practical application. If the problems of environmental dependence and comfort cannot be solved, sleep monitoring technology will be difficult to truly integrate into users' daily lives.

[0005] Therefore, how to achieve efficient and accurate sleep data collection without specific environmental restrictions and affecting user comfort has become a key problem that needs to be overcome in this research. SUMMARY

[0006] The present application provides a sleep quality analysis method and system for unattended care, aiming to solve at least one of the defects in the prior art.

[0007] One aspect of the present application relates to a sleep quality analysis method for unattended care, comprising the following steps: Obtaining sleep-related signals from the user's surrounding environment through non-contact sensor technology, capturing body movement frequency characteristics and breathing rhythm characteristics in real time to obtain preliminary sleep behavior data records; According to the preliminary sleep behavior data record, the signal denoising processing method is adopted to filter the body motion frequency signal and the breathing rhythm signal, the environmental noise interference is eliminated, and the optimized sleep signal data set is determined; For the optimized sleep signal data set, the pre-established sleep stage classification model is used for feature extraction, the regularity change of body motion mode and breathing regularity is analyzed, and the sleep stage distribution of the user is judged; According to the sleep stage distribution, the sleep quality of the user is graded and evaluated in combination with the preset threshold range, the duration of each stage and the conversion frequency are obtained, and the sleep quality evaluation result is obtained; For the sleep quality evaluation result, the sleep mode of the user is tracked for a long time through the time series analysis method, the periodic change of the sleep stage distribution is identified, and the sleep habit characteristics of the user are determined; According to the sleep habit characteristics, the signal acquisition parameters of the non-contact sensor are dynamically optimized by using the adaptive adjustment mechanism, and the monitoring configuration scheme conforming to the daily habits of the user is obtained; For the monitoring configuration scheme, the sleep data of the user is uploaded to the remote server for storage and backup through the cloud data synchronization technology, and the real-time updated sleep archive record is obtained; According to the real-time updated sleep archive record, the sleep quality evaluation report of the user is generated by using the data visualization technology, the sleep stage distribution and the sleep quality evaluation result are graphically displayed, and the sleep improvement direction that can be referred by the user is determined.

[0008] Further, according to the preliminary sleep behavior data record, the signal denoising processing method is adopted to filter the body motion frequency signal and the breathing rhythm signal, the environmental noise interference is eliminated, and the optimized sleep signal data set is determined. According to the preliminary sleep behavior data record, the original sleep signal containing body motion frequency and breathing rhythm is obtained, the preliminary cleaning sleep signal set is obtained by using the preset threshold to preliminarily screen the environmental noise interference in the original sleep signal; For the preliminary cleaning sleep signal set, the signal filtering tool is used to denoise the body motion frequency signal and the breathing rhythm signal, if the environmental noise interference exceeds the preset threshold, the noise component is separated by using the frequency domain conversion tool, and the denoised sleep signal combination is obtained; According to the denoised sleep signal combination, the time domain analysis tool is used to adjust the data accuracy of the body motion frequency characteristics and the breathing rhythm characteristics, whether there is abnormal fluctuation is judged by comparing the pre-established normal range value, and the fine-tuned sleep signal data set is determined; For the fine-tuned sleep signal dataset, the body motion frequency and breathing rhythm are classified and stored by the data integration tool. According to the real-time recording requirements, the data characteristics related to behavior analysis are obtained, and the final optimized sleep signal dataset is obtained.

[0009] Further, according to the preliminary sleep behavior data record, the original sleep signal containing body motion frequency and breathing rhythm is obtained. For the environmental noise interference in the original sleep signal, a preset threshold is used for preliminary screening to obtain a preliminary cleaned sleep signal set. In the step of obtaining the original sleep signal:

[0010] Among them, represents the original sleep signal at time , represents the body motion frequency component, represents the breathing rhythm component, represents the environmental noise interference component; The determination criterion for environmental noise screening is:

[0011] Among them, represents the preset noise screening threshold, represents the mean of the sleep signal, represents the standard deviation of the sleep signal, represents the threshold coefficient; The judgment logic of signal cleaning is:

[0012] Among them, represents the sleep signal after preliminary cleaning, represents the original sleep signal, represents the filtering threshold.

[0013] Further, for the optimized sleep signal dataset, a pre-established sleep stage classification model is used for feature extraction to analyze the regularity of body motion pattern and breathing regularity, and to determine the sleep stage distribution of the user. The steps include: According to the optimized sleep signal dataset, for the body motion pattern and breathing regularity therein, a preset feature extraction tool is used for data decomposition to obtain a plurality of characteristic values related to sleep stage classification, and a preliminary extracted feature combination is obtained. For the preliminary extracted feature combination, a data integration tool is used to classify and process a plurality of characteristic values. In the classification, a pre-established stage division standard is combined. If the characteristic value meets the regularity analysis condition of a certain stage, it is classified into the corresponding category, and a classified feature set is determined. According to the classified feature set, for the mode change and behavior association, a time series comparison tool is used for dynamic matching to obtain the sleep stage distribution judgment basis corresponding to the sleep signal, and a preliminary mapping result of the sleep stage distribution is obtained; For the preliminary mapping result of the sleep stage distribution, a data verification tool is used to perform secondary confirmation on the matching degree of the body movement mode and the breathing rule, and if the matching degree exceeds a preset threshold, the corresponding relationship with the sleep stage classification is confirmed, and the final sleep stage distribution data is determined.

[0014] Further, according to the sleep stage distribution, the sleep quality of the user is graded and evaluated in combination with a preset threshold range, the duration and conversion frequency of each stage are obtained, and the steps of obtaining the sleep quality evaluation result include: According to the sleep stage distribution, for the stage duration and conversion frequency, a time series recording tool is used to record the duration and conversion times of each stage in detail, and dynamic change data related to sleep quality is obtained from the sleep stage distribution, and a preliminary time distribution record table is obtained. According to the preliminary time distribution record table, in combination with a preset threshold range, a data classification tool is used to grade and process the duration and conversion times of each stage, and if the duration or conversion times exceeds the corresponding threshold range, it is classified into an abnormal category, and the classified sleep stage distribution data is determined. For the classified sleep stage distribution data, in combination with the abnormal category and behavior mode, a data comparison tool is used to match the pre-established sleep quality standard, and the basis related to the grading evaluation is obtained from the sleep stage distribution, and the sleep quality level distribution of each stage is determined. For the sleep quality level distribution of each stage, in combination with the distribution, a data integration tool is used to summarize and process the sleep quality evaluation result, and a final sleep quality evaluation report is obtained, and the overall sleep quality grading of the user is determined.

[0015] Further, for the sleep quality evaluation result, the sleep mode of the user is tracked for a long time through a time series analysis method, and the steps of identifying the periodic change of the sleep stage distribution and determining the sleep habit characteristics of the user include: According to the sleep quality evaluation result, a time series recording tool is used to collect the sleep stage distribution data of the user for a long time, and the starting and ending time points of each stage are labeled, and continuous data related to the periodic change is obtained from the sleep stage distribution, and a complete sleep stage time distribution record table is obtained. For the sleep stage time distribution record table, a data comparison tool is used to match a pre-established periodic pattern library, and if the sleep stage distribution data matches the time interval of a certain pattern in the periodic pattern library, it is classified as the corresponding periodic characteristic, and the sleep stage periodic distribution characteristics of the user are determined. According to the periodic distribution characteristics of the sleep stages, the data integration tool is used to associate the periodic changes and the duration of each stage, obtain the conversion rule between different stages, judge whether there is regular fluctuation in the sleep pattern, and obtain the long-term change trend.

[0016] For the long-term change trend, the data classification tool is used to classify the fluctuation amplitude in combination with the preset threshold range. If the fluctuation amplitude exceeds the threshold range, it is marked as an abnormal pattern, and the user's sleep habit characteristics are determined.

[0017] Further, according to the sleep habit characteristics, an adaptive adjustment mechanism is used to dynamically optimize the signal acquisition parameters of the non-contact sensor, and the steps of obtaining a monitoring configuration scheme that conforms to the user's daily habits include: According to the sleep habit characteristics, the data recording tool is used to continuously track the user's daily work and rest time period, obtain the behavior data related to sleep, and preliminarily set the acquisition frequency in combination with the environmental adaptation attribute, and obtain an initial monitoring scheme that matches the daily habits; For the initial monitoring scheme, the data comparison tool is used to match the acquisition frequency with the habit tracking database established in advance. If the deviation between the acquisition frequency and the habit data in the habit tracking database exceeds the preset threshold range, the adaptive adjustment mechanism is used to optimize the parameter adjustment of signal acquisition, and the acquisition rule that conforms to the user's characteristics is determined; According to the acquisition rule, the parameter configuration tool is used to dynamically adjust the signal acquisition mode of the non-contact sensor in combination with the environmental adaptation attribute, obtain monitoring data consistent with the daily habits, and determine a monitoring frequency range suitable for the user's current state; For the monitoring frequency range, the data integration tool is used to associate the signal acquisition results with the sleep habit data, and the monitoring configuration is finally calibrated through the data matching attribute, and a personalized monitoring configuration scheme that matches the user's daily habits is obtained.

[0018] Further, for the monitoring configuration scheme, the user's sleep data is uploaded to a remote server for storage and backup through cloud data synchronization technology, and the steps of obtaining a real-time updated sleep archive record include: According to the sleep data, the data sorting tool is used to classify the collected content, set the upload priority for different categories of data, and if the priority exceeds the preset threshold, the uploading mechanism is used to transmit to the remote server, and the classified data set is obtained; For the classified data set, the data synchronization tool is used in combination with the synchronization frequency to detect the transmission process at regular intervals. If the transmission is interrupted, the backup path is used to reconnect, and complete transmission logs are obtained; Based on the complete transmission log, the data uploaded to the remote server is compared using a data verification tool. The consistency is verified by combining the data integrity attributes. Any inconsistencies are adjusted using a repair tool to determine the verified data content. For the verified data content, an access control tool is used in conjunction with user permissions to restrict the access scope, and a copy is generated in the storage space through a backup tool to obtain a real-time updated sleep archive record.

[0019] Furthermore, based on real-time updated sleep profiles, data visualization technology is used to generate sleep quality assessment reports for users, graphically displaying the distribution of sleep stages and sleep quality assessment results. The steps to determine areas for sleep improvement that users can refer to include: Based on sleep records, data processing tools were used to classify the updated content and extract data in layers according to the stage distribution to obtain the classified sleep stage distribution dataset and obtain preliminary organized sleep stage information. Based on the preliminary sleep stage information, an assessment standard tool is used to calculate the scores of the sleep quality assessment content. If the score is lower than the preset threshold, the parameters are adjusted and the calculation is repeated to determine the sleep quality assessment result that meets the standard. Based on the sleep quality assessment results, a graphical display tool is used in combination with the display format to visualize the stage distribution and assessment data, thereby generating an intuitive sleep quality assessment report and obtaining graphical content for presentation. For graphical content, a directional guidance tool is used to match improvement directions. If the matching result matches the low score in the sleep quality assessment report, relevant suggestions are extracted from a pre-established suggestion library to determine the sleep improvement directions that the user can refer to.

[0020] Another aspect of the present invention relates to a sleep quality analysis system for unattended sleep, for performing the above-described sleep quality analysis method for unattended sleep, comprising: The sleep behavior data recording and acquisition module is used to acquire sleep-related signals from the user's surrounding environment through non-contact sensor technology, and to capture body movement frequency characteristics and breathing rhythm characteristics in real time to obtain preliminary sleep behavior data records. The sleep signal dataset determination module is used to filter body motion frequency signals and respiratory rhythm signals based on preliminary sleep behavior data records, using signal denoising processing methods to remove environmental noise interference and determine an optimized sleep signal dataset. The sleep stage distribution judgment module is used to extract features from an optimized sleep signal dataset using a pre-established sleep stage classification model, analyze the regular changes in body movement patterns and breathing patterns, and judge the user's sleep stage distribution. The sleep quality evaluation result acquisition module is configured to grade and evaluate the sleep quality of the user according to the sleep stage distribution and in combination with a preset threshold range, acquire the duration of each stage and the conversion frequency, and obtain the sleep quality evaluation result; The sleep habit feature determination module is configured to track the sleep mode of the user for a long time, identify the periodic change of the sleep stage distribution, and determine the sleep habit feature of the user by using a time series analysis method according to the sleep quality evaluation result; The monitoring configuration scheme acquisition module is configured to dynamically optimize the signal acquisition parameters of the non-contact sensor by using an adaptive adjustment mechanism according to the sleep habit feature, and acquire a monitoring configuration scheme that conforms to the daily habits of the user; The sleep record acquisition module is configured to upload the sleep data of the user to a remote server for storage and backup by using a cloud data synchronization technology according to the monitoring configuration scheme, and obtain a real-time updated sleep record; The sleep improvement direction determination module is configured to generate a sleep quality evaluation report of the user by using a data visualization technology according to the real-time updated sleep record, perform graphical display on the sleep stage distribution and the sleep quality evaluation result, and determine the sleep improvement direction that can be referred to by the user.

[0021] The present application has the following beneficial effects: The application provides a sleep quality analysis method and system for unattended care, which obtains sleep-related signals from the user's surrounding environment through non-contact sensor technology, captures body movement frequency characteristics and breathing rhythm characteristics in real time to obtain preliminary sleep behavior data records; according to the preliminary sleep behavior data records, a signal denoising processing method is used to filter the body movement frequency signal and the breathing rhythm signal, eliminate environmental noise interference, and determine an optimized sleep signal data set; for the optimized sleep signal data set, a pre-established sleep stage classification model is used for feature extraction, analysis of the regularity of body movement patterns and breathing regularity, and judgment of the sleep stage distribution of the user; according to the sleep stage distribution, the sleep quality of the user is graded and evaluated in combination with a pre-set threshold range, the duration of each stage and the conversion frequency are obtained, and a sleep quality evaluation result is obtained; for the sleep quality evaluation result, a time series analysis method is used to track the sleep pattern of the user for a long time, identify the periodic changes of the sleep stage distribution, and determine the sleep habit characteristics of the user; according to the sleep habit characteristics, an adaptive adjustment mechanism is used to dynamically optimize the signal acquisition parameters of the non-contact sensor, and a monitoring configuration scheme conforming to the daily habits of the user is obtained; for the monitoring configuration scheme, the sleep data of the user is uploaded to a remote server for storage and backup through cloud data synchronization technology, and a real-time updated sleep archive record is obtained; according to the real-time updated sleep archive record, a sleep quality evaluation report of the user is generated by using data visualization technology, the sleep stage distribution and the sleep quality evaluation result are graphically displayed, and the sleep improvement direction that can be referred by the user is determined. The sleep quality analysis method and system for unattended care have the following beneficial effects: I. Improve monitoring convenience and comfort: non-contact sensor technology is used, without the need for the user to wear any equipment, avoiding the bondage and discomfort caused by contact monitoring, not interfering with the normal sleep of the user, especially suitable for natural capture of sleep state in unattended care scenarios, improving user acceptance and monitoring continuity.

[0022] II. Ensure the accuracy and reliability of sleep data: the body movement frequency signal and the breathing rhythm signal are filtered by a signal denoising processing method, effectively eliminating environmental noise interference, obtaining an optimized sleep signal data set, providing a high-quality data basis for subsequent sleep stage analysis and sleep quality evaluation, and reducing analysis errors caused by noise.

[0023] III. Accurate sleep stage judgment and sleep quality evaluation: the pre-established sleep stage classification model is used for feature extraction of the optimized signal, which can accurately analyze the regularity of body movement patterns and breathing regularity, and accurately judge the sleep stage distribution of the user. Combined with the pre-set threshold for grading evaluation, the duration of each stage and the conversion frequency can be clearly presented, so that the user can fully understand the sleep quality status.

[0024] Four, assist long-term sleep management and habit formation: through time series analysis of user sleep patterns, identify periodic changes in sleep stage distribution, determine sleep habit characteristics. Based on this, adaptive adjustment mechanism is used to optimize the signal acquisition parameters of non-contact sensor, form a monitoring configuration scheme that conforms to the user's daily habits, which helps users to improve sleep habits and improve long-term sleep quality.

[0025] Five, ensure data security and accessibility: sleep data is uploaded to remote server storage and backup through cloud data synchronization technology, forming a real-time updated sleep archive record, which not only ensures the security of data, but also makes it convenient for users to check at any time. At the same time, data visualization technology is used to generate graphical sleep quality evaluation report, which intuitively displays sleep stage distribution and sleep quality evaluation results, and provides clear and understandable sleep improvement direction guidance for users.

[0026] Six, suitable for a variety of unattended scenarios: the system can complete a series of processes such as sleep monitoring, analysis and evaluation without human intervention, and is particularly suitable for scenarios such as elderly people, infants, hospitalized patients and other unattended scenarios. It can provide timely and accurate sleep information for relevant personnel (such as children, parents, medical staff) to better monitor the health status of the monitored person. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 is a flowchart of an embodiment of the sleep quality analysis method for unattended scenarios of the present application; Figure 2 is a functional block diagram of an embodiment of the sleep quality analysis system for unattended scenarios of the present application.

[0028] REFERENCE NUMERALS 10, sleep behavior data record acquisition module; 20, sleep signal data set determination module; 30, sleep stage distribution condition determination module; 40, sleep quality evaluation result acquisition module; 50, sleep habit characteristic determination module; 60, monitoring configuration scheme acquisition module; 70, sleep archive record acquisition module; 80, sleep improvement direction determination module. DETAILED DESCRIPTION

[0029] In order to better understand the above technical solutions, the above technical solutions will be described in detail in conjunction with the drawings in the specification and specific embodiments.

[0030] As shown in Figure 1 , the first embodiment of the present application proposes a sleep quality analysis method for unattended scenarios, including the following steps: Step S100, acquire sleep-related signals from the user's surrounding environment through non-contact sensor technology, capture body motion frequency characteristics and breathing rhythm characteristics in real time, and obtain preliminary sleep behavior data records.

[0031] Through non-contact sensor technology, sleep-related signals are collected from the user's surrounding environment without physical contact, and core physiological and behavioral indicators (mainly body motion frequency characteristics and breathing rhythm characteristics) are monitored, captured, and analyzed in real time, and finally the technical process and results of generating preliminary sleep behavior data records are generated.

[0032] Non-contact sensor technology refers to a technology that does not require direct contact with the user's body, and achieves signal acquisition by sensing changes in physical quantities in the surrounding environment (such as air pressure, sound waves, electromagnetic waves, infrared radiation, etc.).

[0033] Step S200, according to the preliminary sleep behavior data records, adopt signal denoising processing method to filter body motion frequency signal and breathing rhythm signal, eliminate environmental noise interference, and determine the optimized sleep signal data set.

[0034] The optimized sleep signal data set is based on the preliminary sleep behavior data records, and through the targeted signal denoising processing method, the body motion frequency signal and the breathing rhythm signal are filtered and interference is eliminated, and finally the body motion and breathing rhythm signal set with better signal quality and environmental noise interference is obtained. The optimized sleep signal data set is an optimization and upgrade of the original preliminary data, and provides a more reliable basis for subsequent sleep analysis (such as sleep staging, abnormal event recognition, etc.).

[0035] Step S300, for the optimized sleep signal data set, use the pre-established sleep stage classification model to extract features, analyze the regularity of body motion pattern and breathing regularity, and determine the sleep stage distribution of the user.

[0036] With the optimized sleep signal data set (body motion and breathing rhythm signal after denoising processing) as input, and with the help of the pre-constructed sleep stage classification model, the key features in the body motion and breathing rhythm signal are extracted, the regularity of the key features in the time dimension (such as frequency, amplitude, periodic fluctuation of rhythm) is analyzed, and finally the analysis process and results of the time distribution and conversion rule of each stage (such as wakefulness, light sleep, deep sleep, rapid eye movement sleep) in the user's sleep process are output.

[0037] Step S400, according to the sleep stage distribution, combined with the pre-set threshold range, the sleep quality of the user is graded and evaluated, the duration of each stage and the conversion frequency are obtained, and the sleep quality evaluation result is obtained.

[0038] With the sleep stage distribution situation (each stage duration, conversion frequency, etc. Data) as the core basis, combined with the pre-set sleep quality evaluation threshold range (such as each stage time length standard, conversion frequency reasonable interval, etc.), the overall sleep quality of the user is graded (such as excellent, good, medium, poor), and the comprehensive evaluation process and results of the specific parameters (duration, conversion times and frequency) of each sleep stage are output synchronously.

[0039] Step S500, for the sleep quality evaluation result, the sleep mode of the user is tracked by time series analysis method for a long time, the periodic change of sleep stage distribution is identified, and the sleep habit characteristics of the user are determined.

[0040] With the sleep quality evaluation results of the user for a period of time (such as weeks, months) as the core data, the sleep mode of the user is tracked for a long time by time series analysis technology, the periodicity of sleep stage distribution (such as deep sleep time, REM sleep proportion, sleep / wake time, etc.) in time dimension (such as weekly, monthly or seasonal fluctuation) is identified, and the stable and personalized sleep behavior mode of the user is refined combined with these regular changes, and finally the analysis process and conclusion of the sleep habit characteristics of the user are determined.

[0041] Step S600, according to the sleep habit characteristics, the signal acquisition parameters of the non-contact sensor are dynamically optimized by using the adaptive adjustment mechanism, and the monitoring configuration scheme conforming to the daily habits of the user is obtained.

[0042] With the sleep habit characteristics of the user as the core basis, the signal acquisition parameters (such as sampling frequency, sensitivity, monitoring area range, etc.) of the non-contact sensor (such as millimeter wave radar, infrared sensor, sound sensor, etc.) are dynamically adjusted in real time or periodically by the pre-set adaptive adjustment mechanism, and finally the process and result of generating the personalized monitoring configuration scheme which adapts to the daily sleep habits of the user and can stably capture high-quality sleep signals are generated.

[0043] Step S700, for the monitoring configuration scheme, the sleep data of the user is uploaded to the remote server for storage and backup by using the cloud data synchronization technology, and the real-time updated sleep archive record is obtained.

[0044] Based on the personalized monitoring configuration scheme of the non-contact sensor, the original sleep data (such as body movement, breathing rhythm signal) and derivative analysis results (such as sleep stage, sleep quality grade score) of the user collected by the sensor are automatically uploaded to the remote server by using the cloud data synchronization technology (such as real-time data transmission protocol, incremental synchronization algorithm, etc.), after storage, integration and backup, the process and result of forming the complete and traceable personal sleep archive record which is dynamically updated with the user's sleep process are formed.

[0045] Step S800, according to the real-time updated sleep record, a sleep quality evaluation report of the user is generated by using data visualization technology, the sleep stage distribution and sleep quality evaluation results are graphically displayed, and the sleep improvement direction that can be referred by the user is determined.

[0046] Based on the cloud real-time updated sleep record, the sleep core indicators (including sleep stage distribution, sleep quality evaluation results, etc.) are converted into intuitive and easy-to-understand graphical reports by using data visualization technology (such as line chart, column chart, heat map, time axis, etc.), and through the analysis of abnormal data or regular characteristics in the report, the process and results of the individual and operable sleep improvement suggestions for the user are refined.

[0047] Further, the sleep quality analysis method for unattended care provided by the embodiment includes: Step S110, sleep-related signals are captured from the user's surrounding environment by a non-contact sensor; the body movement frequency characteristics and breathing rhythm characteristics are collected in real time, the combination of infrared sensors and acoustic sensors is used to obtain the original data of body movement frequency and breathing rhythm when the user is sleeping, and preliminary sleep behavior data records are obtained.

[0048] In the scenario of capturing sleep-related signals by a non-contact sensor, the combination of infrared sensors and acoustic sensors can effectively obtain the body movement frequency and breathing rhythm data of the user.

[0049] The infrared sensor mainly senses body movement by detecting changes in the user's body surface temperature. The principle is that when a person turns over or moves slightly during sleep, the heat distribution between the body and the surrounding environment will change. The sensor captures these changes and converts them into body movement frequency data. Assuming that in a sleep monitoring, the infrared sensor records 5 times of user body movement per minute, combined with time axis analysis, it can be preliminarily judged whether the user is in deep sleep or light sleep stage. This way does not need to contact the user's body, avoiding the discomfort that may be caused by traditional contact-type devices, which helps to improve the user experience.

[0050] The acoustic sensor monitors the breathing rhythm by emitting and receiving ultrasonic signals. Its working principle is to use the propagation characteristics of ultrasonic waves in the air to detect the slight distance changes caused by the user's chest and abdomen, and then calculate the breathing frequency. For example, in an 8-hour sleep cycle, the acoustic sensor records that the user breathes 15 times per minute on average, and the breathing frequency is relatively stable between 2am and 3am, about 12 times per minute, which may indicate that the user is in deep sleep stage.

[0051] Through this non-invasive way, not only the breathing data can be accurately captured, but also the sleep disturbance caused by the device wearing can be avoided, significantly improving the naturalness and accuracy of the monitoring.

[0052] Data acquisition from infrared and acoustic sensors is performed synchronously in real time to create a more comprehensive record of sleep behavior. For example, in a monitoring session, the infrared sensor detects frequent body movements at 1 AM, averaging 8 movements per minute, while the acoustic sensor simultaneously records a sudden increase in respiratory rate to 20 breaths per minute. Cross-validation of these two sets of data can infer that the user may have experienced a brief awakening or dreaming phase. This multi-dimensional data fusion method improves the reliability of sleep analysis, provides a solid basis for subsequent sleep quality assessments, and offers users more scientific suggestions for sleep improvement.

[0053] Infrared sensors are positioned at the head of the bed, covering the user's upper body area to ensure sensitivity in capturing heat changes; while acoustic sensors are placed on the side of the bed, focusing on detecting subtle movements in the chest and abdomen area. This rational spatial distribution reduces data interference and significantly improves acquisition accuracy.

[0054] Furthermore, in the sleep quality analysis method for unattended sleep provided in this embodiment, step S200 includes: Step S210: Based on the preliminary sleep behavior data record, obtain the original sleep signal containing body movement frequency and breathing rhythm. For the environmental noise interference in the original sleep signal, use a preset threshold to perform preliminary screening to obtain a set of sleep signals that have been preliminarily cleaned.

[0055] The original sleep signals are:

[0056] In formula (1), Indicates time The original sleep signals, Represents the frequency component of body motion. Indicates respiratory rhythm components, The environmental noise interference component is represented. Formula (1) describes the basic structure of the original sleep signal, which consists of three parts: body movement frequency, breathing rhythm, and environmental noise.

[0057] The criteria for determining environmental noise screening are as follows:

[0058] In formula (2), This indicates the preset noise filtering threshold. This represents the mean of sleep signals. The standard deviation of sleep signals This represents the threshold coefficient. Formula (2) is used to determine the judgment criteria for environmental noise screening. Signals exceeding this threshold are considered noise interference.

[0059] The judgment logic of signal cleaning is:

[0060] In formula (3), represents the sleep signal after preliminary cleaning, represents the original sleep signal, represents the filtering threshold. Formula (3) describes the judgment logic of signal cleaning. When the amplitude of the original signal does not exceed the filtering threshold, the original signal is retained, and when the threshold is exceeded, the signal is set to zero to remove noise interference.

[0061] In processing the preliminary sleep behavior data record, for the original sleep signal of body motion frequency and breathing rhythm, a preliminary screening is performed through a preset threshold to reduce environmental noise interference. Assuming that in a sleep monitoring scene, the original signal is mixed with low-frequency noise of the air conditioner running in the room, which may affect the accuracy of the body motion frequency. By setting a frequency threshold, such as considering signals below 2Hz as noise and eliminating them, a relatively pure sleep signal set is preliminarily cleaned. This way can effectively reduce irrelevant interference and lay a foundation for subsequent processing.

[0062] Step S220, for the sleep signal set after preliminary cleaning, a signal filtering tool is used to perform noise reduction processing on the body motion frequency signal and the breathing rhythm signal, and if it is detected that the environmental noise interference exceeds the preset threshold, a frequency domain conversion tool is used to separate the noise component, and a denoised sleep signal combination is obtained.

[0063] The sleep signal after filtering processing is:

[0064] In formula (4), represents the sleep signal after filtering processing, represents the original body motion frequency signal and the breathing rhythm signal, represents the frequency response function of the th filter, represents the weight coefficient of the th filter, represents the total number of filters, represents the time variable, represents the frequency variable. Formula (4) describes the process of using multiple filters to perform noise reduction processing on the sleep signal.

[0065] For the pre-processed sleep signal set, it is particularly important to use signal filtering tools to denoise the body motion frequency and respiratory rhythm signals. If environmental noise interference is detected to exceed a preset threshold, such as when the noise intensity reaches more than 30% of the original signal during a certain period, the noise component can be separated using frequency domain transformation tools. For example, if continuous vehicle noise is recorded outside the window while the user is sleeping during a monitoring session, the frequency domain transformation tool can separate this high-frequency noise from the respiratory rhythm signal, resulting in a clearer denoised signal combination. This method helps preserve core data characteristics and avoids noise interference in subsequent analysis.

[0066] Step S230: Based on the noise-reduced sleep signal combination, time-domain analysis tools are used to adjust the data precision for the body movement frequency characteristics and breathing rhythm characteristics. By comparing with the pre-established normal range values, it is determined whether there are abnormal fluctuations, and the fine-tuned sleep signal dataset is determined.

[0067] The body motion frequency signal after precision adjustment is:

[0068] In formula (5), This represents the body motion frequency signal after precision adjustment. This represents the original body motion frequency signal. Indicates time The characteristics of body movement amplitude, This represents the mean amplitude of body movement. The standard deviation of the amplitude of body movement. This represents the accuracy adjustment coefficient. Formula (5) corrects the accuracy of the original signal by the degree of deviation of the body motion amplitude.

[0069] When performing time-domain analysis on the denoised sleep signal combination, precision adjustments are made to body movement frequency and respiratory rhythm characteristics, and abnormal fluctuations are identified by comparing them with pre-established normal range values. Assuming a normal respiratory rate range of 12 to 20 breaths per minute, if a user's respiratory rate suddenly rises to 25 breaths per minute during a monitoring period, and this rise is short-lived, body movement frequency analysis suggests it might be due to a brief awakening. Time-domain analysis tools are used to fine-tune this type of data, removing outliers and obtaining a more realistic sleep signal dataset. This adjustment method improves data reliability and provides a more accurate foundation for subsequent analysis.

[0070] Step S240: For the finely tuned sleep signal dataset, body movement frequency and breathing rhythm are classified and stored using a data integration tool. During storage, data features related to behavioral analysis are obtained according to real-time recording requirements to obtain the final optimized sleep signal dataset.

[0071] The optimized sleep signal dataset is:

[0072] In formula (6), represents the optimized sleep signal dataset, represents the total number of data samples, represents the weight coefficient of the th sample, represents the body movement frequency characteristic value of the th individual, represents the respiratory rhythm data characteristic value of the th individual, represents the classification weight parameter of the body movement frequency, represents the classification weight parameter of the respiratory rhythm data.

[0073] For the fine-tuned sleep signal dataset, when the data integration tool is used to classify and store the body movement frequency and respiratory rhythm, the features related to behavior analysis are extracted according to the real-time recording requirements. Assuming that in an 8-hour sleep monitoring, the system stores the body movement frequency according to the frequency change per hour, and at the same time, the respiratory rhythm data is classified and saved according to the deep sleep and light sleep stages. Through this classification method, the system can quickly extract the behavior characteristics of a specific period, such as the respiratory stable data from 2am to 3am, thereby providing support for sleep quality evaluation. This storage method not only improves the efficiency of data management, but also facilitates subsequent targeted analysis.

[0074] Preferably, the sleep quality analysis method for unattended use provided in the embodiment comprises the following steps: Step S310, according to the optimized sleep signal dataset, the body movement pattern and the breathing rule are decomposed by using the preset feature extraction tool, and a plurality of characteristic values related to sleep stage classification are obtained, and a preliminary extraction characteristic combination is obtained.

[0075] For example, when analyzing based on the optimized sleep signal dataset, the features of the body movement pattern and the breathing rule are extracted by decomposing the data through the preset tool to obtain characteristic values related to sleep stage classification. The body movement pattern is the frequency and amplitude of the user turning over in sleep, and the breathing rule is the number of breaths per minute and the change in depth. Assuming that in a night monitoring, the system records that the user turns over 3 times per hour in a certain period of time, and the breathing frequency is stable at 15 times per minute, through the feature extraction tool, these data can be decomposed into characteristic values related to the light sleep stage, forming a preliminary extraction characteristic combination. This decomposition method helps to convert complex signals into analyzable indicators, providing a basis for subsequent classification.

[0076] Step S320: For the initially extracted feature combination, use a data integration tool to classify multiple feature values. When classifying, combine the pre-established stage division criteria. If a feature value meets the pattern analysis conditions of a certain stage, it is classified into the corresponding category to determine the classified feature set.

[0077] When classifying the initially extracted feature combinations, the system uses pre-established stage division criteria to assign feature values ​​to corresponding categories. For example, if the criteria define a respiratory rate of 12 to 18 breaths per minute and a low body movement frequency as deep sleep, then in a monitoring session, if a data point has a respiratory rate of 14 breaths per minute and a body movement frequency of once per hour, the system will classify it as belonging to the deep sleep feature set. This classification method integrates scattered feature values ​​into meaningful categories, facilitating further analysis of the distribution of sleep structure.

[0078] Step S330: Based on the classified feature set, a time series comparison tool is used to dynamically match the pattern changes and behavioral associations to obtain the basis for judging the sleep stage distribution corresponding to the sleep signal, and obtain the preliminary mapping result of the sleep stage distribution.

[0079] When performing dynamic matching based on the categorized feature set, time-series comparison tools are used to help identify the correlation between body movement patterns and breathing patterns and sleep stages. For example, in an 8-hour sleep monitoring session, the system detects a sudden increase in body movement frequency to 5 times per hour and a fluctuating respiratory rate to 20 breaths per minute between 1 AM and 2 AM. Using time-series comparison tools, this period can be preliminarily determined to correspond to either brief awakenings or light sleep stages, yielding a mapping result of the sleep stage distribution. This dynamic matching method can capture pattern changes during sleep, providing a more detailed basis for stage segmentation.

[0080] Step S340: Based on the preliminary mapping results of sleep stage distribution, a data verification tool is used to confirm the matching degree of body movement patterns and breathing patterns. If the matching degree exceeds the preset threshold, the correspondence between the matching degree and the sleep stage classification is confirmed, and the final sleep stage distribution data is determined.

[0081] The final sleep stage classification results are as follows:

[0082] In formula (7), Indicates the first The final sleep stage classification results for each time period, This indicates the initial mapping of the sleep stage. Indicates the corrected sleep stage. Indicates the first The verification confidence level for each time period. represents a confidence determination threshold.

[0083] When the preliminary mapping result of the sleep stage distribution is confirmed again, the data verification tool evaluates the matching degree to ensure the accuracy of the classification. Assuming that the preset matching degree threshold is 80%, in a monitoring, the body movement frequency feature and the breathing rhythm feature of a certain segment of data match the light sleep stage by 85%, the system will confirm that it is classified as light sleep, and finally determine the sleep stage distribution data. Conversely, if the matching degree is only 60%, the data needs to be reanalyzed. This secondary confirmation mechanism can improve the credibility of the classification and provide reliable support for subsequent sleep quality evaluation. Through the above multi-level processing method, the system can gradually build a structured sleep analysis framework from the original data to the final stage distribution.

[0084] Further, the sleep quality analysis method for unattended care provided by the embodiment comprises the following steps: Step S410, according to the sleep stage distribution, the time series recording tool is used to record the duration of each stage and the number of transitions for the duration of each stage and the transition frequency, and the dynamic change data related to the sleep quality is obtained, and a preliminary time distribution record table is obtained.

[0085] When analyzing the sleep stage distribution, the recording of the duration of each stage and the transition frequency can capture the dynamic changes of the user's night sleep through the time series recording tool. The basic principle of the time series recording tool is to mark the start and end time points of each stage in the sleep process to form continuous time axis data. Specifically, assuming that in a night sleep monitoring of 8 hours, the system records that the deep sleep stage lasts for 2 hours, the light sleep stage lasts for 4 hours, the rapid eye movement sleep stage lasts for 1.5 hours, and the remaining 0.5 hours is in a wakeful state, and the number of stage transitions is 5 times. These data are arranged into a preliminary time distribution record table to lay a foundation for subsequent analysis.

[0086] Step S420, according to the preliminary time distribution record table, combining the preset threshold range, using the data classification tool to classify the duration of each stage and the number of transitions, if the duration or the number of transitions exceeds the corresponding threshold range, it will be classified into an abnormal category, and the classified sleep stage distribution data is determined.

[0087] The abnormal determination result of the number of transitions of each stage is:

[0088] In formula (8), represents the abnormal determination result of the number of transitions of the i-th stage, represents the abnormal determination result of the number of transitions of the i-th stage, ​the number of transitions in a stage, the average of the number of transitions, the standard deviation of the number of transitions, the normalized threshold value for anomaly detection. When the normalized deviation is greater than the threshold value, it is determined to be abnormal.

[0089] For example, when the preliminary time distribution record table is processed in stages, in combination with the preset threshold range, the data classification tool can make abnormal judgments on the duration and the number of transitions. Assuming that the preset normal range of deep sleep duration is 1.5 to 3 hours, and the normal range of the number of transitions is 3 to 6 times, if the deep sleep in a certain monitoring lasts only 1 hour, and the number of transitions is as high as 8 times, the system will classify it into the abnormal category. This classification method helps to quickly screen out key data points that may affect sleep quality.

[0090] Step S430, for the classified sleep stage distribution data, in combination with the abnormal category and the behavior pattern, the data comparison tool is used to match the pre-established sleep quality standard to obtain the basis related to the grading evaluation, and the sleep quality level distribution of each stage is judged.

[0091] The sleep quality level score of a single sleep stage is:

[0092] In formula (9), denotes the sleep quality level score of a single sleep stage, denotes the total number of abnormal behaviors detected in the stage, denotes the weight coefficient of the th abnormal behavior denotes the actual detection value of the th abnormal behavior, denotes the corresponding sleep quality standard threshold value, the function denotes the matching degree calculation function of the abnormal behavior and the standard.

[0093] The comprehensive evaluation value of the sleep quality level distribution is:

[0094] In formula (10), denotes the comprehensive evaluation value of the sleep quality level distribution, denotes the total number of time periods analyzed, denotes the normal behavior pattern intensity in the th time period, denotes the abnormal behavior pattern intensity in the th time period, denotes the contribution coefficient of the normal behavior, A penalty coefficient representing abnormal behavior.

[0095] For example, when matching the sleep quality level based on the classified sleep stage distribution data, the data comparison tool will compare the abnormal category and behavior pattern with the pre-established sleep quality standard. Assuming that the standard stipulates that the sleep quality level is low when the deep sleep duration is less than 1.2 hours and the conversion times exceed 7 times, and in a certain monitoring, if the user's deep sleep duration is 1 hour and the conversion times are 8 times, the system will judge that the sleep quality level of this stage is low. This matching process can evaluate the performance of each stage from multiple dimensions and provide a basis for the overall sleep quality classification.

[0096] Step S440, for the sleep quality level distribution of each stage, the data integration tool is used to summarize and process the sleep quality evaluation results in combination with the distribution, to obtain the final sleep quality evaluation report and determine the overall sleep quality classification of the user.

[0097] When the sleep quality level distribution of each stage is summarized and processed, the data integration tool integrates the sleep quality evaluation results of all stages into the final sleep quality evaluation report. Assuming that in a certain monitoring, the deep sleep quality level is low, and the light sleep and rapid eye movement sleep are medium, the system will comprehensively calculate that the overall sleep quality classification is medium. This summary method can directly reflect the user's sleep condition and facilitate subsequent targeted adjustment. Through the above multi-level analysis and processing, from the stage duration to the final overall sleep quality classification, the system can build a complete evaluation framework and provide valuable reference information for the user.

[0098] Further, the sleep quality analysis method for unattended care provided in the embodiment comprises the following steps: Step S510, according to the sleep quality evaluation result, the time series record tool is used to collect the sleep stage distribution data of the user for a long time, and the starting and ending time points of each stage are labeled to obtain continuous data related to periodic changes and obtain a complete sleep stage time distribution record table.

[0099] When using a time series recording tool to collect the user's sleep stage distribution data for a long time, the portable sleep monitoring device continuously records the changes of each stage of the user's night sleep. The principle of the time series recording tool is to accurately mark the starting and ending time points of each stage in the sleep process, such as deep sleep, light sleep and rapid eye movement sleep, to form a continuous time axis data stream. Assuming that in a 30-day monitoring, the device records that the user's deep sleep starts at 23:00 and ends at 01:00, lasting 2 hours, and the light sleep is concentrated in 01:00-04:00. This long-term collection method can capture the subtle changes of the user's sleep stage, providing detailed data basis for subsequent analysis.

[0100] Step S520, for the sleep stage time distribution record table, using the data comparison tool to match with the pre-established periodic pattern library, if the sleep stage distribution data is consistent with the time interval of a certain pattern in the periodic pattern library, it is classified as the corresponding periodic feature, and the sleep stage periodic distribution feature of the user is determined.

[0101] The sleep periodicity feature of the user is determined by a weighted voting mechanism:

[0102] In formula (11), represents the determined sleep stage periodicity distribution feature category of the user, represents the total number of periodicity feature categories, represents the number of sleep stage indicators, represents the weight of the th sleep stage indicator, represents the time interval of the th sleep stage of the user, represents the th stage standard time interval of the th periodicity feature, is an indicator function, represents the allowed time interval deviation threshold.

[0103] For the matching process of the sleep stage time distribution record table and the periodic pattern library, the automatic analysis is realized by the data comparison tool. Assuming that the periodic pattern library stores a variety of common sleep periodicity features, such as the 90-minute deep sleep and rapid eye movement sleep alternation pattern. In an analysis, if the user's sleep data of a week shows that the alternation time interval of deep sleep and rapid eye movement sleep is stable between 85 and 95 minutes, the system will classify it as close to the standard periodicity feature. This matching method helps to identify the user's sleep regularity and provides a basis for personalized adjustment.

[0104] Step S530: Based on the periodic distribution characteristics of sleep stages, use data integration tools to correlate the periodic changes and duration of each stage, obtain the transition patterns between different stages, determine whether there are regular fluctuations in sleep patterns, and obtain long-term trends.

[0105] For example, when performing correlation processing based on the periodic distribution characteristics of sleep stages, data integration tools combine the periodic changes and duration of each stage for analysis to explore potential patterns in stage transitions. For instance, if long-term monitoring reveals that a user typically enters REM sleep directly after deep sleep, with the transition time consistently around 1:00 AM, the system would determine that their sleep pattern exhibits certain regular fluctuations. This analysis can help reveal the stability of a user's sleep structure, thus providing a reference for improving sleep habits.

[0106] Step S540: For long-term trends, use data classification tools combined with preset threshold ranges to classify the fluctuation amplitude. If the fluctuation amplitude exceeds the threshold range, it is marked as an abnormal pattern to determine the user's sleep habit characteristics.

[0107] The fluctuation range is calculated using the following formula:

[0108] In formula (12), Indicates a point in time The volatility index Indicates the length of the time window. Indicates the first Sleep data values ​​at different times. Indicates the first Sleep data values ​​at any given time. Formula (12) quantifies the degree of fluctuation by calculating the average absolute difference of sleep data at consecutive time points.

[0109] For example, the data classification tool categorizes fluctuations in long-term trends and uses preset threshold ranges to mark anomalies. Assuming normal fluctuation is defined as a change in deep sleep duration of no more than 0.5 hours, if monitoring in a given week shows a user's deep sleep duration suddenly dropping from 2 hours to 1 hour, exceeding the threshold, the system will mark it as an abnormal pattern. This categorization method can quickly identify abnormal situations that may affect sleep quality, providing data support for subsequent interventions. Through this multi-dimensional analysis and processing, the system can extract users' sleep habit characteristics from long-term data, providing valuable guidance for optimizing the sleep environment and adjusting behaviors.

[0110] Furthermore, in the sleep quality analysis method for unattended sleep provided in this embodiment, step S600 includes: Step S610, according to the sleep habit characteristics, the data recording tool is used to continuously track the daily work and rest time period of the user, the sleep-related behavior data is obtained therefrom, and the collection frequency is preliminarily set in combination with the environmental adaptation attribute, so as to obtain an initial monitoring scheme matched with the daily habit.

[0111] The long-term change trend and periodic component of the sleep pattern are described as:

[0112] In formula (13), represents the long-term change trend value of the sleep cycle, represents the linear trend coefficient, represents the baseline level, represents the amplitude of the periodic fluctuation, represents the cycle length, represents the random error term, represents the serial number of the sleep cycle.

[0113] When the data recording tool is used to continuously track the daily work and rest time period of the user, the behavior data related to sleep is obtained by recording each time node of the user from getting up to falling asleep through a portable device. Assuming that the daily habit of a user is to get up at 7:00 in the morning and fall asleep at 11:00 at night, the system will preliminarily set the collection frequency to record key behavior data once an hour, and at the same time, in combination with the environmental adaptation attribute, such as the user being in a quiet indoor environment, the frequency is appropriately reduced to reduce interference, so as to obtain an initial monitoring scheme.

[0114] Step S620, for the initial monitoring scheme, the data comparison tool is used to match the collection frequency with the habit tracking database preliminarily established, if the deviation of the collection frequency from the habit data in the habit tracking database exceeds the preset threshold range, the parameter adjustment of the signal collection is optimized through the self-adaptive adjustment mechanism, and the collection rule conforming to the user characteristics is determined.

[0115] The relative deviation degree of the actual collection frequency from the frequency in the habit tracking database preliminarily established is calculated by the following formula:

[0116] In formula (14), represents the deviation percentage of the collection frequency from the habit data in the habit tracking database, represents the current collection frequency value, represents the habit tracking frequency reference value preliminarily established in the habit tracking database.

[0117] ​For the matching process of the initial monitoring scheme and the habit tracking database, the rationality of the collection frequency is analyzed through the data comparison tool. Assuming that the typical habit data stored in the habit tracking database suggests a collection frequency of every 2 hours for the night rest period, and the user's initial scheme is every hour, the deviation exceeds the preset threshold range, the system will adjust the frequency to every 1.5 hours through the adaptive adjustment mechanism, ensuring that the collection rule is more in line with the user's characteristics. This adjustment can effectively balance data accuracy and user comfort.

[0118] Step S630, according to the collection rule, the signal collection mode of the non-contact sensor is dynamically adjusted by the parameter configuration tool combined with the environmental adaptation attribute, and the monitoring data consistent with the daily habits is obtained, and the monitoring frequency range suitable for the current state of the user is determined.

[0119] When dynamically adjusting the signal collection mode of the non-contact sensor according to the collection rule, the parameter configuration tool is optimized and set in combination with the environmental adaptation attribute. Assuming that the user's night sleep environment is relatively stable, the sensor can be adjusted to a low-power mode, collecting heart rate and body motion frequency every 2 hours, and determining that the monitoring frequency range suitable for the current state is between 1.5 and 2.5 hours. This way can not only obtain consistent monitoring data, but also avoid potential interference caused by frequent collection on the user.

[0120] Step S640, for the monitoring frequency range, the signal collection result is associated with the sleep habit data by using the data integration tool, and the monitoring configuration is finally calibrated through the data matching attribute, to obtain a personalized monitoring configuration scheme consistent with the user's daily habits.

[0121] For the association processing of the monitoring frequency range and the sleep habit data, the signal collection result is compared with the user's daily habits through the data integration tool, and the monitoring configuration is finally calibrated. Assuming that the user's long-term data shows that the period from 11:00 at night to 1:00 in the morning is the key period for falling asleep, the system will preferentially increase the collection frequency of this period to every 30 minutes, while keeping the collection frequency of other periods at 2 hours, forming a personalized monitoring configuration scheme. This calibration can more accurately capture the user's sleep behavior characteristics and provide a reliable basis for subsequent optimization.

[0122] Further, the sleep quality analysis method for unattended care provided by the embodiment comprises the following steps: Step S710, according to the sleep data, the data collection is classified and processed by using the data arrangement tool, and the upload priority is set for different categories of data. If the priority exceeds the preset threshold, the data set after classification is transmitted to the remote server through the upload mechanism.

[0123] The upload priority value of the data is calculated by the following formula:

[0124] In formula (15), represents the upload priority value of the first data, represents the data importance weight factor, represents the importance score of the first data, represents the maximum importance score, represents the time sensitivity weight factor, represents the current timestamp, represents the data generation timestamp, represents the time window length, represents the sleep quality weight factor, represents the sleep quality evaluation value of the first data.

[0125] In processing sleep data, using a data sorting tool to classify the collected content is a key link. For different categories of data, such as heart rate, respiratory rate and body movement information, different upload priorities are set. Suppose the heart rate data is set as high priority because its change directly reflects the user's sleep quality, while the body movement frequency has lower priority. If the priority of the heart rate data exceeds the preset threshold, such as setting the threshold to 80, and the actual score is 85, the system will transmit it to the remote server through the upload mechanism to form a classified data set. This way can ensure that important data is processed and analyzed in time.

[0126] Step S720, for the classified data set, the data synchronization tool is combined with the synchronization frequency to detect the transmission process in time, if the transmission is interrupted, the complete transmission log is obtained through the backup path reconnection.

[0127] For the classified data set, the data synchronization tool is combined with the synchronization frequency to detect the transmission process in time. Suppose the synchronization frequency is set to detect once every 30 minutes, if transmission interruption is found in a certain detection, the system will automatically switch to the backup path to reconnect, and generate a complete transmission log. This mechanism can effectively avoid data loss and ensure the continuity of transmission.

[0128] Step S730, according to the complete transmission log, the data verification tool is used to compare the data uploaded to the remote server, and the consistency is verified combining with the data integrity attribute, the inconsistent part is adjusted through the repair tool, and the verified data content is determined.

[0129] In data verification according to the transmission log, the data verification tool compares the data uploaded to the remote server, verifies the consistency in combination with the data integrity attribute. Assuming that some heart rate data is found to be missing after transmission, the system will adjust the inconsistent part through the repair tool, such as extracting the missing data from the local cache to complete, and finally determine the content of the verified data. This verification process can ensure the accuracy and reliability of the data, and provide a solid foundation for subsequent analysis.

[0130] Step S740, for the verified data content, the permission management tool limits the access range in combination with the user permission, generates a copy in the storage space through the backup tool, and obtains the real-time updated sleep archive record.

[0131] For the verified data content, the permission management tool limits the access range in combination with the user permission, which is an important security measure. Assuming that a user is only authorized to view his own heart rate data, the system will limit his access to other types of data. At the same time, by generating a copy in the storage space through the backup tool, the data is ensured to be updated in real time, forming an archive record. Assuming that the backup frequency is once a day, the system will automatically generate a copy at dawn to prevent accidental data loss. This permission management and backup mechanism can effectively protect user privacy while improving data traceability.

[0132] Further, the sleep quality analysis method for unattended care provided by the embodiment comprises the following steps: Step S810, according to the sleep archive, the data arrangement tool classifies and processes the updated record, and extracts the data in layers according to the stage distribution, to obtain the classified sleep stage distribution data set, and obtains the preliminarily arranged sleep stage information.

[0133] In processing the sleep archive, the data arrangement tool classifies and processes the updated record, which is an important link. For the data of sleep stage distribution, the data is divided into different stages such as light sleep, deep sleep and rapid eye movement sleep through the way of layer extraction. Assuming that the sleep archive of a user shows that the total night length is 8 hours, of which light sleep accounts for 40%, deep sleep accounts for 30%, and rapid eye movement sleep accounts for 30%, the data arrangement tool will extract these information in layers to form the classified sleep stage distribution data set. This classification method can provide clear stage information for subsequent analysis, and facilitate further mining of potential problems.

[0134] Step S820, for the preliminarily arranged sleep stage information, the evaluation standard tool calculates the score of the sleep quality evaluation content, and if the score is lower than the preset threshold, the sleep quality evaluation result meeting the standard is determined by recalculating the adjusted parameters.

[0135] The adjusted evaluation parameter value is:

[0136] In formula (16), represents the adjusted evaluation parameter value, represents the original evaluation parameter value, represents the parameter adjustment coefficient, represents the deviation from the standard value. Formula (16) is used for dynamic adjustment of parameters when the evaluation score is below the threshold.

[0137] The compliance determination of the sleep quality evaluation result is achieved by the following formula:

[0138] In formula (17), represents the compliance determination of the sleep quality evaluation result, represents the current calculated sleep quality evaluation value, represents the preset minimum threshold. When the score reaches or exceeds the threshold, it is determined to meet the standard, otherwise the parameters need to be adjusted again.

[0139] For the preliminarily sorted sleep stage information, the evaluation standard tool will calculate the score of sleep quality. Assuming that the evaluation standard includes sleep continuity, deep sleep proportion, and stage transition frequency, the system will assign weights to each indicator and calculate the total score. If the preset threshold is 75 points, and the sleep quality evaluation result is only 60 points, the system will adjust the parameters, such as increasing the weight of deep sleep proportion, to recalculate the score until the sleep quality evaluation result meets the standard. This dynamic adjustment mechanism helps to more accurately reflect the actual situation of sleep quality.

[0140] Step S830, according to the sleep quality evaluation result, using the graphical display tool to visualize the stage distribution and evaluation data in combination with the display form, and generating an intuitive sleep quality evaluation report, obtaining graphical content for presentation.

[0141] When generating the sleep quality evaluation report, the graphical display tool will visualize the stage distribution and evaluation data. Assuming that the system displays the proportion of each stage duration in a column chart, and presents the trend of sleep quality score change in a line chart, the user can directly see that the low deep sleep time may be the reason for the low total score. Such graphical content can help users quickly understand their own sleep status and provide intuitive basis for subsequent improvement.

[0142] Step S840, for the graphical content, using the direction guide tool to match the improvement direction, if the matching result is consistent with the low score item in the sleep quality evaluation report, then extracting the relevant suggestions from the pre-established suggestion library to determine the sleep improvement direction that the user can refer to.

[0143] The following formula is used to calculate the match between the graphical content and the low-scoring items in the sleep quality assessment report:

[0144] In formula (18), Representing graphical content Low scores in sleep quality assessment report The degree of matching, This represents the total number of matching features. Indicates the first The weight coefficients of each feature, The graphical content is in the first... The values ​​that can be taken on each feature This indicates that the low score in the sleep quality assessment report is in the [number]th category. The values ​​of each feature This represents the similarity function between two feature values.

[0145] The sleep improvement suggestions extracted from the suggestion library are scored as follows:

[0146] In formula (19), This indicates the score of sleep improvement suggestions extracted from the suggestion library. This represents a pre-established set of suggestions. This indicates the total number of suggestions in the suggestion library. Indicate suggestions Priority weights, Indicate suggestions For specific problem types The correlation coefficient.

[0147] The optimal direction for improving sleep is:

[0148] In formula (20), This indicates the optimal direction for improving sleep. Represents the set of all possible directions for improvement. Indicates directions for improvement Sleep quality assessment value, Indicates directions for improvement The feasibility assessment value, Indicates directions for improvement The applicability assessment value for users, , , The weighting coefficients represent the sleep quality, feasibility, and applicability of the improvement direction, respectively.

[0149] For graphical content, the direction guide tool provides the user with an improvement direction through matching processing. Assuming that the sleep quality assessment report shows that the deep sleep proportion is insufficient, the system extracts relevant suggestions from the pre-established suggestion library, such as adjusting the pre-sleep activity or optimizing the sleep environment. Specifically, if the user is used to using electronic devices before sleeping, the suggestions may include reducing the screen time to 1 hour before sleeping, and trying to use a warm light lamp to create a relaxing atmosphere. This matching mechanism can provide personalized guidance for the user.

[0150] See Figure 2The embodiment also provides a sleep quality analysis system for unattended care, which is used for executing the sleep quality analysis method for unattended care, and comprises a sleep behavior data record acquisition module 10, a sleep signal data set determination module 20, a sleep stage distribution condition judgment module 30, a sleep quality evaluation result acquisition module 40, a sleep habit feature determination module 50, a monitoring configuration scheme acquisition module 60, a sleep archive record acquisition module 70 and a sleep improvement direction determination module 80. The sleep behavior data record acquisition module 10 is used for acquiring sleep-related signals from the surrounding environment of a user by a non-contact sensor technology, capturing body movement frequency features and breathing rhythm features in real time, and obtaining preliminary sleep behavior data records. The sleep signal data set determination module 20 is used for filtering body movement frequency signals and breathing rhythm signals by a signal denoising processing method according to the preliminary sleep behavior data records, eliminating environmental noise interference, and determining an optimized sleep signal data set. The sleep stage distribution condition judgment module 30 is used for extracting features by using a pre-established sleep stage classification model for the optimized sleep signal data set, analyzing regular changes of body movement patterns and breathing rules, and judging sleep stage distribution conditions of the user. The sleep quality evaluation result acquisition module 40 is used for grading and evaluating sleep quality of the user according to the sleep stage distribution conditions and in combination with a preset threshold range, obtaining duration and conversion frequency of each stage, and obtaining sleep quality evaluation results. The sleep habit feature determination module 50 is used for long-term tracking of a sleep pattern of the user by a time series analysis method for the sleep quality evaluation results, identifying periodic changes of sleep stage distribution, and determining sleep habit features of the user. The monitoring configuration scheme acquisition module 60 is used for dynamically optimizing signal acquisition parameters of the non-contact sensor by an adaptive adjustment mechanism according to the sleep habit features, and obtaining a monitoring configuration scheme conforming to daily habits of the user. The sleep archive record acquisition module 70 is used for uploading sleep data of the user to a remote server for storage and backup by cloud data synchronization technology for the monitoring configuration scheme, and obtaining real-time updated sleep archive records. The sleep improvement direction determination module 80 is used for generating a sleep quality evaluation report of the user by using data visualization technology according to the real-time updated sleep archive records, graphically displaying sleep stage distribution and sleep quality evaluation results, and determining a sleep improvement direction that can be referred by the user.

[0151] The sleep quality analysis method and system for unattended care have the following beneficial effects. 1. Improve the convenience and comfort of monitoring: Adopt non-contact sensor technology, without the need for users to wear any equipment, avoiding the restraint and discomfort brought by contact monitoring, not interfering with normal sleep, especially suitable for natural capture of sleep state in unattended scenarios, improving user acceptance and monitoring continuity.

[0152] 2. Ensure the accuracy and reliability of sleep data: Through signal denoising processing method to filter body motion frequency signal and respiratory rhythm signal, effectively eliminate environmental noise interference, get optimized sleep signal data set, provide high quality data basis for subsequent sleep stage analysis and sleep quality evaluation, reduce analysis error caused by noise.

[0153] 3. Realize accurate sleep stage judgment and sleep quality evaluation: Use the pre-established sleep stage classification model to extract the features of the optimized signal, accurately analyze the regularity of body motion pattern and breathing rhythm, and accurately judge the sleep stage distribution of the user. Combined with the preset threshold for grading evaluation, the duration and conversion frequency of each stage can be clearly presented, so that the user can fully understand their own sleep quality.

[0154] 4. Help long-term sleep management and habit formation: Through time series analysis to track the user's sleep pattern for a long time, identify the periodic changes of sleep stage distribution, and determine the sleep habit characteristics. Based on this, use adaptive adjustment mechanism to optimize the signal acquisition parameters of non-contact sensor, form a monitoring configuration scheme that meets the user's daily habits, which helps users to improve their sleep habits and improve long-term sleep quality.

[0155] 5. Ensure data security and accessibility: Use cloud data synchronization technology to upload sleep data to remote server storage and backup, form real-time updated sleep archive records, which not only ensures the security of data, but also makes it convenient for users to check at any time. At the same time, use data visualization technology to generate graphical sleep quality evaluation report, intuitively display sleep stage distribution and sleep quality evaluation results, and provide clear and understandable sleep improvement direction guide for users.

[0156] 6. Suitable for various unattended scenarios: This system can complete a series of processes such as sleep monitoring, analysis and evaluation without human intervention, especially suitable for scenarios such as elderly people, infants, hospitalized patients and other unattended scenarios, which can provide timely and accurate sleep information for relevant personnel (such as children, parents, medical staff), so as to better pay attention to the health status of the monitored person.

[0157] While the preferred embodiments of the application have been described, additional variations and modifications can be made to these embodiments by those skilled in the art once they have the benefit of the present disclosure without departing from the spirit and scope of the application. Accordingly, it is intended that the appended claims include all such variations and modifications as fall within the scope of the present application. It is apparent that those skilled in the art can modify and adapt the present application in various ways without departing from the spirit and scope of the present application. It is therefore intended that the present application encompass all such modifications and variations as fall within the scope of the claims and their equivalents.

Claims

1. A method for sleep quality analysis for unattended sleep, characterized in that, The method comprises the following steps: Obtaining sleep-related signals from the user's surrounding environment through non-contact sensor technology, capturing body motion frequency characteristics and breathing rhythm characteristics in real time, and obtaining preliminary sleep behavior data records; According to the preliminary sleep behavior data records, a signal denoising processing method is used to filter the body motion frequency signal and the breathing rhythm signal, eliminate environmental noise interference, and determine an optimized sleep signal data set; For the optimized sleep signal data set, a pre-established sleep stage classification model is used for feature extraction, analysis of the regularity of body motion patterns and breathing regularity, and determination of the sleep stage distribution of the user; According to the sleep stage distribution, the sleep quality of the user is graded and evaluated in combination with a pre-set threshold range, the duration of each stage and the conversion frequency are obtained, and a sleep quality evaluation result is obtained; According to the sleep quality evaluation result, the sleep mode of the user is tracked in the long term through a time series analysis method, the periodic change of the sleep stage distribution is identified, and the sleep habit characteristics of the user are determined; According to the sleep habit characteristics, an adaptive adjustment mechanism is used to dynamically optimize the signal acquisition parameters of the non-contact sensor, and a monitoring configuration scheme conforming to the user's daily habits is obtained; According to the monitoring configuration scheme, the sleep data of the user is uploaded to a remote server for storage and backup through cloud data synchronization technology, and a real-time updated sleep archive record is obtained; According to the real-time updated sleep archive record, a sleep quality evaluation report of the user is generated by using data visualization technology, the sleep stage distribution and the sleep quality evaluation result are graphically displayed, and the sleep improvement direction that the user can refer to is determined.

2. The method for sleep quality analysis without a guardian according to claim 1, wherein, The step of determining the optimized sleep signal data set according to the preliminary sleep behavior data records and using the signal denoising processing method to filter the body motion frequency signal and the breathing rhythm signal and eliminate environmental noise interference comprises: According to the preliminary sleep behavior data records, the original sleep signals containing body motion frequency and breathing rhythm are obtained, the environmental noise interference in the original sleep signals is preliminarily screened by using a pre-set threshold, and a sleep signal set that has been preliminarily cleaned is obtained; For the sleep signal set that has been preliminarily cleaned, a signal filtering tool is used to denoise the body motion frequency signal and the breathing rhythm signal, if it is detected that the environmental noise interference exceeds the pre-set threshold, a frequency domain conversion tool is used to separate the noise component, and a denoised sleep signal combination is obtained; According to the denoised sleep signal combination, the body motion frequency characteristics and the breathing rhythm characteristics are analyzed by using a time domain analysis tool to adjust the data accuracy, whether there is abnormal fluctuation is judged by comparing the pre-established normal range value, and an adjusted sleep signal data set is determined; For the adjusted sleep signal data set, a data integration tool is used to classify and store the body motion frequency and the breathing rhythm, and the data characteristics related to behavior analysis are obtained according to the real-time record requirements during storage, and a final optimized sleep signal data set is obtained. 3.The method for sleep quality analysis without a guardian according to claim 1, wherein, The step of obtaining the preliminary cleaned sleep signal set by preliminarily screening the original sleep signal according to the preliminary sleep behavior data record and the environmental noise interference in the original sleep signal by using a preset threshold includes: wherein, denotes the time instant of the original sleep signal, denotes the body movement frequency component, denotes the respiration rhythm component, denotes the ambient noise disturbance component; The judgment standard for environmental noise screening is determined as: wherein, denotes a preset noise screening threshold, denotes a mean value of the sleep signal, denotes a standard deviation of the sleep signal, denotes a threshold coefficient; The judgment logic for signal cleaning is: wherein, denotes the sleep signal after preliminary cleaning, denotes the original sleep signal, denotes the filter threshold. 4.The method for sleep quality analysis without a guardian according to claim 1, wherein, The step of judging the sleep stage distribution of the user by using a pre-established sleep stage classification model to extract features, analyze the regularity change of the body movement mode and the breathing rule, and analyze the optimized sleep signal data set includes: According to the optimized sleep signal data set, the body movement mode and the breathing rule are extracted by using a preset feature extraction tool, and a plurality of feature values related to sleep stage classification are obtained, to obtain a preliminary feature combination; According to the preliminary feature combination, a data integration tool is used to classify a plurality of feature values, and a pre-established stage division standard is combined during classification. If the feature value meets the regularity analysis condition of a certain stage, it is classified into the corresponding category, and a classified feature set is determined; According to the classified feature set, the mode change and behavior association are matched by using a time sequence comparison tool to obtain sleep stage distribution judgment basis corresponding to the sleep signal, and a preliminary mapping result of the sleep stage distribution is obtained; According to the preliminary mapping result of the sleep stage distribution, a data verification tool is used to confirm the matching degree of the body movement mode and the breathing rule. If the matching degree exceeds a preset threshold, the corresponding relationship with the sleep stage classification is confirmed, and the final sleep stage distribution data is determined. 5.The method for sleep quality analysis without a guardian according to claim 1, wherein, According to the sleep stage distribution, the sleep quality of the user is graded and evaluated by combining a preset threshold range, the duration of each stage and the conversion frequency are obtained, and the sleep quality evaluation result is obtained. The step includes: According to the sleep stage distribution, the duration of each stage and the conversion frequency are recorded in detail by using a time sequence record tool, and dynamic change data related to sleep quality are obtained, to obtain a preliminary time distribution record table; According to the preliminary time distribution record table, the duration of each stage and the conversion frequency are classified by using a data classification tool by combining a preset threshold range. If the duration or the conversion frequency exceeds the corresponding threshold range, it is classified into an abnormal category, and the classified sleep stage distribution data is determined; According to the classified sleep stage distribution data, the data comparison tool is matched with the pre-established sleep quality standard by combining the abnormal category and the behavior mode, and the basis related to the grading evaluation is obtained, and the sleep quality grade distribution of each stage is judged; According to the sleep quality grade distribution of each stage, the sleep quality evaluation result is summarized by using a data integration tool by combining the distribution, and a final sleep quality evaluation report is obtained, and the overall sleep quality grade of the user is determined. 6.The method for sleep quality analysis without a guardian according to claim 1, wherein, According to the sleep quality evaluation result, the long-term tracking of the sleep mode of the user is performed through a time series analysis method, the periodic change of the sleep stage distribution is identified, and the steps of determining the sleep habit characteristics of the user include: According to the sleep quality evaluation result, the long-term collection of the sleep stage distribution data of the user is performed using a time series recording tool, the starting and ending time points of each stage are labeled, the continuous data related to the periodic change is obtained, and a complete sleep stage time distribution record table is obtained; According to the sleep stage time distribution record table, a data comparison tool is used to match the periodic pattern library established in advance, if the sleep stage distribution data matches the time interval of a certain pattern in the periodic pattern library, it is classified as the corresponding periodic characteristic, and the sleep stage periodic distribution characteristics of the user are determined; According to the sleep stage periodic distribution characteristics, the data integration tool is used to associate the periodic change and the duration of each stage, obtain the conversion rule between different stages, judge whether there is a regular fluctuation in the sleep mode, and obtain the long-term change trend; According to the long-term change trend, the data classification tool is used to classify the fluctuation amplitude in combination with the preset threshold range, if the fluctuation amplitude exceeds the threshold range, it is marked as an abnormal mode, and the sleep habit characteristics of the user are determined. 7.The method for sleep quality analysis without care according to claim 1, wherein, According to the sleep habit characteristics, the signal acquisition parameters of the non-contact sensor are dynamically optimized using an adaptive adjustment mechanism, and a monitoring configuration scheme consistent with the user's daily habits is obtained. The steps include: According to the sleep habit characteristics, the data recording tool is used to continuously track the daily work and rest time period of the user, the sleep-related behavior data is obtained, the collection frequency is preliminarily set in combination with the environmental adaptation attribute, and an initial monitoring scheme matching the daily habits is obtained; According to the initial monitoring scheme, the data comparison tool is used to match the collection frequency with the habit tracking database established in advance, if the deviation between the collection frequency and the habit data in the habit tracking database exceeds the preset threshold range, the parameter adjustment of signal acquisition is optimized through the adaptive adjustment mechanism, and the collection rule consistent with the user characteristics is determined; According to the collection rule, the parameter configuration tool is used to dynamically adjust the signal acquisition mode of the non-contact sensor in combination with the environmental adaptation attribute, the monitoring data consistent with the daily habits is obtained, and the monitoring frequency range suitable for the current state of the user is determined; According to the monitoring frequency range, the data integration tool is used to associate the signal acquisition result with the sleep habit data, the monitoring configuration is finally calibrated through the data matching attribute, and a personalized monitoring configuration scheme consistent with the daily habits of the user is obtained. 8.The method for sleep quality analysis without care according to claim 1, wherein, According to the monitoring configuration scheme, the sleep data of the user is uploaded to a remote server for storage and backup through cloud data synchronization technology, and a real-time updated sleep archive record is obtained. The steps include: According to the sleep data, the collected content is classified and processed by using a data arrangement tool, and an upload priority is set for different categories of data. If the priority exceeds a preset threshold, the data is transmitted to a remote server through an upload mechanism to obtain a classified data set; For the classified data set, a data synchronization tool is used to synchronize the frequency and detect the transmission process at regular intervals. If the transmission is interrupted, the backup path is used to reconnect and obtain the complete transmission log; According to the complete transmission log, a data verification tool is used to compare the data uploaded to the remote server, verify the consistency of the data integrity attribute, and adjust the inconsistent part through a repair tool to determine the verified data content; For the verified data content, a permission management tool is used to limit the access range according to the user's permission, and a backup tool is used to generate a copy in the storage space to obtain a real-time updated sleep archive record. 9.The method for sleep quality analysis without care according to claim 1, wherein, According to the real-time updated sleep archive record, a sleep quality evaluation report of the user is generated by using data visualization technology, and the sleep stage distribution and sleep quality evaluation results are graphically displayed to determine the sleep improvement direction that the user can refer to, including: According to the sleep archive, a data arrangement tool is used to classify and process the updated content, and the stage distribution data is extracted to obtain a classified sleep stage distribution data set, and the preliminary sleep stage information is obtained; For the preliminary sleep stage information, an evaluation standard tool is used to calculate the sleep quality evaluation score. If the score is lower than the preset threshold, the parameters are adjusted to recalculate the sleep quality evaluation result that meets the standard; According to the sleep quality evaluation result, a graphical display tool is used to visually process the stage distribution and evaluation data according to the display form to generate an intuitive sleep quality evaluation report, and the graphical content is obtained for presentation; For the graphical content, a direction guide tool is used to match the improvement direction. If the matching result is consistent with the low score item in the sleep quality evaluation report, the relevant suggestions are extracted from the pre-established suggestion library to determine the sleep improvement direction that the user can refer to.

10. An unattended sleep quality analysis system for performing the unattended sleep quality analysis method according to any one of claims 1 to 9, characterized by It includes: A sleep behavior data record acquisition module (10) is used to obtain sleep-related signals from the user's surrounding environment by using non-contact sensor technology, and the body motion frequency characteristics and breathing rhythm characteristics are captured in real time to obtain preliminary sleep behavior data records; A sleep signal data set determination module (20) is used to filter the body motion frequency signal and breathing rhythm signal by using a signal denoising processing method according to the preliminary sleep behavior data records, and the environmental noise interference is removed to determine an optimized sleep signal data set; A sleep stage distribution condition judgment module (30) is used to extract features from the optimized sleep signal data set by using a pre-established sleep stage classification model, analyze the regularity of body motion patterns and breathing patterns, and determine the sleep stage distribution condition of the user. The sleep quality evaluation result obtaining module (40) is configured to evaluate the sleep quality of the user according to the sleep stage distribution and in combination with a preset threshold range, obtain the duration of each stage and the conversion frequency, and obtain a sleep quality evaluation result; The sleep habit feature determining module (50) is configured to track the sleep mode of the user for a long time by using a time series analysis method, identify the periodic change of the sleep stage distribution, and determine the sleep habit feature of the user according to the sleep quality evaluation result; The monitoring configuration scheme obtaining module (60) is configured to dynamically optimize the signal acquisition parameters of the non-contact sensor by using an adaptive adjustment mechanism according to the sleep habit feature, and obtain a monitoring configuration scheme that conforms to the daily habit of the user; The sleep record obtaining module (70) is configured to upload the sleep data of the user to a remote server for storage and backup by using a cloud data synchronization technology according to the monitoring configuration scheme, and obtain a real-time updated sleep record; The sleep improvement direction determining module (80) is configured to generate a sleep quality evaluation report of the user by using a data visualization technology according to the real-time updated sleep record, perform graphical display on the sleep stage distribution and the sleep quality evaluation result, and determine a sleep improvement direction that can be referred to by the user.

Images