More preferably, the sleep trend adjusting peripheral device at least comprises a vocal music stimulation device, an ultrasonic stimulation device, a light stimulation device, an electric stimulation device, a magnetic stimulation device, a temperature stimulation device, a humidity stimulation device, a touch stimulation device and a touch stimulation device

Any one of the regulating devices is regulated and is determined by the specific regulating mode.

More preferably, the specific calculation formula of the dynamic adjustment effect coefficient is specifically:

；

wherein ,

the effect coefficient is dynamically adjusted; / >

The target regulating value, the sleep trend predicting characteristic value and the sleep trend characteristic value in the sleep trend characteristic curve obtained by quantization after dynamic regulation in the personalized time phase scene sleep regulating strategy are respectively; />

And correcting the coefficient for the preset user personality related to the biological state information of the user.

More preferably, the user biological status information includes at least gender, age, occupation, health status, disease status, and education level.

More preferably, the dynamic adjustment effect coefficient is reversely applied to the filtering frequency in the time-frequency analysis parameter, the dynamic adjustment of the window length, the adjustment of the weight of the decomposed signal in the frequency weight fusion calculation, and the generation of the personalized time phase scene sleep adjustment strategy.

More preferably, the specific steps of dynamically calculating and extracting the sleep trend index, the time phase trend correlation coefficient, the time phase trend distribution characteristic and the comprehensive regulation effect index according to the sleep stage curve, the sleep depth characteristic curve, the sleep trend characteristic curve and the dynamic regulation effect further include:

according to the sleep depth characteristic curve and the sleep trend characteristic curve, calculating to obtain the sleep trend index;

Performing correlation calculation according to the sleep stage curve and the sleep trend characteristic curve to obtain the time phase trend correlation coefficient;

based on the sleep stage curve, carrying out time phase distribution statistics on the sleep trend characteristic curve to obtain time phase trend distribution characteristics;

and extracting the average value and/or root mean square of the dynamic adjustment effect curve to obtain the comprehensive adjustment effect index.

More preferably, the method for calculating the sleep trend index specifically comprises the following steps:

1) Acquiring the sleep depth characteristic curve and the sleep trend characteristic curve;

2) Squaring the sleep depth characteristic curve and the sleep trend characteristic curve respectively to obtain a sleep depth characteristic square curve and a sleep trend characteristic square curve;

3) Calculating a sample point ratio of the sleep depth characteristic square curve to the sleep trend characteristic square curve to obtain a sleep trend curve;

4) Solving an evolution value of the average value of the sleep trend curve to obtain a sleep trend characteristic value;

5) And calculating the product of the sleep trend characteristic value, a preset method correction coefficient corresponding to a time-frequency analysis method and a preset user personality correction coefficient related to the user biological state information to generate the sleep trend index.

More preferably, the formula for calculating the sleep trend index specifically includes:

；

wherein ,

for the sleep trend index, +.>

The user personal correction coefficient and the method correction coefficient are preset respectively, and the user personal correction coefficient and the method correction coefficient are +.>

、/>

And respectively obtaining an ith value in the sleep trend characteristic curve and the sleep depth characteristic curve, wherein N is the data length of the sleep trend characteristic curve.

More preferably, the correlation calculation method at least comprises any one of coherence analysis, pearson correlation analysis, jacquard similarity analysis, linear mutual information analysis, linear correlation analysis, euclidean distance analysis, manhattan distance analysis and Chebyshev distance analysis.

More preferably, the time phase distribution statistics specifically includes performing numerical distribution statistical analysis on the sleep trend characteristic values in the sleep trend characteristic curve according to sleep time phase period in the sleep period curve, so as to obtain numerical distribution statistical characteristics of the sleep trend characteristic curve.

More preferably, the time-phase trend distribution characteristic includes at least any one of an average value, a root mean square, a maximum value, a minimum value, a variance, a standard deviation, a coefficient of variation, kurtosis and a skewness.

More preferably, the specific steps of establishing and updating the personalized sleep trend database of the user and dynamically optimizing the personalized time phase scene sleep adjustment strategy, and generating the sleep trend quantification and adjustment report according to the preset report period further comprise:

Carrying out dynamic detection quantification and dynamic intervention adjustment on sleep trends of users under different sleep scenes, and establishing and updating a personalized sleep trend database of the users;

dynamically updating key process data of dynamic detection quantification and dynamic intervention adjustment into the personalized sleep trend database of the user, and dynamically optimizing the personalized time phase scene sleep adjustment strategy;

and generating the sleep trend quantification and adjustment report according to a preset report period.

More preferably, the user personalized sleep trend database at least comprises the user biological state information, a sleep scene, the sleep stage curve, the sleep depth characteristic curve, the sleep trend index, the time phase trend correlation coefficient, the time phase trend distribution characteristic, an adjusting mode, an executing mode, an adjusting method, a target adjusting value curve, an adjusting intensity curve and the dynamic adjusting effect curve.

More preferably, the sleep trend quantification and adjustment report at least includes the user biological status information, a sleep scene, the sleep stage curve, the sleep depth characteristic curve, the sleep trend index, the time phase trend correlation coefficient, the time phase trend distribution characteristic, the comprehensive adjustment effect index, the dynamic adjustment effect curve, and a user sleep trend quantification and adjustment summary.

According to the purpose of the invention, the invention provides a sleep stability quantifying and regulating system based on time-frequency analysis, which comprises the following modules:

the state detection analysis module is used for carrying out acquisition and monitoring treatment, characteristic analysis and time phase identification on the physiological state of the sleeping process of the user to obtain a sleeping depth characteristic curve and a sleeping stage curve;

the characteristic curve extraction module is used for selecting time-frequency analysis parameters according to the characteristics of the sleep depth characteristic curve, performing time-frequency analysis on the time-frequency analysis parameters, identifying ultralow frequency trend components and extracting a sleep trend characteristic curve;

the trend dynamic regulation module is used for dynamically predicting the sleep trend characteristic curve, extracting a personalized time phase scene sleep regulation strategy and dynamically regulating and evaluating the sleep process of the user;

the trend dynamic quantization module is used for dynamically calculating and extracting sleep trend indexes, time phase trend correlation coefficients, time phase trend distribution characteristics and comprehensive regulation effect indexes according to the sleep stage curves, the sleep depth characteristic curves, the sleep trend characteristic curves and the dynamic regulation effects;

the trend data optimization module is used for establishing and updating a personalized sleep trend database of a user and dynamically optimizing the personalized time phase scene sleep adjustment strategy, and generating a sleep trend quantification and adjustment report according to a preset report period;

And the data operation management module is used for carrying out visual management, unified storage and operation management on all process data of the system.

More preferably, the state detection and analysis module further comprises the following functional units:

the state detection processing unit is used for carrying out acquisition monitoring and signal processing on the physiological state of the sleeping process of the user to obtain sleeping physiological state time frame data of the user;

the depth feature extraction unit is used for carrying out feature analysis and feature fusion on the frame data of the user in the sleep physiological state to obtain the sleep depth feature curve;

and the sleep phase identification unit is used for carrying out phase identification and sleep stage on the frame data of the user in the sleep physiological state to obtain the sleep stage curve.

More preferably, the characteristic curve extraction module further comprises the following functional units:

the time-frequency analysis unit is used for selecting time-frequency analysis parameters according to the characteristics of the sleep depth characteristic curve and performing time-frequency analysis on the time-frequency analysis parameters to obtain a TF time-frequency signal set;

and the trend component identification unit is used for identifying trend components from the TF time frequency signal set and generating the sleep trend characteristic curve.

More preferably, the trend dynamic adjustment module further comprises the following functional units:

The sleep state prediction unit is used for carrying out dynamic trend prediction on the sleep trend characteristic curve, extracting sleep trend prediction characteristic values and generating a sleep trend prediction characteristic curve;

the regulation strategy generation unit is used for dynamically generating the personalized time phase scene sleep regulation strategy according to a preset dynamic regulation period and according to a preset sleep stability regulation knowledge base, the sleep trend prediction characteristic value, a current specific sleep scene and a dynamic regulation history;

the dynamic regulation and control execution unit is used for connecting and regulating sleep trend regulation peripheral equipment according to the individual time phase scene sleep regulation strategy to dynamically intervene and regulate the sleep process of the user;

the dynamic effect evaluation unit is used for carrying out dynamic tracking evaluation on the effect of dynamic intervention adjustment, extracting the dynamic adjustment effect coefficient and generating a dynamic adjustment effect curve.

More preferably, the trend dynamic quantization module further comprises the following functional units:

the trend index quantization unit is used for calculating the sleep trend index according to the sleep depth characteristic curve and the sleep trend characteristic curve;

the correlation calculation unit is used for carrying out correlation calculation according to the sleep stage curve and the sleep trend characteristic curve to obtain the time phase trend correlation coefficient;

The time phase distribution statistics unit is used for carrying out time phase distribution statistics on the sleep trend characteristic curve based on the sleep stage curve to obtain the time phase trend distribution characteristics;

and the comprehensive regulation analysis unit is used for extracting the average value and/or root mean square of the dynamic regulation effect curve to obtain the comprehensive regulation effect index.

More preferably, the trend data optimization module further comprises the following functional units:

the trend data management unit is used for carrying out dynamic detection quantification and dynamic intervention adjustment on sleep trends of the user under different sleep scenes, and establishing and updating the personalized sleep trend database of the user;

the trend data application unit is used for dynamically updating key process data of dynamic detection quantification and dynamic intervention adjustment into the personalized sleep trend database of the user and dynamically optimizing the personalized time phase scene sleep adjustment strategy;

the user report generating unit is used for generating the sleep trend quantification and adjustment report according to a preset report period;

and the user report output unit is used for uniformly managing the format output and the presentation form of the sleep trend quantification and regulation report.

More preferably, the data operation management module further comprises the following functional units:

a user information management unit for registering input, editing, inquiry, output and deletion of user basic information;

the data visual management unit is used for visual display management of all data in the system;

the data storage management unit is used for uniformly storing and managing all data in the system;

and the data operation management unit is used for backing up, migrating and exporting all data in the system.

According to the purpose of the invention, the invention provides a sleep stability quantifying and adjusting device based on time-frequency analysis, which comprises the following modules:

the trend dynamic regulation module is used for dynamically predicting the sleep trend characteristic curve, extracting a personalized time phase scene sleep regulation strategy, and dynamically regulating and evaluating the sleep process of the user;

The trend dynamic quantization module is used for dynamically calculating and extracting sleep trend index, time phase trend correlation coefficient, time phase trend distribution characteristic and comprehensive regulation effect index according to the sleep stage curve, the sleep depth characteristic curve, the sleep trend characteristic curve and the dynamic regulation effect;

the trend data optimizing module is used for establishing and updating a personalized sleep trend database of a user and dynamically optimizing the personalized time phase scene sleep regulating strategy, and generating a sleep trend quantification and regulating report according to a preset report period;

the data visualization module is used for carrying out unified visual display management on all process data and/or result data in the device;

and the data management center module is used for uniformly storing and managing data operation of all process data and/or result data in the device.

The invention provides a sleep stability quantification and adjustment method, a system and a device based on time-frequency analysis, which acquire and monitor physiological states of a user sleep process, perform characteristic analysis and time phase identification to obtain a sleep depth characteristic curve, perform the time-frequency analysis, extract a sleep trend characteristic curve to perform dynamic prediction, extract a personalized time phase scene sleep adjustment strategy, and realize the real-time dynamic interference adjustment of the sleep stability of the user in a multi-means mode; further extracting sleep trend index, time phase trend correlation coefficient and time phase trend distribution characteristic to finish detection and quantification of sleep trend of the user; and (3) incorporating the key process data of detection quantification and intervention regulation into a database, establishing and continuously updating a user personalized sleep trend database, and continuously improving the detection quantification efficiency and the intervention regulation effect of user personalization.

The invention further optimizes the specific design of stability quantification on the basis of the prior research of the applicant, applies time-frequency analysis to the extraction of trend information, considers the state characteristics of the complete sleep period, and has more comprehensive and wide adaptability; the method further improves the calculation mode of the stability index, and improves the fine granularity and sensitivity of evaluation; the corresponding effect coefficient calculation scheme is also provided, so that a powerful basis is provided for controlling the adjustment process. The invention can provide a more scientific and efficient implementation method for detecting, quantifying, intervening and adjusting sleep stability and a landing scheme. In an actual application scene, the sleep stability quantification and adjustment method, system and device based on modal decomposition can enable related sleep quantified or adjusted products and services, meet different user scene requirements and assist a user in sleeping.

Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.

Drawings

The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate and do not limit the invention.

FIG. 1 is a schematic diagram of the steps of a method for quantifying and adjusting sleep stability based on time-frequency analysis according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of a module composition of a sleep stability quantification and adjustment system based on time-frequency analysis according to an embodiment of the present invention;

fig. 3 is a schematic diagram of a module configuration of a sleep stability quantifying and adjusting device based on time-frequency analysis according to an embodiment of the present invention.

Detailed Description

In order to more clearly illustrate the objects and technical solutions of the present invention, the present invention will be further described below with reference to the accompanying drawings in the embodiments of the present invention. It will be apparent that the embodiments described below are only some, but not all, embodiments of the invention. Other embodiments, which are derived from the embodiments of the invention by a person skilled in the art without creative efforts, shall fall within the protection scope of the invention. It should be noted that, in the case of no conflict, the embodiments and features in the embodiments may be arbitrarily combined with each other.

The applicant found that in general, the human sleep physiology is a non-stationary time series process, and the characterization curve is also a non-stationary time series data curve. In order to extract the information therein, the applicant applies a time-frequency analysis to the extraction of the sleep trend information, further optimizing the quantization and adjustment process. The time-frequency transformation is a method for analyzing the time-frequency distribution characteristics of signals, provides joint distribution information of a time domain and a frequency domain, and clearly describes the relation of the frequency of the signals with time; common time-frequency transformation methods include standard Fourier transformation FT, fast Fourier transformation FFT, short-time Fourier transformation STFT, S transformation, gabor transformation, hilbert yellow transformation HHT, wigner-Ville distribution WVD, smooth Wigner-Ville distribution SWVD, pseudo-Wigner-Ville distribution PWVD, smooth pseudo-Wigner-Ville distribution SPWVD, choi-Williams distribution CWD and the like, and are widely used for time-frequency analysis of linear stationary signals and nonlinear stationary signals. The time-frequency filtering is a method for analyzing and extracting time-frequency characteristics of signals, is divided into time-domain filtering and frequency-domain filtering, and is widely applied to various fields of signal processing; the time domain filtering may be classified into average filtering, median filtering, kalman filtering, savitzky-Golay filtering, etc., and the frequency domain filtering may be classified into low-pass filtering, high-pass filtering, band-stop filtering, etc.

Referring to fig. 1, the method for quantifying and adjusting sleep stability based on time-frequency analysis provided by the embodiment of the invention includes the following steps:

p100: and (3) carrying out acquisition and monitoring processing, feature analysis and time phase identification on the physiological state of the sleeping process of the user to obtain a sleeping depth feature curve and a sleeping stage curve.

The first step, the physiological state of the user in the sleeping process is collected, monitored and processed to obtain the time frame data of the physiological state of the user sleeping.

In this embodiment, the signal processing at least includes AD digital-to-analog conversion, resampling, re-referencing, noise reduction, artifact removal, power frequency notch, low-pass filtering, high-pass filtering, band-stop filtering, band-pass filtering, correction processing, and time frame segmentation; the correction processing specifically includes signal correction, signal prediction and smoothing processing on signal data segments containing artifacts or distortion in a signal, and the time frame segmentation refers to continuous sliding segmentation on target signal data according to a signal sampling rate and with a preset time window length and a preset time translation step length.

In this embodiment, the frame data at least includes any one of brain center state data and autonomic nerve state data when the user sleeps in a physiological state; wherein, the brain center state data at least comprises any one of an electroencephalogram signal, a magnetoencephalography signal and an oxygen level dependent signal, and the autonomic nerve state data at least comprises any one of a blood oxygen level dependent signal, an electrocardiosignal, a pulse signal, a respiratory signal, an oxygen signal, a body temperature signal and a skin electric signal.

In this embodiment, the specific implementation process of the technical scheme is stated by collecting the electroencephalogram signals and the electrocardiograph signals for monitoring the sleeping process of the user as the sleeping physiological state.

Firstly, acquiring and recording sleeping electroencephalogram of a user by an electroencephalograph, wherein the sampling rate is 1024Hz, the recording electrodes are F3, F4, C3 and C4, and the reference electrodes M1 and M2; the electroencephalogram signals are subjected to unified signal processing, including re-referencing, artifact removal, wavelet noise reduction, 50Hz frequency doubling power frequency notch, 0.5-80Hz band-pass filtering and signal correction processing by M1/2, and pure electroencephalogram signals are obtained. Collecting and extracting electrocardiosignals of a user through a portable single-lead electrocardiograph, wherein the collecting position is above the left chest, and the sampling rate is 512Hz; and performing unified signal processing on the electrocardiosignal, including artifact removal, wavelet noise reduction, 0.5-40hz band-pass filtering and signal correction processing, so as to obtain a pure electrocardiosignal.

Secondly, extracting signal frequency bands of the pure brain electrical signals sequentially, wherein the signal frequency bands comprise delta rhythms (0.5-4 Hz), theta rhythms (4-8 Hz), alpha rhythms (8-12 Hz), beta rhythms (12-30 Hz) and gamma rhythms (30-80 Hz), and obtaining frequency band brain electrical signals; and further, continuously sliding and dividing the pure electroencephalogram signal, the frequency band electroencephalogram signal and the pure electrocardiosignal according to a preset time window length 30s and a preset time shift step length 15s to obtain the sleep physiological state time frame data of the user.

And secondly, carrying out feature analysis and feature fusion on the frame data of the user in the sleep physiological state to obtain a sleep depth feature curve.

In this embodiment, the feature analysis includes at least numerical analysis, envelope analysis, time-frequency analysis, entropy analysis, fractal analysis, and complexity analysis. Feature fusion refers to the step of selecting target features with preset feature quantity from a target feature set obtained by feature analysis, and performing weighted calculation to generate a sleep depth feature curve.

In this embodiment, the sleep depth characteristic curve is a characteristic curve representing the sleep depth and the phase state of the user in a period before sleeping, a sleep duration, and a period after sleeping is preset, and the calculation and generation method specifically includes:

1) Carrying out feature analysis on time frame data in the sleep physiological state time frame data of the user one by one, and splicing the time frame data according to a time sequence to obtain a sleep physiological state time frame feature curve set;

2) And screening target related characteristic curves from the sleep physiological state time frame characteristic curve set, and carrying out weighted calculation according to preset characteristic fusion weights to generate a sleep depth characteristic curve.

In the embodiment, time-frequency analysis (frequency band power, frequency band power duty ratio), entropy analysis (sample entropy) and complexity analysis (LZC index: lempel-Ziv complexity index) are carried out on the electroencephalogram data of the user sleep physiological state time frame data frame by frame; and carrying out numerical analysis on the electrocardio data of the frame data of the sleep physiological state of the user frame by frame, and extracting heart rate variation characteristics (heart rate average value and heart rate variation coefficient) of the user. Further, the delta-theta (delta rhythm + theta rhythm) joint band power ratio, normalized sample entropy reciprocal, and mean value of normalized heart rate mean reciprocal (feature fusion process) of the F4-M1 channel are selected as the sleep depth feature curve. In general, the deeper the user sleeps, the larger the delta-theta combined band power ratio, the larger the normalized sample entropy reciprocal and the normalized heart rate average reciprocal (the smaller the heart rate average reciprocal on the contrary), and the more stable the user sleep state, the cortical electrophysiology and the autonomic neurophysiologic performance.

And thirdly, carrying out phase recognition and sleep stage on the frame data of the user in the sleep physiological state to obtain a sleep stage curve.

In this embodiment, the method for generating the sleep stage curve specifically includes:

1) Performing learning training and data modeling on user sleep physiological state time frame data of a scale sleep user sample and corresponding sleep stage data through a deep learning algorithm to obtain a sleep time phase automatic stage model;

2) Inputting the frame data of the current user in the sleep physiological state into a sleep time phase automatic stage model to obtain the corresponding sleep time phase stage and generating a sleep stage curve according to the time sequence.

In an actual use scene, the accuracy of the sleep phase automatic stage model is higher and higher through data accumulation of a user sample and deep learning of the stage model.

P200: and selecting time-frequency analysis parameters according to the characteristics of the sleep depth characteristic curve, performing time-frequency analysis on the parameters, identifying ultralow frequency trend components, and extracting the sleep trend characteristic curve.

The first step, selecting time-frequency analysis parameters according to the characteristics of the sleep depth characteristic curve and performing time-frequency analysis on the time-frequency analysis parameters to obtain a TF time-frequency signal set.

In this embodiment, the selection of the time-frequency analysis parameters depends on the feature source and combination mode of the sleep depth feature curve, the specific method of time-frequency analysis, and the preset time window length of time frame segmentation, which at least includes one of the filtering frequency, the window length, and the window function.

In this embodiment, the method of time-frequency analysis at least includes any one of time-frequency transformation and time-frequency filtering.

In this embodiment, the time-frequency transformation at least includes any one of standard fourier transform FT, fast fourier transform FFT, short-time fourier transform STFT, S transform, gabor transform, hilbert yellow transform HHT, wigner-Ville distribution WVD, smooth Wigner-Ville distribution SWVD, pseudo Wigner-Ville distribution PWVD, smooth pseudo Wigner-Ville distribution SPWVD, choi-Williams distribution CWD.

In this embodiment, the time-frequency filtering includes at least any one of time-domain filtering and frequency-domain filtering; the time domain filtering at least comprises any one of average value filtering, median filtering, kalman filtering and Savitzky-Golay filtering, and the frequency domain filtering at least comprises any one of low-pass filtering, high-pass filtering, band-pass filtering and band-stop filtering.

In this embodiment, S transform is selected as a method of time-frequency analysis. The S transformation overcomes the defect of the width of a fixed window function of the short-time Fourier transformation, adopts a Gaussian window function which changes along with the frequency, the width of the Gaussian window function is in direct proportion to the reciprocal of the frequency, and a narrow window and a wide window are used at high frequency. The definition is specifically as follows:

For signals

S transform of (2)

；

Wherein the window function is a Gaussian window function

。

In this embodiment, the sleep depth characteristic curve is calculated from δ - θ (δ rhythm+θ rhythm) combined band power ratio, normalized sample entropy reciprocal, normalized heart rate average reciprocal, and the TF time-frequency signal set of the sleep depth characteristic curve is extracted by S-transformation.

Secondly, identifying trend components from the TF time-frequency signal set, and generating a sleep trend characteristic curve.

In this embodiment, a method for identifying an ultralow frequency trending component specifically includes:

if time-frequency transformation is adopted, extracting time domain signals with different frequencies through inverse transformation to obtain TF component signals;

if time-frequency filtering is adopted, the TF component signals are obtained by extracting signals with different frequency band intervals or window lengths;

2) And carrying out frequency weighting fusion calculation on the trend TF time-frequency signal set to generate a sleep trend characteristic curve.

In this embodiment, the preset ultralow frequency threshold depends on the sampling rate of the target signal, the preset time window length of the time frame segmentation, and the dynamic adjustment effect.

In this embodiment, the frequency weighted fusion calculation specifically uses the inverse proportion of the weighted weight of the signal and the center frequency as the calculation principle, and performs weighted fusion on the target signal set to generate the signal frequency weighted characteristic description curve.

P300: and dynamically predicting the sleep trend characteristic curve, extracting a personalized time phase scene sleep regulation strategy, and dynamically regulating and evaluating the sleep process of the user.

The first step, dynamically predicting the sleep trend characteristic curve, extracting a sleep trend prediction characteristic value and generating a sleep trend prediction characteristic curve.

In this embodiment, the prediction method of the sleep trend prediction feature value at least includes any one of AR, MR, ARMA, ARIMA, SARIMA, VAR and deep learning.

In this embodiment, the MR method is used to perform trend prediction analysis on the sleep depth characteristic curve to obtain a sleep trend prediction characteristic value, and generate or update the sleep trend prediction characteristic curve.

In the actual adaptation scene, the time-frequency analysis and the index prediction may adopt a commonly used time-sequential prediction method such as AR, MR, ARMA, ARIMA, SARIMA, VAR, and the prediction calculation of the sleep trend prediction characteristic value can also be completed through a deep learning model.

And secondly, dynamically generating a personalized time phase scene sleep regulation strategy according to a preset dynamic regulation period and a preset sleep stability regulation knowledge base, a sleep trend prediction characteristic value, a current specific sleep scene and a dynamic regulation history.

In this embodiment, the individual time phase scene sleep adjustment policy at least includes a sleep scene, a sleep time phase, an adjustment mode, an execution mode, an adjustment method, an adjustment intensity, an adjustment time point, a duration, a target adjustment value, and a device control parameter; wherein the modulation means comprises at least vocal stimulation, ultrasound stimulation, optical stimulation, electrical stimulation, magnetic stimulation, temperature stimulation, humidity stimulation, tactile stimulation and/or

The implementation mode of any mode of regulation at least comprises any mode of separation type and contact type.

In the actual use scene, the regulation effect is ensured by selecting ex-vivo vocal stimulation, optical stimulation, temperature stimulation, humidity stimulation and

the regulation and control has small interference to the sleep of the user and good user experience.

Thirdly, according to a personalized time phase scene sleep regulation strategy, connecting and regulating sleep trend regulation peripheral equipment, and dynamically intervening and regulating the sleep process of a user.

In this embodiment, the sleep trend regulating peripheral device regulates at leastIncluding vocal stimulation device, ultrasonic stimulation device, optical stimulation device, electrical stimulation device, magnetic stimulation device, temperature stimulation device, humidity stimulation device, tactile stimulation device, and tactile stimulation device

Any of the control devices is regulated and determined by the specific regulation mode.

And fourthly, dynamically tracking and evaluating the effect of dynamic intervention adjustment, extracting the dynamic adjustment effect coefficient and generating a dynamic adjustment effect curve.

In this embodiment, the specific calculation formula of the dynamic adjustment effect coefficient is specifically:

；

wherein ,

the effect coefficient is dynamically adjusted; />

The method comprises the steps of respectively obtaining a target regulation value, a sleep trend prediction characteristic value and a sleep trend characteristic value in a sleep trend characteristic curve obtained by quantization after dynamic regulation in a personalized time phase scene sleep regulation strategy; />

In this embodiment, the user biological status information includes at least gender, age, occupation, health status, disease status, and education level.

In this embodiment, the dynamic adjustment effect coefficient is reversely applied to the filtering frequency in the time-frequency analysis parameter, the dynamic adjustment of the window length, the adjustment of the weight of the decomposed signal in the frequency weight fusion calculation, and the generation of the sleep adjustment strategy of the individual time phase scene, so as to continuously optimize the closed loop circulation efficiency of detection quantization and intervention adjustment.

In an actual use scene, the effect of dynamic intervention adjustment can be realized through the correlation calculation, curve distance characteristic calculation and comprehensive evaluation of a sleep trend prediction characteristic curve, a sleep trend characteristic curve and a target adjustment value curve; the dynamic adjustment effect can be accurately estimated, for example, by averaging the euclidean distance of the sleep trend characteristic curve and the target adjustment value curve.

P400: and dynamically calculating and extracting sleep trend indexes, time phase trend correlation coefficients, time phase trend distribution characteristics and comprehensive regulation effect indexes according to the sleep stage curves, the sleep depth characteristic curves, the sleep trend characteristic curves and the dynamic regulation effects.

And step one, calculating to obtain a sleep trend index according to the sleep depth characteristic curve and the sleep trend characteristic curve.

In this embodiment, the method for calculating the sleep trend index specifically includes:

1) Acquiring a sleep depth characteristic curve and a sleep trend characteristic curve;

4) Calculating the average value of the sleep trend curve to obtain a sleep trend characteristic value;

5) And calculating the product of the sleep trend characteristic value, a preset method correction coefficient corresponding to the time-frequency analysis method and a preset user individual correction coefficient related to the biological state information of the user to generate a sleep trend index.

In this embodiment, the calculation formula of the sleep trend index is specifically:

；

wherein ,

is a sleep trend fingerCount (n)/(l)>

、/>

In general, the correction coefficient of the preset method of S transformation is 0.85, and the correction coefficient of the preset user personality of a normal healthy user is 1.0.

And step two, carrying out correlation calculation according to the sleep stage curve and the sleep trend characteristic curve to obtain a time phase trend correlation coefficient.

In this embodiment, the correlation calculation method at least includes any one of coherence analysis, pearson correlation analysis, jaccard similarity analysis, linear mutual information analysis, linear correlation analysis, euclidean distance analysis, manhattan distance analysis, and chebyshev distance analysis.

In the embodiment, firstly, performing curve smoothing operation on a sleep stage curve; and then selecting linear correlation analysis as correlation calculation and obtaining a linear correlation coefficient as a time phase trend correlation coefficient of the sleep stage curve and the sleep trend characteristic curve.

And thirdly, carrying out time phase distribution statistics on the sleep trend characteristic curve based on the sleep stage curve to obtain time phase trend distribution characteristics.

In this embodiment, the phase distribution statistics specifically refers to performing numerical distribution statistical analysis on the sleep trend feature values in the sleep trend feature curve according to the sleep phase period in the sleep period curve, so as to obtain the numerical distribution statistical feature of the sleep trend feature curve. The time phase trend distribution characteristics at least comprise any one of average value, root mean square, maximum value, minimum value, variance, standard deviation, variation coefficient, kurtosis and skewness.

And step four, extracting the average value and/or root mean square of the dynamic adjustment effect curve to obtain the comprehensive adjustment effect index.

In this embodiment, the average value of the dynamic adjustment effect curve is taken as the overall adjustment effect index.

P500: establishing and updating a personalized sleep trend database of a user, dynamically optimizing the personalized time phase scene sleep regulation strategy, and generating a sleep trend quantification and regulation report according to a preset report period.

The method comprises the steps of firstly, carrying out dynamic detection quantification and dynamic intervention adjustment on sleep trends of a user in different sleep scenes, and establishing and updating a personalized sleep trend database of the user.

In this embodiment, the user personalized sleep trend database at least includes user biological status information, a sleep scene, a sleep stage curve, a sleep depth characteristic curve, a sleep trend index, a time phase trend correlation coefficient, a time phase trend distribution characteristic, an adjustment mode, an execution mode, an adjustment method, a target adjustment value curve, an adjustment intensity curve, and a dynamic adjustment effect curve.

In the actual use scene, different scene combinations are selected according to the basic situation of the user, and the dynamic detection quantification and dynamic intervention adjustment are carried out on the sleep trend of the user under multiple scenes such as different sleep pressures, different sleep environments, different health states and the like, so that more comprehensive personalized sleep trend data of the user can be obtained.

And step two, dynamically updating key process data of dynamic detection quantification and dynamic intervention adjustment into a user personalized sleep trend database, and dynamically optimizing a personalized time phase scene sleep adjustment strategy.

In this embodiment, key process data such as a user key physiological index curve, a physiological event, a sleep scene, a sleep stage curve, a sleep depth characteristic curve, a sleep trend index, a time phase trend correlation coefficient, a time phase trend distribution characteristic, an adjusting mode, an executing mode, an adjusting method, a target adjusting value curve, an adjusting intensity curve, a dynamic adjusting effect curve and the like are required to be updated into a user personalized sleep trend database in real time, and an accurate data basis is provided for a personalized time phase scene sleep adjusting strategy generated in real time.

In the actual use scene, along with the continuous accumulation of the user personalized related data and scene adjustment feedback, the data richness of the user personalized sleep trend database is increased, so that the sleep stability of the user can be further and accurately detected and quantified, and the quality and effect of the user sleep stability dynamic adjustment intervention can be continuously improved.

Thirdly, generating a sleep trend quantification and adjustment report according to a preset report period.

In this embodiment, the sleep trend quantifying and adjusting report at least includes user biological status information, a sleep scene, a sleep stage curve, a sleep depth characteristic curve, a sleep trend index, a time phase trend correlation coefficient, a time phase trend distribution characteristic, a comprehensive adjusting effect index, a dynamic adjusting effect curve, and user sleep trend quantifying and adjusting summary.

In the actual use scene, the sleep trend quantification and adjustment report can be generated and output according to different time periods to meet different scene demands of different crowds, and health data statistics and strategy improvement basis are provided for sleep health management of users.

As shown in fig. 2, an embodiment of the present invention provides a sleep stability quantification and adjustment system based on time-frequency analysis, which is configured to perform the above-mentioned method steps. The system comprises the following modules:

the state detection analysis module S100 is used for carrying out acquisition and monitoring processing, characteristic analysis and time phase identification on the physiological state of the sleeping process of the user to obtain a sleeping depth characteristic curve and a sleeping stage curve;

the characteristic curve extraction module S200 is used for selecting time-frequency analysis parameters according to the characteristics of the sleep depth characteristic curve, performing time-frequency analysis on the time-frequency analysis parameters, identifying ultralow frequency trend components and extracting the sleep trend characteristic curve;

the trend dynamic adjustment module S300 is used for dynamically predicting a sleep trend characteristic curve, extracting a personalized time phase scene sleep adjustment strategy, and dynamically adjusting and evaluating the sleep process of a user;

the trend dynamic quantization module S400 is used for dynamically calculating and extracting sleep trend indexes, time phase trend correlation coefficients, time phase trend distribution characteristics and comprehensive regulation effect indexes according to the sleep stage curves, the sleep depth characteristic curves, the sleep trend characteristic curves and the dynamic regulation effects;

The trend data optimization module S500 is used for establishing and updating a personalized sleep trend database of a user and dynamically optimizing a personalized time phase scene sleep adjustment strategy, and generating a sleep trend quantification and adjustment report according to a preset report period;

and the data operation management module S600 is used for carrying out visual management, unified storage and operation management on all process data of the system.

In this embodiment, the state detection and analysis module S100 further includes the following functional units:

the depth feature extraction unit is used for carrying out feature analysis and feature fusion on the frame data of the user in the sleep physiological state to obtain a sleep depth feature curve;

and the sleep phase identification unit is used for carrying out phase identification and sleep stage on the frame data of the user in the sleep physiological state to obtain a sleep stage curve.

In this embodiment, the feature curve extraction module S200 further includes the following functional units:

And the trend component recognition unit is used for recognizing trend components from the TF time frequency signal set and generating a sleep trend characteristic curve.

In this embodiment, the trend dynamic adjustment module S300 further includes the following functional units:

the regulation strategy generation unit is used for dynamically generating a personalized time phase scene sleep regulation strategy according to a preset dynamic regulation period and according to a preset sleep stability regulation knowledge base, a sleep trend prediction characteristic value, a current specific sleep scene and a dynamic regulation history;

the dynamic regulation and control execution unit is used for connecting and regulating sleep trend regulating peripheral equipment according to the individual time phase scene sleep regulating strategy and carrying out dynamic intervention regulation on the sleep process of the user;

In this embodiment, the trend dynamic quantization module S400 further includes the following functional units:

the trend index quantization unit is used for calculating and obtaining a sleep trend index according to the sleep depth characteristic curve and the sleep trend characteristic curve;

The correlation calculation unit is used for carrying out correlation calculation according to the sleep stage curve and the sleep trend characteristic curve to obtain a time phase trend correlation coefficient;

the time phase distribution statistics unit is used for carrying out time phase distribution statistics on the sleep trend characteristic curve based on the sleep stage curve to obtain time phase trend distribution characteristics;

and the comprehensive regulation analysis unit is used for extracting the average value and/or root mean square of the dynamic regulation effect curve to obtain a comprehensive regulation effect index.

In this embodiment, the trend data optimization module S500 further includes the following functional units:

the trend data management unit is used for carrying out dynamic detection quantification and dynamic intervention adjustment on sleep trends of the user in different sleep scenes, and establishing and updating a personalized sleep trend database of the user;

the trend data application unit is used for dynamically updating key process data of dynamic detection quantification and dynamic intervention adjustment into a user personalized sleep trend database and dynamically optimizing a personalized time phase scene sleep adjustment strategy;

the user report generating unit is used for generating a sleep trend quantification and adjustment report according to a preset report period;

and the user report output unit is used for carrying out unified management on the format output and the presentation form of the sleep trend quantification and regulation report.

In this embodiment, the data operation management module S600 further includes the following functional units:

As shown in fig. 3, the device for quantifying and adjusting sleep stability based on time-frequency analysis provided by the embodiment of the invention comprises the following modules:

the state detection analysis module M100 is used for carrying out acquisition and monitoring processing, characteristic analysis and time phase identification on the physiological state of the sleeping process of the user to obtain a sleeping depth characteristic curve and a sleeping stage curve;

the characteristic curve extraction module M200 is used for selecting time-frequency analysis parameters according to the characteristics of the sleep depth characteristic curve, performing time-frequency analysis on the time-frequency analysis parameters, identifying ultralow frequency trend components and extracting the sleep trend characteristic curve;

the trend dynamic adjustment module M300 is used for dynamically predicting a sleep trend characteristic curve, extracting a personalized time phase scene sleep adjustment strategy, and dynamically adjusting and evaluating the sleep process of a user;

The trend dynamic quantization module M400 is used for dynamically calculating and extracting sleep trend indexes, time phase trend correlation coefficients, time phase trend distribution characteristics and comprehensive regulation effect indexes according to sleep stage curves, sleep depth characteristic curves, sleep trend characteristic curves and dynamic regulation effects;

the trend data optimization module M500 is used for establishing and updating a personalized sleep trend database of a user and dynamically optimizing a personalized time phase scene sleep adjustment strategy, and generating a sleep trend quantification and adjustment report according to a preset report period;

the data visualization module M600 is used for carrying out unified visual display management on all process data and/or result data in the device;

the data management center module M700 is used for unified storage and data operation management of all process data and/or result data in the device.

The apparatus is configured to correspondingly perform the steps of the method of fig. 1, and will not be described in detail herein.

The present invention also provides various types of programmable processors (FPGA, ASIC or other integrated circuit) for running a program, wherein the program when run performs the steps of the embodiments described above.

The invention also provides corresponding computer equipment, comprising a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the memory realizes the steps in the embodiment when the program is executed.

Although the embodiments of the present invention are described above, the embodiments are only used for facilitating understanding of the present invention, and are not intended to limit the present invention. Any person skilled in the art to which the present invention pertains may make any modifications, changes, equivalents, etc. in form and detail of the implementation without departing from the spirit and principles of the present invention disclosed herein, which are within the scope of the present invention. Accordingly, the scope of the invention should be determined from the following claims.

Claims

1. The sleep stability quantification and adjustment method based on time-frequency analysis is characterized by comprising the following steps of:

2. The method of claim 1, wherein the specific steps of acquiring, processing and analyzing the physiological state of the sleep process of the user to obtain a sleep depth characteristic curve and a sleep stage characteristic curve further comprise:

3. The method of claim 2, wherein: the signal processing at least comprises AD digital-to-analog conversion, resampling, re-referencing, noise reduction, artifact removal, power frequency notch, low-pass filtering, high-pass filtering, band-stop filtering, band-pass filtering, correction processing and time frame segmentation; the correction processing specifically includes signal correction, signal prediction and smoothing processing on signal data segments containing artifacts or distortion in a signal, and the time frame segmentation refers to continuous sliding segmentation on target signal data according to a signal sampling rate and with a preset time window length and a preset time translation step length.

4. The method of claim 2, wherein: the frame data at least comprises any one of brain center state data and autonomic nerve state data when the user sleeps in a physiological state; wherein the brain center state data at least comprises any one of an electroencephalogram signal, a magnetoencephalography signal and an oxygen level dependent signal, and the autonomic nerve state data at least comprises any one of a medium oxygen level dependent signal, an electrocardio signal, a pulse signal, a respiratory signal, an oxygen blood signal, a body temperature signal and a skin electric signal.

5. The method of claim 2 or 4, wherein: the feature analysis at least comprises numerical analysis, envelope analysis, time-frequency analysis, entropy analysis, fractal analysis and complexity analysis.

6. The method of claim 2 or 4, wherein: the feature fusion is to select target features with preset feature quantity from a target feature set obtained by the feature analysis and perform weighted calculation to generate the sleep depth feature curve.

7. The method according to claim 2 or 4, wherein the sleep depth characteristic curve is a characteristic curve representing the sleep depth and the phase state of the user in a preset pre-sleep period, a sleep duration and a preset post-sleep period, and the calculation generation method is as follows:

8. The method according to claim 2 or 4, wherein the sleep stage curve is generated by the following steps:

9. The method according to claim 1 or 2, wherein the specific steps of selecting a time-frequency analysis parameter according to the characteristics of the sleep depth characteristic curve and performing time-frequency analysis on the time-frequency analysis parameter, identifying an ultralow frequency trend component, and extracting the sleep trend characteristic curve further include:

10. The method of claim 9, wherein: the selection of the time-frequency analysis parameters depends on the characteristic source and combination mode of the sleep depth characteristic curve, the specific method of time-frequency analysis and the preset time window length of time frame segmentation, and at least comprises one of filtering frequency, window length and window function.

11. The method of claim 9, wherein: the method for time-frequency analysis at least comprises any one of time-frequency transformation and time-frequency filtering.

12. The method of claim 11, wherein the time-frequency transform comprises at least any one of a standard fourier transform FT, a fast fourier transform FFT, a short-time fourier transform STFT, an S-transform, a Gabor transform, a hilbert yellow transform HHT, a Wigner-Ville distribution WVD, a smoothed Wigner-Ville distribution SWVD, a pseudo-Wigner-Ville distribution PWVD, a smoothed pseudo-Wigner-Ville distribution SPWVD, a Choi-Williams distribution CWD.

13. The method of claim 11, wherein the time-frequency filtering comprises at least any one of time-domain filtering and frequency-domain filtering; wherein the time domain filtering at least comprises any one of average value filtering, median filtering, kalman filtering and Savitzky-Golay filtering, and the frequency domain filtering at least comprises any one of low-pass filtering, high-pass filtering, band-pass filtering and band-stop filtering.

14. The method of claim 11, wherein the method for identifying the ultralow frequency trend component specifically comprises:

15. The method as recited in claim 14, wherein: the preset ultralow frequency threshold depends on the sampling rate of the target signal, the preset time window length of time frame segmentation and the dynamic regulation effect.

16. The method of claim 14 or 15, wherein: the frequency weighted fusion calculation specifically uses the principle that the weighted weight of the signal is inversely proportional to the center frequency of the signal as a calculation principle, and performs weighted fusion on the target signal set to generate a signal frequency weighted characteristic description curve.

17. The method of claim 14, wherein the specific steps of dynamically predicting the sleep trend profile, extracting a personalized temporal scene sleep adjustment strategy, and dynamically adjusting and assessing the effect of the user's sleep process further comprise:

18. The method of claim 17, wherein: the prediction method of the sleep trend prediction characteristic value at least comprises any one of AR, MR, ARMA, ARIMA, SARIMA, VAR and deep learning.

19. The method of claim 17, wherein: the individual time phase scene sleep regulation strategy at least comprises a sleep scene, a sleep time phase, a regulation mode, an execution mode, a regulation method, a regulation intensity, a regulation time point, a duration, a target regulation value and device controlParameters; wherein the modulation means comprises at least vocal stimulation, ultrasound stimulation, light stimulation, electrical stimulation, magnetic stimulation, temperature stimulation, humidity stimulation, tactile stimulation, and/or the like

20. The method of claim 19, wherein: the sleep trend regulating peripheral equipment at least comprises a vocal music stimulation equipment, an ultrasonic stimulation equipment, a light stimulation equipment, an electric stimulation equipment, a magnetic stimulation equipment, a temperature stimulation equipment, a humidity stimulation equipment, a touch stimulation equipment and a touch stimulation equipment

21. The method of any one of claims 17-20, wherein: the specific calculation formula of the dynamic adjustment effect coefficient is specifically as follows:

；

wherein ,

the effect coefficient is dynamically adjusted; />

For information of biological state of userAnd the related preset user personality correction coefficients.

22. The method of claim 21, wherein: the user biological status information includes at least gender, age, occupation, health status, disease status, and education level.

23. The method of claim 17, wherein: the dynamic adjustment effect coefficient is reversely applied to the filtering frequency in the time-frequency analysis parameter, the dynamic adjustment of window length, the adjustment of the weight of the decomposed signal in the frequency weight fusion calculation and the generation of the personalized time phase scene sleep adjustment strategy.

24. The method of claim 1 or 14, wherein: the specific steps of dynamically calculating and extracting the sleep trend index, the time phase trend correlation coefficient, the time phase trend distribution characteristic and the comprehensive regulation effect index according to the sleep stage curve, the sleep depth characteristic curve, the sleep trend characteristic curve and the dynamic regulation effect further comprise the following steps:

25. The method of claim 24, wherein the method for calculating the sleep trend index specifically comprises:

5) And calculating the product of the sleep trend characteristic value, a preset method correction coefficient corresponding to a time-frequency analysis method and a preset user individual correction coefficient related to the user biological state information to generate the sleep trend index.

26. The method of claim 25, wherein the sleep trend index is calculated by the formula:

；

wherein ,

for the sleep trend index, +.>

、/>

27. The method of claim 24, wherein the method of correlation calculation comprises at least any one of a coherence analysis, a pearson correlation analysis, a jaccard similarity analysis, a linear mutual information analysis, a linear correlation analysis, a euclidean distance analysis, a manhattan distance analysis, and a chebyshev distance analysis.

28. The method of claim 24, wherein the phase distribution statistics are obtained by performing a numerical distribution statistical analysis on the sleep trend feature values in the sleep trend feature curve according to sleep phase intervals in the sleep interval curve.

29. The method of claim 28, wherein the phase trend distribution characteristics include at least any one of mean, root mean square, maximum, minimum, variance, standard deviation, coefficient of variation, kurtosis, and skewness.

30. The method according to claim 1 or 14, wherein the specific steps of creating and updating a user personalized sleep trend database and dynamically optimizing the personalized time phase scene sleep adjustment strategy, and generating a sleep trend quantification and adjustment report according to a preset report period further comprise:

31. The method of claim 30, wherein the user personalized sleep trend database comprises at least user biological status information, sleep scenes, sleep stage curves, sleep depth profiles, sleep trend indices, phase trend correlation coefficients, phase trend distribution profiles, adjustment patterns, execution patterns, adjustment methods, target adjustment value curves, adjustment intensity curves, and dynamic adjustment effect curves.

32. The method of claim 30, wherein: the sleep trend quantification and adjustment report at least comprises user biological state information, a sleep scene, a sleep stage curve, a sleep depth characteristic curve, a sleep trend index, a time phase trend correlation coefficient, a time phase trend distribution characteristic, a comprehensive adjustment effect index, a dynamic adjustment effect curve, user sleep trend quantification and adjustment summary.

33. The sleep stability quantification and adjustment system based on time-frequency analysis is characterized by comprising the following modules:

34. The system of claim 33, wherein the status detection analysis module further comprises the following functional units:

35. The system of claim 33, wherein the feature curve extraction module further comprises the following functional units:

36. The system of claim 33 or 35, wherein the trend dynamic adjustment module further comprises the following functional units:

37. The system of claim 33 or 35, wherein the trend dynamic quantization module further comprises the following functional units:

38. The system of claim 37, wherein the trend data optimization module further comprises the following functional units:

39. The system of claim 33, wherein the data operation management module further comprises the following functional units:

40. The invention provides a sleep stability quantifying and adjusting device based on time-frequency analysis, which is characterized by comprising the following modules: