Sleep monitoring method, program product, and electronic device

Abstract

The invention discloses a sleep monitoring method, a program product and electronic equipment, and the method comprises the steps: controlling a millimeter-wave radar sensor to transmit an electromagnetic wave signal to a target space region where a target user is located, and obtaining a target echo signal within a target duration; determining target sleep index data related to sleep according to the target echo signal; based on the target sleep index data, forming input of a sleep pressure global detection model, and outputting target sleep pressure global data corresponding to the target sleep index data through the sleep pressure global detection model; and determining a target sleep index value of the target user based on the target sleep pressure global data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, in particular to a sleep monitoring method, a computer program product and an electronic device. BACKGROUND

[0002] Sleep, as a core physiological process of human beings, its quality objective evaluation has been a core issue in the field of sleep science, clinical medicine and health management. Traditionally, sleep evaluation is heavily dependent on subjective reports (such as sleep diary, PSQI questionnaire), which is easily affected by recall bias and subjective perception, and cannot quantify the microstructure of sleep. Therefore, it is necessary to develop objective and accurate detection technology.

[0003] In the field of objective detection, polysomnography, which has been regarded as the gold standard for a long time, can accurately stage sleep (N1, N2, N3, REM) and diagnose sleep disorders by synchronously recording multiple physiological signals such as electroencephalogram, electrooculogram, electromyogram, electrocardiogram, respiration and blood oxygen in a laboratory environment. However, it is high in cost, strange in environment, needs professional technical personnel to operate, is limited to single or short-term monitoring, and is difficult to reflect the long-term sleep pattern of individuals in daily environment.

[0004] With the progress of microelectronics and sensor technology, portable and wearable devices (such as PPG-based optical heart rate bracelet, accelerometer) realize long-term monitoring at home. They mainly use actigraphy and heart rate variability derived parameters to estimate sleep-wake cycle and sleep quality. This kind of technology greatly improves the convenience and ecological validity of monitoring, and becomes the cornerstone of consumer health market. However, its limitation is that the indicators are homogeneous, most devices provide general and rough sleep scores, and lack personalized background and deep physiological insight. SUMMARY

[0005] In order to solve the existing technical problems, the present application provides a sleep monitoring method, a computer program product and an electronic device, which can accurately quantify sleep quality.

[0006] In a first aspect, a sleep monitoring method is provided, comprising: controlling the millimeter wave radar sensor to send electromagnetic wave signals to a target space region where a target user is located, and acquiring target echo signals within a target time period; determining target sleep index data related to sleep according to the target echo signals; forming an input of a sleep pressure global detection model based on the target sleep index data, and outputting target sleep pressure global data corresponding to the target sleep index data through the sleep pressure global detection model; determining a target sleep index value of the target user based on the target sleep pressure global data.

[0007] In a second aspect, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the sleep monitoring method according to any of the embodiments of the present application.

[0008] In a third aspect, an electronic device is provided, comprising a memory and a processor, wherein the memory stores a computer program which, when executed by the processor, causes the processor to perform the sleep monitoring method according to any of the embodiments of the present application.

[0009] The present application controls the millimeter wave radar sensor to send electromagnetic wave signals to a target space region where a target user is located, and acquires target echo signals in a target time period; according to the target echo signals, target sleep index data related to sleep is determined; based on the target sleep index data, input of a sleep pressure global detection model is formed, and target sleep pressure global data corresponding to the target sleep index data is output through the sleep pressure global detection model. Since the model is fused through multiple indexes, the sleep state of the target user in the target time period is comprehensively evaluated, and the evaluation result represented by the target sleep pressure global data is more comprehensive and reliable. Based on the target sleep pressure global data, a target sleep index value of the target user is determined, which is more accurate. Therefore, the sleep quality can be accurately quantified. BRIEF DESCRIPTION OF DRAWINGS

[0010] Figure 1 FIG. 1 is a diagram of an application environment of the sleep monitoring method in an embodiment; Figure 2 FIG. 2 is a flowchart of the sleep monitoring method in an embodiment; Figure 3 FIG. 3 is a flowchart of outputting target sleep pressure global data corresponding to target sleep index data in the sleep monitoring method in an embodiment; Figure 4 FIG. 4 is a schematic diagram of fitting the number of times of turning over sampling data using different statistical models in an embodiment; Figure 5 FIG. 5 is a schematic diagram of fitting the length of time of being awake sampling data using different statistical models in an embodiment; Figure 6 FIG. 6 is a schematic diagram of a sleep monitoring device in an embodiment; Figure 7 FIG. 7 is a schematic diagram of an electronic device in an embodiment. DETAILED DESCRIPTION

[0011] The technical solutions of the present application will be further described in detail below in combination with the accompanying drawings and specific embodiments.

[0012] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description herein is for describing particular embodiments only and is not intended to be limiting of the application. The use herein of the terms "and / or" includes a set of one or more associated listed items.

[0013] In the following description, references are made to "some embodiments" which describe a subset of all possible embodiments, but it is to be understood that "some embodiments" can be the same subset or a different subset of all possible embodiments and can be combined with each other, without conflicts.

[0014] Referring to Figure 1 , an application environment diagram of a sleep monitoring method in an embodiment. The sleep monitoring method is applied in an electronic device 10, the electronic device 10 comprising a millimeter wave radar sensor 12 and a processor 13. The millimeter wave radar sensor 12 is used to send electromagnetic wave signals to a target space area. Since the target space area may have a human body, other objects and the like, the multiple targets will reflect the electromagnetic wave signals, and the millimeter wave radar sensor 12 can receive the echo signals reflected by the multiple targets. The processor 13 detects the sleep quality of the user in the field of view of the millimeter wave radar sensor according to the echo signals detected by the millimeter wave radar sensor 12.

[0015] The millimeter wave radar sensor 12 is a radar system that uses millimeter wave frequency band electromagnetic waves to detect targets. The wavelength of the millimeter wave is between 1 mm and 10 mm, between the microwave and the terahertz wave. The millimeter wave radar sensor transmits millimeter wave signals and receives signals reflected by targets, thereby obtaining sleep detection data of the target user.

[0016] The processor 13 can be one or more. When there are multiple processors 13, the multiple processors can be integrated on one chip or independently arranged on each chip. The electronic device 10 is a device installed with a millimeter wave radar sensor, which can include computing devices (e.g., desktop computers, laptop computers, tablet computers, handheld computers, smart speakers, servers, etc.), terminal devices (e.g., mobile phones, etc.), mobile devices, monitoring devices, wearable devices (e.g., a pair of smart glasses or a smart watch), various infrared imaging devices, smart table lamps or the like.

[0017] The technical frontier of sleep monitoring is moving from generalizability to individualization. This has triggered two key shifts: first, from single metrics (e.g., total sleep time) to multi-dimensional index fusion analysis (e.g., sleep stress index, sleep fragmentation index, cardiopulmonary coupling coordination), to more comprehensively characterize sleep stability and recovery. Second, from comparison with population norms to establishing individual dynamic baselines. Research has found that individual normal sleep intervals vary significantly (e.g., one person's normal deep sleep percentage is 15%, while another's is 25%), and that departing from individual baselines using universal threshold assessments has limited significance.

[0018] Therefore, the core challenge and opportunity of the next generation of sleep quality detection technology is: how to use multi-modal data obtained through long-term and convenient monitoring, through advanced signal processing and machine learning algorithms, to build personalized sleep physiological models and dynamic baselines, and on this basis, to achieve early and accurate identification and attribution of sleep abnormalities. This requires technology not only to record sleep, but also to understand an individual's unique sleep patterns, thereby providing truly individualized health insights and intervention guidance.

[0019] See Figure 2 The flowchart of the sleep monitoring method provided by an embodiment of the present application. The sleep monitoring method is applied to an electronic device, and the sleep monitoring method includes the following steps: S11, controlling the millimeter wave radar sensor to send an electromagnetic wave signal to a target space area where a target user is located, and acquiring a return signal within a target time length.

[0020] In this embodiment, the millimeter wave radar sensor is controlled to send an electromagnetic wave signal to the target space area and receive a return signal reflected by the target space area. The target space area indicates the area to which the signal of the millimeter wave radar sensor can be transmitted. Since the installation position of the millimeter wave radar sensor in the electronic device, the target space area is related to the installation position and the parameters of the millimeter wave radar sensor itself. When the electronic device is in an open state, the millimeter wave radar sensor is controlled to send an electromagnetic wave signal.

[0021] The target time length can be a preset time length, and the sleep monitoring method provided by the present application is executed to detect sleep quality every preset target time length after the electronic device is in an open state. For example, the target time length can be 15 minutes, half an hour, 1 hour, one night, etc. The target return signal is the return data detected within the target time length. By analyzing the target return signal, the user's sleep can be detected.

[0022] The millimeter wave radar sensor actively emits a series of continuous frequency modulation electromagnetic waves of specific frequency (usually 30 GHz to 300 GHz) to the target space region where the target user is located through the radio frequency front end. These millimeter wave band radio signals propagate at the speed of light, and after penetrating light and thin media such as bedclothes, they are reflected when encountering the surface of the human body (especially the chest and abdomen) and the surrounding static environment objects. The sensor synchronously receives mixed echo signals containing rich time delay and phase information, which include not only strong static background reflections produced by walls, furniture, and the like, but also weak Doppler modulation signals caused by human body breathing, body movement, and even heartbeat micro-motions. Through mixing, filtering, and digital signal processing of the echo signals, the system can accurately separate the dynamic components modulated by life activities, thereby realizing non-contact, high-sensitivity sensing of respiratory frequency, heartbeat waveform, and limb movement.

[0023] S12, determining target sleep index data related to sleep according to the target echo signal.

[0024] In the embodiment, the target sleep index data represents sleep-related index data extracted from the target echo signal. Sleep detection data related to sleep can be extracted from the target echo signal data first, wherein the sleep detection data includes but is not limited to at least one of the following: respiratory data, body movement data, heart function data, in-bed and out-of-bed data, and the like. Through decoupling and analysis of the target echo signal, multi-dimensional structured physiological and behavioral data directly related to sleep are extracted and calculated. This process first converts the original radio frequency signal into a three-dimensional data cube containing time, distance, and velocity information using signal processing techniques such as continuous wave Doppler analysis or frequency modulation continuous wave distance calculation. Then, through a static clutter suppression algorithm, the background reflections of static objects such as walls and furniture are filtered out, and the dynamic signal components produced by human life activities are retained, thereby obtaining sleep detection data. Since the sleep detection data represents the detection data within a target time period, analyzing the sleep detection data can obtain the target sleep index data.

[0025] The target sleep index data includes, but is not limited to, the number of times of turning over, the number of times of gross movement, the sleep ratio, the deep sleep duration, the sleep-in time, the wake-up duration, the number of times of waking up, the sleep-in duration, the position standard deviation, and the REM time in the target duration. The sleep ratio represents the ratio of the total duration of the deep sleep duration and the shallow sleep duration to the in-bed duration. The number of times of gross movement represents the number of times of body movement with an amplitude greater than a preset body movement amplitude. The position standard deviation represents the statistical dispersion of the position movement of the body center of mass or the main reflection point in space during sleep. The greater the position standard deviation, the greater the range of body movement in space and the greater the change in position. The deep sleep duration represents the total duration of the deep sleep stage in the target duration. The sleep-in time represents the clock time of actually entering the sleep state. The wake-up duration represents the total duration of being in the state of going to bed and being in the wake-up state. The number of times of waking up represents the number of times of intermediate awakening in the target duration. The sleep-in duration represents the duration from the time of starting to go to bed to the sleep-in time. The rapid eye movement (REM) time represents the total duration of the rapid eye movement sleep stage in the target duration.

[0026] S13, based on the target sleep index data, forming an input of a sleep pressure global detection model, and outputting target sleep pressure global data corresponding to the target sleep index data through the sleep pressure global detection model.

[0027] In this embodiment, the sleep pressure global detection model is a sleep detection model obtained based on historical data of people of different sleep quality types. The sleep quality types include, but are not limited to, normal sleep type, light sleep type, and heavy sleep type. The light sleep type represents shallow sleep, short sleep time, long sleep-in duration, etc. The heavy sleep type represents deep sleep, long sleep time, short sleep-in duration, etc. Since the historical data of people of different sleep types is a kind of priori data, the sleep pressure global detection model represents a comprehensive evaluation system based on data driving and multi-index fusion. The sleep pressure global detection model learns the historical data of people of different sleep quality types to establish a mapping relationship from multi-dimensional sleep features (such as the number of times of turning over and the deep sleep duration) to the overall sleep pressure grade value, thereby realizing objective and quantitative evaluation of the sleep pressure level of an individual. The target sleep pressure global data represents the sleep pressure data of the target user obtained according to the sleep pressure global detection model, which is quantitative data quantified based on a general model. The target sleep pressure global data represents a comprehensive, quantitative, and interpretable evaluation result of the sleep state of the target user in the target duration.

[0028] S14, determining a target sleep index value of the target user based on the target sleep pressure global data.

[0029] In the embodiment, since the target sleep stress global data represents the quantitative data of the sleep state of the target user in the target time length, the overall sleep state of the user can be determined according to the quantitative data. The target sleep index value represents the global sleep state of the target user in the target time length.

[0030] In the above embodiment, the millimeter wave radar sensor is controlled to send an electromagnetic wave signal to a target space region where the target user is located, and a target echo signal in a target time length is acquired; target sleep index data related to sleep is determined according to the target echo signal; input of a sleep stress global detection model is formed based on the target sleep index data, and target sleep index data corresponding to target sleep stress global data is output by the sleep stress global detection model. Since the model is fused by multiple indexes, the sleep state of the target user in the target time length is comprehensively evaluated, the evaluation result represented by the target sleep stress global data is more comprehensive and more reliable, and the target sleep index value of the target user is more accurate based on the target sleep stress global data, so that the sleep quality can be accurately quantified.

[0031] In some embodiments, the target sleep index data includes multiple target sleep global index data, and the sleep stress global detection model includes multiple global index detection models, wherein each global index corresponds to a global index detection model. As shown in Figure 3 Figure 3 is a flowchart of an embodiment of the sleep monitoring method for outputting target sleep stress global data corresponding to target sleep index data. The output of the target sleep stress global data corresponding to the target sleep index data by the sleep stress global detection model includes the following steps: S31, for each global index, target sleep global index data corresponding to each global index is acquired from the target sleep index data.

[0032] In the embodiment, the global index is used for detection of the global index detection model. The global index detection model is a general standard model. The global index includes but is not limited to at least one of the following: number of times of turning over, number of times of gross movement, sleep ratio, and deep sleep time length. The target sleep global index data represents data of the global index in the target time length. The target sleep global index data includes but is not limited to at least one of the following: target number of times of turning over, target number of times of gross movement, target sleep ratio, and target deep sleep time length.

[0033] S32, the target sleep global index data corresponding to each global index is taken as input of the global index detection model corresponding to each global index, and target global index evaluation data corresponding to each global index is output.

[0034] ​In this embodiment, each global indicator corresponds to a global indicator detection model. For example, the number of rollovers corresponds to a rollover detection model, the number of large body movements corresponds to a large body movement detection model, and so on. The target global indicator evaluation data corresponding to the global indicator represents the quantitative data of that global indicator. The target global indicator evaluation data includes at least one of the following: the target global indicator level value, and the confidence level corresponding to the target global indicator level value.

[0035] Optionally, the method further includes: The method further includes: Obtain a sampling dataset, wherein the sampling dataset includes global indicator sampling data of various global indicators for people with different sleep quality types; For any global indicator detection model, the global indicator sampling data set of the global indicator is obtained from the sampling dataset. Based on the global indicator sampling data set of the global indicator, histogram statistics are performed to obtain the sampling histogram distribution of the global indicator. Based on multiple statistical models, the sampling histogram distribution of the global index is fitted to obtain multiple estimated global index probability density models and fitting evaluation indicators corresponding to each estimated global index probability density model. Based on the fitting evaluation index corresponding to each estimated global indicator probability density model, multiple estimated global indicator probability density models are evaluated, and the estimated global indicator probability density model with the highest evaluation value is taken as the global indicator detection model corresponding to the global indicator. The global indicator detection model includes a global indicator level value and a confidence interval corresponding to the global indicator level value.

[0036] In this embodiment, the global indicator sampling data set includes multiple global indicator sampling data sets for that global indicator. The global indicator sampling data represents the quantified value of that global indicator. For example, the sleep ratio corresponds to the sleep ratio sampling value, and the number of times a person turns over corresponds to the number of times they turn over sampling value. The statistical model includes, but is not limited to, at least one of the following: normal distribution, Poisson distribution, negative binomial distribution, gamma distribution, Weibull distribution, log-normal distribution, etc. Figure 4 As shown, Figure 4 This is a schematic diagram illustrating the fitting of rolling-over frequency sampling data using different statistical models in one embodiment. Normal distribution, lognormal distribution, and gamma distribution were used to fit the rolling-over frequency sampling data set, resulting in the following... Figure 4 The diagram shows three probability density models for the number of times a child rolls over. Each model has a corresponding fit evaluation index. For example... Figure 4As shown, the fitting evaluation metrics are KS_Statistic and P_Value. KS_Statistic is the KS statistic, the core statistic of the Kolmogorov-Smirnov test, and P_Value (Probability Value) represents the significance probability value. A larger P_Value indicates a more accurate probability distribution, and a smaller KS_Statistic indicates a more accurate probability distribution. When multiple fitting evaluation metrics exist, different weights can be assigned to each metric. Based on the multiple fitting evaluation metrics corresponding to each estimated global indicator probability density model, the evaluation value of each estimated global indicator probability density model is calculated. The estimated global indicator probability density model with the highest evaluation value is taken as the global indicator detection model corresponding to the global indicator.

[0037] The global indicator detection model corresponding to a global indicator includes multiple global indicator level values ​​for that global indicator, and the confidence intervals corresponding to each global indicator level value. Taking the number of rollovers as an example, the representation of the rollover detection model is as follows: As shown in the table above, the rolling-over frequency detection model is divided into 7 rolling-over frequency levels. Each rolling-over frequency level corresponds to a rolling-over frequency level value, and each rolling-over frequency level value corresponds to its own confidence interval. By inputting the target sleep rolling-over frequency data corresponding to the rolling-over frequency into the rolling-over frequency detection model, the target rolling-over frequency level value and the corresponding confidence interval can be determined.

[0038] Based on the sampled dataset, through histogram statistics, multi-model fitting and optimization, a rigorous, interpretable and robust statistical modeling system was constructed to establish a detection model with confidence intervals for each global indicator. The data itself, through histograms, reveals the true distribution of the global indicator's sampled data set. Subsequently, multiple candidate statistical models were introduced for competitive fitting, and objective evaluation and optimization were performed using quantitative fitting indices (such as AIC, BIC, and KS tests). Ultimately, the global indicator detection model for each global indicator represents the optimal mathematical description of the inherent random law of that global indicator. This method maximizes the theoretical fit between the global indicator detection model and the global indicator's sampled data set, laying a solid mathematical foundation for all subsequent analyses.

[0039] The global indicator detection model established using the above method outputs far more than a simple average. The global indicator level, combined with confidence intervals, constitutes a comprehensive evaluation scale. For example, it not only determines that someone's sleep duration falls within a level of 5, but also indicates the precise range of this judgment through confidence intervals (e.g., a 95% certainty that their true level lies between level 4.5 and 5.5). This output format upgrades point estimation to interval estimation, making the evaluation results contain robust conclusions based on reliability measures. This enables refined quantitative comparisons between different individuals and at different time points.

[0040] For the target users, the global indicator detection model obtained based on this method possesses interpretability and reliability. When the system provides a global indicator level value, it is backed by a clear statistical distribution, an optimal probability model, and a calculable confidence level.

[0041] S33. Based on the target global indicator evaluation data corresponding to each global indicator, determine the target sleep pressure global data corresponding to the target sleep indicator data.

[0042] In this embodiment, the method further includes: generating the global sleep pressure detection model based on the global indicator detection model corresponding to each of the global indicators. The global sleep pressure detection model includes multiple global level values, a global sleep index range corresponding to each global level value, and a sleep index difference corresponding to each global level value. For a given global level value, the global sleep index range can be determined based on the confidence interval corresponding to the global indicator level value at the same level as that global level value. The sleep index difference represents the difference between the maximum and minimum values ​​within the global sleep index range.

[0043] For example, the generated global sleep pressure detection model is shown in the table below: As shown in the table above, the global detection model is divided into seven global levels, and each global level corresponds to a global level value.

[0044] Optionally, the target global indicator evaluation data includes at least one of the following: target global indicator level value; the target sleep pressure global data includes at least one of the following: target global level value; the sleep pressure global detection model further includes at least one of the following: multiple global level values, global sleep index range corresponding to each global level value, and sleep index difference corresponding to each global level value. The determination of the target sleep pressure global data corresponding to the target sleep index data based on the target global index evaluation data corresponding to each of the global indices includes at least one of the following: The target global indicator level values ​​corresponding to each of the global indicators are accumulated to obtain a first accumulated level value. The first accumulated level value is divided by a preset number to obtain a target remainder. The first accumulated level value is added to the target remainder and then divided by the total number of global indicators to obtain the target global level value. Based on the target global level value and the global sleep index range corresponding to the target global level value, the target global sleep index corresponding to the target global level value is determined, and the target sleep index difference is determined based on the target global level value.

[0045] In this embodiment, the global indicators, including the number of times the body turns over, the number of gross motor movements, the sleep ratio, and the deep sleep duration, are used as examples to illustrate the above steps, i.e., the total number of global indicators. The indicators have been determined. The formula for calculating the target global level value is as follows, where the preset value is 2.

[0046] By inputting the target number of turning over data corresponding to the number of turning over during sleep into the turning over frequency detection model, the target turning over frequency level value and its corresponding confidence interval can be determined. Similarly, by inputting the target number of large body movements data corresponding to the number of large body movements during sleep into the large body movement frequency detection model, the target number of large body movements level value and its corresponding confidence interval can be determined. Likewise, by inputting the target sleep ratio data corresponding to the sleep ratio detection model, the target sleep ratio level value and its corresponding confidence interval can be determined. Finally, by inputting the target deep sleep duration data corresponding to the deep sleep duration detection model, the target deep sleep duration level value and its corresponding confidence interval can be determined. Substituting the target turning over frequency level value, target large body movement level value, target sleep ratio level value, and target deep sleep duration level value into the above formulas, the target global level value can be obtained. Based on the global detection model, the target global sleep index and target sleep index difference corresponding to the target global level value can be obtained.

[0047] In the above embodiments, for group data of different sleep quality types in the sampled dataset, histogram statistics, multi-model fitting and optimization are used to finally establish a global indicator detection model with confidence intervals for each global indicator. This reduces human experience. Through big data analysis, various global indicator detection models are obtained, thus obtaining a global detection model suitable for various sleep quality types. When the global indicator detection model and the global detection model are applied to detection, more accurate global data on target sleep pressure can be obtained. Multiple candidate statistical models are introduced for competitive fitting, and objective evaluation and optimization are carried out through quantitative fitting indicators (such as AIC, BIC, KS test). Finally, the global indicator detection model corresponding to the global indicator is the optimal mathematical description of the inherent random law of the global indicator, which facilitates subsequent accurate and refined quantification of sleep data of different individuals and different time points.

[0048] In some embodiments, determining the sleep quality of the target user based on the target sleep pressure global data includes: Based on the target sleep index data, the input of the personalized sleep pressure detection model is formed, and the personalized sleep pressure data corresponding to the target sleep index data is output through the personalized sleep pressure detection model. Based on the global data and personalized data of the target sleep pressure, the sleep quality of the target user is determined.

[0049] In this embodiment, the global sleep pressure detection model is obtained by fitting group data of different sleep quality types, but it cannot truly reflect the personalized sleep data of individual users. The personalized sleep pressure detection model is obtained by fitting the historical sleep data of the target user. The personalized sleep pressure detection model represents a dynamic, evolving digital profile and assessment system of personal sleep health that is unique to a specific user. The personalized sleep pressure detection model is not a general model, but a unique sleep pressure detection model built for the target user by continuously learning from the user's personal historical sleep data.

[0050] Personalized target sleep pressure data represents a personalized quantification of a target user's sleep patterns within a target duration. Global target sleep pressure data is a quantitative description of a target user's sleep patterns within a target duration based on a universal sleep pressure detection model. By integrating both group and individual reference dimensions, global target sleep pressure data calibrates users within a broad range of healthy population norms, providing an objective and stable external benchmark that reveals the target user's global sleep index within the universal model. Personalized data focuses on the dynamic changes in a user's own sleep patterns, depicting their fluctuation trajectory relative to their personal historical baseline. The combination of the two effectively avoids the shortcomings of a single perspective: looking only at the global sleep index may overlook a user's unique physiological baseline (for example, even with improvement, a chronic insomniac's score may still lag behind the general population); looking only at personalized data may fail to detect long-term risks that are slowly deteriorating but still within the individual's normal range.

[0051] Optionally, the target sleep index data includes: multiple target individual index data, and the personalized sleep pressure detection model includes multiple individual index detection models, wherein each individual index corresponds to one individual index detection model; The step of outputting personalized target sleep pressure data corresponding to the target sleep index data through the personalized sleep pressure detection model includes: For various individual indicators, the target individual indicator data corresponding to each individual indicator is obtained from the target sleep indicator data; The target individual indicator data corresponding to each individual indicator is used as the input to the individual indicator detection model corresponding to each individual indicator, and the target individual indicator evaluation data corresponding to each individual indicator is output.

[0052] In this embodiment, individual metrics include, but are not limited to, at least one of the following: sleep onset time, awakening duration, number of awakenings, sleep onset duration, and location standard deviation. The total number of individual metrics is N². Individual metrics are used to measure personalized sleep profile data of users. These individual metrics can be changed for different target users. For example, by statistically analyzing the data of various individual metrics in the sleep data of a target user, metrics with fluctuations exceeding a preset fluctuation value are used as the target user's individual metrics. Larger fluctuation data better reflects the user's sleep fluctuations within the sampling period. Each individual metric corresponds to an individual metric detection model used to detect the individual metric level value.

[0053] Optionally, the method further includes: Acquire the sleep data of the target user within the target time period to obtain a sleep sample set, which includes individual indicator sampling data corresponding to each individual indicator; For any of the individual indicator detection models, the individual indicator sampling data set of the individual indicator is obtained from the sleep sample set, and histogram statistics are performed based on the individual indicator sampling data set to obtain the sampling histogram distribution of the individual indicator. Based on multiple statistical analysis models, the sampling histogram distribution of the individual indicators is fitted to obtain multiple estimated individual indicator probability density models and fitting evaluation indicators corresponding to various individual global indicator probability density models. Based on the fitting evaluation index corresponding to each estimated individual indicator probability density model, multiple estimated individual indicator probability density models are evaluated, and the estimated individual indicator probability density model with the highest evaluation value is taken as the individual indicator detection model corresponding to the individual indicator. The individual indicator detection model includes the individual indicator index value and the confidence interval corresponding to the individual indicator index value.

[0054] In this embodiment, the target time period may include multiple different time periods, such as multiple different nights. The sleep sample set represents the target user's personalized historical sleep data. The target user's own sleep sample set is utilized. The individual indicator sampling data set includes multiple sets of individual indicator sampling values. For example, sleep latency corresponds to multiple sets of sleep latency sampling values. The statistical model includes, but is not limited to, at least one of the following: normal distribution, Poisson distribution, negative binomial distribution, gamma distribution, Weibull distribution, log-normal distribution, etc. Figure 5 As shown, Figure 5 This is a schematic diagram illustrating the use of different statistical models to fit lucidity duration sampling data in one embodiment. Normal distribution, lognormal distribution, and gamma distribution were used to fit the lucidity duration sampling data set, resulting in the following... Figure 5 The diagram shows three probability density models of lucid duration. Each lucid duration probability density model corresponds to a fit evaluation index. For example... Figure 5 As shown, the fitting evaluation metrics are KS_Statistic and P_Value. KS_Statistic is the KS statistic, the core statistic of the Kolmogorov-Smirnov test, and P_Value (Probability Value) represents the significance probability value. A larger P_Value indicates a more accurate probability distribution, and a smaller KS_Statistic indicates a more accurate probability distribution. When multiple fitting evaluation metrics exist, different weights can be assigned to each metric. Based on the multiple fitting evaluation metrics corresponding to each estimated individual indicator probability density model, the evaluation value of each estimated individual indicator probability density model is calculated. The estimated individual indicator probability density model with the highest evaluation value is used as the individual indicator detection model corresponding to the individual indicator.

[0055] The individual indicator detection model for an individual indicator includes multiple individual indicator level values ​​and the confidence intervals corresponding to each individual indicator level value. Taking lucidity duration as an example, the lucidity duration detection model is represented as follows: As shown in the table above, the wakefulness duration detection model is divided into three wakefulness duration levels, each with a corresponding wakefulness duration level value and a corresponding confidence interval. By inputting the target sleep wakefulness duration data corresponding to the wakefulness duration into the wakefulness duration detection model, the target wakefulness duration level value and the corresponding confidence interval can be determined.

[0056] Unlike global models based on group data, this step utilizes the target user's own historical sleep data to build the model. Therefore, the resulting individual indicator detection model essentially quantifies the user's personal normal range or typical pattern of sleep indicators. It captures the user's unique circadian rhythms and habits (such as the user's inherent sleep onset duration distribution, typical deep sleep ratio, etc.), thereby establishing a truly personalized sleep health baseline. This provides an indispensable reference standard for subsequent detection of personalized deviations in sleep status (i.e., abnormalities relative to their own baseline).

[0057] The entire process, from data collection and histogram analysis to multi-model fitting and evaluation, can be automated. This ensures that building a personalized detection model for each user is efficient, standardized, and repeatable, requiring no manual intervention or expert experience for individual configuration. This systematic approach makes it possible to provide accurate and personalized sleep health services to users on a large scale, while ensuring consistency in model quality across different users, thus providing technical feasibility for productization.

[0058] In summary, the advantages of this approach are: it uses objective data to build a probabilistic model that best fits each user's individual sleep characteristics and produces quantitative indicators that include uncertainty measures, thus providing core technical support for achieving truly accurate, robust, and scalable personalized sleep stress assessment.

[0059] Optionally, the target sleep pressure global data includes the target global sleep index and the target sleep index difference, and the target individual indicator evaluation data corresponding to the individual indicator includes the target individual indicator level value corresponding to the individual indicator; determining the target sleep index value of the target user based on the target sleep pressure global data includes: Based on the target individual indicator level value corresponding to each individual indicator, the average individual indicator index value is calculated, and the target sleep index difference is multiplied by the average individual indicator index value to obtain the total target individual indicator index value. The target sleep index value is obtained by subtracting the total index value of the target individual index from the target global sleep index.

[0060] In this embodiment, we will use six individual indicators—sleep onset time, awakening duration, number of awakenings, sleep onset duration, positional standard deviation, and REM sleep time—as an example for explanation. The total number of individual indicators is N²=6. The target sleep index value can be calculated using the following formula.

[0061] in Indicates the target sleep index value. Indicates the target global sleep index. This indicates a poor target sleep index. This indicates the target sleep time level value corresponding to the time of falling asleep. This indicates the target lucidity level value corresponding to the duration of lucidity. This indicates the target level of alertness corresponding to the number of times the person regained consciousness. This indicates the target sleep duration level value corresponding to the sleep latency. This represents the target position standard deviation level value corresponding to the position standard deviation. This indicates the target REM time level value corresponding to the REM time. This represents the average individual indicator index value.

[0062] In the above embodiments, the individual indicator detection model corresponding to each individual indicator is fitted based on the target user's historical sleep data, which better reflects the sleep profile of each user. Based on the individual indicator detection model, the target individual indicator level value corresponding to various individual indicators can be obtained. Based on the target global sleep index, the target sleep index difference, and the target individual indicator level values ​​corresponding to various individual indicators, the target sleep index value is obtained. Since the target global sleep index and the target sleep index difference are detected based on a general sleep pressure global detection model, the sleep pressure global detection model provides an objective and stable external yardstick, revealing the target user's global sleep index on the general model. The target individual indicator level value corresponding to various individual indicators is a personalized data, which focuses on the dynamic changes of the user's own sleep patterns, depicting its fluctuation trajectory relative to the individual's historical baseline. The combination of the two effectively avoids the defects of a single perspective: looking only at the global sleep index may ignore the user's unique physiological baseline; looking only at personalized data may fail to detect long-term risks that are slowly deteriorating but still within the individual's normal range, thus making the quantified target sleep index value more objective and realistic, and improving the accuracy of sleep monitoring.

[0063] In some embodiments, the method further includes at least one of the following: The target sleep index value is displayed on the display terminal.

[0064] Understandably, the above sleep monitoring methods can be applied to products such as sleep monitors, sleep aids, and baby monitors.

[0065] In another aspect, this application provides a computer program product, including a computer program that, when executed by a processor, implements the sleep monitoring method described in any embodiment of this application.

[0066] In the computer program product, the optional implementation form of the program module architecture of the computer program that implements each step of the target recognition method can be a sleep monitoring device.

[0067] Please see Figure 6 One embodiment of this application provides a sleep monitoring device, including: a control module 61, used to control the millimeter-wave radar sensor to send electromagnetic wave signals to the target space area where the target user is located, and to acquire target echo signals within the target duration; a determination module 62, used to determine target sleep index data related to sleep based on the target echo signals; a calculation module 63, used to form the input of a global sleep pressure detection model based on the target sleep index data, and to output target sleep pressure global data corresponding to the target sleep index data through the global sleep pressure detection model; the calculation module 63 is also used to determine the target sleep index value of the target user based on the target sleep pressure global data.

[0068] Optionally, the target sleep index data includes: multiple target sleep global index data, and the sleep pressure global detection model includes multiple global index detection models, wherein each global index corresponds to one global index detection model, and the calculation module 63 is further used for: For each of the global indicators, obtain the target sleep global indicator data corresponding to each of the global indicators from the target sleep indicator data; The target sleep global indicator data corresponding to each of the global indicators are used as the input to the global indicator detection model corresponding to each of the global indicators, and the target global indicator evaluation data corresponding to each of the global indicators are output. Based on the target global indicator evaluation data corresponding to each of the global indicators, the target sleep pressure global data corresponding to the target sleep indicator data is determined.

[0069] Optionally, the target global indicator evaluation data includes at least one of the following: target global indicator level value; the target sleep pressure global data includes at least one of the following: target global level value; the sleep pressure global detection model further includes at least one of the following: multiple global level values, global sleep index range corresponding to each global level value, and sleep index difference corresponding to each global level value. The calculation module 63 is also used for: The target global indicator level values ​​corresponding to each of the global indicators are accumulated to obtain a first accumulated level value. The first accumulated level value is divided by a preset number to obtain a target remainder. The first accumulated level value is added to the target remainder and then divided by the total number of global indicators to obtain the target global level value. Based on the target global level value and the global sleep index range corresponding to the target global level value, the target global sleep index corresponding to the target global level value is determined, and the target sleep index difference is determined based on the target global level value.

[0070] Optionally, the calculation module 63 is also used for: Obtain a sampling dataset, wherein the sampling dataset includes global indicator sampling data of various global indicators for people with different sleep quality types; For any global indicator detection model, the global indicator sampling data set of the global indicator is obtained from the sampling dataset. Based on the global indicator sampling data set of the global indicator, histogram statistics are performed to obtain the sampling histogram distribution of the global indicator. Based on multiple statistical models, the sampling histogram distribution of the global index is fitted to obtain multiple estimated global index probability density models and fitting evaluation indicators corresponding to each estimated global index probability density model. Based on the fitting evaluation index corresponding to each estimated global indicator probability density model, multiple estimated global indicator probability density models are evaluated, and the estimated global indicator probability density model with the highest evaluation value is taken as the global indicator detection model corresponding to the global indicator. The global indicator detection model includes multiple global indicator level values ​​and confidence intervals corresponding to each global indicator level value. The sleep pressure global detection model is generated based on the global indicator detection model corresponding to each of the global indicators.

[0071] Optionally, the calculation module 63 is also used for: Based on the target sleep index data, the input of the personalized sleep pressure detection model is formed, and the personalized sleep pressure data corresponding to the target sleep index data is output through the personalized sleep pressure detection model. Based on the global data and personalized data of target sleep pressure, the target sleep index value of the target user is determined.

[0072] Optionally, the target sleep index data includes: multiple target individual index data, and the personalized sleep pressure detection model includes multiple individual index detection models, wherein each individual index corresponds to one individual index detection model; The calculation module 63 is also used for: For various individual indicators, the target individual indicator data corresponding to each individual indicator is obtained from the target sleep indicator data; The target individual indicator data corresponding to each individual indicator is used as the input to the individual indicator detection model corresponding to each individual indicator, and the target individual indicator evaluation data corresponding to each individual indicator is output.

[0073] Optionally, the target sleep pressure global data includes the target global sleep index and the target sleep index difference, and the target individual indicator evaluation data corresponding to the individual indicator includes the target individual indicator level value corresponding to the individual indicator; the calculation module 63 is further used for: Based on the target individual indicator level value corresponding to each individual indicator, the average individual indicator index value is calculated, and the target sleep index difference is multiplied by the average individual indicator index value to obtain the total target individual indicator index value. The target sleep index value is obtained by subtracting the total index value of the target individual index from the target global sleep index.

[0074] Optionally, the calculation module 63 is also used for: Acquire the sleep data of the target user within the target time period to obtain a sleep sample set, which includes individual indicator sampling data corresponding to each individual indicator; For any of the individual indicator detection models, the individual indicator sampling data set of the individual indicator is obtained from the sleep sample set, and histogram statistics are performed based on the individual indicator sampling data set to obtain the sampling histogram distribution of the individual indicator. Based on multiple statistical analysis models, the sampling histogram distribution of the individual indicators is fitted to obtain multiple estimated individual indicator probability density models and fitting evaluation indicators corresponding to various individual global indicator probability density models. Based on the fitting evaluation index corresponding to each estimated individual indicator probability density model, multiple estimated individual indicator probability density models are evaluated, and the estimated individual indicator probability density model with the highest evaluation value is taken as the individual indicator detection model corresponding to the individual indicator. The individual indicator detection model includes the individual indicator index value and the confidence interval corresponding to the individual indicator index value.

[0075] It will be understood by those skilled in the art that Figure 6 The structure of the sleep monitoring device does not constitute a limitation on the sleep monitoring device. Each module can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module. In other embodiments, the sleep monitoring device may include more or fewer modules than shown in the figures.

[0076] Please see Figure 7 In another aspect of this application, an electronic device 10 is also provided, including a processor 13 and a memory 14. The memory 14 stores a computer program, which, when executed by the processor, causes the processor 13 to perform the steps of the sleep monitoring method provided in any of the above embodiments of this application. The electronic device may include computing devices (e.g., desktop computers, laptop computers, tablet computers, handheld computers, smart speakers, servers, etc.), terminal devices (e.g., mobile phones), wearable devices (e.g., a pair of smart glasses or a smartwatch), smart lamps, or similar devices.

[0077] The processor 13 serves as the control center, connecting various parts of the computer device via various interfaces and lines. It executes software programs and / or modules stored in the memory 14, and accesses data stored in the memory 14 to perform various functions and process data. Optionally, the processor 13 may include one or more processing cores; preferably, the processor 13 may integrate an application processor and a modem processor, wherein the application processor primarily handles the operating system, user pages, and applications, while the modem processor primarily handles wireless communication. It is understood that the modem processor may not be integrated into the processor 13.

[0078] The memory 14 can be used to store software programs and modules. The processor 13 executes various functional applications and data processing by running the software programs and modules stored in the memory 14. The memory 14 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 14 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 14 may also include a memory processor to provide the processor 13 with access to the memory 14.

[0079] In another aspect, this application also provides a storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the sleep monitoring method provided in any of the above embodiments of this application.

[0080] Those skilled in the art will understand that all or part of the processes in the methods provided in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0081] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A sleep monitoring method, characterized by, The application is applied to an electronic device provided with a millimeter wave radar sensor, comprising: controlling the millimeter wave radar sensor to send electromagnetic wave signals to a target space region where a target user is located, and acquiring target echo signals within a target time length; determining target sleep index data related to sleep according to the target echo signals; forming an input of a sleep pressure global detection model based on the target sleep index data, and outputting target sleep pressure global data corresponding to the target sleep index data through the sleep pressure global detection model; determining a target sleep index value of the target user based on the target sleep pressure global data.

2. The sleep monitoring method of claim 1, wherein, The target sleep index data comprises a plurality of target sleep global index data, and the sleep pressure global detection model comprises a plurality of global index detection models, wherein each global index corresponds to a global index detection model, and the outputting of the target sleep pressure global data corresponding to the target sleep index data through the sleep pressure global detection model comprises: for each global index, acquiring target sleep global index data corresponding to each global index from the target sleep index data; taking the target sleep global index data corresponding to each global index as an input of the global index detection model corresponding to each global index, and outputting target global index evaluation data corresponding to each global index; determining the target sleep pressure global data corresponding to the target sleep index data based on the target global index evaluation data corresponding to each global index.

3. The sleep monitoring method of claim 2, wherein, The target global index evaluation data comprises at least one of a target global index level value, and the target sleep pressure global data comprises at least one of a target global level value; the sleep pressure global detection model further comprises at least one of a plurality of global level values, a global sleep index range corresponding to each global level value, and a sleep index difference corresponding to each global level value; The determination of the target sleep pressure global data corresponding to the target sleep index data based on the target global index evaluation data corresponding to each global index comprises at least one of the following: adding up the target global index level values corresponding to each global index to obtain a first accumulated level value, dividing the first accumulated level value by a preset number to obtain a target remainder, adding the first accumulated level value and the target remainder, and then dividing by the total number of global indexes to obtain a target global level value; determining a target global sleep index corresponding to the target global level value according to the target global level value and a global sleep index range corresponding to the target global level value, and determining a target sleep index difference according to the target global level value.

4. The sleep monitoring method of any one of claims 1 to 3, wherein, The method further comprises: acquiring a sampling data set, wherein the sampling data set comprises global index sampling data of various global indexes of people of different sleep quality types; The global index detection model of any global index obtains a global index sampling data set of the global index from the sampling data set, and performs histogram statistics based on the global index sampling data set of the global index to obtain a sampling histogram distribution of the global index; Based on multiple statistical models, the sampling histogram distribution of the global index is fitted to obtain multiple estimated global index probability density models and fitting evaluation indexes corresponding to each estimated global index probability density model; Based on the fitting evaluation indexes corresponding to each estimated global index probability density model, the multiple estimated global index probability density models are evaluated, and the estimated global index probability density model with the highest evaluation value is taken as the global index detection model corresponding to the global index, wherein the global index detection model includes multiple global index level values and confidence intervals corresponding to each global index level value; According to the global index detection model corresponding to each global index, the global sleep stress detection model is generated.

5. The sleep monitoring method of claim 1, wherein, The target sleep index value of the target user is determined based on the target sleep stress global data, which includes: Based on the target sleep index data, an input of a sleep stress personalized detection model is formed, and the target sleep stress personalized data corresponding to the target sleep index data is output through the sleep stress personalized detection model; Based on the target sleep stress global data and the target sleep stress personalized data, the target sleep index value of the target user is determined.

6. The sleep monitoring method of claim 5, wherein, The target sleep index data includes multiple target individual index data, and the sleep stress personalized detection model includes multiple individual index detection models, wherein each individual index corresponds to an individual index detection model; The target sleep stress personalized data corresponding to the target sleep index data is output through the sleep stress personalized detection model, which includes: For each individual index, the target individual index data corresponding to each individual index is obtained from the target sleep index data; The target individual index data corresponding to each individual index is taken as the input of the individual index detection model corresponding to each individual index, and the target individual index evaluation data corresponding to each individual index is output.

7. The sleep monitoring method of claim 6, wherein, The target sleep stress global data includes a target global sleep index and a target sleep index difference, and the target individual index evaluation data corresponding to the individual index includes a target individual index level value corresponding to the individual index; the target sleep index value of the target user is determined based on the target sleep stress global data, which includes: Based on the target individual index level value corresponding to each individual index, an average individual index index value is calculated, the target sleep index difference is multiplied by the average individual index index value to obtain a target individual index total index value; The target global sleep index is subtracted from the target individual index total index value to obtain the target sleep index value.

8. The sleep monitoring method of claim 6, wherein, The method further includes: Obtaining sleep data of the target user in a target time period to obtain a sleep sample set, and the sleep sample set includes individual index sampling data corresponding to each individual index; For the individual index detection model of any of the individual indexes, obtain an individual index sample data set of the individual index from the sleep sample set, perform histogram statistics based on the individual index sample data set to obtain a sample histogram distribution of the individual index; Based on multiple statistical analysis models, fit the sample histogram distribution of the individual index to obtain multiple estimated individual index probability density models and fitting evaluation indexes corresponding to various individual global index probability density models; Based on the fitting evaluation indexes corresponding to each estimated individual index probability density model, evaluate the multiple estimated individual index probability density models, and take the estimated individual index probability density model with the highest evaluation value as the individual index detection model corresponding to the individual index, wherein the individual index detection model comprises an individual index index value and a confidence interval corresponding to the individual index index value.

9. A computer program product comprising a computer program, characterized in that, The computer program is executed by a processor to implement the sleep monitoring method of any of claims 1-8.

10. An electronic device, comprising: A computer program is stored in a memory and executed by a processor to make the processor execute the sleep monitoring method of any of claims 1-8.

Images