A non-invasive sleep quality detection method and system fusing autonomic nervous monitoring

Abstract

The application provides a non-invasive sleep quality detection method and system fused with autonomic nerve monitoring, and relates to the technical field of biomedical sensing and sleep health monitoring. The method comprises the following steps: collecting night non-invasive physiological data of a to-be-detected object, wherein the night non-invasive physiological data comprises an electrocardiosignal and a body movement posture signal; constructing a mean heart beat prediction model based on Laguerre according to the electrocardiosignal and the body movement posture signal; performing parameter estimation on the mean heart beat prediction model to determine a time-varying Laguerre coefficient; calculating an autonomic nerve balance index according to the time-varying Laguerre coefficient; performing feature extraction on the night non-invasive physiological data to obtain movement features and heart rate statistical features; combining the movement features, the heart rate statistical features and the autonomic nerve balance index to obtain an autonomic nerve enhancement feature vector; inputting the autonomic nerve enhancement feature vector into a sleep staging analysis module to output a sleep stage; and generating a non-invasive sleep quality detection report based on the sleep stage.

Description

Technical Field

[0001] This invention relates to the field of biomedical sensing and sleep health monitoring technology, and in particular to a non-invasive sleep quality detection method and system that integrates autonomic nervous system monitoring. Background Technology

[0002] With the accelerating pace of modern life and increased awareness of health management, nighttime sleep quality has become a crucial indicator for health monitoring. Non-invasive physiological signal monitoring technology offers a new approach to sleep research, enabling the acquisition of continuous electrocardiogram (ECG), body movement and posture signals, and other nocturnal physiological parameters without disturbing the subjects. By collecting and analyzing multimodal physiological data, this technology provides a rich data foundation for studying autonomic nervous activity, further supporting sleep stage identification and health status assessment, and offering data support for personalized health management and disease prevention.

[0003] Existing technologies classify sleep stages and assess sleep quality through electrocardiogram (ECG) signal analysis, motion monitoring, and autonomic nervous system (ASS) index calculation. By constructing a heart rate mean prediction model based on the Laguerre function and combining it with ASS balance index and motion characteristics, continuous nighttime sleep stages can be determined. This method shows good adaptability in analyzing heart rate variability and calculating sympathetic and parasympathetic nervous system activity indices. Furthermore, through multimodal feature fusion and a deep learning classifier, the accuracy and reliability of sleep staging are improved, which is of great significance for sleep health research and nighttime physiological monitoring.

[0004] However, existing technologies still have room for improvement in terms of continuity and refinement of autonomic nervous system indicators and multimodal feature fusion when processing nocturnal multimodal physiological data. Further optimization of heart rate prediction models, feature fusion strategies, and the continuity of sleep stage determination is needed. Summary of the Invention

[0005] To address the shortcomings of existing technologies in processing multimodal physiological data at night, such as the lack of continuity and refinement in the fusion of autonomic nervous system indicators and multimodal features, and the need for further optimization of heart rate prediction models, feature fusion strategies, and the continuity of sleep stage determination, this invention provides a non-invasive sleep quality detection method and system that integrates autonomic nervous system monitoring.

[0006] The technical solutions provided by the embodiments of the present invention are as follows: The first aspect of this invention provides a non-invasive sleep quality detection method integrating autonomic nervous system monitoring, comprising: S1: Collect the non-invasive physiological data of the subject at night, including electrocardiogram signals and body movement posture signals; S2: Construct a Laguerre-based heart rate mean prediction model based on electrocardiogram signals and body posture signals; S3: Estimate the parameters of the heart rate mean prediction model and determine the time-varying Laguerre coefficient; S4: Calculate the autonomic nervous system balance index based on the time-varying Laguerre coefficient; S5: Extract features from non-invasive physiological data at night to obtain motion features and heart rate statistical features; S6: Combining motion characteristics, heart rate statistics, and autonomic balance index, we obtain the autonomic enhancement feature vector; S7: Input the autonomic nervous system enhancement feature vector into the sleep stage analysis module and output the sleep stage; S8: Generates a non-invasive sleep quality test report based on sleep stages.

[0007] A second aspect of this invention provides a non-invasive sleep quality detection system integrating autonomic nervous system monitoring, comprising: processor; A memory storing computer-readable instructions, which, when executed by the processor, implement the non-invasive sleep quality detection method integrating autonomic nervous system monitoring as described in the first aspect.

[0008] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: In this invention, addressing the need for continuity and refinement in processing multimodal physiological data at night, a heart rate mean prediction model based on the Laguerre function is constructed to preprocess electrocardiogram signals and calculate the autonomic nervous system balance index. Motion features and heart rate statistical features are extracted from nighttime physiological data according to a time window, and these are combined with the autonomic nervous system balance index to generate an autonomic nervous system enhancement feature vector. This feature vector is input into a sleep stage analysis module and analyzed using a deep learning classifier or structural attention mechanism to determine continuous nighttime sleep stages. This generates a non-invasive sleep quality report that accurately reflects the dynamics of nighttime autonomic nervous activity, while simultaneously achieving continuous fusion and refined processing of multimodal features, thereby improving the accuracy of sleep health monitoring and assessment. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1This is a flowchart illustrating a non-invasive sleep quality detection method that integrates autonomic nervous system monitoring, as provided in an embodiment of the present invention.

[0011] Figure 2 This is a schematic diagram of a non-invasive sleep quality detection system that integrates autonomic nervous system monitoring, provided as an embodiment of the present invention. Detailed Implementation

[0012] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0013] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0014] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0015] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0016] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0017] Reference manual attached Figure 1 The diagram shows a flowchart of a non-invasive sleep quality detection method integrating autonomic nervous system monitoring provided by an embodiment of the present invention.

[0018] This invention provides a non-invasive sleep quality detection method integrating autonomic nervous system monitoring. This method can be implemented using a non-invasive sleep quality detection device integrating autonomic nervous system monitoring, which can be a terminal or a server. The processing flow of the non-invasive sleep quality detection method integrating autonomic nervous system monitoring may include the following steps: S1: Collect the non-invasive physiological data of the subject at night, including electrocardiogram signals and body posture signals.

[0019] Among them, non-invasive physiological data at night are physiological signals that can be collected without invasive operation during the subject's sleep. Electrocardiogram signals record the electrical signals of the heart's electrical activity changing over time. Body posture signals record the subject's body posture and movement state through accelerometers and gyroscopes.

[0020] Specifically, by simultaneously acquiring electrocardiogram (ECG) signals and body posture signals, synchronous information on cardiac activity and motion disturbances can be obtained, providing a complete data foundation for the subsequent construction of a mean heart rate prediction model, thereby improving the accuracy of autonomic nervous system index calculation.

[0021] For example, nighttime micro-awakening or body movement may cause changes in heart rate waveforms. These fluctuations can be effectively captured through real-time acquisition, thus providing reliable data for subsequent prediction of mean heart rate and calculation of the autonomic nervous system balance index.

[0022] In this embodiment of the invention, the advantage of collecting non-invasive physiological data at night is that it can continuously acquire physiological signals throughout the night. Furthermore, by recording data synchronously through multiple channels, it helps to accurately analyze changes in heart rate and exercise status at night.

[0023] S2: Construct a Laguerre-based heart rate mean prediction model based on electrocardiogram signals and body posture signals.

[0024] Among them, the mean heart rate prediction model is a mathematical model that uses the periodic variation characteristics of electrocardiogram signals for prediction and modeling. The Laguerre function is a set of orthogonal polynomial functions used to expand the heart rate sequence into multi-order components for dynamic analysis.

[0025] Specifically, by using the Laguerre function to perform multi-order expansion of the heartbeat sequence, the dynamic characteristics of the heartbeat can be captured, providing a stable input for subsequent parameter estimation and autonomic nervous system index calculation.

[0026] In one possible implementation, S2 specifically includes sub-steps S201 to S204: S201: Preprocess the electrocardiogram (ECG) signal to generate an effective ECG signal sequence.

[0027] The effective ECG signal sequence is the ECG data sequence after filtering, abnormal RR correction and artifact removal. Abnormal RR correction is to adjust the RR interval that deviates from the normal range, and artifact removal is to remove invalid signals caused by movement or external interference.

[0028] Optionally, the preprocessing specifically includes filtering, anomaly RR correction, and artifact removal. Filtering, anomaly RR correction, and artifact removal are all mature existing technologies, and will not be described in detail here.

[0029] It should be noted that preprocessing the ECG signal can remove noise and abnormal RR interference, thereby ensuring that the generated effective ECG signal sequence is accurate and reliable.

[0030] S202: Construct the Laguerre function based on the effective electrocardiogram signal sequence.

[0031] Furthermore, the Laguerre function formula is constructed as follows:

[0032] in, L j ( n ) indicates the first j The first Laguerre basis function is at the 1st order. n The values ​​of each sampling point are used to represent the multi-order components of the heartbeat sequence. α This represents the time decay coefficient of the Laguerre function, controlling the influence of historical data on the current value. n Indicates the sampling point number. j ( k ) indicates the first j The first Laguerre basis function is at the 1st order. k The value of each sampling point.

[0033] It should be noted that by constructing the Laguerre function, the dynamic characteristics of the heartbeat signal at different order components can be captured, thereby improving the stability of the subsequent heartbeat mean prediction model and ensuring the reliability of the calculation of autonomic nerve indicators under nighttime signal fluctuations.

[0034] S203: The effective ECG signal sequence is convolved using the Laguerre function, and the body posture signal is introduced as a dynamic weighting coefficient to participate in the calculation, and the Laguerre filter output is calculated.

[0035] The Laguerre filter output is a feature sequence obtained by convolving the Laguerre basis function with the heartbeat sequence after weighted modulation of the body posture signal, which is used to reflect the dynamic response of the heartbeat on different Laguerre components.

[0036] Furthermore, the specific formula for calculating the Laguerre filter output is as follows:

[0037] in, Y j ( n ) indicates the first j The output of the first Laguerre filter is at the 1st order. n The value at each point in time. L j ( k ) indicates the first j The first Laguerre basis function is at the 1st order.k The value of each sampling point, RR corr ( n - k ) indicates that the sequence is extracted from a valid electrocardiogram signal sequence and corrected for abnormalities. n - k The RR interval value corresponding to the sampling point W act ( n - k ) indicates the first n - k The weighting coefficients of the body motion attitude signal at each sampling point are calculated from the body motion data collected by the accelerometer and gyroscope.

[0038] It should be noted that convolution operations can combine heartbeat signals with multi-order Laguerre basis functions.

[0039] Furthermore, the output feature sequence retains the main dynamic information of the heartbeat, thereby improving the stability of the heartbeat mean prediction model and the accuracy of subsequent autonomic nervous system index calculations.

[0040] S204: Construct a heart rate mean prediction model based on the Laguerre filter output.

[0041] Among them, the mean heartbeat prediction model generates the predicted average heartbeat interval for each heartbeat cycle, providing input for subsequent parameter estimation and autonomic nervous system index calculation.

[0042] It should be noted that the "mean heartbeat interval prediction" refers to an estimate of the RR interval length for the next heartbeat cycle. Here, "mean" does not refer to a statistical average, but rather emphasizes the model's smooth prediction of the heartbeat cycle to reduce random fluctuations in individual heartbeats. This prediction is numerically equal to the output of the linear combination described above, and is expressed in milliseconds.

[0043] Specifically, the Laguerre filter output is obtained by convolving the ECG signal with multiple Laguerre functions, with each order reflecting the dynamic characteristics of the heartbeat sequence at different time scales. The heartbeat mean prediction model is constructed by linearly combining the above-mentioned Laguerre filter outputs according to certain weights, and the combined result is the predicted heartbeat mean value corresponding to the current sampling point.

[0044] It should be noted that by using the output of the Laguerre filter to construct a heart rate mean prediction model, changes in the heart rate waveform can be smoothed.

[0045] Furthermore, it provides a continuous and reliable data sequence for calculating the autonomic nervous system balance index, thereby improving the robustness of the model under nocturnal physiological fluctuations.

[0046] In this embodiment of the invention, the advantage of constructing a heart rate mean prediction model based on the Laguerre function is that it can continuously describe the dynamic changes of the heart rate signal. The model output is not only used for heart rate mean prediction, but can also generate time series data that can be used to calculate the sympathetic and parasympathetic active components, thereby improving the calculation accuracy of the autonomic nervous system balance index.

[0047] S3: Estimate the parameters of the mean heart rate prediction model and determine the time-varying Laguerre coefficient.

[0048] Optionally, the parameter estimation method may include one of the following: maximum likelihood estimation, Kalman filtering, or recursive least squares.

[0049] Furthermore, the specific calculation formula for the Kalman filtering method is as follows:

[0050] in, Indicates time t The updated time-varying Laguerre coefficients, Indicates time t The prediction coefficient of -1 K t Indicates time t The Kalman gain is used to balance the weights of predicted and observed values. RR obs ( t () indicates time t Observations RR Interval, H t Indicates time t The observation matrix maps the model parameters to the observation space.

[0051] In this embodiment of the invention, the Laguerre coefficient can be accurately updated through parameter estimation. This operation enables the mean heart rate prediction model to maintain high-precision output under continuous changes in heart rate waveforms throughout the night. By obtaining accurate time-varying coefficients, dynamic weights can be used when subsequently calculating the active components of the sympathetic and parasympathetic nervous systems, making the calculation of autonomic nervous system indicators more stable and reliable. In addition, high-precision time-varying coefficients help capture subtle fluctuations in heart rate at night, ensuring that the model can reflect the trend of heart rate changes under different physiological states, thereby providing a reliable basis for sleep staging analysis and nighttime physiological assessment.

[0052] S4: Calculate the autonomic nervous system balance index based on the time-varying Laguerre coefficient.

[0053] Specifically, the time-varying Laguerre coefficient can be used to calculate the sympathetic and parasympathetic active components, which can be used to quantify the intensity and dynamic changes of autonomic nervous activity at night.

[0054] In one possible implementation, S4 specifically includes sub-steps S401 to S403: S401: Calculate the autonomic nervous system index based on the time-varying Laguerre coefficient. The autonomic nervous system index specifically includes the nocturnal sympathetic nervous system activity component and the parasympathetic nervous system activity component.

[0055] Furthermore, the specific formulas for calculating the nocturnal sympathetic and parasympathetic active components are as follows:

[0056]

[0057] in, SAI ( t () indicates time t The sympathetic nervous system's active components, P S0 Indicates the sympathetic baseline bias value. P Sj Indicates the first j The weights of the Laguerre coefficients, N 1 indicates the Laguerre order used to calculate the active components of the sympathetic nervous system. i j ( t ) indicates the first j Laguerre coefficient at time t The value, PAI ( t () indicates time t The parasympathetic nervous system's active components, P P0 Indicates the parasympathetic baseline bias value. P Pj Indicates the first j The weights of the Laguerre coefficients, N 2 represents the Laguerre order used to calculate the active components of the parasympathetic nervous system. Indicates the first j + N The first-order Laguerre coefficient at time t The value of .

[0058] It should be noted that by calculating the sympathetic and parasympathetic neural activity components, changes in heart rate waveforms can be mapped to quantitative indicators of autonomic nervous activity. This method can accurately reflect the changing trends of sympathetic and parasympathetic activity at night and provide reliable input for dynamic weighting and autonomic nervous balance indices. Furthermore, these indicators provide a stable foundation for subsequent multimodal feature fusion and sleep stage determination, ensuring the continuity of nighttime physiological signal analysis.

[0059] S402: Dynamically weight the nocturnal sympathetic and parasympathetic nervous system activity components.

[0060] Furthermore, the specific formula for dynamically weighting the nocturnal sympathetic and parasympathetic nervous system activity components is as follows:

[0061] in, ANW ( t () indicates time t The weighted index of autonomic nervous activity, w s Indicates the weight of the sympathetic nervous system. w p This indicates the parasympathetic nerve weight.

[0062] It should be noted that dynamic weighting can balance the contributions of sympathetic and parasympathetic components in the calculation of the autonomic nervous system balance index, thus more accurately reflecting the state of autonomic nervous activity at night. Furthermore, the weighting operation enables the subsequent feature fusion stage to utilize stable dynamic indicators, improving the reliability of multimodal feature fusion and the accuracy of sleep staging analysis.

[0063] S403: Calculate the autonomic nervous system balance index based on the dynamic weighting results.

[0064] Furthermore, the specific formula for calculating the autonomic nervous system balance index is as follows:

[0065] in, ANS _ balance ( t () indicates time t The autonomic nervous system balance index.

[0066] It should be noted that the dynamically weighted autonomic balance index accurately reflects the balance of autonomic activity at night, providing high-quality input for subsequent feature fusion and sleep stage determination. This method not only ensures the continuity of autonomic activity at night but also improves the accuracy of sleep quality assessment and nighttime physiological state monitoring.

[0067] Optionally, after S4, the following may also be included: When the autonomic nervous system balance index is less than the first preset balance index, the autonomic nervous system balance index is smoothed and then proceeds to step S5.

[0068] When the autonomic nervous system balance index is greater than or equal to the first preset balance index and less than the second preset balance index, proceed to step S5.

[0069] When the autonomic nervous system balance index is greater than or equal to the second preset balance index, the autonomic nervous system balance index is weighted and enhanced, and then the process proceeds to step S5.

[0070] It should be noted that those skilled in the art can set the magnitude of the first preset balance index and the second preset balance index according to actual needs, and this invention does not limit this.

[0071] Specifically, when the autonomic balance index is less than the first preset balance index, it indicates that the subject's autonomic activity is low. To improve the stability of subsequent feature extraction, the index is smoothed to reduce the impact of instantaneous fluctuations on the extraction of motion features and heart rate statistics, before proceeding to step S5 for feature extraction. Secondly, when the autonomic balance index is greater than or equal to the first preset balance index and less than the second preset balance index, it indicates that the subject's autonomic activity is in the intermediate range. In this case, the system requires no additional processing and can directly proceed to step S5 for feature extraction. Finally, when the autonomic balance index is greater than or equal to the second preset balance index, it indicates that the subject's autonomic activity is high. To highlight the importance of motion features in a high balance state, the autonomic balance index is weighted and enhanced to improve the representational power of subsequent feature vectors, before proceeding to step S5 for feature extraction.

[0072] It should be noted that the purpose of this branch design is to improve the adaptability of feature extraction to different autonomic nervous states and enhance the robustness and accuracy of non-invasive sleep quality detection methods. Those skilled in the art can adjust the division of the balance index, the smoothing method, and the weighted enhancement strategy according to actual needs without departing from the core solution of this invention.

[0073] In this embodiment of the invention, the advantage of calculating sympathetic and parasympathetic nerve activity indicators based on time-varying Laguerre coefficients and generating an autonomic balance index through dynamic weighting lies in its ability to continuously and accurately reflect changes in nocturnal autonomic nerve activity. This balance index provides stable and reliable quantitative input for subsequent multimodal feature fusion and sleep stage analysis. This method improves the continuity and accuracy of nocturnal autonomic nerve monitoring, providing a quantifiable and reliable data foundation for sleep stage determination and sleep quality assessment, thereby enhancing the overall sleep health analysis capability.

[0074] S5: Extract features from non-invasive physiological data at night to obtain motion features and heart rate statistical features.

[0075] Among them, motion features are used to describe the intensity and frequency of physical activity of subjects at night, and heart rate statistics are used to describe the statistical changes in heart rate intervals, such as mean heart rate, RR interval variance, and short-term heart rate variability.

[0076] In one possible implementation, S5 specifically includes sub-steps S501 and S502: S501: Divide the non-invasive physiological data at night into time windows to obtain multiple physiological data segments.

[0077] The time window division divides the continuous nighttime physiological signals into several fixed-length time periods, which facilitates the independent calculation of features for each time period.

[0078] It should be noted that dividing the time window can improve the accuracy of feature extraction, enabling motion features and heart rate statistics to reflect the physiological state at different time periods. Furthermore, processing each window independently can reduce the impact of abnormal signals on the overall analysis and improve the stability of subsequent feature fusion and sleep staging analysis.

[0079] S502: Within each time window, extract motion features and heart rate statistics from the physiological data of each segment.

[0080] Specifically, regarding the extraction of motion features: Based on the collected body posture signals (including data from triaxial accelerometers and gyroscopes), the synthetic acceleration amplitude within each time window is first calculated, which is the square root of the sum of the squares of the three axial accelerations. Then, based on the change of the synthetic acceleration amplitude over time, the following motion features are extracted: the average motion intensity within the window, the standard deviation of the motion intensity, the number of body movements (the number of events where the synthetic acceleration exceeds a preset threshold), and the maximum duration of a single body movement. Furthermore, by analyzing the angular velocity changes in the body posture signals, features such as the number of body rolls and the duration of posture maintenance can be extracted to reflect the subject's overall activity level at night.

[0081] Furthermore, regarding the extraction method of heart rate statistical features: based on the effective ECG signal sequence after preprocessing and correction for abnormal RR intervals, relevant indicators of the RR interval are statistically analyzed within each time window. Specifically, these include: the mean RR interval (i.e., the mean heart rate cycle), the standard deviation of the RR interval (to reflect the overall fluctuation of heart rate), the root mean square of the difference between adjacent RR intervals (to characterize short-term heart rate variability), and the approximate energy ratio of the high-frequency and low-frequency components of heart rate variability. In addition, the fastest heart rate, the slowest heart rate, and the coefficient of variation of heart rate (the ratio of the standard deviation to the mean) within the window can be calculated to comprehensively describe the dynamic changes in heart rate at night.

[0082] It should be noted that these motion features and heart rate statistics provide comprehensive information on nighttime heartbeats and motion status, providing high-quality input for the construction of autonomic neural enhancement feature vectors and ensuring the continuity and accuracy of multimodal feature analysis.

[0083] S6: Combining motion characteristics, heart rate statistics, and autonomic balance index, we obtain the autonomic enhancement feature vector.

[0084] In one possible implementation, S6 specifically includes sub-steps S601 and S602: S601: Combines the autonomic balance index, motor characteristics, and heart rate statistical characteristics to obtain combined features.

[0085] Among them, the combined features form a multi-dimensional vector by fusing different types of physiological indicators, which is used to reflect the subject's comprehensive physiological state at night.

[0086] Furthermore, the specific formula for calculating the combined features is as follows:

[0087] in, F fusion ( t () indicates time t The combined features, f ( ) represents the feature fusion function, which can be a vector concatenation or weighted fusion. X w This represents the motion features and heart rate statistics extracted within the time window. ANS _ balance ( t () indicates time t The autonomic nervous system balance index.

[0088] It should be noted that combining features from different sources to form composite features can provide complete multimodal information. By fusing autonomic nervous system indicators and exercise heart rate features, the sensitivity and stability of changes in nocturnal physiological state can be enhanced, providing high-quality input for sleep stage analysis.

[0089] S602: Normalize and standardize the combined features to generate an autonomic neural enhancement feature vector.

[0090] Among them, the autonomic nervous system enhancement feature vector is a multi-dimensional vector formed by unifying the scale of combined features, which can be directly input into the sleep stage analysis module.

[0091] It should be noted that this normalization and standardization process eliminates differences in feature dimensions, ensures balanced input to the classifier, thereby improving the accuracy of model training and prediction, and enhancing the stability and robustness of nighttime sleep staging analysis.

[0092] S7: Input the autonomic nervous system enhancement feature vector into the sleep stage analysis module and output the sleep stage.

[0093] In one possible implementation, S7 specifically includes sub-steps S701 to S704: S701: Input the autonomic nervous system enhancement feature vector into the sleep staging analysis module.

[0094] The sleep staging analysis module uses a deep learning classifier with an attention mechanism to process the input feature vector and determine the sleep stage for each time window.

[0095] It should be noted that inputting the autonomic neural enhancement feature vector can ensure that the classifier obtains complete multimodal physiological information, which enables the sleep stage determination to accurately reflect the physiological state at night and improve the accuracy and stability of the determination of continuous sleep stages at night.

[0096] S702: Calculate the reliability of the feature vector based on the autonomic neural enhancement feature vector.

[0097] Furthermore, the specific formula for calculating the credibility of the generated feature vector is as follows:

[0098] in, SCI ( t () indicates time t The feature vector credibility index is used to measure the time window. t The reliability of the autonomic neural enhancement feature vector, with values ​​ranging from 0 to 1. SCI ( t The larger the value, the more reliable the feature vector is for classifying sleep stages. s This represents the Sigmoid function, which maps the input to continuous values ​​in the interval [0,1]. It is used to map the ratio of feature variance to mean to a confidence index. F fusion ( t ) represents the autonomic neural enhancement feature vector. Var ( F fusion ( t )) indicates time. t eigenvectors F fusion ( t The variance of a feature reflects its dispersion or fluctuation within a time window and is used to assess its stability. Mean ( F fusion ( t )) indicates time. t eigenvectors F fusion ( t The mean of ) represents the overall average level of the feature within the time window and is used as a normalized variance index.

[0099] S703: Based on the credibility of the feature vector, the autonomic neural enhancement feature vector is processed by a deep learning classifier containing an attention mechanism to generate a sleep state probability vector for each time window.

[0100] Furthermore, the specific formula for calculating the sleep state probability vector for each time window is as follows:

[0101] in, P sleep ( t () indicates time t The probability vector of sleep stages, Softmax This represents the Softmax activation function. W Represents the classifier weight matrix. b This represents the bias vector.

[0102] It should be noted that by introducing feature vector credibility for weighting, the classifier's sensitivity to reliable features can be improved, making the sleep state probability vector more accurately reflect nocturnal physiological changes and providing a reliable basis for the final sleep stage determination.

[0103] S704: Based on the probability vectors of each sleep state, output the sleep stage of the current time window through the extreme value method. The sleep stages include the wakefulness stage, light sleep stage, deep sleep stage, and REM sleep stage.

[0104] Furthermore, the process for determining sleep stages is as follows:

[0105] in, S ( t () indicates time t The determined sleep stage indicates the final sleep state category in the output. This represents a function operator used to return the index that maximizes the probability value. i That is, the sleep stage with the highest probability is selected as the judgment result. i This represents the sleep stage index, indicating the numbers of all possible sleep stages, typically 1. i =1,2,3,4, corresponding to the stages of wakefulness, light sleep, deep sleep, and REM sleep.

[0106] It should be noted that selecting the sleep stage with the highest probability through the extreme value method can quickly determine the current state. Furthermore, this method combines probability vectors and confidence levels to make the determination of continuous sleep stages at night accurate and stable.

[0107] S8: Generates a non-invasive sleep quality test report based on sleep stages.

[0108] Optionally, the non-invasive sleep quality test report may include the proportion of sleep structure, the nocturnal autonomic rhythm curve, and the type of sleep disorder.

[0109] It should be noted that by generating sleep quality test reports, multidimensional physiological signals can be quantified into easily understandable indicators. Furthermore, this helps in the monitoring of nighttime sleep health and the analysis of autonomic nervous activity, providing a reliable reference for clinical or health management.

[0110] In this embodiment of the invention, the advantage of combining non-invasive nocturnal physiological data processing, mean heart rate prediction models, Laguerre function expansion, autonomic nervous system index calculation, multimodal feature fusion, and deep learning classifiers to form a complete analysis process is that it can continuously and accurately capture nocturnal heart rate dynamics and autonomic nervous system activity. Dynamic weighting and feature reliability enhance the stability and reliability of the data, enabling highly accurate sleep stage determination and sleep quality assessment, while providing reliable basic data for subsequent health monitoring and personalized sleep intervention.

[0111] Reference manual attached Figure 2 The diagram shows a structural schematic of a non-invasive sleep quality detection system that integrates autonomic nervous system monitoring, provided by the present invention.

[0112] The present invention also provides a non-invasive sleep quality detection system 20 integrating autonomic nervous system monitoring, applied to the above-mentioned non-invasive sleep quality detection method integrating autonomic nervous system monitoring, comprising: Processor 201.

[0113] The memory 202 stores computer-readable instructions, which, when executed by the processor 201, implement the non-invasive sleep quality detection method incorporating autonomic nervous system monitoring as described in the method embodiment.

[0114] The non-invasive sleep quality detection system 20 integrating autonomic nervous system monitoring provided by the present invention can perform the above-mentioned non-invasive sleep quality detection method integrating autonomic nervous system monitoring and achieve the same or similar technical effects. To avoid duplication, the present invention will not elaborate further.

[0115] It should be understood that the processor in the embodiments of the present invention can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0116] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0117] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0118] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0119] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0120] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0121] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0122] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0123] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0124] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0125] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0126] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0127] This invention provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the non-invasive sleep quality detection method integrating autonomic nervous system monitoring as described in the method embodiment.

[0128] The present invention provides a computer-readable storage medium that can implement the steps and effects of the non-invasive sleep quality detection method integrating autonomic nervous system monitoring in the above-described method embodiments. To avoid repetition, the present invention will not repeat the details.

[0129] 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. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0130] The following points need to be explained: (1) The accompanying drawings of the embodiments of the present invention only involve the structures involved in the embodiments of the present invention. Other structures can refer to the general design.

[0131] (2) For clarity, the thickness of layers or regions is enlarged or reduced in the drawings used to describe embodiments of the invention, i.e., these drawings are not drawn to scale. It is understood that when an element such as a layer, film, region or substrate is referred to as being “above” or “below” another element, the element may be “directly” located “above” or “below” the other element or there may be intermediate elements.

[0132] (3) Where there is no conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0133] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A non-invasive sleep quality detection method integrating autonomic nervous system monitoring, characterized in that, include: S1: Collect non-invasive physiological data of the subject at night, wherein the non-invasive physiological data at night includes electrocardiogram signals and body posture signals; S2: Construct a Laguerre-based heart rate mean prediction model based on the electrocardiogram signal and the body posture signal; S3: Estimate the parameters of the mean heart rate prediction model and determine the time-varying Laguerre coefficient; S4: Calculate the autonomic nervous system balance index based on the time-varying Laguerre coefficient; S5: Extract features from the non-invasive physiological data at night to obtain motion features and heart rate statistical features; S6: Combining the motion characteristics, the heart rate statistics, and the autonomic nervous balance index, we obtain the autonomic nervous enhancement feature vector; S7: Input the autonomic nervous system enhancement feature vector into the sleep stage analysis module and output the sleep stage; S8: Based on the sleep stage, generate a non-invasive sleep quality test report.

2. The non-invasive sleep quality detection method integrating autonomic nervous system monitoring according to claim 1, characterized in that, S2 specifically includes: S201: Preprocess the electrocardiogram (ECG) signal to generate an effective ECG signal sequence; S202: Construct the Laguerre function based on the effective electrocardiogram signal sequence; S203: Convolve the effective ECG signal sequence using the Laguerre function, and incorporate the body posture signal as a dynamic weighting coefficient to participate in the calculation, and calculate the Laguerre filter output; S204: Construct the mean heart rate prediction model based on the Laguerre filter output.

3. The non-invasive sleep quality detection method integrating autonomic nervous system monitoring according to claim 2, characterized in that, The preprocessing specifically includes filtering, anomaly correction (RR) and artifact removal.

4. The non-invasive sleep quality detection method integrating autonomic nervous system monitoring according to claim 1, characterized in that, The parameter estimation methods specifically include one of the following: maximum likelihood estimation, Kalman filtering, and recursive least squares.

5. The non-invasive sleep quality detection method integrating autonomic nervous system monitoring according to claim 1, characterized in that, S4 specifically includes: S401: Calculate the autonomic nervous system index based on the time-varying Laguerre coefficient, wherein the autonomic nervous system index specifically includes the nocturnal sympathetic nervous system activity component and the parasympathetic nervous system activity component. S402: Dynamically weight the nighttime sympathetic nerve activity component and the parasympathetic nerve activity component; S403: Calculate the autonomic nervous system balance index based on the dynamic weighting results.

6. The non-invasive sleep quality detection method integrating autonomic nervous system monitoring according to claim 1, characterized in that, S5 specifically includes: S501: Divide the nighttime non-invasive physiological data into time windows to obtain multiple segmented physiological data; S502: Within each time window, extract the motion features and heart rate statistical features from each segmented physiological data.

7. The non-invasive sleep quality detection method integrating autonomic nervous system monitoring according to claim 1, characterized in that, S6 specifically includes: S601: Combine the autonomic balance index, the movement characteristics, and the heart rate statistical characteristics to obtain a combined feature; S602: Normalize and standardize the combined features to obtain the autonomic neural enhancement feature vector.

8. The non-invasive sleep quality detection method integrating autonomic nervous system monitoring according to claim 1, characterized in that, Specifically, S7 includes: S701: Input the autonomic nervous system enhancement feature vector into the sleep staging analysis module; S702: Calculate the confidence level of the feature vector based on the aforementioned autonomic neural enhancement feature vector; S703: Based on the credibility of the feature vector, the autonomic neural enhancement feature vector is processed by a deep learning classifier containing an attention mechanism to generate a sleep state probability vector for each time window. S704: Based on the probability vectors of each sleep state, output the sleep stage of the current time window using the extreme value method, wherein the sleep stage includes the wakefulness stage, light sleep stage, deep sleep stage and REM sleep stage.

9. The non-invasive sleep quality detection method integrating autonomic nervous system monitoring according to claim 1, characterized in that, The non-invasive sleep quality test report specifically includes the proportion of sleep structure, the nocturnal autonomic nervous rhythm curve, and the type of sleep disorder.

10. A non-invasive sleep quality detection system integrating autonomic nervous system monitoring, characterized in that, include: processor; A memory storing computer-readable instructions, which, when executed by the processor, implement the non-invasive sleep quality detection method incorporating autonomic nervous system monitoring as described in any one of claims 1 to 9.

Images